Effectiveness of interventions to improve the public’s antimicrobial resistance awareness and behaviours associated with prudent use of antimicrobials: a systematic review

Effectiveness of interventions to improve the public’s antimicrobial resistance awareness and... Abstract Background A global antimicrobial resistance (AMR) awareness intervention targeting the general public has been prioritized. Objectives To evaluate the effectiveness of interventions that aim to change AMR awareness and subsequent stewardship behaviours amongst the public. Methods Five databases were searched between 2000 and 2016 for interventions to change the public’s AMR awareness and/or antimicrobial stewardship behaviours. Study designs meeting the Cochrane Effective Practice and Organization of Care (EPOC) criteria, non-controlled before-and-after studies and prospective cohort studies were considered eligible. Participants recruited from healthcare settings and studies measuring stewardship behaviours of healthcare professionals were excluded. Quality of studies was assessed using EPOC risk of bias criteria. Data were extracted and synthesized narratively. Registration: PROSPERO international prospective register of systematic reviews (PROSPERO 2016: CRD42016050343). Results Twenty studies were included in the review with nine meeting the EPOC criteria. The overall risk of bias was high. Nineteen studies were conducted in high-income countries. Mass media interventions were most common (n = 7), followed by school-based (n = 6) and printed material interventions (n = 6). Seventeen studies demonstrated a significant effect on changing knowledge, attitudes or the public’s antimicrobial stewardship behaviours. Analysis showed that interventions targeting schoolchildren and parents have notable potential, but for the general public the picture is less clear. Conclusions Our work provides an in-depth examination of the effectiveness of AMR interventions for the public. However, the studies were heterogeneous and the quality of evidence was poor. Well-designed, experimental studies on behavioural outcomes of such interventions are required. Introduction The rise of antimicrobial resistance (AMR) is a rapidly developing global threat that greatly affects our ability to deliver effective healthcare and results in a financial burden.1 AMR refers to the ability of a microorganism to adapt and grow despite the presence of antimicrobials. AMR threatens effective treatment of an ever-increasing range of infections.1 Therefore, increasing AMR is becoming a major public health concern. Although AMR is a naturally occurring phenomenon, inappropriate use of antimicrobials is the main driver of AMR.1 The demands for the use of antimicrobials are increasing worldwide and because of suboptimal management of these demands, huge quantities of antimicrobials are being misused.2 Together these highlight the need for effective strategies encouraging prudent use of antimicrobials. The O’Neill report emphasizes the need for AMR awareness interventions directed towards the public and development of a uniform, globally consistent set of AMR messages that could be then tailored to meet the specific demands of local settings.2 However, the report does not provide recommendations on the components of such interventions.2 Previous evidence syntheses show that the overall levels of knowledge and understanding of AMR amongst the public are generally low and members of the public often lack an understanding of their potential contribution to the development of AMR.3–5 Although high-level evidence demonstrating the effectiveness of interventions in increasing public understanding of AMR exists,5–7 these evaluations are methodologically diverse. It is therefore challenging to identify what interventions work and why and for whom they work, in order to inform future interventions. Thus, the aim of this systematic review is to provide the best-quality evidence regarding the effectiveness of AMR interventions that change public awareness and their subsequent antimicrobial stewardship behaviours. Although antimicrobial stewardship is most commonly thought of in medical settings, the word ‘stewardship’ means ‘taking care of’, particularly on behalf of others. Furthermore, a One Health perspective requires the collaborative effort of all stakeholders to take the responsibility for the prudent use of antimicrobials. Therefore, within this work, we use the term ‘antimicrobial stewardship’ to explore the public’s behaviours related to their prudent use of antimicrobials (such as, but not limited to, adhering to prescribers’ directions, not taking or demanding antimicrobial prescription for colds and flu and safe disposal of leftover antimicrobials). We believe an understanding of the public’s antimicrobial stewardship is central to engaging them with their part in reducing the drivers of AMR on behalf of future generations, other key stakeholders, such as prescribers, and the global community. Methods This review was prospectively registered on the PROSPERO international prospective register of systematic reviews (http://www.crd.york.ac.uk/PROSPERO/display_record.php?ID=CRD42016050343) and is reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement.8 Search strategy CINAHL, Cochrane Library, EMBASE, MEDLINE and PsycINFO databases were searched for articles published between 2000 and 2016 using keywords associated with the following four concept areas: (i) population—general public; (ii) intervention—interventions designed to increase antimicrobial awareness and/or to improve antimicrobial stewardship behaviour amongst the general public; (iii) context—AMR or the public’s antimicrobial stewardship; and (iv) outcomes—all relevant short-, medium- or long-term outcomes related to the public’s AMR and/or antimicrobial stewardship behaviours (knowledge/awareness, learning, public behavioural and cognition outcomes). The search strategy incorporated controlled vocabulary thesaurus terms and free text words contained in titles and abstracts. No restrictions were applied to language or publication status. The search strategy was amended according to the functionality of each of the databases. An example of the search strategy applied to MEDLINE is presented in Table S1 (available as Supplementary data at JAC Online). In addition to the database search, the reference lists of included papers and previous systematic reviews were searched manually and citation searches were conducted through Web of Science in order to identify additional records. Study selection Cochrane’s Effective Practice and Organization of Care (EPOC) recommendations9 were used to initially select studies for inclusion in the review. Although EPOC guidelines suggest the inclusion of randomized controlled trials, non-randomized controlled trials, controlled before-and-after studies, interrupted time series studies and repeated measures studies exclusively,9 because of the limited number of eligible studies meeting the EPOC criteria, non-controlled before-and-after studies and prospective cohort studies were also deemed eligible for the review. Interventions targeting the general public population and designed to increase public antimicrobial awareness and/or to improve the public’s antimicrobial stewardship were deemed eligible for the review. Furthermore, time-bound geographical controls or no exposure comparators and all relevant short-, medium- or long-term outcomes related to AMR and/or the public’s antimicrobial stewardship behaviours were included, whereas those related to antimicrobial prescribing were excluded as this was regarded as the behaviour of healthcare professionals rather than of the members of the general public. Eligibility criteria applied in this study are presented in detail in Table 1. Table 1. Review’s eligibility criteria Inclusion criteria Exclusion criteria Design randomized controlled trials, non-randomized trials, interrupted time series studies, controlled before-and-after studies, non-controlled before-and-after studies and cohort studies – Population members of the public participants recruited from healthcare settings Intervention intervention designed to increase public antimicrobial awareness and/or to improve antimicrobial stewardship (through mass media, social marketing or printed media campaigns) – Comparator time bound, geographical controls or no exposure – Context non-healthcare settings; AMR or the public’s antimicrobial stewardship – Outcomes all relevant short-, medium- or long-term outcomes related to AMR and/or antimicrobial stewardship behaviours (knowledge/awareness, learning, public behavioural and cognition outcomes) antimicrobial prescribing Publication date published after January 2000 published before January 2000 Inclusion criteria Exclusion criteria Design randomized controlled trials, non-randomized trials, interrupted time series studies, controlled before-and-after studies, non-controlled before-and-after studies and cohort studies – Population members of the public participants recruited from healthcare settings Intervention intervention designed to increase public antimicrobial awareness and/or to improve antimicrobial stewardship (through mass media, social marketing or printed media campaigns) – Comparator time bound, geographical controls or no exposure – Context non-healthcare settings; AMR or the public’s antimicrobial stewardship – Outcomes all relevant short-, medium- or long-term outcomes related to AMR and/or antimicrobial stewardship behaviours (knowledge/awareness, learning, public behavioural and cognition outcomes) antimicrobial prescribing Publication date published after January 2000 published before January 2000 Table 1. Review’s eligibility criteria Inclusion criteria Exclusion criteria Design randomized controlled trials, non-randomized trials, interrupted time series studies, controlled before-and-after studies, non-controlled before-and-after studies and cohort studies – Population members of the public participants recruited from healthcare settings Intervention intervention designed to increase public antimicrobial awareness and/or to improve antimicrobial stewardship (through mass media, social marketing or printed media campaigns) – Comparator time bound, geographical controls or no exposure – Context non-healthcare settings; AMR or the public’s antimicrobial stewardship – Outcomes all relevant short-, medium- or long-term outcomes related to AMR and/or antimicrobial stewardship behaviours (knowledge/awareness, learning, public behavioural and cognition outcomes) antimicrobial prescribing Publication date published after January 2000 published before January 2000 Inclusion criteria Exclusion criteria Design randomized controlled trials, non-randomized trials, interrupted time series studies, controlled before-and-after studies, non-controlled before-and-after studies and cohort studies – Population members of the public participants recruited from healthcare settings Intervention intervention designed to increase public antimicrobial awareness and/or to improve antimicrobial stewardship (through mass media, social marketing or printed media campaigns) – Comparator time bound, geographical controls or no exposure – Context non-healthcare settings; AMR or the public’s antimicrobial stewardship – Outcomes all relevant short-, medium- or long-term outcomes related to AMR and/or antimicrobial stewardship behaviours (knowledge/awareness, learning, public behavioural and cognition outcomes) antimicrobial prescribing Publication date published after January 2000 published before January 2000 Titles and abstracts of identified records were screened against the eligibility criteria (Table 1) by one of three reviewers (M. Y., L. G. or F. S.) with a 30% subset of excluded studies independently checked by another reviewer (M. Y., L. G. or F. S.). The level of agreement on this subset was 99%. Disagreements were resolved with the involvement of another, experienced reviewer (L. P.). Full texts of papers that appeared to meet the inclusion criteria or those with insufficient information within the title and abstract were screened by two out of three independent reviewers (L. G., M. Y. and F. S.), with a fourth reviewer (L. P.) checking all decisions and resolving any discrepancies. Whenever possible, foreign-language papers were translated by members of the team who have a command of foreign languages or were translated using Google Translate. Data extraction and quality assessment Two out of three reviewers (L. G., M. Y. and F. S.) independently extracted data from eligible studies using a standardized tool, designed for the purpose of the study (Table S2). For studies that met the EPOC study design criteria (randomized controlled trials, non-randomized controlled trials, controlled before-and-after studies and interrupted time series studies),9 risk of bias was assessed across domains by one reviewer (M. Y., L. G. or F. S.) and checked by a second reviewer (M. Y., L. G. or F. S.) using standard EPOC risk of bias criteria.9 Disagreements were resolved through consensus or, if necessary, in consultation with a fourth reviewer (L. P.). Risk of bias assessments were not conducted for non-controlled before-and-after studies as it was assumed that the risk of bias of these studies was high. No studies were excluded based on quality assessment. Data analysis Given the heterogeneity of the study designs, populations, interventions and outcome measures, it was not possible to pool the results in a meta-analysis. Therefore, we applied an alternative, systematic approach to assessing complex interventions and carried out a narrative synthesis of evidence following the Cochrane Consumers and Communication Review Group’s guidelines.10 Individual study characteristics and findings were summarized and similarities, differences and patterns identified. Studies were grouped into those meeting or not meeting the EPOC criteria and categorized according to the target population. To identify discernible patterns of effectiveness, identified studies were mapped across five categories of intervention effectiveness. These categories were based upon both the strength of the evidence and the position of the primary outcome within the causal chain linking antecedents of behaviour to actual behaviour change. For example, knowledge is understood to be a necessary, but insufficient, predictor of behaviour as people can develop good awareness and understanding of AMR yet still fail to implement the public’s AMR stewardship. The five categories of a relative measure of effectiveness included: (i) interventions indicative of clear positive behaviour change in the desired direction; (ii) interventions indicative of some positive behaviour change in the desired direction; (iii) interventions indicative of a positive effect on the antecedent of behaviour, such as knowledge or awareness, in the desired direction; (iv) interventions indicative of no effect on behaviour or antecedents of behaviour; and (v) interventions indicative of a negative effect on behaviour or antecedents of behaviour in a non-desired direction. Results An electronic search resulted in the total of 17 312 records. An additional 31 records were identified through reference and citation searching of the included papers. A total number of 60 studies that did not meet eligibility criteria were excluded during the full-text reviewing stage. Articles were excluded for not meeting study design criteria, study participants being recruited from healthcare settings, context other than AMR, study outcomes not related to the public’s AMR awareness or antimicrobial stewardship, a full-text record being unavailable and other reasons, such as the record being a study protocol, a conference abstract of an already identified study, a short report of an already identified study, inability to translate a non-English paper or the majority of participants recruited for the study being healthcare workers. A detailed list of excluded papers is presented in Table S3. Following screening, 20 studies that matched the eligibility criteria were included in the review. A detailed process of study selection is presented in Figure 1. Figure 1. View largeDownload slide Study selection flowchart. RCT, randomized controlled trial; NCT, non-randomized controlled trial. Figure 1. View largeDownload slide Study selection flowchart. RCT, randomized controlled trial; NCT, non-randomized controlled trial. Study characteristics As shown in Table 2, study designs of the 20 reviewed studies included randomized controlled trials (n = 2),11,12 non-randomized controlled trials (n = 3),13–15 controlled before-and-after studies (n = 4),16–19 non-controlled before-and-after studies (n = 10)20–29 and a prospective cohort study (n = 1).30 Table 2. Study characteristics and results of the included studies Study Country Design Sample Nature of intervention(s) Outcome measures Significant results Azevedo et al. (2013)20 Braga, Portugal non-controlled before-and-after study n = 82 schoolchildren School-based presentation followed by discussion. knowledge and attitudes Knowledge of the correct use of antibiotics for bacterial diseases rather than viral diseases rose from 43% to 76% in the post-test (P < 0.01). Knowledge of the risk of bacterial resistance to antibiotics from their incorrect use rose from 48% to 74% in the post-test (P < 0.05). Cebotarenco and Bush (2008)16 Chisinau, Moldova controlled before-and-after study n = 3586 schoolchildren and n = 2716 parents Educational intervention about the use of antibiotics delivered by student volunteers trained as peer leaders delivered to their classmates and the classmates’ parents. beliefs and behaviour Students in both the intervention district and the post-intervention phase were 3.2 (95%  CI = 2.065–4.909) times more likely than other students to indicate they had not taken an antibiotic. Croft et al. (2007)11 WI, USA randomized controlled trial n = 300 parents Distribution of printed materials to parents by childcare staff; slide presentation delivered to staff. knowledge In parents who were college graduates, the median knowledge scores were 7.0 at intervention centres and 6.5 at control centres (P < 0.01). Curry et al. (2006)21 Auckland, New Zealand non-controlled before-and-after study n = 400 general public National campaign ‘Wise use of antibiotics’. Posters and leaflets delivered to the public attending pharmacies. knowledge and attitudes and behaviour Patients who had ever been to the doctor for a common cold significantly decreased (45% versus 62%; P = 0.0006). They were significantly less likely to feel positive about antibiotics in 2003 for the treatment of a cold (16% versus 33%; P = 0.00001). The perception that antibiotics were beneficial for cold/flu symptoms significantly reduced from 1998 to 2003 (P < 0.05) and the perceived benefit of antibiotics for tonsillitis increased from 83% to 91% in 2003 (P = 0.014). Significantly fewer people reported ever attending a doctor for a cold in 2003 versus 1998 (45% versus 62%; P = 0.0006) and the number of people who would usually see a doctor for a cold decreased from 24% to 15% (P = 0.026). Farrell et al. (2011)22 Glasgow, Gloucester and London, UK non-controlled before-and-after study n = 1736 children e-Bug web game. knowledge and attitudes No overall change in knowledge. Significant knowledge change in 3 out of 21 questions (P  ≤  0.02). Formoso et al. (2013)13 Emilia-Romagna, Italy non-randomized controlled trial n = 1200 general public Local mass media campaign (posters, brochures and advertisements on local media) delivered to general population to raise awareness of inappropriate use of antibiotics. knowledge After the intervention, consistency with campaign messages worsened (or did not improve) similarly in both intervention and control areas, the only exception being knowledge of the presumptive antiviral activity of antibiotics, worsening in the intervention area more than the control area. Gonzales et al. (2008)14 CO, USA non-randomized controlled trial n = 1503 general public Mass media intervention (outdoor and radio advertisements) delivered to general public about use of antibiotics. behaviour Linear regression analysis showed significant net differences in monthly paediatric office visit rates between mass media and comparison communities before and after the campaign (P = 0.01). Huang et al. (2007)12 MA, USA randomized controlled trial n = 3142 parents Community-based educational intervention occurred through three successive cold and flu seasons. Printed materials: mailed newsletters, posters, pamphlets and fact sheets in the waiting rooms of local paediatric providers, pharmacies and childcare centres. knowledge and attitudes The proportion of parents who answered ≥7 of 10 knowledge questions correctly increased significantly in both intervention (from 52% to 64%; P < 0.001) and control (from 54% to 61%; P < 0.01) communities. Substantial improvements in the percentage of correct answers were seen for items on middle ear fluid (41% in 2000; 50% in 2003; P < 0.001) and the general question of whether antibiotics were needed for colds and flu (66% in 2000; 77% in 2003; P < 0.001). Larson et al. (2009)23 Upper Manhattan, USA non-controlled before-and-after study n = 422 households Targeted Latino households. Educational materials (colouring book, pamphlets) based on knowledge, attitudes and practices regarding prevention and treatment of upper respiratory tract infections. Programme was delivered during home visits every 2 months. knowledge and attitudes and behaviour After the intervention, the mean composite knowledge scores at baseline and end of study were 5.19 (SD = 1.60) and 5.91 (SD = 1.71) (P < 0.001), respectively. With regard to reported practices, significantly more participants after the intervention reported using alcohol hand sanitizers (1.4% baseline and 66.8% post-intervention; P = 0.001). Significantly more also reported that one or more members of their household had received the influenza vaccination after the intervention (63.7% at baseline and 73.9% post-intervention; P = 0.001). Lecky et al. (2010)17 Gloucestershire and London, England; Nice and Bordeaux, France; Prague and Ostrava, Czech Republic controlled before-and-after study n = 2724 school students School-based educational intervention (e-Bug) regarding inappropriate antibiotic use delivered to classes of 9- to 11-year-old (junior) and 12- to 15-year-old (senior) students in state schools. Included 45 min lesson, handouts, worksheets, factsheets, interactive activity and a follow-up plenary question-and-answer session. knowledge Junior school: significant change in knowledge and significant change in retention 6 weeks post-intervention, across countries. Little significant difference in knowledge change between intervention and control, with the exception of the Czech Republic. Senior school: significant improvement in knowledge 6 weeks post-intervention in the Czech Republic. Significant improvement in knowledge and knowledge retention between control and intervention in England and the Czech Republic Madle et al. (2004)24 London, UK non-controlled before-and-after study n = 177 general public Open access to the National Electronic Library of Infection Antimicrobial Resistance website on the use of antibiotics and antibiotic resistance. The site comprises frequently asked questions and links to evidence-based resources. knowledge and attitudes Significant improvements in knowledge about the use of antibiotics and antibiotic resistance in 2 out of 3 statements: (i) ‘people cannot become resistant to antibiotics’ (P < 0.001, χ2 = 60.357, 95% CI of change = 27.47–44.53); (ii) ‘antibiotics do not cure most sore throats’ (P < 0.001, χ2 = 19.22, 95% CI of change = 8.62–27.38). Significant changes in the scores assigned by users for 3 out of 4 statements designed to test users’ attitudes to the information on the site (P ≤ 0.003). Expectations that antibiotics should be prescribed were significantly reduced after using the website (P < 0.001). Non-healthcare workers continued to have higher expectations of antibiotics being prescribed than healthcare workers (P = 0.0046 before and P = 0.0098 after using the website). Mainous et al. (2009)30 SC, USA Prospective cohort study n = 691 self-identified Latinos Mass media educational intervention (pamphlets, radio, newspapers) delivered to local Latino communities about use of antibiotics. behaviour Numbers in the intervention group reporting that they had bought antibiotics without a prescription increased following the intervention compared with baseline (χ2 test reported as significant, but P value not given). The regression analysis showed the strongest predictor of purchase of antibiotics without a prescription in the previous 12 months was past purchase of antibiotics without a prescription outside the USA (OR = 5.72; 95% CI = 3.12–10.48). The regression analysis also showed the strongest predictor of likelihood of importing antibiotics into the USA was past purchase of antibiotics without a prescription outside the USA (OR = 3.01; 95% CI = 1.95–4.65). Mazińska and Hryniewicz (2010)25 Poland non-controlled before-and-after study n = 1000 general public Mass media educational intervention (posters, leaflets, billboards, TV, cinemas, radio, press, magazines, thematic exhibitions, Internet) implemented across the country. knowledge, attitudes and behaviour Significant increase in the percentage of people who have limited the use of antibiotics (27% to 43%), have become more disciplined and cautious in their use (3% to 24%) and pay attention to the correct dosage (6% to 18%) (no P values given). McNulty et al. (2001)27 Gloucester, UK non-controlled before-and-after study n = 38 year 5 schoolchildren School-based intervention to children aged 9–10 years at a state school. Included two 90 min interactive workshops entitled ‘Antibiotics and your good bugs’. knowledge Before the workshops 23% and 26% knew antibiotics do not kill viruses, but kill good bacteria, compared with 47% and 69% afterwards (P = 0.03 and 0.0001). 