Marked increase in measles vaccination coverage among young adults in Switzerland: a campaign or cohort effect?

Marked increase in measles vaccination coverage among young adults in Switzerland: a campaign or... Objectives To evaluate the impact of the Swiss measles elimination strategy—including a mass media campaign—on vaccination coverage and awareness among young adults aged 20–29 years. Methods Comparison of the results of two cross-sectional population surveys in 2012 and 2015. Results Documented vaccination coverage increased from 77 to 88% for two doses of measles vaccine. Major determinants of complete vaccination were survey year, birth cohort, sex and the absence of prior measles disease. If birth cohort and prior history of measles disease are included as factors in a multivariate model, the difference between 2012 and 2015 vanishes. Conclusions The marked increase in complete measles vaccination coverage is due to a cohort effect, owing to the introduction of the second dose of vaccine in 1996. Most of the vaccinations were administered before the national strategy was implemented and vaccination catch-ups did not increase during the campaign in young adults. Nevertheless, this study provides evidence of an improvement in the awareness of measles and measles vaccination in young adults, which may result in an impact on measles vaccination coverage in the near future. Keywords Measles  Vaccination  Vaccination coverage  Epidemiology  Survey Introduction Despite a relatively high vaccination coverage of 93% with two doses during the years 2014–2016 in the 16-year- Measles is a highly transmissible infectious disease. Its olds, measles is still an endemic disease in Switzerland. complications are particularly frequent and severe in The annual baseline incidence is low at 3–10 cases per infants and adults. However, it is possible to prevent the million people. However, a large outbreak between 2007 disease effectively with vaccination. The Swiss Federal and 2009 led to over 4000 cases (Richard and Masserey Office of Public Health (FOPH) has recommended the Spicher 2009). On this occasion, Switzerland also exported vaccination of 12-month-old children against measles since measles, causing outbreaks in various European countries 1976, and has recommended the combined vaccine against and the United States. In response to this national outbreak measles, rubella and mumps since 1985. In 1996, a second and to achieve the WHO’s goal of eliminating measles in dose was added to the vaccination schedule for 4–7-year- the European region by 2015, the FOPH together with the old children; in 2001, the recommended age for this second cantons (states) and other public health partners launched a dose was reduced to 15–24 months. national strategy for 2011–2015 to eliminate measles (Bundesamt fu ¨ r Gesundheit 2013, 2016a). One of the major objectives of this strategy was to raise public awareness of measles and vaccination against this disease to sustainably Electronic supplementary material The online version of this article increase two-dose vaccination coverage to at least 95% in (https://doi.org/10.1007/s00038-018-1102-x) contains supplementary material, which is available to authorized users. children and adolescents, including through catch-up vac- cinations for individuals born after 1963. The strategy & Ekkehardt Altpeter consisted of six areas of intervention, including facilitating ekkehardt.altpeter@bag.admin.ch and encouraging catch-up vaccinations in the over 2-year- Division of Communicable Diseases, Federal Office of Public olds to fill vaccination gaps, as well as communication and Health, Berne, Switzerland 123 590 E. Altpeter et al. promotion to raise public awareness of measles and vac- Participants, recruitment and sample size cination against this disease. Although vaccination coverage for measles has contin- Inclusion criteria for participation in the study were: age ually increased in children of all age groups since 2000, the 20–29 years, address within Switzerland recorded in a national target of 95% of people vaccinated with two doses commercial directory, and ability to answer questions in has not yet been reached. According to the last cantonal German, French or Italian. The survey company used a surveys of 2-, 8- and 16-year-olds during the period commercial household directory, which covers about 95% 2014–2016 (Swiss National Vaccination Coverage Sur- of all Swiss private households. Stratified random sampling vey—SNVCS), national coverage for a single dose was based on two (baseline) or three (endline) language regions 94–96% depending on age group, but only 87–93% for two and municipality size was applied. Each selected household doses (Bundesamt fu ¨ r Gesundheit 2016b). However, a few received an invitation letter from the FOPH to announce cantons did reach 95% for two doses. Vaccination coverage the study and its background. A total of 8645 invitation of adults is not routinely monitored in Switzerland and is letters were sent. For households without a registered unknown. telephone number, the letter requested for it. A free tele- The aim of this study was to assess the effect of the phone hotline for participants who wished to know more measles campaign in regard to improved vaccination cov- about the study was in place during both study periods. erage in young adults by catch-up vaccinations and The telephone contact started with a screening for increased awareness. With two consecutive surveys, we willingness to participate and to assess eligibility by age. If assessed the vaccination coverage as well as awareness of several persons within a household fulfilled the age criteria, measles in 20–29-year-olds in 2012 before the campaign one was selected at random. At the end of the telephone (baseline) and in 2015 afterwards (endline). interview, respondents were asked to send their vaccination card or a copy of it to the FOPH. The originals were returned after data collection. Those consenting to send in Methods their vaccination card received an instruction letter with a stamped and addressed envelope and an incentive of CHF Study design 10. If the FOPH did not receive the vaccination card within 3 weeks, a