45% before and 73% after the workshops correctly answered all the questions (P < 0.0001). Children thought antibiotics helped hay fever; this improved significantly after the workshop [correct answer 28% before, 77% after (P < 0.0001)]. Overall score for 7 questions in the ‘where are bugs found’ section was increased significantly from an average of 80.5% success to 93.2% success (P = 0.0002). The overall score improvement in the ‘How do bugs spread’ section was significant (P = 0.00001). McNulty et al. (2007)26 Gloucestershire, UK non-controlled before-and-after study n = 198 year 5 and 6 schoolchildren School-based intervention. ‘Bug Investigators’ pack about microorganisms, hygiene and antibiotics. The pack included 11 activity sheets, teachers’ guide, poster and website. knowledge Children’s knowledge improved in all topic areas and was significant in 6 out of the 7 topic areas (P < 0.005). Improved knowledge was most significant for what antibiotics do and how to use them [improvement = 27% (95% CI = 22.8–31.1) and 31% (95% CI = 23.4–37.7), respectively] and the value of our own good bugs (16% improvement). McNulty et al. (2010)18 England and Scotland, UK controlled before-and-after study n = 3718 general public Mass media campaign about antibiotic use involving posters displayed in magazines and newspapers. knowledge and attitudes and behaviour No positive effect of the campaigns. Pontes and Pontes (2005)15 Mid-Atlantic region, USA non-randomized controlled trials n = 105 university students University-based educational intervention (information booklet) to increase young adult consumers’ preference for physicians who do not unnecessarily prescribe antibiotics for simple acute upper respiratory tract infections. attitudes Exposure to the intervention significantly increased the mean preferred start date for antibiotics after initiation of an infection from 2.3 to 3.9 days (P < 0.01). Respondents’ preferences were significantly greater for the physician who indicated he would not prescribe antibiotics in the intervention (mean = 4.84) compared with the control (mean = 3.91) (P < 0.01). Stockwell et al. (2010)28 New York City, USA non-controlled before-and-after study n = 10 parents Health literacy intervention regarding upper respiratory tract infection delivered to parents on an ‘Early Head Start’ programme. Involved 3 × 1.5 h interactive sessions and provision of kit for care of a child with such an infection. knowledge and attitudes and behaviour The mean composite knowledge/attitude score increased from 4.1 (total possible: 10) to 6.6 (P < 0.05). The number of parents reporting that the last time their child was sick they sought antibiotics without a prescription instead of, or in addition to, seeing their healthcare provider decreased from 6 to 1 (P = 0.06). Trepka et al. (2001)19 Northern WI, USA controlled before-and-after study n = 365 parents Nurse educators delivered parent-oriented presentations in community organizations, distributed information pamphlets and displayed posters. Topics covered included antibiotic resistance and use. knowledge and attitudes and behaviour From baseline to post-intervention the percentage of parents with high antibiotic resistance awareness significantly increased in the intervention group (change = 14.3%; 95% CI = 6.6%–22.0%), but not in the control group (change = 4.3%; 95% CI = −4.1% to 12.7%; P = 0.015). The proportion of parents who expected an antibiotic for their child and did not receive one declined in the intervention area (14% to 9%), while it increased in the control area (7% to 10%). The difference between the two area changes was −8.4% (95% CI = −13.9% to −2.8%; P = 0.003). The percentage of parents in the intervention area who brought their child to another physician because they did not receive an antibiotic decreased (5% to 2%), whereas it increased in the control area (2% to 4%). The difference between the two area changes was −4.5% (95% CI=−8.0% to −0.9%; P = 0.02). Wutzke et al. (2007)29 Australia non-controlled before-and-after study n = 6217 general public National mass media intervention for consumers delivered during winter months in 2001, 2002, 2003 and 2004. About the inappropriate use of antibiotics for upper respiratory tract infection. Strategies included newsletters and brochures, mass media activity using billboards, television, radio and magazines and small grants to promote local community education. knowledge and attitudes and behaviour There was a significant decline in those who believe taking antibiotics for cold and flu is appropriate, from 28.7% pre-programme in 2002 to 21.7% in 2004 (percentage point change = 7.0; 95% CI = 3.5–10.5). Significant decrease in self-reported use of antibiotics to treat cough, cold or flu, from 10.8% in 1999 to 7.4% in 2004 (percentage point change = 3.4; 95% CI = 1.3–5.5). Study Country Design Sample Nature of intervention(s) Outcome measures Significant results Azevedo et al. (2013)20 Braga, Portugal non-controlled before-and-after study n = 82 schoolchildren School-based presentation followed by discussion. knowledge and attitudes Knowledge of the correct use of antibiotics for bacterial diseases rather than viral diseases rose from 43% to 76% in the post-test (P < 0.01). Knowledge of the risk of bacterial resistance to antibiotics from their incorrect use rose from 48% to 74% in the post-test (P < 0.05). Cebotarenco and Bush (2008)16 Chisinau, Moldova controlled before-and-after study n = 3586 schoolchildren and n = 2716 parents Educational intervention about the use of antibiotics delivered by student volunteers trained as peer leaders delivered to their classmates and the classmates’ parents. beliefs and behaviour Students in both the intervention district and the post-intervention phase were 3.2 (95%  CI = 2.065–4.909) times more likely than other students to indicate they had not taken an antibiotic. Croft et al. (2007)11 WI, USA randomized controlled trial n = 300 parents Distribution of printed materials to parents by childcare staff; slide presentation delivered to staff. knowledge In parents who were college graduates, the median knowledge scores were 7.0 at intervention centres and 6.5 at control centres (P < 0.01). Curry et al. (2006)21 Auckland, New Zealand non-controlled before-and-after study n = 400 general public National campaign ‘Wise use of antibiotics’. Posters and leaflets delivered to the public attending pharmacies. knowledge and attitudes and behaviour Patients who had ever been to the doctor for a common cold significantly decreased (45% versus 62%; P = 0.0006). They were significantly less likely to feel positive about antibiotics in 2003 for the treatment of a cold (16% versus 33%; P = 0.00001). The perception that antibiotics were beneficial for cold/flu symptoms significantly reduced from 1998 to 2003 (P < 0.05) and the perceived benefit of antibiotics for tonsillitis increased from 83% to 91% in 2003 (P = 0.014). Significantly fewer people reported ever attending a doctor for a cold in 2003 versus 1998 (45% versus 62%; P = 0.0006) and the number of people who would usually see a doctor for a cold decreased from 24% to 15% (P = 0.026). Farrell et al. (2011)22 Glasgow, Gloucester and London, UK non-controlled before-and-after study n = 1736 children e-Bug web game. knowledge and attitudes No overall change in knowledge. Significant knowledge change in 3 out of 21 questions (P  ≤  0.02). Formoso et al. (2013)13 Emilia-Romagna, Italy non-randomized controlled trial n = 1200 general public Local mass media campaign (posters, brochures and advertisements on local media) delivered to general population to raise awareness of inappropriate use of antibiotics. knowledge After the intervention, consistency with campaign messages worsened (or did not improve) similarly in both intervention and control areas, the only exception being knowledge of the presumptive antiviral activity of antibiotics, worsening in the intervention area more than the control area. Gonzales et al. (2008)14 CO, USA non-randomized controlled trial n = 1503 general public Mass media intervention (outdoor and radio advertisements) delivered to general public about use of antibiotics. behaviour Linear regression analysis showed significant net differences in monthly paediatric office visit rates between mass media and comparison communities before and after the campaign (P = 0.01). Huang et al. (2007)12 MA, USA randomized controlled trial n = 3142 parents Community-based educational intervention occurred through three successive cold and flu seasons. Printed materials: mailed newsletters, posters, pamphlets and fact sheets in the waiting rooms of local paediatric providers, pharmacies and childcare centres. knowledge and attitudes The proportion of parents who answered ≥7 of 10 knowledge questions correctly increased significantly in both intervention (from 52% to 64%; P < 0.001) and control (from 54% to 61%; P < 0.01) communities. Substantial improvements in the percentage of correct answers were seen for items on middle ear fluid (41% in 2000; 50% in 2003; P < 0.001) and the general question of whether antibiotics were needed for colds and flu (66% in 2000; 77% in 2003; P < 0.001). Larson et al. (2009)23 Upper Manhattan, USA non-controlled before-and-after study n = 422 households Targeted Latino households. Educational materials (colouring book, pamphlets) based on knowledge, attitudes and practices regarding prevention and treatment of upper respiratory tract infections. Programme was delivered during home visits every 2 months. knowledge and attitudes and behaviour After the intervention, the mean composite knowledge scores at baseline and end of study were 5.19 (SD = 1.60) and 5.91 (SD = 1.71) (P < 0.001), respectively. With regard to reported practices, significantly more participants after the intervention reported using alcohol hand sanitizers (1.4% baseline and 66.8% post-intervention; P = 0.001). Significantly more also reported that one or more members of their household had received the influenza vaccination after the intervention (63.7% at baseline and 73.9% post-intervention; P = 0.001). Lecky et al. (2010)17 Gloucestershire and London, England; Nice and Bordeaux, France; Prague and Ostrava, Czech Republic controlled before-and-after study n = 2724 school students School-based educational intervention (e-Bug) regarding inappropriate antibiotic use delivered to classes of 9- to 11-year-old (junior) and 12- to 15-year-old (senior) students in state schools. Included 45 min lesson, handouts, worksheets, factsheets, interactive activity and a follow-up plenary question-and-answer session. knowledge Junior school: significant change in knowledge and significant change in retention 6 weeks post-intervention, across countries. Little significant difference in knowledge change between intervention and control, with the exception of the Czech Republic. Senior school: significant improvement in knowledge 6 weeks post-intervention in the Czech Republic. Significant improvement in knowledge and knowledge retention between control and intervention in England and the Czech Republic Madle et al. (2004)24 London, UK non-controlled before-and-after study n = 177 general public Open access to the National Electronic Library of Infection Antimicrobial Resistance website on the use of antibiotics and antibiotic resistance. The site comprises frequently asked questions and links to evidence-based resources. knowledge and attitudes Significant improvements in knowledge about the use of antibiotics and antibiotic resistance in 2 out of 3 statements: (i) ‘people cannot become resistant to antibiotics’ (P < 0.001, χ2 = 60.357, 95% CI of change = 27.47–44.53); (ii) ‘antibiotics do not cure most sore throats’ (P < 0.001, χ2 = 19.22, 95% CI of change = 8.62–27.38). Significant changes in the scores assigned by users for 3 out of 4 statements designed to test users’ attitudes to the information on the site (P ≤ 0.003). Expectations that antibiotics should be prescribed were significantly reduced after using the website (P < 0.001). Non-healthcare workers continued to have higher expectations of antibiotics being prescribed than healthcare workers (P = 0.0046 before and P = 0.0098 after using the website). Mainous et al. (2009)30 SC, USA Prospective cohort study n = 691 self-identified Latinos Mass media educational intervention (pamphlets, radio, newspapers) delivered to local Latino communities about use of antibiotics. behaviour Numbers in the intervention group reporting that they had bought antibiotics without a prescription increased following the intervention compared with baseline (χ2 test reported as significant, but P value not given). The regression analysis showed the strongest predictor of purchase of antibiotics without a prescription in the previous 12 months was past purchase of antibiotics without a prescription outside the USA (OR = 5.72; 95% CI = 3.12–10.48). The regression analysis also showed the strongest predictor of likelihood of importing antibiotics into the USA was past purchase of antibiotics without a prescription outside the USA (OR = 3.01; 95% CI = 1.95–4.65). Mazińska and Hryniewicz (2010)25 Poland non-controlled before-and-after study n = 1000 general public Mass media educational intervention (posters, leaflets, billboards, TV, cinemas, radio, press, magazines, thematic exhibitions, Internet) implemented across the country. knowledge, attitudes and behaviour Significant increase in the percentage of people who have limited the use of antibiotics (27% to 43%), have become more disciplined and cautious in their use (3% to 24%) and pay attention to the correct dosage (6% to 18%) (no P values given). McNulty et al. (2001)27 Gloucester, UK non-controlled before-and-after study n = 38 year 5 schoolchildren School-based intervention to children aged 9–10 years at a state school. Included two 90 min interactive workshops entitled ‘Antibiotics and your good bugs’. knowledge Before the workshops 23% and 26% knew antibiotics do not kill viruses, but kill good bacteria, compared with 47% and 69% afterwards (P = 0.03 and 0.0001). 45% before and 73% after the workshops correctly answered all the questions (P < 0.0001). Children thought antibiotics helped hay fever; this improved significantly after the workshop [correct answer 28% before, 77% after (P < 0.0001)]. Overall score for 7 questions in the ‘where are bugs found’ section was increased significantly from an average of 80.5% success to 93.2% success (P = 0.0002). The overall score improvement in the ‘How do bugs spread’ section was significant (P = 0.00001). McNulty et al. (2007)26 Gloucestershire, UK non-controlled before-and-after study n = 198 year 5 and 6 schoolchildren School-based intervention. ‘Bug Investigators’ pack about microorganisms, hygiene and antibiotics. The pack included 11 activity sheets, teachers’ guide, poster and website. knowledge Children’s knowledge improved in all topic areas and was significant in 6 out of the 7 topic areas (P < 0.005). Improved knowledge was most significant for what antibiotics do and how to use them [improvement = 27% (95% CI = 22.8–31.1) and 31% (95% CI = 23.4–37.7), respectively] and the value of our own good bugs (16% improvement). McNulty et al. (2010)18 England and Scotland, UK controlled before-and-after study n = 3718 general public Mass media campaign about antibiotic use involving posters displayed in magazines and newspapers. knowledge and attitudes and behaviour No positive effect of the campaigns. Pontes and Pontes (2005)15 Mid-Atlantic region, USA non-randomized controlled trials n = 105 university students University-based educational intervention (information booklet) to increase young adult consumers’ preference for physicians who do not unnecessarily prescribe antibiotics for simple acute upper respiratory tract infections. attitudes Exposure to the intervention significantly increased the mean preferred start date for antibiotics after initiation of an infection from 2.3 to 3.9 days (P < 0.01). Respondents’ preferences were significantly greater for the physician who indicated he would not prescribe antibiotics in the intervention (mean = 4.84) compared with the control (mean = 3.91) (P < 0.01). Stockwell et al. (2010)28 New York City, USA non-controlled before-and-after study n = 10 parents Health literacy intervention regarding upper respiratory tract infection delivered to parents on an ‘Early Head Start’ programme. Involved 3 × 1.5 h interactive sessions and provision of kit for care of a child with such an infection. knowledge and attitudes and behaviour The mean composite knowledge/attitude score increased from 4.1 (total possible: 10) to 6.6 (P < 0.05). The number of parents reporting that the last time their child was sick they sought antibiotics without a prescription instead of, or in addition to, seeing their healthcare provider decreased from 6 to 1 (P = 0.06). Trepka et al. (2001)19 Northern WI, USA controlled before-and-after study n = 365 parents Nurse educators delivered parent-oriented presentations in community organizations, distributed information pamphlets and displayed posters. Topics covered included antibiotic resistance and use. knowledge and attitudes and behaviour From baseline to post-intervention the percentage of parents with high antibiotic resistance awareness significantly increased in the intervention group (change = 14.3%; 95% CI = 6.6%–22.0%), but not in the control group (change = 4.3%; 95% CI = −4.1% to 12.7%; P = 0.015). The proportion of parents who expected an antibiotic for their child and did not receive one declined in the intervention area (14% to 9%), while it increased in the control area (7% to 10%). The difference between the two area changes was −8.4% (95% CI = −13.9% to −2.8%; P = 0.003). The percentage of parents in the intervention area who brought their child to another physician because they did not receive an antibiotic decreased (5% to 2%), whereas it increased in the control area (2% to 4%). The difference between the two area changes was −4.5% (95% CI=−8.0% to −0.9%; P = 0.02). Wutzke et al. (2007)29 Australia non-controlled before-and-after study n = 6217 general public National mass media intervention for consumers delivered during winter months in 2001, 2002, 2003 and 2004. About the inappropriate use of antibiotics for upper respiratory tract infection. Strategies included newsletters and brochures, mass media activity using billboards, television, radio and magazines and small grants to promote local community education. knowledge and attitudes and behaviour There was a significant decline in those who believe taking antibiotics for cold and flu is appropriate, from 28.7% pre-programme in 2002 to 21.7% in 2004 (percentage point change = 7.0; 95% CI = 3.5–10.5). Significant decrease in self-reported use of antibiotics to treat cough, cold or flu, from 10.8% in 1999 to 7.4% in 2004 (percentage point change = 3.4; 95% CI = 1.3–5.5). Table 2. Study characteristics and results of the included studies Study Country Design Sample Nature of intervention(s) Outcome measures Significant results Azevedo et al. (2013)20 Braga, Portugal non-controlled before-and-after study n = 82 schoolchildren School-based presentation followed by discussion. knowledge and attitudes Knowledge of the correct use of antibiotics for bacterial diseases rather than viral diseases rose from 43% to 76% in the post-test (P < 0.01). Knowledge of the risk of bacterial resistance to antibiotics from their incorrect use rose from 48% to 74% in the post-test (P < 0.05). Cebotarenco and Bush (2008)16 Chisinau, Moldova controlled before-and-after study n = 3586 schoolchildren and n = 2716 parents Educational intervention about the use of antibiotics delivered by student volunteers trained as peer leaders delivered to their classmates and the classmates’ parents. beliefs and behaviour Students in both the intervention district and the post-intervention phase were 3.2 (95%  CI = 2.065–4.909) times more likely than other students to indicate they had not taken an antibiotic. Croft et al. (2007)11 WI, USA randomized controlled trial n = 300 parents Distribution of printed materials to parents by childcare staff; slide presentation delivered to staff. knowledge In parents who were college graduates, the median knowledge scores were 7.0 at intervention centres and 6.5 at control centres (P < 0.01). Curry et al. (2006)21 Auckland, New Zealand non-controlled before-and-after study n = 400 general public National campaign ‘Wise use of antibiotics’. Posters and leaflets delivered to the public attending pharmacies. knowledge and attitudes and behaviour Patients who had ever been to the doctor for a common cold significantly decreased (45% versus 62%; P = 0.0006). They were significantly less likely to feel positive about antibiotics in 2003 for the treatment of a cold (16% versus 33%; P = 0.00001). The perception that antibiotics were beneficial for cold/flu symptoms significantly reduced from 1998 to 2003 (P < 0.05) and the perceived benefit of antibiotics for tonsillitis increased from 83% to 91% in 2003 (P = 0.014). Significantly fewer people reported ever attending a doctor for a cold in 2003 versus 1998 (45% versus 62%; P = 0.0006) and the number of people who would usually see a doctor for a cold decreased from 24% to 15% (P = 0.026). Farrell et al. (2011)22 Glasgow, Gloucester and London, UK non-controlled before-and-after study n = 1736 children e-Bug web game. knowledge and attitudes No overall change in knowledge. Significant knowledge change in 3 out of 21 questions (P  ≤  0.02). Formoso et al. (2013)13 Emilia-Romagna, Italy non-randomized controlled trial n = 1200 general public Local mass media campaign (posters, brochures and advertisements on local media) delivered to general population to raise awareness of inappropriate use of antibiotics. knowledge After the intervention, consistency with campaign messages worsened (or did not improve) similarly in both intervention and control areas, the only exception being knowledge of the presumptive antiviral activity of antibiotics, worsening in the intervention area more than the control area. Gonzales et al. (2008)14 CO, USA non-randomized controlled trial n = 1503 general public Mass media intervention (outdoor and radio advertisements) delivered to general public about use of antibiotics. behaviour Linear regression analysis showed significant net differences in monthly paediatric office visit rates between mass media and comparison communities before and after the campaign (P = 0.01). Huang et al. (2007)12 MA, USA randomized controlled trial n = 3142 parents Community-based educational intervention occurred through three successive cold and flu seasons. Printed materials: mailed newsletters, posters, pamphlets and fact sheets in the waiting rooms of local paediatric providers, pharmacies and childcare centres. knowledge and attitudes The proportion of parents who answered ≥7 of 10 knowledge questions correctly increased significantly in both intervention (from 52% to 64%; P < 0.001) and control (from 54% to 61%; P < 0.01) communities. Substantial improvements in the percentage of correct answers were seen for items on middle ear fluid (41% in 2000; 50% in 2003; P < 0.001) and the general question of whether antibiotics were needed for colds and flu (66% in 2000; 77% in 2003; P < 0.001). Larson et al. (2009)23 Upper Manhattan, USA non-controlled before-and-after study n = 422 households Targeted Latino households. Educational materials (colouring book, pamphlets) based on knowledge, attitudes and practices regarding prevention and treatment of upper respiratory tract infections. Programme was delivered during home visits every 2 months. knowledge and attitudes and behaviour After the intervention, the mean composite knowledge scores at baseline and end of study were 5.19 (SD = 1.60) and 5.91 (SD = 1.71) (P < 0.001), respectively. With regard to reported practices, significantly more participants after the intervention reported using alcohol hand sanitizers (1.4% baseline and 66.8% post-intervention; P = 0.001). Significantly more also reported that one or more members of their household had received the influenza vaccination after the intervention (63.7% at baseline and 73.9% post-intervention; P = 0.001). Lecky et al. (2010)17 Gloucestershire and London, England; Nice and Bordeaux, France; Prague and Ostrava, Czech Republic controlled before-and-after study n = 2724 school students School-based educational intervention (e-Bug) regarding inappropriate antibiotic use delivered to classes of 9- to 11-year-old (junior) and 12- to 15-year-old (senior) students in state schools. Included 45 min lesson, handouts, worksheets, factsheets, interactive activity and a follow-up plenary question-and-answer session. knowledge Junior school: significant change in knowledge and significant change in retention 6 weeks post-intervention, across countries. Little significant difference in knowledge change between intervention and control, with the exception of the Czech Republic. Senior school: significant improvement in knowledge 6 weeks post-intervention in the Czech Republic. Significant improvement in knowledge and knowledge retention between control and intervention in England and the Czech Republic Madle et al. (2004)24 London, UK non-controlled before-and-after study n = 177 general public Open access to the National Electronic Library of Infection Antimicrobial Resistance website on the use of antibiotics and antibiotic resistance. The site comprises frequently asked questions and links to evidence-based resources. knowledge and attitudes Significant improvements in knowledge about the use of antibiotics and antibiotic resistance in 2 out of 3 statements: (i) ‘people cannot become resistant to antibiotics’ (P < 0.001, χ2 = 60.357, 95% CI of change = 27.47–44.53); (ii) ‘antibiotics do not cure most sore throats’ (P < 0.001, χ2 = 19.22, 95% CI of change = 8.62–27.38). Significant changes in the scores assigned by users for 3 out of 4 statements designed to test users’ attitudes to the information on the site (P ≤ 0.003). Expectations that antibiotics should be prescribed were significantly reduced after using the website (P < 0.001). Non-healthcare workers continued to have higher expectations of antibiotics being prescribed than healthcare workers (P = 0.0046 before and P = 0.0098 after using the website). Mainous et al. (2009)30 SC, USA Prospective cohort study n = 691 self-identified Latinos Mass media educational intervention (pamphlets, radio, newspapers) delivered to local Latino communities about use of antibiotics. behaviour Numbers in the intervention group reporting that they had bought antibiotics without a prescription increased following the intervention compared with baseline (χ2 test reported as significant, but P value not given). The regression analysis showed the strongest predictor of purchase of antibiotics without a prescription in the previous 12 months was past purchase of antibiotics without a prescription outside the USA (OR = 5.72; 95% CI = 3.12–10.48). The regression analysis also showed the strongest predictor of likelihood of importing antibiotics into the USA was past purchase of antibiotics without a prescription outside the USA (OR = 3.01; 95% CI = 1.95–4.65). Mazińska and Hryniewicz (2010)25 Poland non-controlled before-and-after study n = 1000 general public Mass media educational intervention (posters, leaflets, billboards, TV, cinemas, radio, press, magazines, thematic exhibitions, Internet) implemented across the country. knowledge, attitudes and behaviour Significant increase in the percentage of people who have limited the use of antibiotics (27% to 43%), have become more disciplined and cautious in their use (3% to 24%) and pay attention to the correct dosage (6% to 18%) (no P values given). McNulty et al. (2001)27 Gloucester, UK non-controlled before-and-after study n = 38 year 5 schoolchildren School-based intervention to children aged 9–10 years at a state school. Included two 90 min interactive workshops entitled ‘Antibiotics and your good bugs’. knowledge Before the workshops 23% and 26% knew antibiotics do not kill viruses, but kill good bacteria, compared with 47% and 69% afterwards (P = 0.03 and 0.0001). 45% before and 73% after the workshops correctly answered all the questions (P < 0.0001). Children thought antibiotics helped hay fever; this improved significantly after the workshop [correct answer 28% before, 77% after (P < 0.0001)]. Overall score for 7 questions in the ‘where are bugs found’ section was increased significantly from an average of 80.5% success to 93.2% success (P = 0.0002). The overall score improvement in the ‘How do bugs spread’ section was significant (P = 0.00001). McNulty et al. (2007)26 Gloucestershire, UK non-controlled before-and-after study n = 198 year 5 and 6 schoolchildren School-based intervention. ‘Bug Investigators’ pack about microorganisms, hygiene and antibiotics. The pack included 11 activity sheets, teachers’ guide, poster and website. knowledge Children’s knowledge improved in all topic areas and was significant in 6 out of the 7 topic areas (P < 0.005). Improved knowledge was most significant for what antibiotics do and how to use them [improvement = 27% (95% CI = 22.8–31.1) and 31% (95% CI = 23.4–37.7), respectively] and the value of our own good bugs (16% improvement). McNulty et al. (2010)18 England and Scotland, UK controlled before-and-after study n = 3718 general public Mass media campaign about antibiotic use involving posters displayed in magazines and newspapers. knowledge and attitudes and behaviour No positive effect of the campaigns. Pontes and Pontes (2005)15 Mid-Atlantic region, USA non-randomized controlled trials n = 105 university students University-based educational intervention (information booklet) to increase young adult consumers’ preference for physicians who do not unnecessarily prescribe antibiotics for simple acute upper respiratory tract infections. attitudes Exposure to the intervention significantly increased the mean preferred start date for antibiotics after initiation of an infection from 2.3 to 3.9 days (P < 0.01). Respondents’ preferences were significantly greater for the physician who indicated he would not prescribe antibiotics in the intervention (mean = 4.84) compared with the control (mean = 3.91) (P < 0.01). Stockwell et al. (2010)28 New York City, USA non-controlled before-and-after study n = 10 parents Health literacy intervention regarding upper respiratory tract infection delivered to parents on an ‘Early Head Start’ programme. Involved 3 × 1.5 h interactive sessions and provision of kit for care of a child with such an infection. knowledge and attitudes and behaviour The mean composite knowledge/attitude score increased from 4.1 (total possible: 10) to 6.6 (P < 0.05). The number of parents reporting that the last time their child was sick they sought antibiotics without a prescription instead of, or in addition to, seeing their healthcare provider decreased from 6 to 1 (P = 0.06). Trepka et al. (2001)19 Northern WI, USA controlled before-and-after study n = 365 parents Nurse educators delivered parent-oriented presentations in community organizations, distributed information pamphlets and displayed posters. Topics covered included antibiotic resistance and use. knowledge and attitudes and behaviour From baseline to post-intervention the percentage of parents with high antibiotic resistance awareness significantly increased in the intervention group (change = 14.3%; 95% CI = 6.6%–22.0%), but not in the control group (change = 4.3%; 95% CI = −4.1% to 12.7%; P = 0.015). The proportion of parents who expected an antibiotic for their child and did not receive one declined in the intervention area (14% to 9%), while it increased in the control area (7% to 10%). The difference between the two area changes was −8.4% (95% CI = −13.9% to −2.8%; P = 0.003). The percentage of parents in the intervention area who brought their child to another physician because they did not receive an antibiotic decreased (5% to 2%), whereas it increased in the control area (2% to 4%). The difference between the two area changes was −4.5% (95% CI=−8.0% to −0.9%; P = 0.02). Wutzke et al. (2007)29 Australia non-controlled before-and-after study n = 6217 general public National mass media intervention for consumers delivered during winter months in 2001, 2002, 2003 and 2004. About the inappropriate use of antibiotics for upper respiratory tract infection. Strategies included newsletters and brochures, mass media activity using billboards, television, radio and magazines and small grants to promote local community education. knowledge and attitudes and behaviour There was a significant decline in those who believe taking antibiotics for cold and flu is appropriate, from 28.7% pre-programme in 2002 to 21.7% in 2004 (percentage point change = 7.0; 95% CI = 3.5–10.5). Significant decrease in self-reported use of antibiotics to treat cough, cold or flu, from 10.8% in 1999 to 7.4% in 2004 (percentage point change = 3.4; 95% CI = 1.3–5.5). Study Country Design Sample Nature of intervention(s) Outcome measures Significant results Azevedo et al. (2013)20 Braga, Portugal non-controlled before-and-after study n = 82 schoolchildren School-based presentation followed by discussion. knowledge and attitudes Knowledge of the correct use of antibiotics for bacterial diseases rather than viral diseases rose from 43% to 76% in the post-test (P < 0.01). Knowledge of the risk of bacterial resistance to antibiotics from their incorrect use rose from 48% to 74% in the post-test (P < 0.05). Cebotarenco and Bush (2008)16 Chisinau, Moldova controlled before-and-after study n = 3586 schoolchildren and n = 2716 parents Educational intervention about the use of antibiotics delivered by student volunteers trained as peer leaders delivered to their classmates and the classmates’ parents. beliefs and behaviour Students in both the intervention district and the post-intervention phase were 3.2 (95%  CI = 2.065–4.909) times more likely than other students to indicate they had not taken an antibiotic. Croft et al. (2007)11 WI, USA randomized controlled trial n = 300 parents Distribution of printed materials to parents by childcare staff; slide presentation delivered to staff. knowledge In parents who were college graduates, the median knowledge scores were 7.0 at intervention centres and 6.5 at control centres (P < 0.01). Curry et al. (2006)21 Auckland, New Zealand non-controlled before-and-after study n = 400 general public National campaign ‘Wise use of antibiotics’. Posters and leaflets delivered to the public attending pharmacies. knowledge and attitudes and behaviour Patients who had ever been to the doctor for a common cold significantly decreased (45% versus 62%; P = 0.0006). They were significantly less likely to feel positive about antibiotics in 2003 for the treatment of a cold (16% versus 33%; P = 0.00001). The perception that antibiotics were beneficial for cold/flu symptoms significantly reduced from 1998 to 2003 (P < 0.05) and the perceived benefit of antibiotics for tonsillitis increased from 83% to 91% in 2003 (P = 0.014). Significantly fewer people reported ever attending a doctor for a cold in 2003 versus 1998 (45% versus 62%; P = 0.0006) and the number of people who would usually see a doctor for a cold decreased from 24% to 15% (P = 0.026). Farrell et al. (2011)22 Glasgow, Gloucester and London, UK non-controlled before-and-after study n = 1736 children e-Bug web game. knowledge and attitudes No overall change in knowledge. Significant knowledge change in 3 out of 21 questions (P  ≤  0.02). Formoso et al. (2013)13 Emilia-Romagna, Italy non-randomized controlled trial n = 1200 general public Local mass media campaign (posters, brochures and advertisements on local media) delivered to general population to raise awareness of inappropriate use of antibiotics. knowledge After the intervention, consistency with campaign messages worsened (or did not improve) similarly in both intervention and control areas, the only exception being knowledge of the presumptive antiviral activity of antibiotics, worsening in the intervention area more than the control area. Gonzales et al. (2008)14 CO, USA non-randomized controlled trial n = 1503 general public Mass media intervention (outdoor and radio advertisements) delivered to general public about use of antibiotics. behaviour Linear regression analysis showed significant net differences in monthly paediatric office visit rates between mass media and comparison communities before and after the campaign (P = 0.01). Huang et al. (2007)12 MA, USA randomized controlled trial n = 3142 parents Community-based educational intervention occurred through three successive cold and flu seasons. Printed materials: mailed newsletters, posters, pamphlets and fact sheets in the waiting rooms of local paediatric providers, pharmacies and childcare centres. knowledge and attitudes The proportion of parents who answered ≥7 of 10 knowledge questions correctly increased significantly in both intervention (from 52% to 64%; P < 0.001) and control (from 54% to 61%; P < 0.01) communities. Substantial improvements in the percentage of correct answers were seen for items on middle ear fluid (41% in 2000; 50% in 2003; P < 0.001) and the general question of whether antibiotics were needed for colds and flu (66% in 2000; 77% in 2003; P < 0.001). Larson et al. (2009)23 Upper Manhattan, USA non-controlled before-and-after study n = 422 households Targeted Latino households. Educational materials (colouring book, pamphlets) based on knowledge, attitudes and practices regarding prevention and treatment of upper respiratory tract infections. Programme was delivered during home visits every 2 months. knowledge and attitudes and behaviour After the intervention, the mean composite knowledge scores at baseline and end of study were 5.19 (SD = 1.60) and 5.91 (SD = 1.71) (P < 0.001), respectively. With regard to reported practices, significantly more participants after the intervention reported using alcohol hand sanitizers (1.4% baseline and 66.8% post-intervention; P = 0.001). Significantly more also reported that one or more members of their household had received the influenza vaccination after the intervention (63.7% at baseline and 73.9% post-intervention; P = 0.001). Lecky et al. (2010)17 Gloucestershire and London, England; Nice and Bordeaux, France; Prague and Ostrava, Czech Republic controlled before-and-after study n = 2724 school students School-based educational intervention (e-Bug) regarding inappropriate antibiotic use delivered to classes of 9- to 11-year-old (junior) and 12- to 15-year-old (senior) students in state schools. Included 45 min lesson, handouts, worksheets, factsheets, interactive activity and a follow-up plenary question-and-answer session. knowledge Junior school: significant change in knowledge and significant change in retention 6 weeks post-intervention, across countries. Little significant difference in knowledge change between intervention and control, with the exception of the Czech Republic. Senior school: significant improvement in knowledge 6 weeks post-intervention in the Czech Republic. Significant improvement in knowledge and knowledge retention between control and intervention in England and the Czech Republic Madle et al. (2004)24 London, UK non-controlled before-and-after study n = 177 general public Open access to the National Electronic Library of Infection Antimicrobial Resistance website on the use of antibiotics and antibiotic resistance. The site comprises frequently asked questions and links to evidence-based resources. knowledge and attitudes Significant improvements in knowledge about the use of antibiotics and antibiotic resistance in 2 out of 3 statements: (i) ‘people cannot become resistant to antibiotics’ (P < 0.001, χ2 = 60.357, 95% CI of change = 27.47–44.53); (ii) ‘antibiotics do not cure most sore throats’ (P < 0.001, χ2 = 19.22, 95% CI of change = 8.62–27.38). Significant changes in the scores assigned by users for 3 out of 4 statements designed to test users’ attitudes to the information on the site (P ≤ 0.003). Expectations that antibiotics should be prescribed were significantly reduced after using the website (P < 0.001). Non-healthcare workers continued to have higher expectations of antibiotics being prescribed than healthcare workers (P = 0.0046 before and P = 0.0098 after using the website). Mainous et al. (2009)30 SC, USA Prospective cohort study n = 691 self-identified Latinos Mass media educational intervention (pamphlets, radio, newspapers) delivered to local Latino communities about use of antibiotics. behaviour Numbers in the intervention group reporting that they had bought antibiotics without a prescription increased following the intervention compared with baseline (χ2 test reported as significant, but P value not given). The regression analysis showed the strongest predictor of purchase of antibiotics without a prescription in the previous 12 months was past purchase of antibiotics without a prescription outside the USA (OR = 5.72; 95% CI = 3.12–10.48). The regression analysis also showed the strongest predictor of likelihood of importing antibiotics into the USA was past purchase of antibiotics without a prescription outside the USA (OR = 3.01; 95% CI = 1.95–4.65). Mazińska and Hryniewicz (2010)25 Poland non-controlled before-and-after study n = 1000 general public Mass media educational intervention (posters, leaflets, billboards, TV, cinemas, radio, press, magazines, thematic exhibitions, Internet) implemented across the country. knowledge, attitudes and behaviour Significant increase in the percentage of people who have limited the use of antibiotics (27% to 43%), have become more disciplined and cautious in their use (3% to 24%) and pay attention to the correct dosage (6% to 18%) (no P values given). McNulty et al. (2001)27 Gloucester, UK non-controlled before-and-after study n = 38 year 5 schoolchildren School-based intervention to children aged 9–10 years at a state school. Included two 90 min interactive workshops entitled ‘Antibiotics and your good bugs’. knowledge Before the workshops 23% and 26% knew antibiotics do not kill viruses, but kill good bacteria, compared with 47% and 69% afterwards (P = 0.03 and 0.0001). 45% before and 73% after the workshops correctly answered all the questions (P < 0.0001). Children thought antibiotics helped hay fever; this improved significantly after the workshop [correct answer 28% before, 77% after (P < 0.0001)]. Overall score for 7 questions in the ‘where are bugs found’ section was increased significantly from an average of 80.5% success to 93.2% success (P = 0.0002). The overall score improvement in the ‘How do bugs spread’ section was significant (P = 0.00001). McNulty et al. (2007)26 Gloucestershire, UK non-controlled before-and-after study n = 198 year 5 and 6 schoolchildren School-based intervention. ‘Bug Investigators’ pack about microorganisms, hygiene and antibiotics. The pack included 11 activity sheets, teachers’ guide, poster and website. knowledge Children’s knowledge improved in all topic areas and was significant in 6 out of the 7 topic areas (P < 0.005). Improved knowledge was most significant for what antibiotics do and how to use them [improvement = 27% (95% CI = 22.8–31.1) and 31% (95% CI = 23.4–37.7), respectively] and the value of our own good bugs (16% improvement). McNulty et al. (2010)18 England and Scotland, UK controlled before-and-after study n = 3718 general public Mass media campaign about antibiotic use involving posters displayed in magazines and newspapers. knowledge and attitudes and behaviour No positive effect of the campaigns. Pontes and Pontes (2005)15 Mid-Atlantic region, USA non-randomized controlled trials n = 105 university students University-based educational intervention (information booklet) to increase young adult consumers’ preference for physicians who do not unnecessarily prescribe antibiotics for simple acute upper respiratory tract infections. attitudes Exposure to the intervention significantly increased the mean preferred start date for antibiotics after initiation of an infection from 2.3 to 3.9 days (P < 0.01). Respondents’ preferences were significantly greater for the physician who indicated he would not prescribe antibiotics in the intervention (mean = 4.84) compared with the control (mean = 3.91) (P < 0.01). Stockwell et al. (2010)28 New York City, USA non-controlled before-and-after study n = 10 parents Health literacy intervention regarding upper respiratory tract infection delivered to parents on an ‘Early Head Start’ programme. Involved 3 × 1.5 h interactive sessions and provision of kit for care of a child with such an infection. knowledge and attitudes and behaviour The mean composite knowledge/attitude score increased from 4.1 (total possible: 10) to 6.6 (P < 0.05). The number of parents reporting that the last time their child was sick they sought antibiotics without a prescription instead of, or in addition to, seeing their healthcare provider decreased from 6 to 1 (P = 0.06). Trepka et al. (2001)19 Northern WI, USA controlled before-and-after study n = 365 parents Nurse educators delivered parent-oriented presentations in community organizations, distributed information pamphlets and displayed posters. Topics covered included antibiotic resistance and use. knowledge and attitudes and behaviour From baseline to post-intervention the percentage of parents with high antibiotic resistance awareness significantly increased in the intervention group (change = 14.3%; 95% CI = 6.6%–22.0%), but not in the control group (change = 4.3%; 95% CI = −4.1% to 12.7%; P = 0.015). The proportion of parents who expected an antibiotic for their child and did not receive one declined in the intervention area (14% to 9%), while it increased in the control area (7% to 10%). The difference between the two area changes was −8.4% (95% CI = −13.9% to −2.8%; P = 0.003). The percentage of parents in the intervention area who brought their child to another physician because they did not receive an antibiotic decreased (5% to 2%), whereas it increased in the control area (2% to 4%). The difference between the two area changes was −4.5% (95% CI=−8.0% to −0.9%; P = 0.02). Wutzke et al. (2007)29 Australia non-controlled before-and-after study n = 6217 general public National mass media intervention for consumers delivered during winter months in 2001, 2002, 2003 and 2004. About the inappropriate use of antibiotics for upper respiratory tract infection. Strategies included newsletters and brochures, mass media activity using billboards, television, radio and magazines and small grants to promote local community education. knowledge and attitudes and behaviour There was a significant decline in those who believe taking antibiotics for cold and flu is appropriate, from 28.7% pre-programme in 2002 to 21.7% in 2004 (percentage point change = 7.0; 95% CI = 3.5–10.5). Significant decrease in self-reported use of antibiotics to treat cough, cold or flu, from 10.8% in 1999 to 7.4% in 2004 (percentage point change = 3.4; 95% CI = 1.3–5.5). Apart from one study conducted in Moldova,16 all studies were conducted in high-income countries, with the majority conducted in the USA (n = 8)11,12,14,15,19,23,28,30 or in the UK (n = 5).18,22,24,26,27 The remaining studies were conducted in Italy,13 Portugal,20 Poland,25 New Zealand21 and Australia,29 and one study was a multisite study conducted in the UK, the Czech Republic and France17 (Table 2). As shown in Table 2, the most common types of interventions were mass media interventions.13,14,18,24,25,29,30 Apart from Madle et al.,24 who used a website only, all mass media interventions were multimodal and used a variety of outlets, such as billboards, radio, television, newspapers, magazines, websites and printed resources such as posters, brochures, pamphlets, leaflets, stickers or badges distributed to community sites or healthcare settings. Six studies used printed material interventions, either alone12,15,21,23 or in combination with educational presentations,11,19 while Stockwell et al.28 delivered taught modules to Latino community parents. Other interventions were school based and included a student peer-taught programme,16 an e-Bug web game,22 interactive workshops,27 school lessons delivered using the Bug Investigators pack26 and presentation followed by discussion,20 while the intervention delivered in the study by Lecky et al.17 involved the delivery of a lesson, printed materials, interactive activities and a question-and-answer session. Comparators were similar across the 10 controlled studies.11–19,30 With the exception of Lecky et al.,17 who compared the educational intervention with the usual school curriculum, control groups were not exposed to the interventions. Table 2 shows that the most common outcome measure was change in knowledge, attitudes or beliefs, measured alone (n = 10)11–13,15,17,20,22,24,26,27 or in combination with change in the public’s antimicrobial stewardship behaviour (n = 8).16,18,19,21,23,25,28,29 Two studies measured the public’s antimicrobial stewardship behaviour outcomes exclusively (n = 2).14,30 Quality of studies Amongst the included studies, nine met the EPOC study design criteria.9 As shown in Table 3, the overall risk of bias of the included studies was generally high. Apart from one study,11 all had at least one item assessed as high risk, with the number of high-risk items ranging from one15 to five.18 High risk of bias was most commonly associated with generation of sequence allocation, risk of contamination and other risks. Table 3. Risk of bias of studies meeting the EPOC criteria Study Allocation: sequence generation Allocation: concealment Baseline primary outcome Baseline secondary outcome Baseline characteristics Incomplete primary outcome data Incomplete secondary outcome data Blinding primary outcome data Blinding secondary outcome Contamination Selective reporting primary outcome Selective reporting secondary outcome Other risks of bias Cebotarenco and Bush (2008)16 H H U U L U U U U L L L L Croft et al. (2007)11 L U U NA U L NA U NA U L NA L Formoso et al. (2013)13 H U U U L U U U U U L L H Gonzales et al. (2008)14 H U U L U U U U L H L L H Huang et al. (2007)12 L U H NA L L NA U NA H H NA H Lecky et al. (2010)17 H U U NA U H NA U NA H L NA H McNulty et al. (2010)18 H H H NA U U NA U NA H L NA H Pontes and Pontes (2005)15 U U U U U U U U U U L L H Trepka et al. (2001)19 H U L NA U L NA U NA H L NA H Study Allocation: sequence generation Allocation: concealment Baseline primary outcome Baseline secondary outcome Baseline characteristics Incomplete primary outcome data Incomplete secondary outcome data Blinding primary outcome data Blinding secondary outcome Contamination Selective reporting primary outcome Selective reporting secondary outcome Other risks of bias Cebotarenco and Bush (2008)16 H H U U L U U U U L L L L Croft et al. (2007)11 L U U NA U L NA U NA U L NA L Formoso et al. (2013)13 H U U U L U U U U U L L H Gonzales et al. (2008)14 H U U L U U U U L H L L H Huang et al. (2007)12 L U H NA L L NA U NA H H NA H Lecky et al. (2010)17 H U U NA U H NA U NA H L NA H McNulty et al. (2010)18 H H H NA U U NA U NA H L NA H Pontes and Pontes (2005)15 U U U U U U U U U U L L H Trepka et al. (2001)19 H U L NA U L NA U NA H L NA H H, high risk of bias; L, low risk of bias; U, unclear risk of bias; NA, not applicable. Table 3. Risk of bias of studies meeting the EPOC criteria Study Allocation: sequence generation Allocation: concealment Baseline primary outcome Baseline secondary outcome Baseline characteristics Incomplete primary outcome data Incomplete secondary outcome data Blinding primary outcome data Blinding secondary outcome Contamination Selective reporting primary outcome Selective reporting secondary outcome Other risks of bias Cebotarenco and Bush (2008)16 H H U U L U U U U L L L L Croft et al. (2007)11 L U U NA U L NA U NA U L NA L Formoso et al. (2013)13 H U U U L U U U U U L L H Gonzales et al. (2008)14 H U U L U U U U L H L L H Huang et al. (2007)12 L U H NA L L NA U NA H H NA H Lecky et al. (2010)17 H U U NA U H NA U NA H L NA H McNulty et al. (2010)18 H H H NA U U NA U NA H L NA H Pontes and Pontes (2005)15 U U U U U U U U U U L L H Trepka et al. (2001)19 H U L NA U L NA U NA H L NA H Study Allocation: sequence generation Allocation: concealment Baseline primary outcome Baseline secondary outcome Baseline characteristics Incomplete primary outcome data Incomplete secondary outcome data Blinding primary outcome data Blinding secondary outcome Contamination Selective reporting primary outcome Selective reporting secondary outcome Other risks of bias Cebotarenco and Bush (2008)16 H H U U L U U U U L L L L Croft et al. (2007)11 L U U NA U L NA U NA U L NA L Formoso et al. (2013)13 H U U U L U U U U U L L H Gonzales et al. (2008)14 H U U L U U U U L H L L H Huang et al. (2007)12 L U H NA L L NA U NA H H NA H Lecky et al. (2010)17 H U U NA U H NA U NA H L NA H McNulty et al. (2010)18 H H H NA U U NA U NA H L NA H Pontes and Pontes (2005)15 U U U U U U U U U U L L H Trepka et al. (2001)19 H U L NA U L NA U NA H L NA H H, high risk of bias; L, low risk of bias; U, unclear risk of bias; NA, not applicable. For all studies that met the EPOC criteria, insufficient information was provided for at least two of the items that were regarded as unclear risk. None of the studies provided information regarding blinding and, in all studies reporting secondary outcomes, the risk of bias for incomplete secondary outcome data could not be assessed.13–16 The number of low-risk items ranged from one17,18 to five,16 with the ‘selective outcome reporting’ item being most commonly assessed as low risk. None of the studies had a low risk score for any of the following items: ‘allocation concealment’, ‘addressing incomplete secondary outcome data’ and ‘blinding primary outcome data’. Risk of bias was not assessed for the non-controlled before-and-after studies20–29 and a prospective cohort study.30 These study designs did not meet the EPOC criteria;9 therefore, it was assumed that the risk of bias of these studies was high. Relative effectiveness of interventions Reviewed interventions were grouped into five categories of relative measure of effectiveness. As shown in Table 4, six studies demonstrated a clear desired behaviour change following the intervention, while two studies resulted in some desired behaviour change. The desired effect on the antecedent of behaviour was reported in nine papers. One study showed no effect, while two studies demonstrated an increase in drivers of AMR following the intervention. Table 4. Patterning of the effectiveness across the type of target population Study Interventions indicative of clear positive behaviour change in the desired direction Interventions indicative of some positive behaviour change in the desired direction Interventions indicative of positive effect on the antecedent of behaviour in the desired direction Interventions indicative of no effect on behaviour or antecedents of behaviour Interventions indicative of negative effect on behaviour or antecedents of behaviour in a non-desired direction Azevedo et al. (2013)20 schoolchildren Cebotarenco and Bush (2008)16 schoolchildren Croft et al. (2007)11 parents; childcare facilities Curry et al. (2006)21 general public Farrell et al. (2011)22 schoolchildren Formoso et al. (2013)13 general public Gonzales et al. (2008)14 general public Huang et al. (2007)12 parents Larson et al. (2009)23 parents Lecky et al. (2010)17 schoolchildren Madle et al. (2004)24 general public Mainous et al. (2009)30 Latino community Mazińska and Hryniewicz (2010)25 general public McNulty et al. (2001)27 schoolchildren McNulty et al. (2007)26 schoolchildren McNulty et al. (2010)18 general public Pontes and Pontes (2005)15 young adults Stockwell et al. (2010)28 Latino community parents Trepka et al. (2001)19 parents Wutzke et al. (2007)29 general public Study Interventions indicative of clear positive behaviour change in the desired direction Interventions indicative of some positive behaviour change in the desired direction Interventions indicative of positive effect on the antecedent of behaviour in the desired direction Interventions indicative of no effect on behaviour or antecedents of behaviour Interventions indicative of negative effect on behaviour or antecedents of behaviour in a non-desired direction Azevedo et al. (2013)20 schoolchildren Cebotarenco and Bush (2008)16 schoolchildren Croft et al. (2007)11 parents; childcare facilities Curry et al. (2006)21 general public Farrell et al. (2011)22 schoolchildren Formoso et al. (2013)13 general public Gonzales et al. (2008)14 general public Huang et al. (2007)12 parents Larson et al. (2009)23 parents Lecky et al. (2010)17 schoolchildren Madle et al. (2004)24 general public Mainous et al. (2009)30 Latino community Mazińska and Hryniewicz (2010)25 general public McNulty et al. (2001)27 schoolchildren McNulty et al. (2007)26 schoolchildren McNulty et al. (2010)18 general public Pontes and Pontes (2005)15 young adults Stockwell et al. (2010)28 Latino community parents Trepka et al. (2001)19 parents Wutzke et al. (2007)29 general public Table 4. Patterning of the effectiveness across the type of target population Study Interventions indicative of clear positive behaviour change in the desired direction Interventions indicative of some positive behaviour change in the desired direction Interventions indicative of positive effect on the antecedent of behaviour in the desired direction Interventions indicative of no effect on behaviour or antecedents of behaviour Interventions indicative of negative effect on behaviour or antecedents of behaviour in a non-desired direction Azevedo et al. (2013)20 schoolchildren Cebotarenco and Bush (2008)16 schoolchildren Croft et al. (2007)11 parents; childcare facilities Curry et al. (2006)21 general public Farrell et al. (2011)22 schoolchildren Formoso et al. (2013)13 general public Gonzales et al. (2008)14 general public Huang et al. (2007)12 parents Larson et al. (2009)23 parents Lecky et al. (2010)17 schoolchildren Madle et al. (2004)24 general public Mainous et al. (2009)30 Latino community Mazińska and Hryniewicz (2010)25 general public McNulty et al. (2001)27 schoolchildren McNulty et al. (2007)26 schoolchildren McNulty et al. (2010)18 general public Pontes and Pontes (2005)15 young adults Stockwell et al. (2010)28 Latino community parents Trepka et al. (2001)19 parents Wutzke et al. (2007)29 general public Study Interventions indicative of clear positive behaviour change in the desired direction Interventions indicative of some positive behaviour change in the desired direction Interventions indicative of positive effect on the antecedent of behaviour in the desired direction Interventions indicative of no effect on behaviour or antecedents of behaviour Interventions indicative of negative effect on behaviour or antecedents of behaviour in a non-desired direction Azevedo et al. (2013)20 schoolchildren Cebotarenco and Bush (2008)16 schoolchildren Croft et al. (2007)11 parents; childcare facilities Curry et al. (2006)21 general public Farrell et al. (2011)22 schoolchildren Formoso et al. (2013)13 general public Gonzales et al. (2008)14 general public Huang et al. (2007)12 parents Larson et al. (2009)23 parents Lecky et al. (2010)17 schoolchildren Madle et al. (2004)24 general public Mainous et al. (2009)30 Latino community Mazińska and Hryniewicz (2010)25 general public McNulty et al. (2001)27 schoolchildren McNulty et al. (2007)26 schoolchildren McNulty et al. (2010)18 general public Pontes and Pontes (2005)15 young adults Stockwell et al. (2010)28 Latino community parents Trepka et al. (2001)19 parents Wutzke et al. (2007)29 general public Effectiveness of interventions delivered to populations through the life cycle In 17 of the studies, the intervention had a significant effect on the outcome of interest amongst the populations through the life cycle. These included schoolchildren, university students, parents and the general public. Schoolchildren All six school-based educational interventions that targeted schoolchildren aged between 9 and 15 years16,17,20,22,26,27 found a significant increase in knowledge following the educational intervention (Table 2). However, Farrell et al.22 found a significant knowledge change in only 3 out of 21 questions (P ≤ 0.02) and no overall change in knowledge. The three questions for which significant improvement was reported related to the valuableness of ‘good microbes’, the presence of microbes despite inability to see them and hand-washing being an effective method of removing microbes from the hands. Only one study16 measured behavioural outcome in addition to beliefs and found that children in the intervention group were 3.2 times more likely than other students to report that they had not taken an antibiotic for a cold or flu (P < 0.001). None of the studies measured long-term outcomes of school-based interventions. Post-intervention outcomes were measured immediately following the intervention,22 or between 1 and 8 weeks after the intervention. In addition, one study17 found that the increase in knowledge was maintained at 6 weeks post-intervention in junior school students, but not senior school students. University students University students were targeted in one experimental study that aimed to investigate whether an educational intervention (information booklet) resulted in an increase in young adult consumers’ preference for physicians who do not unnecessarily prescribe antibiotics for simple acute upper respiratory tract infections.15 This study demonstrated that exposure to the intervention significantly increased the mean preferred start date for antibiotics after the onset of an infection from 2.3 to 3.9 days (P < 0.1) and preference for a physician who would not prescribe antibiotics at day 3 of an infection (P < 0.1). However, this was still well before the recommended time of 10–14 days. Parents The effect of educational interventions delivered to parents on a change in their AMR knowledge, attitudes or beliefs alone11,12 or in combination with parents’ antimicrobial stewardship behaviour outcomes16,19,23,28 was measured in six studies. The majority of these interventions were directed to parents or caregivers of children under the age of 6.11,12,19,28 In the remaining two studies, intervention was delivered to households with at least one child >5 years old23 and parents of children aged 12–13 years.16 As shown in Table 2, all studies showed a significant increase in knowledge following the interventions. In addition, four of the reviewed interventions also had a positive effect on parents’ antimicrobial stewardship behaviour. Cebotarenco and Bush16 found that parents in the intervention group were 5.2 times more likely than other parents to indicate they had not taken an antibiotic for colds or flu (P < 0.001). In Trepka et al.19 the proportion of parents who expected an antibiotic for their child and did not receive one declined in the intervention area from 14% to 9%, whereas it increased from 7% to 10% in the control area (P = 0.003), and the percentage of parents reporting that they brought their child to another physician because they did not receive an antibiotic decreased from 5% to 2% in the intervention area and increased from 2% to 4% in the control area (P = 0.02). Larson et al.23 found that the percentage of participants reporting using alcohol hand sanitizers increased from 1.4% to 66.8% following the intervention (P = 0.001) and the percentage of those reporting that at least one member of their household had been vaccinated against influenza increased from 63.7% to 73.9% (P = 0.001). Stockwell et al.,28 on the other hand, demonstrated that the number of parents reporting that they sought antibiotics without a prescription when their child was sick decreased from six to one (P = 0.06). General public The general public were the population of interest in eight of the included studies.13,14,18,21,24,25,29,30 Apart from Curry et al.,21 who used printed materials in the form of posters and leaflets, all studies were mass media campaigns, including four studies that measured the effects of the national campaign intervention.18,21,25,29 Five studies demonstrated a significant effect on the general public’s knowledge and attitudes14,21,24,25,29 (Table 2). With respect to antimicrobial stewardship behaviour amongst the public, four studies reported a significant effect following the intervention.14,21,25,29 Gonzales et al.14 found that visits to paediatricians declined in the intervention group for all conditions, but mostly for acute respiratory infections (P = 0.01). Similarly, Curry et al.21 demonstrated a significant decrease in the number of respondents who reported consulting a doctor for the common cold (P = 0.026). The results of Wutzke et al.29 showed that significantly fewer participants reported using antibiotics for cough, cold or flu following the intervention (7.4%) in comparison with baseline data (10.8%; percentage point change = 3.4; 95% CI = 1.3–5.5). Mazińska and Hryniewicz,25 on the other hand, demonstrated a significant increase in the percentage of respondents who had limited the use of antibiotics (from 27% at baseline to 43% post-intervention), had become more disciplined and cautious in their use (from 3% to 24%) and who had paid attention to the correct dosage (from 6% to 18%; no P values given). The remaining three studies did not show a significant positive effect on outcomes of interest.13,18,30 In Mainous et al.30 an intervention designed to decrease self-medication with antibiotics surprisingly resulted in a significantly greater percentage of the intervention Latino community group using antibiotics without a prescription in comparison with the control group (OR = 1.81; 95% CI = 1.02–3.22). McNulty et al.,18 on the other hand, found no positive effect on participants’ knowledge or antimicrobial stewardship behaviour following a national campaign and there was a significant increase in the percentage of respondents from the intervention area who reported retaining leftover antibiotics (P < 0.001). Formoso et al.13 reported that knowledge consistency with the national campaign messages either worsened (P < 0.05) or did not improve in both the intervention and control groups after the intervention. Discussion Main findings of this study This systematic review provides an in-depth examination of the effectiveness of interventions that target the public to increase their knowledge and understanding of AMR and engagement with antimicrobial stewardship behaviours. We have also identified patterns between target populations and relative intervention effectiveness. The findings present a complex picture reflecting the heterogeneity of the studies. Our analysis has shown that interventions targeting schoolchildren and parents have notable potential. All interventions that targeted schoolchildren or parents showed a significant effect on the outcome of interest. However, effective school-based interventions tended to only have the effect of increasing knowledge. In addition, these studies measured only short-term outcomes. In contrast, interventions targeting parents demonstrated changes in behaviour in addition to knowledge, with the follow-up period ranging from 2 weeks28 to 3 years.12 With regard to the interventions targeting the general public, the picture is less clear. Although the majority (n = 5) of these studies demonstrated effectiveness of interventions in improving the public’s AMR knowledge or their antimicrobial stewardship behaviour, three studies did not, with two showing a decrease in AMR knowledge13 and in antimicrobial stewardship behaviour.30 These findings highlight the need to examine differences in content between these interventions targeting the general public. Patterning of effectiveness across the types of target population also suggests that different target populations should receive different interventions with different primary outcomes. Nevertheless, targeting children alone is unlikely to make a major contribution to AMR because attitudes and the public’s antimicrobial stewardship behaviours may be passed down through generations. Thus, using the power of familial social influence and parental duty, in which children’s AMR education within school is reinforced and boosted by parental interventions in the home, might be a more appropriate approach for the achievement of desired cultural change. This indicates the potential of a multimodal intervention or programmatic approach to AMR-related interventions. An ideal approach would be to address the entire population simultaneously, but segment it to target sub-populations. Through such segmentation, or stratification of the general public, diverse tailored interventions addressing different sub-populations would be a strategic way to begin the process of cultural change required to reduce the drivers of AMR. The nature of the increase in knowledge that is needed can also be specified by drawing on other evidence syntheses that have shown that the public’s AMR knowledge and understanding of their contribution to AMR are generally poor.3 Therefore, in addition to changing the public’s understanding of appropriate antimicrobial use, interventions should also target the public’s understanding of AMR to enable the public to understand their central role in tackling AMR and the risks for the intervention recipient, their loved ones and the wider population. Findings in relation to other research In their recent paper, Wells and Piddock31 argued that amongst other actions, an urgent review of educational campaigns is required in order to fulfil UK and European AMR action plans. Our review addresses this need. Furthermore, to our knowledge this is the first systematic review that provides such an in-depth examination of the effectiveness of AMR-related interventions that target the public specifically. Previous literature focused on the level of the public’s AMR knowledge and beliefs,3,4 communication interventions or interventions that target both the public and healthcare professionals.6 The latter found that multi-component interventions improve the public’s knowledge of appropriate antimicrobial use, specifically in relation to antibiotics, and that interventions including both physician and public education appear to be effective in reducing antibiotic use.6 Similarly, Cross et al.32 reported that multi-modal communication interventions targeting both the public and clinicians can reduce antibiotic prescribing in high-income countries. Although our review focused on the general public population specifically, the potential of multi-faceted interventions was also highlighted in our work. Another previous systematic review, by King et al.7 reviewed the evidence of effectiveness and cost-effectiveness of interventions changing the public’s risk-related behaviours in relation to antimicrobial use. The review showed that direct contact education interventions were consistently more effective than mass media interventions.7 This appears to explain our findings on the varying effectiveness of interventions targeting the general public, as the majority of these studies used mass media interventions. There is also a body of evidence on large-scale antibiotic campaigns that were not eligible for inclusion in our review as the participants were members of both the public and healthcare professions. A literature review showed that there have been numerous multifaceted antibiotic awareness campaigns launched in high-income countries;33 however, there was substantial heterogeneity in outcomes, including knowledge and awareness, use of antibiotics and AMR, and the interventions themselves often lack a robust grounding in behavioural and social science theory. The majority of campaigns included in the review targeted both the general public and healthcare professionals simultaneously and they appeared to result in a reduction in antibiotic use.33 It therefore appears that targeting different populations at the same time might result in desired outcomes as the healthcare professional’s prescribing decisions might also be influenced by the patient, while the patient’s behaviour might be affected by the prescriber’s advice. One such campaign, conducted in the UK in 2014, simultaneously targeted members of the public and healthcare professionals who pledged as Antibiotic Guardians and showed an increase in AMR knowledge and commitment to pledge behaviour in both surveyed sub-populations.34 Another antibiotic awareness campaign, conducted in Hong Kong, targeted the general public, patients and healthcare professionals in a segmented fashion and resulted in a significant improvement (P ≤ 0.002) in respondents’ knowledge on prudent use of antibiotics following the campaign.35 Yet another successful large-scale antibiotic awareness campaign segmented to target the general public and healthcare professionals was conducted in France.36 Evaluation of the effectiveness of this campaign in reducing the number of antibiotic prescriptions showed a −26.5% (95% CI = −33.5% to −19.6%) decrease in the total number of antibiotic prescriptions following the campaign, with the greatest decrease in prescriptions issued for children (−35.8%; 95% CI = −48.3% to −23.2%) and young adults in the 21–25 years age group (24.1% decrease; CI not provided).36 These findings further emphasize the potential of a programmatic approach to AMR-related interventions segmented to different target sub-populations, as suggested above in the section Main findings of this study. Strengths and limitations We have conducted a rigorous search and systematic review accompanied by a narrative synthesis. Although similar work concerning the effectiveness of interventions aiming to improve antibiotic use has been conducted previously,32 our work focused on interventions targeting the general public population exclusively and did not include outcomes related to healthcare professionals’ AMR awareness or antimicrobial stewardship, such as antibiotic prescribing. Our analysis provides a sense of what is normative within this field, what has been attempted before and what could be repeated. It also provides a unique and valuable contribution to the available literature. However, the study also has limitations. First, because the UK Antimicrobial Resistance Strategy and Action Plan was launched by the Department of Health in 2000,37 followed by the publication of the WHO Global Strategy for Containment of Antimicrobial Resistance in 2001,1 we limited our search to publications from 2000 onwards. This could result in omission of important, older papers. Second, studies from low- and middle-income countries were underrepresented in our review. Thus, relevance and applicability of our findings to different geographical areas or resource contexts is limited. Third, the risk of bias was assessed only for studies that met the EPOC study design criteria. However, a suitable, validated tool for assessing the risk of bias of non-controlled before-and-after studies could not be identified. Furthermore, using different instruments could result in ambiguities in relation to the quality of stronger designs. The overall quality of the evidence was rather low. Major problems were associated with randomization in experimental designs and the evaluation of mass media and other population-level interventions. As these kinds of interventions aim for maximum population reach, it is difficult to attain adequate controls or indeed randomize at this population level. Therefore, good-quality study designs are systematically less likely to be identified within this kind of population-level intervention literature. Notwithstanding this, there was considerable heterogeneity in outcomes. There are no standardized ways of measuring the public’s AMR-related knowledge or associated stewardship behaviours. Furthermore, change in knowledge, awareness or beliefs, which were the most common outcome measures across the included studies, might not necessarily lead to desired behaviour change. As a result, it is particularly challenging to build cumulative knowledge regarding the effectiveness of interventions to increase the public’s engagement with antimicrobial stewardship. Another limitation is that given the problems with the quality of primary research, our measure of relative effectiveness should be treated with caution as this was based on our relative measure and is not equivalent to a strong evidence base within typical evidence-based guidance. Finally, we did not conduct an analysis of the cost-effectiveness of reviewed interventions; however, for the majority of studies included in our review, cost-effectiveness data were not reported. Recommendations for future research Although our work demonstrated the potential of intervention that targets particular sub-populations of the general public, taking into account the low quality of reviewed evidence, lack of cost-effectiveness evaluation and underrepresentation of studies from low- and middle-income countries, these findings must be treated with caution. There is a need for well-designed, randomized experimental studies focusing on behavioural outcomes of the interventions. Furthermore, measures of AMR knowledge and stewardship behaviours need to be standardized and there is a need for improvement of the reporting standards to ensure detailed and transparent reporting of intervention components. Finally, considering the underrepresentation of studies from low- and middle-income countries, there is a need for the development and evaluation of similar interventions within such settings. Conclusions Although some evidence of the effectiveness of interventions that target the general public in engaging with the problem of AMR exists, the public’s understanding of AMR and their role in combating this problem remains poor. Thus, there is a need for a cultural change and effective engagement of the public in addition to other key stakeholders. This need could be addressed through development of well-designed AMR-related interventions robustly grounded within behavioural and social science theory. Our work provided an in-depth examination of the effectiveness of AMR-related interventions targeting the members of the public specifically. We suggest that future policy makers should consider multimodal segmented population-level intervention that tailors its core messages to children, parents and the wider general public alike, particularly in high-income geographical areas. Future interventions should convey messages that elicit the public’s motivation to make their own efforts to address AMR as a growing problem for all and a problem for the present as much as for the future. Funding This work was supported by Health Protection Scotland. Glasgow Caledonian University was responsible for initiating, managing and sponsoring this review (Research, Innovation and Enterprise 15-127). Transparency declarations None to declare. Supplementary data Tables S1 to S3 are available as Supplementary data at JAC Online. References 1 WHO . WHO Global Strategy for Containment of Antimicrobial Resistance. 2001 . http://apps.who.int/iris/bitstream/10665/66860/1/WHO_CDS_CSR_DRS_2001.2.pdf. 2 O’Neill J. Tackling Drug-Resistant Infections Globally: Final Report and Recommendations. https://amr-review.org/sites/default/files/160525_Final%20paper_with%20cover.pdf. 3 Gualano MR , Gili R , Scaioli G et al. General population’s knowledge and attitudes about antibiotics: a systematic review and meta‐analysis . Pharmacoepidemiol Drug Saf 2015 ; 24 : 2 – 10 . Google Scholar CrossRef Search ADS PubMed 4 McCullough A , Parekh S , Rathbone J et al. A systematic review of the public’s knowledge and beliefs about antibiotic resistance . J Antimicrob Chemother 2015 ; 71 : 27 – 33 . Google Scholar CrossRef Search ADS PubMed 5 Pinder R , Sallis A , Berry D et al. Behaviour Change and Antibiotic Prescribing in Healthcare Settings. Literature Review and Behavioural Analysis. PHE. https://www.gov.uk/government/publications/antibiotic-prescribing-and-behaviour-change-in-healthcare-settings. 6 Haynes C , McLeod C. A Review of Reviews of Educational Interventions Designed to Change the Public’s Knowledge and Behaviour in Relation to Antimicrobial Use and Antimicrobial Resistance that Target Healthcare Professionals and Patients. https://www.nice.org.uk/guidance/ng63/documents/antimicrobial-resistance-changing-riskrelated-behaviours-in-the-general-population-evidence-review-32. 7 King S , Exley J , Taylor J et al. Antimicrobial stewardship: the effectiveness of educational interventions to change risk-related behaviours in the general population: a systematic review . Rand Health Q 2016 ; 5 : 2 . Google Scholar PubMed 8 Moher D , Liberati A , Tetzlaff J et al. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement . PLoS Med 2009 ; 6 : e1000097. Google Scholar CrossRef Search ADS PubMed 9 Cochrane Effective Practice and Organisation of Care (EPOC) . EPOC Resources for Review Authors. http://epoc.cochrane.org/resources/epoc-resources-review-authors. 10 Ryan R , Cochrane Consumers and Communication Review Group . Cochrane Consumers and Communication Review Group: Data Synthesis and Analysis. http://cccrg.cochrane.org/sites/cccrg.cochrane.org/files/public/uploads/meta-analysis_revised_december_1st_1_2016.pdf. 11 Croft DR , Knobloch MJ , Chyou P et al. Impact of a child care educational intervention on parent knowledge about appropriate antibiotic use . WMJ 2007 ; 106 : 78 – 84 . Google Scholar PubMed 12 Huang SS , Rifas-Shiman SL , Kleinman K et al. Parental knowledge about antibiotic use: results of a cluster-randomized, multicommunity intervention . Pediatrics 2007 ; 119 : 698 – 706 . Google Scholar CrossRef Search ADS PubMed 13 Formoso G , Paltrinieri B , Marata AM et al. Feasibility and effectiveness of a low cost campaign on antibiotic prescribing in Italy: community level, controlled, non-randomised trial . BMJ 2013 ; 347 : f5391. Google Scholar CrossRef Search ADS PubMed 14 Gonzales R , Corbett KK , Wong S et al. “Get smart Colorado”: impact of a mass media campaign to improve community antibiotic use . Med Care 2008 ; 46 : 597 – 605 . Google Scholar CrossRef Search ADS PubMed 15 Pontes MC , Pontes NM. Debiasing effects of education about appropriate antibiotic use on consumers’ preferences for physicians . Health Care Manage Rev 2005 ; 30 : 9 – 16 . Google Scholar CrossRef Search ADS PubMed 16 Cebotarenco N , Bush PJ. Reducing antibiotics for colds and flu: a student-taught program . Health Educ Res 2008 ; 23 : 146 – 57 . Google Scholar CrossRef Search ADS PubMed 17 Lecky DM , McNulty CA , Touboul P et al. Evaluation of e-Bug, an educational pack, teaching about prudent antibiotic use and hygiene, in the Czech Republic, France and England . J Antimicrob Chemother 2010 ; 65 : 2674 – 84 . Google Scholar CrossRef Search ADS PubMed 18 McNulty CA , Nichols T , Boyle PJ et al. The English antibiotic awareness campaigns: did they change the public’s knowledge of and attitudes to antibiotic use? J Antimicrob Chemother 2010 ; 65 : 1526 – 33 . Google Scholar CrossRef Search ADS PubMed 19 Trepka MJ , Belongia EA , Chyou PH et al. The effect of a community intervention trial on parental knowledge and awareness of antibiotic resistance and appropriate antibiotic use in children . Pediatrics 2001 ; 107 : e6. Google Scholar CrossRef Search ADS PubMed 20 Azevedo MM , Pinheiro C , Yaphe J et al. Assessing the impact of a school intervention to promote students’ knowledge and practices on correct antibiotic use . Int J Environ Res Public Health 2013 ; 10 : 2920 – 31 . Google Scholar CrossRef Search ADS PubMed 21 Curry M , Sung L , Arroll B et al. Public views and use of antibiotics for the common cold before and after an education campaign in New Zealand . N Z Med J 2006 ; 119 : U1957 . Google Scholar PubMed 22 Farrell D , Kostkova P , Weinberg J et al. Computer games to teach hygiene: an evaluation of the e-Bug junior game . J Antimicrob Chemother 2011 ; 66 Suppl 5: v39 – 44 . Google Scholar CrossRef Search ADS PubMed 23 Larson EL , Ferng YH , McLoughlin JW et al. Effect of intensive education on knowledge, attitudes, and practices regarding upper respiratory infections among urban Latinos . Nurs Res 2009 ; 58 : 150 – 7 . Google Scholar CrossRef Search ADS PubMed 24 Madle G , Kostkova P , Mani-Saada J et al. Changing public attitudes to antibiotic prescribing: can the internet help? J Innov Health Inform 2004 ; 12 : 19 – 26 . Google Scholar CrossRef Search ADS 25 Mazińska B , Hryniewicz W. Kampania edukacyjna Europejski Dzień Wiedzy o Antybiotykach–czy wpłynęła na zmianę postaw społeczeństwa w Polsce . Pol Merkur Lekarski 2010 ; 29 : 296 – 303 . Google Scholar PubMed 26 McNulty CA , Bowen J , Gelb D et al. “The Bug Investigators”: assessment of a school teaching resource to improve hygiene and prudent use of antibiotics . Health Educ 2007 ; 107 : 10 – 26 . Google Scholar CrossRef Search ADS 27 McNulty CA , Swan AV , Boland D. Schools’ antimicrobial resistance: National Advice to the Public campaign – a pilot study . Health Educ 2001 ; 101 : 235 – 42 . Google Scholar CrossRef Search ADS 28 Stockwell MS , Catallozzi M , Meyer D et al. Improving care of upper respiratory infections among Latino Early Head Start parents . J Immigrant Minority Health 2010 ; 12 : 925 – 31 . Google Scholar CrossRef Search ADS 29 Wutzke SE , Artist MA , Kehoe LA et al. Evaluation of a national programme to reduce inappropriate use of antibiotics for upper respiratory tract infections: effects on consumer awareness, beliefs, attitudes and behaviour in Australia . Health Promot Int 2007 ; 22 : 53 – 64 . Google Scholar CrossRef Search ADS PubMed 30 Mainous AG , Diaz VA , Carnemolla M. A community intervention to decrease antibiotics used for self-medication among Latino adults . Ann Fam Med 2009 ; 7 : 520. Google Scholar CrossRef Search ADS PubMed 31 Wells V , Piddock LJV. Addressing antimicrobial resistance in the UK and Europe . Lancet Infect Dis 2017 ; 17 : 1230 – 1 . Google Scholar CrossRef Search ADS PubMed 32 Cross ELA , Tolfree R , Kipping R. Systematic review of public-targeted communication interventions to improve antibiotic use . J Antimicrob Chemother 2016 ; 72 : 975 – 87 . 33 Huttner B , Goossens H , Verheij T et al. Characteristics and outcomes of public campaigns aimed at improving the use of antibiotics in outpatients in high-income countries . Lancet Infect Dis 2010 ; 10 : 17 – 31 . Google Scholar CrossRef Search ADS PubMed 34 Chaintarli K , Ingle SM , Bhattacharya A et al. Impact of a United Kingdom-wide campaign to tackle antimicrobial resistance on self-reported knowledge and behaviour change . BMC Public Health 2016 ; 16 : 393 . Google Scholar CrossRef Search ADS PubMed 35 Ho ML , Cowling BJ , Seto WH et al. Determinants of an effective antibiotic campaign: lessons from Hong Kong . J Glob Antimicrob Resist 2014 ; 2 : 334 – 7 . Google Scholar CrossRef Search ADS PubMed 36 Sabuncu E , David J , Bernède-Bauduin C et al. Significant reduction of antibiotic use in the community after a nationwide campaign in France, 2002–2007 . PLoS Med 2009 ; 6 : e1000084. Google Scholar CrossRef Search ADS PubMed 37 Department of Health . UK Antimicrobial Resistance Strategy and Action Plan. http://antibiotic-action.com/wp-content/uploads/2011/07/DH-UK-antimicrobial-resistance-strategy-and-action-plan.pdf. © The Author(s) 2018. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please email: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Antimicrobial Chemotherapy Oxford University Press

Effectiveness of interventions to improve the public’s antimicrobial resistance awareness and behaviours associated with prudent use of antimicrobials: a systematic review

Loading next page...
 
/lp/ou_press/effectiveness-of-interventions-to-improve-the-public-s-antimicrobial-3PzG30rGrN
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please email: journals.permissions@oup.com.
ISSN
0305-7453
eISSN
1460-2091
D.O.I.
10.1093/jac/dky076
Publisher site
See Article on Publisher Site

Abstract

Abstract Background A global antimicrobial resistance (AMR) awareness intervention targeting the general public has been prioritized. Objectives To evaluate the effectiveness of interventions that aim to change AMR awareness and subsequent stewardship behaviours amongst the public. Methods Five databases were searched between 2000 and 2016 for interventions to change the public’s AMR awareness and/or antimicrobial stewardship behaviours. Study designs meeting the Cochrane Effective Practice and Organization of Care (EPOC) criteria, non-controlled before-and-after studies and prospective cohort studies were considered eligible. Participants recruited from healthcare settings and studies measuring stewardship behaviours of healthcare professionals were excluded. Quality of studies was assessed using EPOC risk of bias criteria. Data were extracted and synthesized narratively. Registration: PROSPERO international prospective register of systematic reviews (PROSPERO 2016: CRD42016050343). Results Twenty studies were included in the review with nine meeting the EPOC criteria. The overall risk of bias was high. Nineteen studies were conducted in high-income countries. Mass media interventions were most common (n = 7), followed by school-based (n = 6) and printed material interventions (n = 6). Seventeen studies demonstrated a significant effect on changing knowledge, attitudes or the public’s antimicrobial stewardship behaviours. Analysis showed that interventions targeting schoolchildren and parents have notable potential, but for the general public the picture is less clear. Conclusions Our work provides an in-depth examination of the effectiveness of AMR interventions for the public. However, the studies were heterogeneous and the quality of evidence was poor. Well-designed, experimental studies on behavioural outcomes of such interventions are required. Introduction The rise of antimicrobial resistance (AMR) is a rapidly developing global threat that greatly affects our ability to deliver effective healthcare and results in a financial burden.1 AMR refers to the ability of a microorganism to adapt and grow despite the presence of antimicrobials. AMR threatens effective treatment of an ever-increasing range of infections.1 Therefore, increasing AMR is becoming a major public health concern. Although AMR is a naturally occurring phenomenon, inappropriate use of antimicrobials is the main driver of AMR.1 The demands for the use of antimicrobials are increasing worldwide and because of suboptimal management of these demands, huge quantities of antimicrobials are being misused.2 Together these highlight the need for effective strategies encouraging prudent use of antimicrobials. The O’Neill report emphasizes the need for AMR awareness interventions directed towards the public and development of a uniform, globally consistent set of AMR messages that could be then tailored to meet the specific demands of local settings.2 However, the report does not provide recommendations on the components of such interventions.2 Previous evidence syntheses show that the overall levels of knowledge and understanding of AMR amongst the public are generally low and members of the public often lack an understanding of their potential contribution to the development of AMR.3–5 Although high-level evidence demonstrating the effectiveness of interventions in increasing public understanding of AMR exists,5–7 these evaluations are methodologically diverse. It is therefore challenging to identify what interventions work and why and for whom they work, in order to inform future interventions. Thus, the aim of this systematic review is to provide the best-quality evidence regarding the effectiveness of AMR interventions that change public awareness and their subsequent antimicrobial stewardship behaviours. Although antimicrobial stewardship is most commonly thought of in medical settings, the word ‘stewardship’ means ‘taking care of’, particularly on behalf of others. Furthermore, a One Health perspective requires the collaborative effort of all stakeholders to take the responsibility for the prudent use of antimicrobials. Therefore, within this work, we use the term ‘antimicrobial stewardship’ to explore the public’s behaviours related to their prudent use of antimicrobials (such as, but not limited to, adhering to prescribers’ directions, not taking or demanding antimicrobial prescription for colds and flu and safe disposal of leftover antimicrobials). We believe an understanding of the public’s antimicrobial stewardship is central to engaging them with their part in reducing the drivers of AMR on behalf of future generations, other key stakeholders, such as prescribers, and the global community. Methods This review was prospectively registered on the PROSPERO international prospective register of systematic reviews (http://www.crd.york.ac.uk/PROSPERO/display_record.php?ID=CRD42016050343) and is reported in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) statement.8 Search strategy CINAHL, Cochrane Library, EMBASE, MEDLINE and PsycINFO databases were searched for articles published between 2000 and 2016 using keywords associated with the following four concept areas: (i) population—general public; (ii) intervention—interventions designed to increase antimicrobial awareness and/or to improve antimicrobial stewardship behaviour amongst the general public; (iii) context—AMR or the public’s antimicrobial stewardship; and (iv) outcomes—all relevant short-, medium- or long-term outcomes related to the public’s AMR and/or antimicrobial stewardship behaviours (knowledge/awareness, learning, public behavioural and cognition outcomes). The search strategy incorporated controlled vocabulary thesaurus terms and free text words contained in titles and abstracts. No restrictions were applied to language or publication status. The search strategy was amended according to the functionality of each of the databases. An example of the search strategy applied to MEDLINE is presented in Table S1 (available as Supplementary data at JAC Online). In addition to the database search, the reference lists of included papers and previous systematic reviews were searched manually and citation searches were conducted through Web of Science in order to identify additional records. Study selection Cochrane’s Effective Practice and Organization of Care (EPOC) recommendations9 were used to initially select studies for inclusion in the review. Although EPOC guidelines suggest the inclusion of randomized controlled trials, non-randomized controlled trials, controlled before-and-after studies, interrupted time series studies and repeated measures studies exclusively,9 because of the limited number of eligible studies meeting the EPOC criteria, non-controlled before-and-after studies and prospective cohort studies were also deemed eligible for the review. Interventions targeting the general public population and designed to increase public antimicrobial awareness and/or to improve the public’s antimicrobial stewardship were deemed eligible for the review. Furthermore, time-bound geographical controls or no exposure comparators and all relevant short-, medium- or long-term outcomes related to AMR and/or the public’s antimicrobial stewardship behaviours were included, whereas those related to antimicrobial prescribing were excluded as this was regarded as the behaviour of healthcare professionals rather than of the members of the general public. Eligibility criteria applied in this study are presented in detail in Table 1. Table 1. Review’s eligibility criteria Inclusion criteria Exclusion criteria Design randomized controlled trials, non-randomized trials, interrupted time series studies, controlled before-and-after studies, non-controlled before-and-after studies and cohort studies – Population members of the public participants recruited from healthcare settings Intervention intervention designed to increase public antimicrobial awareness and/or to improve antimicrobial stewardship (through mass media, social marketing or printed media campaigns) – Comparator time bound, geographical controls or no exposure – Context non-healthcare settings; AMR or the public’s antimicrobial stewardship – Outcomes all relevant short-, medium- or long-term outcomes related to AMR and/or antimicrobial stewardship behaviours (knowledge/awareness, learning, public behavioural and cognition outcomes) antimicrobial prescribing Publication date published after January 2000 published before January 2000 Inclusion criteria Exclusion criteria Design randomized controlled trials, non-randomized trials, interrupted time series studies, controlled before-and-after studies, non-controlled before-and-after studies and cohort studies – Population members of the public participants recruited from healthcare settings Intervention intervention designed to increase public antimicrobial awareness and/or to improve antimicrobial stewardship (through mass media, social marketing or printed media campaigns) – Comparator time bound, geographical controls or no exposure – Context non-healthcare settings; AMR or the public’s antimicrobial stewardship – Outcomes all relevant short-, medium- or long-term outcomes related to AMR and/or antimicrobial stewardship behaviours (knowledge/awareness, learning, public behavioural and cognition outcomes) antimicrobial prescribing Publication date published after January 2000 published before January 2000 Table 1. Review’s eligibility criteria Inclusion criteria Exclusion criteria Design randomized controlled trials, non-randomized trials, interrupted time series studies, controlled before-and-after studies, non-controlled before-and-after studies and cohort studies – Population members of the public participants recruited from healthcare settings Intervention intervention designed to increase public antimicrobial awareness and/or to improve antimicrobial stewardship (through mass media, social marketing or printed media campaigns) – Comparator time bound, geographical controls or no exposure – Context non-healthcare settings; AMR or the public’s antimicrobial stewardship – Outcomes all relevant short-, medium- or long-term outcomes related to AMR and/or antimicrobial stewardship behaviours (knowledge/awareness, learning, public behavioural and cognition outcomes) antimicrobial prescribing Publication date published after January 2000 published before January 2000 Inclusion criteria Exclusion criteria Design randomized controlled trials, non-randomized trials, interrupted time series studies, controlled before-and-after studies, non-controlled before-and-after studies and cohort studies – Population members of the public participants recruited from healthcare settings Intervention intervention designed to increase public antimicrobial awareness and/or to improve antimicrobial stewardship (through mass media, social marketing or printed media campaigns) – Comparator time bound, geographical controls or no exposure – Context non-healthcare settings; AMR or the public’s antimicrobial stewardship – Outcomes all relevant short-, medium- or long-term outcomes related to AMR and/or antimicrobial stewardship behaviours (knowledge/awareness, learning, public behavioural and cognition outcomes) antimicrobial prescribing Publication date published after January 2000 published before January 2000 Titles and abstracts of identified records were screened against the eligibility criteria (Table 1) by one of three reviewers (M. Y., L. G. or F. S.) with a 30% subset of excluded studies independently checked by another reviewer (M. Y., L. G. or F. S.). The level of agreement on this subset was 99%. Disagreements were resolved with the involvement of another, experienced reviewer (L. P.). Full texts of papers that appeared to meet the inclusion criteria or those with insufficient information within the title and abstract were screened by two out of three independent reviewers (L. G., M. Y. and F. S.), with a fourth reviewer (L. P.) checking all decisions and resolving any discrepancies. Whenever possible, foreign-language papers were translated by members of the team who have a command of foreign languages or were translated using Google Translate. Data extraction and quality assessment Two out of three reviewers (L. G., M. Y. and F. S.) independently extracted data from eligible studies using a standardized tool, designed for the purpose of the study (Table S2). For studies that met the EPOC study design criteria (randomized controlled trials, non-randomized controlled trials, controlled before-and-after studies and interrupted time series studies),9 risk of bias was assessed across domains by one reviewer (M. Y., L. G. or F. S.) and checked by a second reviewer (M. Y., L. G. or F. S.) using standard EPOC risk of bias criteria.9 Disagreements were resolved through consensus or, if necessary, in consultation with a fourth reviewer (L. P.). Risk of bias assessments were not conducted for non-controlled before-and-after studies as it was assumed that the risk of bias of these studies was high. No studies were excluded based on quality assessment. Data analysis Given the heterogeneity of the study designs, populations, interventions and outcome measures, it was not possible to pool the results in a meta-analysis. Therefore, we applied an alternative, systematic approach to assessing complex interventions and carried out a narrative synthesis of evidence following the Cochrane Consumers and Communication Review Group’s guidelines.10 Individual study characteristics and findings were summarized and similarities, differences and patterns identified. Studies were grouped into those meeting or not meeting the EPOC criteria and categorized according to the target population. To identify discernible patterns of effectiveness, identified studies were mapped across five categories of intervention effectiveness. These categories were based upon both the strength of the evidence and the position of the primary outcome within the causal chain linking antecedents of behaviour to actual behaviour change. For example, knowledge is understood to be a necessary, but insufficient, predictor of behaviour as people can develop good awareness and understanding of AMR yet still fail to implement the public’s AMR stewardship. The five categories of a relative measure of effectiveness included: (i) interventions indicative of clear positive behaviour change in the desired direction; (ii) interventions indicative of some positive behaviour change in the desired direction; (iii) interventions indicative of a positive effect on the antecedent of behaviour, such as knowledge or awareness, in the desired direction; (iv) interventions indicative of no effect on behaviour or antecedents of behaviour; and (v) interventions indicative of a negative effect on behaviour or antecedents of behaviour in a non-desired direction. Results An electronic search resulted in the total of 17 312 records. An additional 31 records were identified through reference and citation searching of the included papers. A total number of 60 studies that did not meet eligibility criteria were excluded during the full-text reviewing stage. Articles were excluded for not meeting study design criteria, study participants being recruited from healthcare settings, context other than AMR, study outcomes not related to the public’s AMR awareness or antimicrobial stewardship, a full-text record being unavailable and other reasons, such as the record being a study protocol, a conference abstract of an already identified study, a short report of an already identified study, inability to translate a non-English paper or the majority of participants recruited for the study being healthcare workers. A detailed list of excluded papers is presented in Table S3. Following screening, 20 studies that matched the eligibility criteria were included in the review. A detailed process of study selection is presented in Figure 1. Figure 1. View largeDownload slide Study selection flowchart. RCT, randomized controlled trial; NCT, non-randomized controlled trial. Figure 1. View largeDownload slide Study selection flowchart. RCT, randomized controlled trial; NCT, non-randomized controlled trial. Study characteristics As shown in Table 2, study designs of the 20 reviewed studies included randomized controlled trials (n = 2),11,12 non-randomized controlled trials (n = 3),13–15 controlled before-and-after studies (n = 4),16–19 non-controlled before-and-after studies (n = 10)20–29 and a prospective cohort study (n = 1).30 Table 2. Study characteristics and results of the included studies Study Country Design Sample Nature of intervention(s) Outcome measures Significant results Azevedo et al. (2013)20 Braga, Portugal non-controlled before-and-after study n = 82 schoolchildren School-based presentation followed by discussion. knowledge and attitudes Knowledge of the correct use of antibiotics for bacterial diseases rather than viral diseases rose from 43% to 76% in the post-test (P < 0.01). Knowledge of the risk of bacterial resistance to antibiotics from their incorrect use rose from 48% to 74% in the post-test (P < 0.05). Cebotarenco and Bush (2008)16 Chisinau, Moldova controlled before-and-after study n = 3586 schoolchildren and n = 2716 parents Educational intervention about the use of antibiotics delivered by student volunteers trained as peer leaders delivered to their classmates and the classmates’ parents. beliefs and behaviour Students in both the intervention district and the post-intervention phase were 3.2 (95%  CI = 2.065–4.909) times more likely than other students to indicate they had not taken an antibiotic. Croft et al. (2007)11 WI, USA randomized controlled trial n = 300 parents Distribution of printed materials to parents by childcare staff; slide presentation delivered to staff. knowledge In parents who were college graduates, the median knowledge scores were 7.0 at intervention centres and 6.5 at control centres (P < 0.01). Curry et al. (2006)21 Auckland, New Zealand non-controlled before-and-after study n = 400 general public National campaign ‘Wise use of antibiotics’. Posters and leaflets delivered to the public attending pharmacies. knowledge and attitudes and behaviour Patients who had ever been to the doctor for a common cold significantly decreased (45% versus 62%; P = 0.0006). They were significantly less likely to feel positive about antibiotics in 2003 for the treatment of a cold (16% versus 33%; P = 0.00001). The perception that antibiotics were beneficial for cold/flu symptoms significantly reduced from 1998 to 2003 (P < 0.05) and the perceived benefit of antibiotics for tonsillitis increased from 83% to 91% in 2003 (P = 0.014). Significantly fewer people reported ever attending a doctor for a cold in 2003 versus 1998 (45% versus 62%; P = 0.0006) and the number of people who would usually see a doctor for a cold decreased from 24% to 15% (P = 0.026). Farrell et al. (2011)22 Glasgow, Gloucester and London, UK non-controlled before-and-after study n = 1736 children e-Bug web game. knowledge and attitudes No overall change in knowledge. Significant knowledge change in 3 out of 21 questions (P  ≤  0.02). Formoso et al. (2013)13 Emilia-Romagna, Italy non-randomized controlled trial n = 1200 general public Local mass media campaign (posters, brochures and advertisements on local media) delivered to general population to raise awareness of inappropriate use of antibiotics. knowledge After the intervention, consistency with campaign messages worsened (or did not improve) similarly in both intervention and control areas, the only exception being knowledge of the presumptive antiviral activity of antibiotics, worsening in the intervention area more than the control area. Gonzales et al. (2008)14 CO, USA non-randomized controlled trial n = 1503 general public Mass media intervention (outdoor and radio advertisements) delivered to general public about use of antibiotics. behaviour Linear regression analysis showed significant net differences in monthly paediatric office visit rates between mass media and comparison communities before and after the campaign (P = 0.01). Huang et al. (2007)12 MA, USA randomized controlled trial n = 3142 parents Community-based educational intervention occurred through three successive cold and flu seasons. Printed materials: mailed newsletters, posters, pamphlets and fact sheets in the waiting rooms of local paediatric providers, pharmacies and childcare centres. knowledge and attitudes The proportion of parents who answered ≥7 of 10 knowledge questions correctly increased significantly in both intervention (from 52% to 64%; P < 0.001) and control (from 54% to 61%; P < 0.01) communities. Substantial improvements in the percentage of correct answers were seen for items on middle ear fluid (41% in 2000; 50% in 2003; P < 0.001) and the general question of whether antibiotics were needed for colds and flu (66% in 2000; 77% in 2003; P < 0.001). Larson et al. (2009)23 Upper Manhattan, USA non-controlled before-and-after study n = 422 households Targeted Latino households. Educational materials (colouring book, pamphlets) based on knowledge, attitudes and practices regarding prevention and treatment of upper respiratory tract infections. Programme was delivered during home visits every 2 months. knowledge and attitudes and behaviour After the intervention, the mean composite knowledge scores at baseline and end of study were 5.19 (SD = 1.60) and 5.91 (SD = 1.71) (P < 0.001), respectively. With regard to reported practices, significantly more participants after the intervention reported using alcohol hand sanitizers (1.4% baseline and 66.8% post-intervention; P = 0.001). Significantly more also reported that one or more members of their household had received the influenza vaccination after the intervention (63.7% at baseline and 73.9% post-intervention; P = 0.001). Lecky et al. (2010)17 Gloucestershire and London, England; Nice and Bordeaux, France; Prague and Ostrava, Czech Republic controlled before-and-after study n = 2724 school students School-based educational intervention (e-Bug) regarding inappropriate antibiotic use delivered to classes of 9- to 11-year-old (junior) and 12- to 15-year-old (senior) students in state schools. Included 45 min lesson, handouts, worksheets, factsheets, interactive activity and a follow-up plenary question-and-answer session. knowledge Junior school: significant change in knowledge and significant change in retention 6 weeks post-intervention, across countries. Little significant difference in knowledge change between intervention and control, with the exception of the Czech Republic. Senior school: significant improvement in knowledge 6 weeks post-intervention in the Czech Republic. Significant improvement in knowledge and knowledge retention between control and intervention in England and the Czech Republic Madle et al. (2004)24 London, UK non-controlled before-and-after study n = 177 general public Open access to the National Electronic Library of Infection Antimicrobial Resistance website on the use of antibiotics and antibiotic resistance. The site comprises frequently asked questions and links to evidence-based resources. knowledge and attitudes Significant improvements in knowledge about the use of antibiotics and antibiotic resistance in 2 out of 3 statements: (i) ‘people cannot become resistant to antibiotics’ (P < 0.001, χ2 = 60.357, 95% CI of change = 27.47–44.53); (ii) ‘antibiotics do not cure most sore throats’ (P < 0.001, χ2 = 19.22, 95% CI of change = 8.62–27.38). Significant changes in the scores assigned by users for 3 out of 4 statements designed to test users’ attitudes to the information on the site (P ≤ 0.003). Expectations that antibiotics should be prescribed were significantly reduced after using the website (P < 0.001). Non-healthcare workers continued to have higher expectations of antibiotics being prescribed than healthcare workers (P = 0.0046 before and P = 0.0098 after using the website). Mainous et al. (2009)30 SC, USA Prospective cohort study n = 691 self-identified Latinos Mass media educational intervention (pamphlets, radio, newspapers) delivered to local Latino communities about use of antibiotics. behaviour Numbers in the intervention group reporting that they had bought antibiotics without a prescription increased following the intervention compared with baseline (χ2 test reported as significant, but P value not given). The regression analysis showed the strongest predictor of purchase of antibiotics without a prescription in the previous 12 months was past purchase of antibiotics without a prescription outside the USA (OR = 5.72; 95% CI = 3.12–10.48). The regression analysis also showed the strongest predictor of likelihood of importing antibiotics into the USA was past purchase of antibiotics without a prescription outside the USA (OR = 3.01; 95% CI = 1.95–4.65). Mazińska and Hryniewicz (2010)25 Poland non-controlled before-and-after study n = 1000 general public Mass media educational intervention (posters, leaflets, billboards, TV, cinemas, radio, press, magazines, thematic exhibitions, Internet) implemented across the country. knowledge, attitudes and behaviour Significant increase in the percentage of people who have limited the use of antibiotics (27% to 43%), have become more disciplined and cautious in their use (3% to 24%) and pay attention to the correct dosage (6% to 18%) (no P values given). McNulty et al. (2001)27 Gloucester, UK non-controlled before-and-after study n = 38 year 5 schoolchildren School-based intervention to children aged 9–10 years at a state school. Included two 90 min interactive workshops entitled ‘Antibiotics and your good bugs’. knowledge Before the workshops 23% and 26% knew antibiotics do not kill viruses, but kill good bacteria, compared with 47% and 69% afterwards (P = 0.03 and 0.0001). 45% before and 73% after the workshops correctly answered all the questions (P < 0.0001). Children thought antibiotics helped hay fever; this improved significantly after the workshop [correct answer 28% before, 77% after (P < 0.0001)]. Overall score for 7 questions in the ‘where are bugs found’ section was increased significantly from an average of 80.5% success to 93.2% success (P = 0.0002). The overall score improvement in the ‘How do bugs spread’ section was significant (P = 0.00001). McNulty et al. (2007)26 Gloucestershire, UK non-controlled before-and-after study n = 198 year 5 and 6 schoolchildren School-based intervention. ‘Bug Investigators’ pack about microorganisms, hygiene and antibiotics. The pack included 11 activity sheets, teachers’ guide, poster and website. knowledge Children’s knowledge improved in all topic areas and was significant in 6 out of the 7 topic areas (P < 0.005). Improved knowledge was most significant for what antibiotics do and how to use them [improvement = 27% (95% CI = 22.8–31.1) and 31% (95% CI = 23.4–37.7), respectively] and the value of our own good bugs (16% improvement). McNulty et al. (2010)18 England and Scotland, UK controlled before-and-after study n = 3718 general public Mass media campaign about antibiotic use involving posters displayed in magazines and newspapers. knowledge and attitudes and behaviour No positive effect of the campaigns. Pontes and Pontes (2005)15 Mid-Atlantic region, USA non-randomized controlled trials n = 105 university students University-based educational intervention (information booklet) to increase young adult consumers’ preference for physicians who do not unnecessarily prescribe antibiotics for simple acute upper respiratory tract infections. attitudes Exposure to the intervention significantly increased the mean preferred start date for antibiotics after initiation of an infection from 2.3 to 3.9 days (P < 0.01). Respondents’ preferences were significantly greater for the physician who indicated he would not prescribe antibiotics in the intervention (mean = 4.84) compared with the control (mean = 3.91) (P < 0.01). Stockwell et al. (2010)28 New York City, USA non-controlled before-and-after study n = 10 parents Health literacy intervention regarding upper respiratory tract infection delivered to parents on an ‘Early Head Start’ programme. Involved 3 × 1.5 h interactive sessions and provision of kit for care of a child with such an infection. knowledge and attitudes and behaviour The mean composite knowledge/attitude score increased from 4.1 (total possible: 10) to 6.6 (P < 0.05). The number of parents reporting that the last time their child was sick they sought antibiotics without a prescription instead of, or in addition to, seeing their healthcare provider decreased from 6 to 1 (P = 0.06). Trepka et al. (2001)19 Northern WI, USA controlled before-and-after study n = 365 parents Nurse educators delivered parent-oriented presentations in community organizations, distributed information pamphlets and displayed posters. Topics covered included antibiotic resistance and use. knowledge and attitudes and behaviour From baseline to post-intervention the percentage of parents with high antibiotic resistance awareness significantly increased in the intervention group (change = 14.3%; 95% CI = 6.6%–22.0%), but not in the control group (change = 4.3%; 95% CI = −4.1% to 12.7%; P = 0.015). The proportion of parents who expected an antibiotic for their child and did not receive one declined in the intervention area (14% to 9%), while it increased in the control area (7% to 10%). The difference between the two area changes was −8.4% (95% CI = −13.9% to −2.8%; P = 0.003). The percentage of parents in the intervention area who brought their child to another physician because they did not receive an antibiotic decreased (5% to 2%), whereas it increased in the control area (2% to 4%). The difference between the two area changes was −4.5% (95% CI=−8.0% to −0.9%; P = 0.02). Wutzke et al. (2007)29 Australia non-controlled before-and-after study n = 6217 general public National mass media intervention for consumers delivered during winter months in 2001, 2002, 2003 and 2004. About the inappropriate use of antibiotics for upper respiratory tract infection. Strategies included newsletters and brochures, mass media activity using billboards, television, radio and magazines and small grants to promote local community education. knowledge and attitudes and behaviour There was a significant decline in those who believe taking antibiotics for cold and flu is appropriate, from 28.7% pre-programme in 2002 to 21.7% in 2004 (percentage point change = 7.0; 95% CI = 3.5–10.5). Significant decrease in self-reported use of antibiotics to treat cough, cold or flu, from 10.8% in 1999 to 7.4% in 2004 (percentage point change = 3.4; 95% CI = 1.3–5.5). Study Country Design Sample Nature of intervention(s) Outcome measures Significant results Azevedo et al. (2013)20 Braga, Portugal non-controlled before-and-after study n = 82 schoolchildren School-based presentation followed by discussion. knowledge and attitudes Knowledge of the correct use of antibiotics for bacterial diseases rather than viral diseases rose from 43% to 76% in the post-test (P < 0.01). Knowledge of the risk of bacterial resistance to antibiotics from their incorrect use rose from 48% to 74% in the post-test (P < 0.05). Cebotarenco and Bush (2008)16 Chisinau, Moldova controlled before-and-after study n = 3586 schoolchildren and n = 2716 parents Educational intervention about the use of antibiotics delivered by student volunteers trained as peer leaders delivered to their classmates and the classmates’ parents. beliefs and behaviour Students in both the intervention district and the post-intervention phase were 3.2 (95%  CI = 2.065–4.909) times more likely than other students to indicate they had not taken an antibiotic. Croft et al. (2007)11 WI, USA randomized controlled trial n = 300 parents Distribution of printed materials to parents by childcare staff; slide presentation delivered to staff. knowledge In parents who were college graduates, the median knowledge scores were 7.0 at intervention centres and 6.5 at control centres (P < 0.01). Curry et al. (2006)21 Auckland, New Zealand non-controlled before-and-after study n = 400 general public National campaign ‘Wise use of antibiotics’. Posters and leaflets delivered to the public attending pharmacies. knowledge and attitudes and behaviour Patients who had ever been to the doctor for a common cold significantly decreased (45% versus 62%; P = 0.0006). They were significantly less likely to feel positive about antibiotics in 2003 for the treatment of a cold (16% versus 33%; P = 0.00001). The perception that antibiotics were beneficial for cold/flu symptoms significantly reduced from 1998 to 2003 (P < 0.05) and the perceived benefit of antibiotics for tonsillitis increased from 83% to 91% in 2003 (P = 0.014). Significantly fewer people reported ever attending a doctor for a cold in 2003 versus 1998 (45% versus 62%; P = 0.0006) and the number of people who would usually see a doctor for a cold decreased from 24% to 15% (P = 0.026). Farrell et al. (2011)22 Glasgow, Gloucester and London, UK non-controlled before-and-after study n = 1736 children e-Bug web game. knowledge and attitudes No overall change in knowledge. Significant knowledge change in 3 out of 21 questions (P  ≤  0.02). Formoso et al. (2013)13 Emilia-Romagna, Italy non-randomized controlled trial n = 1200 general public Local mass media campaign (posters, brochures and advertisements on local media) delivered to general population to raise awareness of inappropriate use of antibiotics. knowledge After the intervention, consistency with campaign messages worsened (or did not improve) similarly in both intervention and control areas, the only exception being knowledge of the presumptive antiviral activity of antibiotics, worsening in the intervention area more than the control area. Gonzales et al. (2008)14 CO, USA non-randomized controlled trial n = 1503 general public Mass media intervention (outdoor and radio advertisements) delivered to general public about use of antibiotics. behaviour Linear regression analysis showed significant net differences in monthly paediatric office visit rates between mass media and comparison communities before and after the campaign (P = 0.01). Huang et al. (2007)12 MA, USA randomized controlled trial n = 3142 parents Community-based educational intervention occurred through three successive cold and flu seasons. Printed materials: mailed newsletters, posters, pamphlets and fact sheets in the waiting rooms of local paediatric providers, pharmacies and childcare centres. knowledge and attitudes The proportion of parents who answered ≥7 of 10 knowledge questions correctly increased significantly in both intervention (from 52% to 64%; P < 0.001) and control (from 54% to 61%; P < 0.01) communities. Substantial improvements in the percentage of correct answers were seen for items on middle ear fluid (41% in 2000; 50% in 2003; P < 0.001) and the general question of whether antibiotics were needed for colds and flu (66% in 2000; 77% in 2003; P < 0.001). Larson et al. (2009)23 Upper Manhattan, USA non-controlled before-and-after study n = 422 households Targeted Latino households. Educational materials (colouring book, pamphlets) based on knowledge, attitudes and practices regarding prevention and treatment of upper respiratory tract infections. Programme was delivered during home visits every 2 months. knowledge and attitudes and behaviour After the intervention, the mean composite knowledge scores at baseline and end of study were 5.19 (SD = 1.60) and 5.91 (SD = 1.71) (P < 0.001), respectively. With regard to reported practices, significantly more participants after the intervention reported using alcohol hand sanitizers (1.4% baseline and 66.8% post-intervention; P = 0.001). Significantly more also reported that one or more members of their household had received the influenza vaccination after the intervention (63.7% at baseline and 73.9% post-intervention; P = 0.001). Lecky et al. (2010)17 Gloucestershire and London, England; Nice and Bordeaux, France; Prague and Ostrava, Czech Republic controlled before-and-after study n = 2724 school students School-based educational intervention (e-Bug) regarding inappropriate antibiotic use delivered to classes of 9- to 11-year-old (junior) and 12- to 15-year-old (senior) students in state schools. Included 45 min lesson, handouts, worksheets, factsheets, interactive activity and a follow-up plenary question-and-answer session. knowledge Junior school: significant change in knowledge and significant change in retention 6 weeks post-intervention, across countries. Little significant difference in knowledge change between intervention and control, with the exception of the Czech Republic. Senior school: significant improvement in knowledge 6 weeks post-intervention in the Czech Republic. Significant improvement in knowledge and knowledge retention between control and intervention in England and the Czech Republic Madle et al. (2004)24 London, UK non-controlled before-and-after study n = 177 general public Open access to the National Electronic Library of Infection Antimicrobial Resistance website on the use of antibiotics and antibiotic resistance. The site comprises frequently asked questions and links to evidence-based resources. knowledge and attitudes Significant improvements in knowledge about the use of antibiotics and antibiotic resistance in 2 out of 3 statements: (i) ‘people cannot become resistant to antibiotics’ (P < 0.001, χ2 = 60.357, 95% CI of change = 27.47–44.53); (ii) ‘antibiotics do not cure most sore throats’ (P < 0.001, χ2 = 19.22, 95% CI of change = 8.62–27.38). Significant changes in the scores assigned by users for 3 out of 4 statements designed to test users’ attitudes to the information on the site (P ≤ 0.003). Expectations that antibiotics should be prescribed were significantly reduced after using the website (P < 0.001). Non-healthcare workers continued to have higher expectations of antibiotics being prescribed than healthcare workers (P = 0.0046 before and P = 0.0098 after using the website). Mainous et al. (2009)30 SC, USA Prospective cohort study n = 691 self-identified Latinos Mass media educational intervention (pamphlets, radio, newspapers) delivered to local Latino communities about use of antibiotics. behaviour Numbers in the intervention group reporting that they had bought antibiotics without a prescription increased following the intervention compared with baseline (χ2 test reported as significant, but P value not given). The regression analysis showed the strongest predictor of purchase of antibiotics without a prescription in the previous 12 months was past purchase of antibiotics without a prescription outside the USA (OR = 5.72; 95% CI = 3.12–10.48). The regression analysis also showed the strongest predictor of likelihood of importing antibiotics into the USA was past purchase of antibiotics without a prescription outside the USA (OR = 3.01; 95% CI = 1.95–4.65). Mazińska and Hryniewicz (2010)25 Poland non-controlled before-and-after study n = 1000 general public Mass media educational intervention (posters, leaflets, billboards, TV, cinemas, radio, press, magazines, thematic exhibitions, Internet) implemented across the country. knowledge, attitudes and behaviour Significant increase in the percentage of people who have limited the use of antibiotics (27% to 43%), have become more disciplined and cautious in their use (3% to 24%) and pay attention to the correct dosage (6% to 18%) (no P values given). McNulty et al. (2001)27 Gloucester, UK non-controlled before-and-after study n = 38 year 5 schoolchildren School-based intervention to children aged 9–10 years at a state school. Included two 90 min interactive workshops entitled ‘Antibiotics and your good bugs’. knowledge Before the workshops 23% and 26% knew antibiotics do not kill viruses, but kill good bacteria, compared with 47% and 69% afterwards (P = 0.03 and 0.0001). 45% before and 73% after the workshops correctly answered all the questions (P < 0.0001). Children thought antibiotics helped hay fever; this improved significantly after the workshop [correct answer 28% before, 77% after (P < 0.0001)]. Overall score for 7 questions in the ‘where are bugs found’ section was increased significantly from an average of 80.5% success to 93.2% success (P = 0.0002). The overall score improvement in the ‘How do bugs spread’ section was significant (P = 0.00001). McNulty et al. (2007)26 Gloucestershire, UK non-controlled before-and-after study n = 198 year 5 and 6 schoolchildren School-based intervention. ‘Bug Investigators’ pack about microorganisms, hygiene and antibiotics. The pack included 11 activity sheets, teachers’ guide, poster and website. knowledge Children’s knowledge improved in all topic areas and was significant in 6 out of the 7 topic areas (P < 0.005). Improved knowledge was most significant for what antibiotics do and how to use them [improvement = 27% (95% CI = 22.8–31.1) and 31% (95% CI = 23.4–37.7), respectively] and the value of our own good bugs (16% improvement). McNulty et al. (2010)18 England and Scotland, UK controlled before-and-after study n = 3718 general public Mass media campaign about antibiotic use involving posters displayed in magazines and newspapers. knowledge and attitudes and behaviour No positive effect of the campaigns. Pontes and Pontes (2005)15 Mid-Atlantic region, USA non-randomized controlled trials n = 105 university students University-based educational intervention (information booklet) to increase young adult consumers’ preference for physicians who do not unnecessarily prescribe antibiotics for simple acute upper respiratory tract infections. attitudes Exposure to the intervention significantly increased the mean preferred start date for antibiotics after initiation of an infection from 2.3 to 3.9 days (P < 0.01). Respondents’ preferences were significantly greater for the physician who indicated he would not prescribe antibiotics in the intervention (mean = 4.84) compared with the control (mean = 3.91) (P < 0.01). Stockwell et al. (2010)28 New York City, USA non-controlled before-and-after study n = 10 parents Health literacy intervention regarding upper respiratory tract infection delivered to parents on an ‘Early Head Start’ programme. Involved 3 × 1.5 h interactive sessions and provision of kit for care of a child with such an infection. knowledge and attitudes and behaviour The mean composite knowledge/attitude score increased from 4.1 (total possible: 10) to 6.6 (P < 0.05). The number of parents reporting that the last time their child was sick they sought antibiotics without a prescription instead of, or in addition to, seeing their healthcare provider decreased from 6 to 1 (P = 0.06). Trepka et al. (2001)19 Northern WI, USA controlled before-and-after study n = 365 parents Nurse educators delivered parent-oriented presentations in community organizations, distributed information pamphlets and displayed posters. Topics covered included antibiotic resistance and use. knowledge and attitudes and behaviour From baseline to post-intervention the percentage of parents with high antibiotic resistance awareness significantly increased in the intervention group (change = 14.3%; 95% CI = 6.6%–22.0%), but not in the control group (change = 4.3%; 95% CI = −4.1% to 12.7%; P = 0.015). The proportion of parents who expected an antibiotic for their child and did not receive one declined in the intervention area (14% to 9%), while it increased in the control area (7% to 10%). The difference between the two area changes was −8.4% (95% CI = −13.9% to −2.8%; P = 0.003). The percentage of parents in the intervention area who brought their child to another physician because they did not receive an antibiotic decreased (5% to 2%), whereas it increased in the control area (2% to 4%). The difference between the two area changes was −4.5% (95% CI=−8.0% to −0.9%; P = 0.02). Wutzke et al. (2007)29 Australia non-controlled before-and-after study n = 6217 general public National mass media intervention for consumers delivered during winter months in 2001, 2002, 2003 and 2004. About the inappropriate use of antibiotics for upper respiratory tract infection. Strategies included newsletters and brochures, mass media activity using billboards, television, radio and magazines and small grants to promote local community education. knowledge and attitudes and behaviour There was a significant decline in those who believe taking antibiotics for cold and flu is appropriate, from 28.7% pre-programme in 2002 to 21.7% in 2004 (percentage point change = 7.0; 95% CI = 3.5–10.5). Significant decrease in self-reported use of antibiotics to treat cough, cold or flu, from 10.8% in 1999 to 7.4% in 2004 (percentage point change = 3.4; 95% CI = 1.3–5.5). Table 2. Study characteristics and results of the included studies Study Country Design Sample Nature of intervention(s) Outcome measures Significant results Azevedo et al. (2013)20 Braga, Portugal non-controlled before-and-after study n = 82 schoolchildren School-based presentation followed by discussion. knowledge and attitudes Knowledge of the correct use of antibiotics for bacterial diseases rather than viral diseases rose from 43% to 76% in the post-test (P < 0.01). Knowledge of the risk of bacterial resistance to antibiotics from their incorrect use rose from 48% to 74% in the post-test (P < 0.05). Cebotarenco and Bush (2008)16 Chisinau, Moldova controlled before-and-after study n = 3586 schoolchildren and n = 2716 parents Educational intervention about the use of antibiotics delivered by student volunteers trained as peer leaders delivered to their classmates and the classmates’ parents. beliefs and behaviour Students in both the intervention district and the post-intervention phase were 3.2 (95%  CI = 2.065–4.909) times more likely than other students to indicate they had not taken an antibiotic. Croft et al. (2007)11 WI, USA randomized controlled trial n = 300 parents Distribution of printed materials to parents by childcare staff; slide presentation delivered to staff. knowledge In parents who were college graduates, the median knowledge scores were 7.0 at intervention centres and 6.5 at control centres (P < 0.01). Curry et al. (2006)21 Auckland, New Zealand non-controlled before-and-after study n = 400 general public National campaign ‘Wise use of antibiotics’. Posters and leaflets delivered to the public attending pharmacies. knowledge and attitudes and behaviour Patients who had ever been to the doctor for a common cold significantly decreased (45% versus 62%; P = 0.0006). They were significantly less likely to feel positive about antibiotics in 2003 for the treatment of a cold (16% versus 33%; P = 0.00001). The perception that antibiotics were beneficial for cold/flu symptoms significantly reduced from 1998 to 2003 (P < 0.05) and the perceived benefit of antibiotics for tonsillitis increased from 83% to 91% in 2003 (P = 0.014). Significantly fewer people reported ever attending a doctor for a cold in 2003 versus 1998 (45% versus 62%; P = 0.0006) and the number of people who would usually see a doctor for a cold decreased from 24% to 15% (P = 0.026). Farrell et al. (2011)22 Glasgow, Gloucester and London, UK non-controlled before-and-after study n = 1736 children e-Bug web game. knowledge and attitudes No overall change in knowledge. Significant knowledge change in 3 out of 21 questions (P  ≤  0.02). Formoso et al. (2013)13 Emilia-Romagna, Italy non-randomized controlled trial n = 1200 general public Local mass media campaign (posters, brochures and advertisements on local media) delivered to general population to raise awareness of inappropriate use of antibiotics. knowledge After the intervention, consistency with campaign messages worsened (or did not improve) similarly in both intervention and control areas, the only exception being knowledge of the presumptive antiviral activity of antibiotics, worsening in the intervention area more than the control area. Gonzales et al. (2008)14 CO, USA non-randomized controlled trial n = 1503 general public Mass media intervention (outdoor and radio advertisements) delivered to general public about use of antibiotics. behaviour Linear regression analysis showed significant net differences in monthly paediatric office visit rates between mass media and comparison communities before and after the campaign (P = 0.01). Huang et al. (2007)12 MA, USA randomized controlled trial n = 3142 parents Community-based educational intervention occurred through three successive cold and flu seasons. Printed materials: mailed newsletters, posters, pamphlets and fact sheets in the waiting rooms of local paediatric providers, pharmacies and childcare centres. knowledge and attitudes The proportion of parents who answered ≥7 of 10 knowledge questions correctly increased significantly in both intervention (from 52% to 64%; P < 0.001) and control (from 54% to 61%; P < 0.01) communities. Substantial improvements in the percentage of correct answers were seen for items on middle ear fluid (41% in 2000; 50% in 2003; P < 0.001) and the general question of whether antibiotics were needed for colds and flu (66% in 2000; 77% in 2003; P < 0.001). Larson et al. (2009)23 Upper Manhattan, USA non-controlled before-and-after study n = 422 households Targeted Latino households. Educational materials (colouring book, pamphlets) based on knowledge, attitudes and practices regarding prevention and treatment of upper respiratory tract infections. Programme was delivered during home visits every 2 months. knowledge and attitudes and behaviour After the intervention, the mean composite knowledge scores at baseline and end of study were 5.19 (SD = 1.60) and 5.91 (SD = 1.71) (P < 0.001), respectively. With regard to reported practices, significantly more participants after the intervention reported using alcohol hand sanitizers (1.4% baseline and 66.8% post-intervention; P = 0.001). Significantly more also reported that one or more members of their household had received the influenza vaccination after the intervention (63.7% at baseline and 73.9% post-intervention; P = 0.001). Lecky et al. (2010)17 Gloucestershire and London, England; Nice and Bordeaux, France; Prague and Ostrava, Czech Republic controlled before-and-after study n = 2724 school students School-based educational intervention (e-Bug) regarding inappropriate antibiotic use delivered to classes of 9- to 11-year-old (junior) and 12- to 15-year-old (senior) students in state schools. Included 45 min lesson, handouts, worksheets, factsheets, interactive activity and a follow-up plenary question-and-answer session. knowledge Junior school: significant change in knowledge and significant change in retention 6 weeks post-intervention, across countries. Little significant difference in knowledge change between intervention and control, with the exception of the Czech Republic. Senior school: significant improvement in knowledge 6 weeks post-intervention in the Czech Republic. Significant improvement in knowledge and knowledge retention between control and intervention in England and the Czech Republic Madle et al. (2004)24 London, UK non-controlled before-and-after study n = 177 general public Open access to the National Electronic Library of Infection Antimicrobial Resistance website on the use of antibiotics and antibiotic resistance. The site comprises frequently asked questions and links to evidence-based resources. knowledge and attitudes Significant improvements in knowledge about the use of antibiotics and antibiotic resistance in 2 out of 3 statements: (i) ‘people cannot become resistant to antibiotics’ (P < 0.001, χ2 = 60.357, 95% CI of change = 27.47–44.53); (ii) ‘antibiotics do not cure most sore throats’ (P < 0.001, χ2 = 19.22, 95% CI of change = 8.62–27.38). Significant changes in the scores assigned by users for 3 out of 4 statements designed to test users’ attitudes to the information on the site (P ≤ 0.003). Expectations that antibiotics should be prescribed were significantly reduced after using the website (P < 0.001). Non-healthcare workers continued to have higher expectations of antibiotics being prescribed than healthcare workers (P = 0.0046 before and P = 0.0098 after using the website). Mainous et al. (2009)30 SC, USA Prospective cohort study n = 691 self-identified Latinos Mass media educational intervention (pamphlets, radio, newspapers) delivered to local Latino communities about use of antibiotics. behaviour Numbers in the intervention group reporting that they had bought antibiotics without a prescription increased following the intervention compared with baseline (χ2 test reported as significant, but P value not given). The regression analysis showed the strongest predictor of purchase of antibiotics without a prescription in the previous 12 months was past purchase of antibiotics without a prescription outside the USA (OR = 5.72; 95% CI = 3.12–10.48). The regression analysis also showed the strongest predictor of likelihood of importing antibiotics into the USA was past purchase of antibiotics without a prescription outside the USA (OR = 3.01; 95% CI = 1.95–4.65). Mazińska and Hryniewicz (2010)25 Poland non-controlled before-and-after study n = 1000 general public Mass media educational intervention (posters, leaflets, billboards, TV, cinemas, radio, press, magazines, thematic exhibitions, Internet) implemented across the country. knowledge, attitudes and behaviour Significant increase in the percentage of people who have limited the use of antibiotics (27% to 43%), have become more disciplined and cautious in their use (3% to 24%) and pay attention to the correct dosage (6% to 18%) (no P values given). McNulty et al. (2001)27 Gloucester, UK non-controlled before-and-after study n = 38 year 5 schoolchildren School-based intervention to children aged 9–10 years at a state school. Included two 90 min interactive workshops entitled ‘Antibiotics and your good bugs’. knowledge Before the workshops 23% and 26% knew antibiotics do not kill viruses, but kill good bacteria, compared with 47% and 69% afterwards (P = 0.03 and 0.0001). 45% before and 73% after the workshops correctly answered all the questions (P < 0.0001). Children thought antibiotics helped hay fever; this improved significantly after the workshop [correct answer 28% before, 77% after (P < 0.0001)]. Overall score for 7 questions in the ‘where are bugs found’ section was increased significantly from an average of 80.5% success to 93.2% success (P = 0.0002). The overall score improvement in the ‘How do bugs spread’ section was significant (P = 0.00001). McNulty et al. (2007)26 Gloucestershire, UK non-controlled before-and-after study n = 198 year 5 and 6 schoolchildren School-based intervention. ‘Bug Investigators’ pack about microorganisms, hygiene and antibiotics. The pack included 11 activity sheets, teachers’ guide, poster and website. knowledge Children’s knowledge improved in all topic areas and was significant in 6 out of the 7 topic areas (P < 0.005). Improved knowledge was most significant for what antibiotics do and how to use them [improvement = 27% (95% CI = 22.8–31.1) and 31% (95% CI = 23.4–37.7), respectively] and the value of our own good bugs (16% improvement). McNulty et al. (2010)18 England and Scotland, UK controlled before-and-after study n = 3718 general public Mass media campaign about antibiotic use involving posters displayed in magazines and newspapers. knowledge and attitudes and behaviour No positive effect of the campaigns. Pontes and Pontes (2005)15 Mid-Atlantic region, USA non-randomized controlled trials n = 105 university students University-based educational intervention (information booklet) to increase young adult consumers’ preference for physicians who do not unnecessarily prescribe antibiotics for simple acute upper respiratory tract infections. attitudes Exposure to the intervention significantly increased the mean preferred start date for antibiotics after initiation of an infection from 2.3 to 3.9 days (P < 0.01). Respondents’ preferences were significantly greater for the physician who indicated he would not prescribe antibiotics in the intervention (mean = 4.84) compared with the control (mean = 3.91) (P < 0.01). Stockwell et al. (2010)28 New York City, USA non-controlled before-and-after study n = 10 parents Health literacy intervention regarding upper respiratory tract infection delivered to parents on an ‘Early Head Start’ programme. Involved 3 × 1.5 h interactive sessions and provision of kit for care of a child with such an infection. knowledge and attitudes and behaviour The mean composite knowledge/attitude score increased from 4.1 (total possible: 10) to 6.6 (P < 0.05). The number of parents reporting that the last time their child was sick they sought antibiotics without a prescription instead of, or in addition to, seeing their healthcare provider decreased from 6 to 1 (P = 0.06). Trepka et al. (2001)19 Northern WI, USA controlled before-and-after study n = 365 parents Nurse educators delivered parent-oriented presentations in community organizations, distributed information pamphlets and displayed posters. Topics covered included antibiotic resistance and use. knowledge and attitudes and behaviour From baseline to post-intervention the percentage of parents with high antibiotic resistance awareness significantly increased in the intervention group (change = 14.3%; 95% CI = 6.6%–22.0%), but not in the control group (change = 4.3%; 95% CI = −4.1% to 12.7%; P = 0.015). The proportion of parents who expected an antibiotic for their child and did not receive one declined in the intervention area (14% to 9%), while it increased in the control area (7% to 10%). The difference between the two area changes was −8.4% (95% CI = −13.9% to −2.8%; P = 0.003). The percentage of parents in the intervention area who brought their child to another physician because they did not receive an antibiotic decreased (5% to 2%), whereas it increased in the control area (2% to 4%). The difference between the two area changes was −4.5% (95% CI=−8.0% to −0.9%; P = 0.02). Wutzke et al. (2007)29 Australia non-controlled before-and-after study n = 6217 general public National mass media intervention for consumers delivered during winter months in 2001, 2002, 2003 and 2004. About the inappropriate use of antibiotics for upper respiratory tract infection. Strategies included newsletters and brochures, mass media activity using billboards, television, radio and magazines and small grants to promote local community education. knowledge and attitudes and behaviour There was a significant decline in those who believe taking antibiotics for cold and flu is appropriate, from 28.7% pre-programme in 2002 to 21.7% in 2004 (percentage point change = 7.0; 95% CI = 3.5–10.5). Significant decrease in self-reported use of antibiotics to treat cough, cold or flu, from 10.8% in 1999 to 7.4% in 2004 (percentage point change = 3.4; 95% CI = 1.3–5.5). Study Country Design Sample Nature of intervention(s) Outcome measures Significant results Azevedo et al. (2013)20 Braga, Portugal non-controlled before-and-after study n = 82 schoolchildren School-based presentation followed by discussion. knowledge and attitudes Knowledge of the correct use of antibiotics for bacterial diseases rather than viral diseases rose from 43% to 76% in the post-test (P < 0.01). Knowledge of the risk of bacterial resistance to antibiotics from their incorrect use rose from 48% to 74% in the post-test (P < 0.05). Cebotarenco and Bush (2008)16 Chisinau, Moldova controlled before-and-after study n = 3586 schoolchildren and n = 2716 parents Educational intervention about the use of antibiotics delivered by student volunteers trained as peer leaders delivered to their classmates and the classmates’ parents. beliefs and behaviour Students in both the intervention district and the post-intervention phase were 3.2 (95%  CI = 2.065–4.909) times more likely than other students to indicate they had not taken an antibiotic. Croft et al. (2007)11 WI, USA randomized controlled trial n = 300 parents Distribution of printed materials to parents by childcare staff; slide presentation delivered to staff. knowledge In parents who were college graduates, the median knowledge scores were 7.0 at intervention centres and 6.5 at control centres (P < 0.01). Curry et al. (2006)21 Auckland, New Zealand non-controlled before-and-after study n = 400 general public National campaign ‘Wise use of antibiotics’. Posters and leaflets delivered to the public attending pharmacies. knowledge and attitudes and behaviour Patients who had ever been to the doctor for a common cold significantly decreased (45% versus 62%; P = 0.0006). They were significantly less likely to feel positive about antibiotics in 2003 for the treatment of a cold (16% versus 33%; P = 0.00001). The perception that antibiotics were beneficial for cold/flu symptoms significantly reduced from 1998 to 2003 (P < 0.05) and the perceived benefit of antibiotics for tonsillitis increased from 83% to 91% in 2003 (P = 0.014). Significantly fewer people reported ever attending a doctor for a cold in 2003 versus 1998 (45% versus 62%; P = 0.0006) and the number of people who would usually see a doctor for a cold decreased from 24% to 15% (P = 0.026). Farrell et al. (2011)22 Glasgow, Gloucester and London, UK non-controlled before-and-after study n = 1736 children e-Bug web game. knowledge and attitudes No overall change in knowledge. Significant knowledge change in 3 out of 21 questions (P  ≤  0.02). Formoso et al. (2013)13 Emilia-Romagna, Italy non-randomized controlled trial n = 1200 general public Local mass media campaign (posters, brochures and advertisements on local media) delivered to general population to raise awareness of inappropriate use of antibiotics. knowledge After the intervention, consistency with campaign messages worsened (or did not improve) similarly in both intervention and control areas, the only exception being knowledge of the presumptive antiviral activity of antibiotics, worsening in the intervention area more than the control area. Gonzales et al. (2008)14 CO, USA non-randomized controlled trial n = 1503 general public Mass media intervention (outdoor and radio advertisements) delivered to general public about use of antibiotics. behaviour Linear regression analysis showed significant net differences in monthly paediatric office visit rates between mass media and comparison communities before and after the campaign (P = 0.01). Huang et al. (2007)12 MA, USA randomized controlled trial n = 3142 parents Community-based educational intervention occurred through three successive cold and flu seasons. Printed materials: mailed newsletters, posters, pamphlets and fact sheets in the waiting rooms of local paediatric providers, pharmacies and childcare centres. knowledge and attitudes The proportion of parents who answered ≥7 of 10 knowledge questions correctly increased significantly in both intervention (from 52% to 64%; P < 0.001) and control (from 54% to 61%; P < 0.01) communities. Substantial improvements in the percentage of correct answers were seen for items on middle ear fluid (41% in 2000; 50% in 2003; P < 0.001) and the general question of whether antibiotics were needed for colds and flu (66% in 2000; 77% in 2003; P < 0.001). Larson et al. (2009)23 Upper Manhattan, USA non-controlled before-and-after study n = 422 households Targeted Latino households. Educational materials (colouring book, pamphlets) based on knowledge, attitudes and practices regarding prevention and treatment of upper respiratory tract infections. Programme was delivered during home visits every 2 months. knowledge and attitudes and behaviour After the intervention, the mean composite knowledge scores at baseline and end of study were 5.19 (SD = 1.60) and 5.91 (SD = 1.71) (P < 0.001), respectively. With regard to reported practices, significantly more participants after the intervention reported using alcohol hand sanitizers (1.4% baseline and 66.8% post-intervention; P = 0.001). Significantly more also reported that one or more members of their household had received the influenza vaccination after the intervention (63.7% at baseline and 73.9% post-intervention; P = 0.001). Lecky et al. (2010)17 Gloucestershire and London, England; Nice and Bordeaux, France; Prague and Ostrava, Czech Republic controlled before-and-after study n = 2724 school students School-based educational intervention (e-Bug) regarding inappropriate antibiotic use delivered to classes of 9- to 11-year-old (junior) and 12- to 15-year-old (senior) students in state schools. Included 45 min lesson, handouts, worksheets, factsheets, interactive activity and a follow-up plenary question-and-answer session. knowledge Junior school: significant change in knowledge and significant change in retention 6 weeks post-intervention, across countries. Little significant difference in knowledge change between intervention and control, with the exception of the Czech Republic. Senior school: significant improvement in knowledge 6 weeks post-intervention in the Czech Republic. Significant improvement in knowledge and knowledge retention between control and intervention in England and the Czech Republic Madle et al. (2004)24 London, UK non-controlled before-and-after study n = 177 general public Open access to the National Electronic Library of Infection Antimicrobial Resistance website on the use of antibiotics and antibiotic resistance. The site comprises frequently asked questions and links to evidence-based resources. knowledge and attitudes Significant improvements in knowledge about the use of antibiotics and antibiotic resistance in 2 out of 3 statements: (i) ‘people cannot become resistant to antibiotics’ (P < 0.001, χ2 = 60.357, 95% CI of change = 27.47–44.53); (ii) ‘antibiotics do not cure most sore throats’ (P < 0.001, χ2 = 19.22, 95% CI of change = 8.62–27.38). Significant changes in the scores assigned by users for 3 out of 4 statements designed to test users’ attitudes to the information on the site (P ≤ 0.003). Expectations that antibiotics should be prescribed were significantly reduced after using the website (P < 0.001). Non-healthcare workers continued to have higher expectations of antibiotics being prescribed than healthcare workers (P = 0.0046 before and P = 0.0098 after using the website). Mainous et al. (2009)30 SC, USA Prospective cohort study n = 691 self-identified Latinos Mass media educational intervention (pamphlets, radio, newspapers) delivered to local Latino communities about use of antibiotics. behaviour Numbers in the intervention group reporting that they had bought antibiotics without a prescription increased following the intervention compared with baseline (χ2 test reported as significant, but P value not given). The regression analysis showed the strongest predictor of purchase of antibiotics without a prescription in the previous 12 months was past purchase of antibiotics without a prescription outside the USA (OR = 5.72; 95% CI = 3.12–10.48). The regression analysis also showed the strongest predictor of likelihood of importing antibiotics into the USA was past purchase of antibiotics without a prescription outside the USA (OR = 3.01; 95% CI = 1.95–4.65). Mazińska and Hryniewicz (2010)25 Poland non-controlled before-and-after study n = 1000 general public Mass media educational intervention (posters, leaflets, billboards, TV, cinemas, radio, press, magazines, thematic exhibitions, Internet) implemented across the country. knowledge, attitudes and behaviour Significant increase in the percentage of people who have limited the use of antibiotics (27% to 43%), have become more disciplined and cautious in their use (3% to 24%) and pay attention to the correct dosage (6% to 18%) (no P values given). McNulty et al. (2001)27 Gloucester, UK non-controlled before-and-after study n = 38 year 5 schoolchildren School-based intervention to children aged 9–10 years at a state school. Included two 90 min interactive workshops entitled ‘Antibiotics and your good bugs’. knowledge Before the workshops 23% and 26% knew antibiotics do not kill viruses, but kill good bacteria, compared with 47% and 69% afterwards (P = 0.03 and 0.0001). 45% before and 73% after the workshops correctly answered all the questions (P < 0.0001). Children thought antibiotics helped hay fever; this improved significantly after the workshop [correct answer 28% before, 77% after (P < 0.0001)]. Overall score for 7 questions in the ‘where are bugs found’ section was increased significantly from an average of 80.5% success to 93.2% success (P = 0.0002). The overall score improvement in the ‘How do bugs spread’ section was significant (P = 0.00001). McNulty et al. (2007)26 Gloucestershire, UK non-controlled before-and-after study n = 198 year 5 and 6 schoolchildren School-based intervention. ‘Bug Investigators’ pack about microorganisms, hygiene and antibiotics. The pack included 11 activity sheets, teachers’ guide, poster and website. knowledge Children’s knowledge improved in all topic areas and was significant in 6 out of the 7 topic areas (P < 0.005). Improved knowledge was most significant for what antibiotics do and how to use them [improvement = 27% (95% CI = 22.8–31.1) and 31% (95% CI = 23.4–37.7), respectively] and the value of our own good bugs (16% improvement). McNulty et al. (2010)18 England and Scotland, UK controlled before-and-after study n = 3718 general public Mass media campaign about antibiotic use involving posters displayed in magazines and newspapers. knowledge and attitudes and behaviour No positive effect of the campaigns. Pontes and Pontes (2005)15 Mid-Atlantic region, USA non-randomized controlled trials n = 105 university students University-based educational intervention (information booklet) to increase young adult consumers’ preference for physicians who do not unnecessarily prescribe antibiotics for simple acute upper respiratory tract infections. attitudes Exposure to the intervention significantly increased the mean preferred start date for antibiotics after initiation of an infection from 2.3 to 3.9 days (P < 0.01). Respondents’ preferences were significantly greater for the physician who indicated he would not prescribe antibiotics in the intervention (mean = 4.84) compared with the control (mean = 3.91) (P < 0.01). Stockwell et al. (2010)28 New York City, USA non-controlled before-and-after study n = 10 parents Health literacy intervention regarding upper respiratory tract infection delivered to parents on an ‘Early Head Start’ programme. Involved 3 × 1.5 h interactive sessions and provision of kit for care of a child with such an infection. knowledge and attitudes and behaviour The mean composite knowledge/attitude score increased from 4.1 (total possible: 10) to 6.6 (P < 0.05). The number of parents reporting that the last time their child was sick they sought antibiotics without a prescription instead of, or in addition to, seeing their healthcare provider decreased from 6 to 1 (P = 0.06). Trepka et al. (2001)19 Northern WI, USA controlled before-and-after study n = 365 parents Nurse educators delivered parent-oriented presentations in community organizations, distributed information pamphlets and displayed posters. Topics covered included antibiotic resistance and use. knowledge and attitudes and behaviour From baseline to post-intervention the percentage of parents with high antibiotic resistance awareness significantly increased in the intervention group (change = 14.3%; 95% CI = 6.6%–22.0%), but not in the control group (change = 4.3%; 95% CI = −4.1% to 12.7%; P = 0.015). The proportion of parents who expected an antibiotic for their child and did not receive one declined in the intervention area (14% to 9%), while it increased in the control area (7% to 10%). The difference between the two area changes was −8.4% (95% CI = −13.9% to −2.8%; P = 0.003). The percentage of parents in the intervention area who brought their child to another physician because they did not receive an antibiotic decreased (5% to 2%), whereas it increased in the control area (2% to 4%). The difference between the two area changes was −4.5% (95% CI=−8.0% to −0.9%; P = 0.02). Wutzke et al. (2007)29 Australia non-controlled before-and-after study n = 6217 general public National mass media intervention for consumers delivered during winter months in 2001, 2002, 2003 and 2004. About the inappropriate use of antibiotics for upper respiratory tract infection. Strategies included newsletters and brochures, mass media activity using billboards, television, radio and magazines and small grants to promote local community education. knowledge and attitudes and behaviour There was a significant decline in those who believe taking antibiotics for cold and flu is appropriate, from 28.7% pre-programme in 2002 to 21.7% in 2004 (percentage point change = 7.0; 95% CI = 3.5–10.5). Significant decrease in self-reported use of antibiotics to treat cough, cold or flu, from 10.8% in 1999 to 7.4% in 2004 (percentage point change = 3.4; 95% CI = 1.3–5.5). Apart from one study conducted in Moldova,16 all studies were conducted in high-income countries, with the majority conducted in the USA (n = 8)11,12,14,15,19,23,28,30 or in the UK (n = 5).18,22,24,26,27 The remaining studies were conducted in Italy,13 Portugal,20 Poland,25 New Zealand21 and Australia,29 and one study was a multisite study conducted in the UK, the Czech Republic and France17 (Table 2). As shown in Table 2, the most common types of interventions were mass media interventions.13,14,18,24,25,29,30 Apart from Madle et al.,24 who used a website only, all mass media interventions were multimodal and used a variety of outlets, such as billboards, radio, television, newspapers, magazines, websites and printed resources such as posters, brochures, pamphlets, leaflets, stickers or badges distributed to community sites or healthcare settings. Six studies used printed material interventions, either alone12,15,21,23 or in combination with educational presentations,11,19 while Stockwell et al.28 delivered taught modules to Latino community parents. Other interventions were school based and included a student peer-taught programme,16 an e-Bug web game,22 interactive workshops,27 school lessons delivered using the Bug Investigators pack26 and presentation followed by discussion,20 while the intervention delivered in the study by Lecky et al.17 involved the delivery of a lesson, printed materials, interactive activities and a question-and-answer session. Comparators were similar across the 10 controlled studies.11–19,30 With the exception of Lecky et al.,17 who compared the educational intervention with the usual school curriculum, control groups were not exposed to the interventions. Table 2 shows that the most common outcome measure was change in knowledge, attitudes or beliefs, measured alone (n = 10)11–13,15,17,20,22,24,26,27 or in combination with change in the public’s antimicrobial stewardship behaviour (n = 8).16,18,19,21,23,25,28,29 Two studies measured the public’s antimicrobial stewardship behaviour outcomes exclusively (n = 2).14,30 Quality of studies Amongst the included studies, nine met the EPOC study design criteria.9 As shown in Table 3, the overall risk of bias of the included studies was generally high. Apart from one study,11 all had at least one item assessed as high risk, with the number of high-risk items ranging from one15 to five.18 High risk of bias was most commonly associated with generation of sequence allocation, risk of contamination and other risks. Table 3. Risk of bias of studies meeting the EPOC criteria Study Allocation: sequence generation Allocation: concealment Baseline primary outcome Baseline secondary outcome Baseline characteristics Incomplete primary outcome data Incomplete secondary outcome data Blinding primary outcome data Blinding secondary outcome Contamination Selective reporting primary outcome Selective reporting secondary outcome Other risks of bias Cebotarenco and Bush (2008)16 H H U U L U U U U L L L L Croft et al. (2007)11 L U U NA U L NA U NA U L NA L Formoso et al. (2013)13 H U U U L U U U U U L L H Gonzales et al. (2008)14 H U U L U U U U L H L L H Huang et al. (2007)12 L U H NA L L NA U NA H H NA H Lecky et al. (2010)17 H U U NA U H NA U NA H L NA H McNulty et al. (2010)18 H H H NA U U NA U NA H L NA H Pontes and Pontes (2005)15 U U U U U U U U U U L L H Trepka et al. (2001)19 H U L NA U L NA U NA H L NA H Study Allocation: sequence generation Allocation: concealment Baseline primary outcome Baseline secondary outcome Baseline characteristics Incomplete primary outcome data Incomplete secondary outcome data Blinding primary outcome data Blinding secondary outcome Contamination Selective reporting primary outcome Selective reporting secondary outcome Other risks of bias Cebotarenco and Bush (2008)16 H H U U L U U U U L L L L Croft et al. (2007)11 L U U NA U L NA U NA U L NA L Formoso et al. (2013)13 H U U U L U U U U U L L H Gonzales et al. (2008)14 H U U L U U U U L H L L H Huang et al. (2007)12 L U H NA L L NA U NA H H NA H Lecky et al. (2010)17 H U U NA U H NA U NA H L NA H McNulty et al. (2010)18 H H H NA U U NA U NA H L NA H Pontes and Pontes (2005)15 U U U U U U U U U U L L H Trepka et al. (2001)19 H U L NA U L NA U NA H L NA H H, high risk of bias; L, low risk of bias; U, unclear risk of bias; NA, not applicable. Table 3. Risk of bias of studies meeting the EPOC criteria Study Allocation: sequence generation Allocation: concealment Baseline primary outcome Baseline secondary outcome Baseline characteristics Incomplete primary outcome data Incomplete secondary outcome data Blinding primary outcome data Blinding secondary outcome Contamination Selective reporting primary outcome Selective reporting secondary outcome Other risks of bias Cebotarenco and Bush (2008)16 H H U U L U U U U L L L L Croft et al. (2007)11 L U U NA U L NA U NA U L NA L Formoso et al. (2013)13 H U U U L U U U U U L L H Gonzales et al. (2008)14 H U U L U U U U L H L L H Huang et al. (2007)12 L U H NA L L NA U NA H H NA H Lecky et al. (2010)17 H U U NA U H NA U NA H L NA H McNulty et al. (2010)18 H H H NA U U NA U NA H L NA H Pontes and Pontes (2005)15 U U U U U U U U U U L L H Trepka et al. (2001)19 H U L NA U L NA U NA H L NA H Study Allocation: sequence generation Allocation: concealment Baseline primary outcome Baseline secondary outcome Baseline characteristics Incomplete primary outcome data Incomplete secondary outcome data Blinding primary outcome data Blinding secondary outcome Contamination Selective reporting primary outcome Selective reporting secondary outcome Other risks of bias Cebotarenco and Bush (2008)16 H H U U L U U U U L L L L Croft et al. (2007)11 L U U NA U L NA U NA U L NA L Formoso et al. (2013)13 H U U U L U U U U U L L H Gonzales et al. (2008)14 H U U L U U U U L H L L H Huang et al. (2007)12 L U H NA L L NA U NA H H NA H Lecky et al. (2010)17 H U U NA U H NA U NA H L NA H McNulty et al. (2010)18 H H H NA U U NA U NA H L NA H Pontes and Pontes (2005)15 U U U U U U U U U U L L H Trepka et al. (2001)19 H U L NA U L NA U NA H L NA H H, high risk of bias; L, low risk of bias; U, unclear risk of bias; NA, not applicable. For all studies that met the EPOC criteria, insufficient information was provided for at least two of the items that were regarded as unclear risk. None of the studies provided information regarding blinding and, in all studies reporting secondary outcomes, the risk of bias for incomplete secondary outcome data could not be assessed.13–16 The number of low-risk items ranged from one17,18 to five,16 with the ‘selective outcome reporting’ item being most commonly assessed as low risk. None of the studies had a low risk score for any of the following items: ‘allocation concealment’, ‘addressing incomplete secondary outcome data’ and ‘blinding primary outcome data’. Risk of bias was not assessed for the non-controlled before-and-after studies20–29 and a prospective cohort study.30 These study designs did not meet the EPOC criteria;9 therefore, it was assumed that the risk of bias of these studies was high. Relative effectiveness of interventions Reviewed interventions were grouped into five categories of relative measure of effectiveness. As shown in Table 4, six studies demonstrated a clear desired behaviour change following the intervention, while two studies resulted in some desired behaviour change. The desired effect on the antecedent of behaviour was reported in nine papers. One study showed no effect, while two studies demonstrated an increase in drivers of AMR following the intervention. Table 4. Patterning of the effectiveness across the type of target population Study Interventions indicative of clear positive behaviour change in the desired direction Interventions indicative of some positive behaviour change in the desired direction Interventions indicative of positive effect on the antecedent of behaviour in the desired direction Interventions indicative of no effect on behaviour or antecedents of behaviour Interventions indicative of negative effect on behaviour or antecedents of behaviour in a non-desired direction Azevedo et al. (2013)20 schoolchildren Cebotarenco and Bush (2008)16 schoolchildren Croft et al. (2007)11 parents; childcare facilities Curry et al. (2006)21 general public Farrell et al. (2011)22 schoolchildren Formoso et al. (2013)13 general public Gonzales et al. (2008)14 general public Huang et al. (2007)12 parents Larson et al. (2009)23 parents Lecky et al. (2010)17 schoolchildren Madle et al. (2004)24 general public Mainous et al. (2009)30 Latino community Mazińska and Hryniewicz (2010)25 general public McNulty et al. (2001)27 schoolchildren McNulty et al. (2007)26 schoolchildren McNulty et al. (2010)18 general public Pontes and Pontes (2005)15 young adults Stockwell et al. (2010)28 Latino community parents Trepka et al. (2001)19 parents Wutzke et al. (2007)29 general public Study Interventions indicative of clear positive behaviour change in the desired direction Interventions indicative of some positive behaviour change in the desired direction Interventions indicative of positive effect on the antecedent of behaviour in the desired direction Interventions indicative of no effect on behaviour or antecedents of behaviour Interventions indicative of negative effect on behaviour or antecedents of behaviour in a non-desired direction Azevedo et al. (2013)20 schoolchildren Cebotarenco and Bush (2008)16 schoolchildren Croft et al. (2007)11 parents; childcare facilities Curry et al. (2006)21 general public Farrell et al. (2011)22 schoolchildren Formoso et al. (2013)13 general public Gonzales et al. (2008)14 general public Huang et al. (2007)12 parents Larson et al. (2009)23 parents Lecky et al. (2010)17 schoolchildren Madle et al. (2004)24 general public Mainous et al. (2009)30 Latino community Mazińska and Hryniewicz (2010)25 general public McNulty et al. (2001)27 schoolchildren McNulty et al. (2007)26 schoolchildren McNulty et al. (2010)18 general public Pontes and Pontes (2005)15 young adults Stockwell et al. (2010)28 Latino community parents Trepka et al. (2001)19 parents Wutzke et al. (2007)29 general public Table 4. Patterning of the effectiveness across the type of target population Study Interventions indicative of clear positive behaviour change in the desired direction Interventions indicative of some positive behaviour change in the desired direction Interventions indicative of positive effect on the antecedent of behaviour in the desired direction Interventions indicative of no effect on behaviour or antecedents of behaviour Interventions indicative of negative effect on behaviour or antecedents of behaviour in a non-desired direction Azevedo et al. (2013)20 schoolchildren Cebotarenco and Bush (2008)16 schoolchildren Croft et al. (2007)11 parents; childcare facilities Curry et al. (2006)21 general public Farrell et al. (2011)22 schoolchildren Formoso et al. (2013)13 general public Gonzales et al. (2008)14 general public Huang et al. (2007)12 parents Larson et al. (2009)23 parents Lecky et al. (2010)17 schoolchildren Madle et al. (2004)24 general public Mainous et al. (2009)30 Latino community Mazińska and Hryniewicz (2010)25 general public McNulty et al. (2001)27 schoolchildren McNulty et al. (2007)26 schoolchildren McNulty et al. (2010)18 general public Pontes and Pontes (2005)15 young adults Stockwell et al. (2010)28 Latino community parents Trepka et al. (2001)19 parents Wutzke et al. (2007)29 general public Study Interventions indicative of clear positive behaviour change in the desired direction Interventions indicative of some positive behaviour change in the desired direction Interventions indicative of positive effect on the antecedent of behaviour in the desired direction Interventions indicative of no effect on behaviour or antecedents of behaviour Interventions indicative of negative effect on behaviour or antecedents of behaviour in a non-desired direction Azevedo et al. (2013)20 schoolchildren Cebotarenco and Bush (2008)16 schoolchildren Croft et al. (2007)11 parents; childcare facilities Curry et al. (2006)21 general public Farrell et al. (2011)22 schoolchildren Formoso et al. (2013)13 general public Gonzales et al. (2008)14 general public Huang et al. (2007)12 parents Larson et al. (2009)23 parents Lecky et al. (2010)17 schoolchildren Madle et al. (2004)24 general public Mainous et al. (2009)30 Latino community Mazińska and Hryniewicz (2010)25 general public McNulty et al. (2001)27 schoolchildren McNulty et al. (2007)26 schoolchildren McNulty et al. (2010)18 general public Pontes and Pontes (2005)15 young adults Stockwell et al. (2010)28 Latino community parents Trepka et al. (2001)19 parents Wutzke et al. (2007)29 general public Effectiveness of interventions delivered to populations through the life cycle In 17 of the studies, the intervention had a significant effect on the outcome of interest amongst the populations through the life cycle. These included schoolchildren, university students, parents and the general public. Schoolchildren All six school-based educational interventions that targeted schoolchildren aged between 9 and 15 years16,17,20,22,26,27 found a significant increase in knowledge following the educational intervention (Table 2). However, Farrell et al.22 found a significant knowledge change in only 3 out of 21 questions (P ≤ 0.02) and no overall change in knowledge. The three questions for which significant improvement was reported related to the valuableness of ‘good microbes’, the presence of microbes despite inability to see them and hand-washing being an effective method of removing microbes from the hands. Only one study16 measured behavioural outcome in addition to beliefs and found that children in the intervention group were 3.2 times more likely than other students to report that they had not taken an antibiotic for a cold or flu (P < 0.001). None of the studies measured long-term outcomes of school-based interventions. Post-intervention outcomes were measured immediately following the intervention,22 or between 1 and 8 weeks after the intervention. In addition, one study17 found that the increase in knowledge was maintained at 6 weeks post-intervention in junior school students, but not senior school students. University students University students were targeted in one experimental study that aimed to investigate whether an educational intervention (information booklet) resulted in an increase in young adult consumers’ preference for physicians who do not unnecessarily prescribe antibiotics for simple acute upper respiratory tract infections.15 This study demonstrated that exposure to the intervention significantly increased the mean preferred start date for antibiotics after the onset of an infection from 2.3 to 3.9 days (P < 0.1) and preference for a physician who would not prescribe antibiotics at day 3 of an infection (P < 0.1). However, this was still well before the recommended time of 10–14 days. Parents The effect of educational interventions delivered to parents on a change in their AMR knowledge, attitudes or beliefs alone11,12 or in combination with parents’ antimicrobial stewardship behaviour outcomes16,19,23,28 was measured in six studies. The majority of these interventions were directed to parents or caregivers of children under the age of 6.11,12,19,28 In the remaining two studies, intervention was delivered to households with at least one child >5 years old23 and parents of children aged 12–13 years.16 As shown in Table 2, all studies showed a significant increase in knowledge following the interventions. In addition, four of the reviewed interventions also had a positive effect on parents’ antimicrobial stewardship behaviour. Cebotarenco and Bush16 found that parents in the intervention group were 5.2 times more likely than other parents to indicate they had not taken an antibiotic for colds or flu (P < 0.001). In Trepka et al.19 the proportion of parents who expected an antibiotic for their child and did not receive one declined in the intervention area from 14% to 9%, whereas it increased from 7% to 10% in the control area (P = 0.003), and the percentage of parents reporting that they brought their child to another physician because they did not receive an antibiotic decreased from 5% to 2% in the intervention area and increased from 2% to 4% in the control area (P = 0.02). Larson et al.23 found that the percentage of participants reporting using alcohol hand sanitizers increased from 1.4% to 66.8% following the intervention (P = 0.001) and the percentage of those reporting that at least one member of their household had been vaccinated against influenza increased from 63.7% to 73.9% (P = 0.001). Stockwell et al.,28 on the other hand, demonstrated that the number of parents reporting that they sought antibiotics without a prescription when their child was sick decreased from six to one (P = 0.06). General public The general public were the population of interest in eight of the included studies.13,14,18,21,24,25,29,30 Apart from Curry et al.,21 who used printed materials in the form of posters and leaflets, all studies were mass media campaigns, including four studies that measured the effects of the national campaign intervention.18,21,25,29 Five studies demonstrated a significant effect on the general public’s knowledge and attitudes14,21,24,25,29 (Table 2). With respect to antimicrobial stewardship behaviour amongst the public, four studies reported a significant effect following the intervention.14,21,25,29 Gonzales et al.14 found that visits to paediatricians declined in the intervention group for all conditions, but mostly for acute respiratory infections (P = 0.01). Similarly, Curry et al.21 demonstrated a significant decrease in the number of respondents who reported consulting a doctor for the common cold (P = 0.026). The results of Wutzke et al.29 showed that significantly fewer participants reported using antibiotics for cough, cold or flu following the intervention (7.4%) in comparison with baseline data (10.8%; percentage point change = 3.4; 95% CI = 1.3–5.5). Mazińska and Hryniewicz,25 on the other hand, demonstrated a significant increase in the percentage of respondents who had limited the use of antibiotics (from 27% at baseline to 43% post-intervention), had become more disciplined and cautious in their use (from 3% to 24%) and who had paid attention to the correct dosage (from 6% to 18%; no P values given). The remaining three studies did not show a significant positive effect on outcomes of interest.13,18,30 In Mainous et al.30 an intervention designed to decrease self-medication with antibiotics surprisingly resulted in a significantly greater percentage of the intervention Latino community group using antibiotics without a prescription in comparison with the control group (OR = 1.81; 95% CI = 1.02–3.22). McNulty et al.,18 on the other hand, found no positive effect on participants’ knowledge or antimicrobial stewardship behaviour following a national campaign and there was a significant increase in the percentage of respondents from the intervention area who reported retaining leftover antibiotics (P < 0.001). Formoso et al.13 reported that knowledge consistency with the national campaign messages either worsened (P < 0.05) or did not improve in both the intervention and control groups after the intervention. Discussion Main findings of this study This systematic review provides an in-depth examination of the effectiveness of interventions that target the public to increase their knowledge and understanding of AMR and engagement with antimicrobial stewardship behaviours. We have also identified patterns between target populations and relative intervention effectiveness. The findings present a complex picture reflecting the heterogeneity of the studies. Our analysis has shown that interventions targeting schoolchildren and parents have notable potential. All interventions that targeted schoolchildren or parents showed a significant effect on the outcome of interest. However, effective school-based interventions tended to only have the effect of increasing knowledge. In addition, these studies measured only short-term outcomes. In contrast, interventions targeting parents demonstrated changes in behaviour in addition to knowledge, with the follow-up period ranging from 2 weeks28 to 3 years.12 With regard to the interventions targeting the general public, the picture is less clear. Although the majority (n = 5) of these studies demonstrated effectiveness of interventions in improving the public’s AMR knowledge or their antimicrobial stewardship behaviour, three studies did not, with two showing a decrease in AMR knowledge13 and in antimicrobial stewardship behaviour.30 These findings highlight the need to examine differences in content between these interventions targeting the general public. Patterning of effectiveness across the types of target population also suggests that different target populations should receive different interventions with different primary outcomes. Nevertheless, targeting children alone is unlikely to make a major contribution to AMR because attitudes and the public’s antimicrobial stewardship behaviours may be passed down through generations. Thus, using the power of familial social influence and parental duty, in which children’s AMR education within school is reinforced and boosted by parental interventions in the home, might be a more appropriate approach for the achievement of desired cultural change. This indicates the potential of a multimodal intervention or programmatic approach to AMR-related interventions. An ideal approach would be to address the entire population simultaneously, but segment it to target sub-populations. Through such segmentation, or stratification of the general public, diverse tailored interventions addressing different sub-populations would be a strategic way to begin the process of cultural change required to reduce the drivers of AMR. The nature of the increase in knowledge that is needed can also be specified by drawing on other evidence syntheses that have shown that the public’s AMR knowledge and understanding of their contribution to AMR are generally poor.3 Therefore, in addition to changing the public’s understanding of appropriate antimicrobial use, interventions should also target the public’s understanding of AMR to enable the public to understand their central role in tackling AMR and the risks for the intervention recipient, their loved ones and the wider population. Findings in relation to other research In their recent paper, Wells and Piddock31 argued that amongst other actions, an urgent review of educational campaigns is required in order to fulfil UK and European AMR action plans. Our review addresses this need. Furthermore, to our knowledge this is the first systematic review that provides such an in-depth examination of the effectiveness of AMR-related interventions that target the public specifically. Previous literature focused on the level of the public’s AMR knowledge and beliefs,3,4 communication interventions or interventions that target both the public and healthcare professionals.6 The latter found that multi-component interventions improve the public’s knowledge of appropriate antimicrobial use, specifically in relation to antibiotics, and that interventions including both physician and public education appear to be effective in reducing antibiotic use.6 Similarly, Cross et al.32 reported that multi-modal communication interventions targeting both the public and clinicians can reduce antibiotic prescribing in high-income countries. Although our review focused on the general public population specifically, the potential of multi-faceted interventions was also highlighted in our work. Another previous systematic review, by King et al.7 reviewed the evidence of effectiveness and cost-effectiveness of interventions changing the public’s risk-related behaviours in relation to antimicrobial use. The review showed that direct contact education interventions were consistently more effective than mass media interventions.7 This appears to explain our findings on the varying effectiveness of interventions targeting the general public, as the majority of these studies used mass media interventions. There is also a body of evidence on large-scale antibiotic campaigns that were not eligible for inclusion in our review as the participants were members of both the public and healthcare professions. A literature review showed that there have been numerous multifaceted antibiotic awareness campaigns launched in high-income countries;33 however, there was substantial heterogeneity in outcomes, including knowledge and awareness, use of antibiotics and AMR, and the interventions themselves often lack a robust grounding in behavioural and social science theory. The majority of campaigns included in the review targeted both the general public and healthcare professionals simultaneously and they appeared to result in a reduction in antibiotic use.33 It therefore appears that targeting different populations at the same time might result in desired outcomes as the healthcare professional’s prescribing decisions might also be influenced by the patient, while the patient’s behaviour might be affected by the prescriber’s advice. One such campaign, conducted in the UK in 2014, simultaneously targeted members of the public and healthcare professionals who pledged as Antibiotic Guardians and showed an increase in AMR knowledge and commitment to pledge behaviour in both surveyed sub-populations.34 Another antibiotic awareness campaign, conducted in Hong Kong, targeted the general public, patients and healthcare professionals in a segmented fashion and resulted in a significant improvement (P ≤ 0.002) in respondents’ knowledge on prudent use of antibiotics following the campaign.35 Yet another successful large-scale antibiotic awareness campaign segmented to target the general public and healthcare professionals was conducted in France.36 Evaluation of the effectiveness of this campaign in reducing the number of antibiotic prescriptions showed a −26.5% (95% CI = −33.5% to −19.6%) decrease in the total number of antibiotic prescriptions following the campaign, with the greatest decrease in prescriptions issued for children (−35.8%; 95% CI = −48.3% to −23.2%) and young adults in the 21–25 years age group (24.1% decrease; CI not provided).36 These findings further emphasize the potential of a programmatic approach to AMR-related interventions segmented to different target sub-populations, as suggested above in the section Main findings of this study. Strengths and limitations We have conducted a rigorous search and systematic review accompanied by a narrative synthesis. Although similar work concerning the effectiveness of interventions aiming to improve antibiotic use has been conducted previously,32 our work focused on interventions targeting the general public population exclusively and did not include outcomes related to healthcare professionals’ AMR awareness or antimicrobial stewardship, such as antibiotic prescribing. Our analysis provides a sense of what is normative within this field, what has been attempted before and what could be repeated. It also provides a unique and valuable contribution to the available literature. However, the study also has limitations. First, because the UK Antimicrobial Resistance Strategy and Action Plan was launched by the Department of Health in 2000,37 followed by the publication of the WHO Global Strategy for Containment of Antimicrobial Resistance in 2001,1 we limited our search to publications from 2000 onwards. This could result in omission of important, older papers. Second, studies from low- and middle-income countries were underrepresented in our review. Thus, relevance and applicability of our findings to different geographical areas or resource contexts is limited. Third, the risk of bias was assessed only for studies that met the EPOC study design criteria. However, a suitable, validated tool for assessing the risk of bias of non-controlled before-and-after studies could not be identified. Furthermore, using different instruments could result in ambiguities in relation to the quality of stronger designs. The overall quality of the evidence was rather low. Major problems were associated with randomization in experimental designs and the evaluation of mass media and other population-level interventions. As these kinds of interventions aim for maximum population reach, it is difficult to attain adequate controls or indeed randomize at this population level. Therefore, good-quality study designs are systematically less likely to be identified within this kind of population-level intervention literature. Notwithstanding this, there was considerable heterogeneity in outcomes. There are no standardized ways of measuring the public’s AMR-related knowledge or associated stewardship behaviours. Furthermore, change in knowledge, awareness or beliefs, which were the most common outcome measures across the included studies, might not necessarily lead to desired behaviour change. As a result, it is particularly challenging to build cumulative knowledge regarding the effectiveness of interventions to increase the public’s engagement with antimicrobial stewardship. Another limitation is that given the problems with the quality of primary research, our measure of relative effectiveness should be treated with caution as this was based on our relative measure and is not equivalent to a strong evidence base within typical evidence-based guidance. Finally, we did not conduct an analysis of the cost-effectiveness of reviewed interventions; however, for the majority of studies included in our review, cost-effectiveness data were not reported. Recommendations for future research Although our work demonstrated the potential of intervention that targets particular sub-populations of the general public, taking into account the low quality of reviewed evidence, lack of cost-effectiveness evaluation and underrepresentation of studies from low- and middle-income countries, these findings must be treated with caution. There is a need for well-designed, randomized experimental studies focusing on behavioural outcomes of the interventions. Furthermore, measures of AMR knowledge and stewardship behaviours need to be standardized and there is a need for improvement of the reporting standards to ensure detailed and transparent reporting of intervention components. Finally, considering the underrepresentation of studies from low- and middle-income countries, there is a need for the development and evaluation of similar interventions within such settings. Conclusions Although some evidence of the effectiveness of interventions that target the general public in engaging with the problem of AMR exists, the public’s understanding of AMR and their role in combating this problem remains poor. Thus, there is a need for a cultural change and effective engagement of the public in addition to other key stakeholders. This need could be addressed through development of well-designed AMR-related interventions robustly grounded within behavioural and social science theory. Our work provided an in-depth examination of the effectiveness of AMR-related interventions targeting the members of the public specifically. We suggest that future policy makers should consider multimodal segmented population-level intervention that tailors its core messages to children, parents and the wider general public alike, particularly in high-income geographical areas. Future interventions should convey messages that elicit the public’s motivation to make their own efforts to address AMR as a growing problem for all and a problem for the present as much as for the future. Funding This work was supported by Health Protection Scotland. Glasgow Caledonian University was responsible for initiating, managing and sponsoring this review (Research, Innovation and Enterprise 15-127). Transparency declarations None to declare. Supplementary data Tables S1 to S3 are available as Supplementary data at JAC Online. References 1 WHO . WHO Global Strategy for Containment of Antimicrobial Resistance. 2001 . http://apps.who.int/iris/bitstream/10665/66860/1/WHO_CDS_CSR_DRS_2001.2.pdf. 2 O’Neill J. Tackling Drug-Resistant Infections Globally: Final Report and Recommendations. https://amr-review.org/sites/default/files/160525_Final%20paper_with%20cover.pdf. 3 Gualano MR , Gili R , Scaioli G et al. General population’s knowledge and attitudes about antibiotics: a systematic review and meta‐analysis . Pharmacoepidemiol Drug Saf 2015 ; 24 : 2 – 10 . Google Scholar CrossRef Search ADS PubMed 4 McCullough A , Parekh S , Rathbone J et al. A systematic review of the public’s knowledge and beliefs about antibiotic resistance . J Antimicrob Chemother 2015 ; 71 : 27 – 33 . Google Scholar CrossRef Search ADS PubMed 5 Pinder R , Sallis A , Berry D et al. Behaviour Change and Antibiotic Prescribing in Healthcare Settings. Literature Review and Behavioural Analysis. PHE. https://www.gov.uk/government/publications/antibiotic-prescribing-and-behaviour-change-in-healthcare-settings. 6 Haynes C , McLeod C. A Review of Reviews of Educational Interventions Designed to Change the Public’s Knowledge and Behaviour in Relation to Antimicrobial Use and Antimicrobial Resistance that Target Healthcare Professionals and Patients. https://www.nice.org.uk/guidance/ng63/documents/antimicrobial-resistance-changing-riskrelated-behaviours-in-the-general-population-evidence-review-32. 7 King S , Exley J , Taylor J et al. Antimicrobial stewardship: the effectiveness of educational interventions to change risk-related behaviours in the general population: a systematic review . Rand Health Q 2016 ; 5 : 2 . Google Scholar PubMed 8 Moher D , Liberati A , Tetzlaff J et al. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement . PLoS Med 2009 ; 6 : e1000097. Google Scholar CrossRef Search ADS PubMed 9 Cochrane Effective Practice and Organisation of Care (EPOC) . EPOC Resources for Review Authors. http://epoc.cochrane.org/resources/epoc-resources-review-authors. 10 Ryan R , Cochrane Consumers and Communication Review Group . Cochrane Consumers and Communication Review Group: Data Synthesis and Analysis. http://cccrg.cochrane.org/sites/cccrg.cochrane.org/files/public/uploads/meta-analysis_revised_december_1st_1_2016.pdf. 11 Croft DR , Knobloch MJ , Chyou P et al. Impact of a child care educational intervention on parent knowledge about appropriate antibiotic use . WMJ 2007 ; 106 : 78 – 84 . Google Scholar PubMed 12 Huang SS , Rifas-Shiman SL , Kleinman K et al. Parental knowledge about antibiotic use: results of a cluster-randomized, multicommunity intervention . Pediatrics 2007 ; 119 : 698 – 706 . Google Scholar CrossRef Search ADS PubMed 13 Formoso G , Paltrinieri B , Marata AM et al. Feasibility and effectiveness of a low cost campaign on antibiotic prescribing in Italy: community level, controlled, non-randomised trial . BMJ 2013 ; 347 : f5391. Google Scholar CrossRef Search ADS PubMed 14 Gonzales R , Corbett KK , Wong S et al. “Get smart Colorado”: impact of a mass media campaign to improve community antibiotic use . Med Care 2008 ; 46 : 597 – 605 . Google Scholar CrossRef Search ADS PubMed 15 Pontes MC , Pontes NM. Debiasing effects of education about appropriate antibiotic use on consumers’ preferences for physicians . Health Care Manage Rev 2005 ; 30 : 9 – 16 . Google Scholar CrossRef Search ADS PubMed 16 Cebotarenco N , Bush PJ. Reducing antibiotics for colds and flu: a student-taught program . Health Educ Res 2008 ; 23 : 146 – 57 . Google Scholar CrossRef Search ADS PubMed 17 Lecky DM , McNulty CA , Touboul P et al. Evaluation of e-Bug, an educational pack, teaching about prudent antibiotic use and hygiene, in the Czech Republic, France and England . J Antimicrob Chemother 2010 ; 65 : 2674 – 84 . Google Scholar CrossRef Search ADS PubMed 18 McNulty CA , Nichols T , Boyle PJ et al. The English antibiotic awareness campaigns: did they change the public’s knowledge of and attitudes to antibiotic use? J Antimicrob Chemother 2010 ; 65 : 1526 – 33 . Google Scholar CrossRef Search ADS PubMed 19 Trepka MJ , Belongia EA , Chyou PH et al. The effect of a community intervention trial on parental knowledge and awareness of antibiotic resistance and appropriate antibiotic use in children . Pediatrics 2001 ; 107 : e6. Google Scholar CrossRef Search ADS PubMed 20 Azevedo MM , Pinheiro C , Yaphe J et al. Assessing the impact of a school intervention to promote students’ knowledge and practices on correct antibiotic use . Int J Environ Res Public Health 2013 ; 10 : 2920 – 31 . Google Scholar CrossRef Search ADS PubMed 21 Curry M , Sung L , Arroll B et al. Public views and use of antibiotics for the common cold before and after an education campaign in New Zealand . N Z Med J 2006 ; 119 : U1957 . Google Scholar PubMed 22 Farrell D , Kostkova P , Weinberg J et al. Computer games to teach hygiene: an evaluation of the e-Bug junior game . J Antimicrob Chemother 2011 ; 66 Suppl 5: v39 – 44 . Google Scholar CrossRef Search ADS PubMed 23 Larson EL , Ferng YH , McLoughlin JW et al. Effect of intensive education on knowledge, attitudes, and practices regarding upper respiratory infections among urban Latinos . Nurs Res 2009 ; 58 : 150 – 7 . Google Scholar CrossRef Search ADS PubMed 24 Madle G , Kostkova P , Mani-Saada J et al. Changing public attitudes to antibiotic prescribing: can the internet help? J Innov Health Inform 2004 ; 12 : 19 – 26 . Google Scholar CrossRef Search ADS 25 Mazińska B , Hryniewicz W. Kampania edukacyjna Europejski Dzień Wiedzy o Antybiotykach–czy wpłynęła na zmianę postaw społeczeństwa w Polsce . Pol Merkur Lekarski 2010 ; 29 : 296 – 303 . Google Scholar PubMed 26 McNulty CA , Bowen J , Gelb D et al. “The Bug Investigators”: assessment of a school teaching resource to improve hygiene and prudent use of antibiotics . Health Educ 2007 ; 107 : 10 – 26 . Google Scholar CrossRef Search ADS 27 McNulty CA , Swan AV , Boland D. Schools’ antimicrobial resistance: National Advice to the Public campaign – a pilot study . Health Educ 2001 ; 101 : 235 – 42 . Google Scholar CrossRef Search ADS 28 Stockwell MS , Catallozzi M , Meyer D et al. Improving care of upper respiratory infections among Latino Early Head Start parents . J Immigrant Minority Health 2010 ; 12 : 925 – 31 . Google Scholar CrossRef Search ADS 29 Wutzke SE , Artist MA , Kehoe LA et al. Evaluation of a national programme to reduce inappropriate use of antibiotics for upper respiratory tract infections: effects on consumer awareness, beliefs, attitudes and behaviour in Australia . Health Promot Int 2007 ; 22 : 53 – 64 . Google Scholar CrossRef Search ADS PubMed 30 Mainous AG , Diaz VA , Carnemolla M. A community intervention to decrease antibiotics used for self-medication among Latino adults . Ann Fam Med 2009 ; 7 : 520. Google Scholar CrossRef Search ADS PubMed 31 Wells V , Piddock LJV. Addressing antimicrobial resistance in the UK and Europe . Lancet Infect Dis 2017 ; 17 : 1230 – 1 . Google Scholar CrossRef Search ADS PubMed 32 Cross ELA , Tolfree R , Kipping R. Systematic review of public-targeted communication interventions to improve antibiotic use . J Antimicrob Chemother 2016 ; 72 : 975 – 87 . 33 Huttner B , Goossens H , Verheij T et al. Characteristics and outcomes of public campaigns aimed at improving the use of antibiotics in outpatients in high-income countries . Lancet Infect Dis 2010 ; 10 : 17 – 31 . Google Scholar CrossRef Search ADS PubMed 34 Chaintarli K , Ingle SM , Bhattacharya A et al. Impact of a United Kingdom-wide campaign to tackle antimicrobial resistance on self-reported knowledge and behaviour change . BMC Public Health 2016 ; 16 : 393 . Google Scholar CrossRef Search ADS PubMed 35 Ho ML , Cowling BJ , Seto WH et al. Determinants of an effective antibiotic campaign: lessons from Hong Kong . J Glob Antimicrob Resist 2014 ; 2 : 334 – 7 . Google Scholar CrossRef Search ADS PubMed 36 Sabuncu E , David J , Bernède-Bauduin C et al. Significant reduction of antibiotic use in the community after a nationwide campaign in France, 2002–2007 . PLoS Med 2009 ; 6 : e1000084. Google Scholar CrossRef Search ADS PubMed 37 Department of Health . UK Antimicrobial Resistance Strategy and Action Plan. http://antibiotic-action.com/wp-content/uploads/2011/07/DH-UK-antimicrobial-resistance-strategy-and-action-plan.pdf. © The Author(s) 2018. Published by Oxford University Press on behalf of the British Society for Antimicrobial Chemotherapy. All rights reserved. For Permissions, please email: journals.permissions@oup.com. This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Journal of Antimicrobial ChemotherapyOxford University Press

Published: Mar 14, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off