reminder was sent. To get the vaccination cards, We designed two identical nationally representative cross- a total of 2717 letters were sent followed by 1373 sectional telephone surveys before and after the campaign reminders. in 2012 and 2015, respectively, conducted by the same We anticipated a ratio of one vaccination card received commercial telephone survey company, using identical per two completed telephone interviews. The number of sampling methods and assessment tools for both baseline attempted interviews was calculated on this basis. The and endline survey. The same professional telephone sur- target sample size was 350 interviews with matching vac- vey company conducted recruitment and data collection for cination cards per language region. We aimed at a total both surveys. The interview included the same questions sample size for interviews with matching vaccination cards regarding: participants’ self-reported measles vaccination of 700 in the baseline with two language regions and 1050 status, reasons for receiving measles catch-up vaccination, in the endline with three language regions. willingness to be vaccinated if vaccinations were missing and measles disease history, as well as knowledge of the Statistical analyses disease and the elimination campaign (for the wording of the questions see Table 3). All participants were asked to We calculated sampling weights according to the same send a copy of their vaccination card to the Federal Office protocol for both surveys: The sampling frame was a of Public Health. The average interview time was about database that covers 95% of all households in Switzerland. 7 min. The questionnaire was available in German, French Weights accounted for each language region and were then and Italian. The sampling procedure and questionnaire calibrated to the population total considering sex and were piloted with 38 completed interviews. The baseline canton (Lumley 2004, 2014a). survey took place between April 23 and July 22, 2012 and We combined the data of the two surveys and defined an the second survey took place between November 9 and indicator variable identifying the 2012 survey as zero and December 18, 2015. the 2015 survey as one. We will call this variable ‘survey’ The two surveys were conducted under the Swiss Epi- in the following. Since we did not sample data from the demics Act of December 18, 1970, thus ethical committee Italian-speaking part of Switzerland in the 2012 survey, we approval was not required. discarded these data from the 2015 survey for this analysis. 123 Marked increase in measles vaccination coverage among young adults in Switzerland: a… 591 Only datasets with a vaccination card were considered for We evaluated the primary endpoint in a univariate the analysis, resulting in an overall total N of 1851 (Fig. 1). analysis and then included all variables with a significant The primary endpoint was defined as documented vac- effect in a multivariate logistic regression model consid- cination with at least two doses of measles vaccine. We ering the survey weights and calculating odds ratios (OR) defined two cohorts based on the median birth year of the and their 95% CI. This analysis assumes ‘missing com- complete study population: one born in 1990 or later pletely at random’ (MCAR). formed the reference group, while the other cohort included All logistic regression models (weighted as well as respondents born before 1990. In a sensitivity analysis, we unweighted) were estimated with and without multiple included annual birth cohort as a factor into the analyses. imputations assuming a joint model (Quartagno and Car- We recoded nationality for 2015 to the standard of 2012, penter 2016). The model included all categorical variables since in 2012 we had two groups (Swiss and foreign as well as age and weight as continuous variables. We nationality) and in 2015 three groups (Swiss, foreign and estimated five imputed datasets and combined the results both). We defined people with both nationalities as Swiss. by means of Rubin’s rule (Lumley 2014b). We analysed the data excluding and including the We tabulated the documented vaccine coverage by sampling weights. For unweighted analysis, we used cohort and the variable ‘survey’. We determined the vac- standard tabulations and standard inference. The level of cination coverage at age 16-years by cohort and compared significance was set to alpha \ 0.05. We calculated 95% these estimates with coverage of the SNVCS for that age. confidence intervals (CI) where applicable. For the We further derive the catch-up vaccination by subtracting weighted analysis, we described the data by means of the overall cohort-specific vaccine coverage at the age of tabulation and calculated proportions (Thomas Lumley 16 years from the overall cohort-specific vaccine coverage 2004, 2014a). at the time of the interview. We restricted our analysis to the birth cohorts 1986–1992 which overlap between the randomly selected households based on an address directory with sex and age attributes 2012 N = 4519 2015 N = 4126 phone number invalid (out of phone number valid service, fax) 2012 N = 4095 2012 N = 424 2015 N = 3798 2015 N = 328 contact with household no contact established (no answer, line established occupied, answering machine) 2012 N = 886 2012 N = 3209 2015 N = 2864 2015 N = 934 communication not possible (language communication possible barrier) 2012 N = 3054 2012 N = 155 2015 N = 2827 2015 N = 37 eligible for age not eligible for age 2012 N = 2132 2012 N = 922 2015 N = 2294 2015 N = 533 interview not completed interview completed 2012 N = 711 2012 N = 1421 2015 N = 771 2015 N = 1521 vaccinaon card vaccinaon card exploitable not exploitable 2012 N = 803 2012 N = 618 2015 N = 1048 2015 N = 473 Fig. 1 Data flow chart, measles surveys 2012 and 2015, Switzerland 123 592 E. Altpeter et al. two surveys and compared the vaccination coverage for at Table 1 Demographic characteristics, measles surveys 2012 and 2015, Switzerland least two doses at the time of interview between the survey 2012 and 2015. We included annual birth cohort as a factor Survey Survey Total p value in a logistic regression model. 2012 2015 To evaluate the effect of absent vaccination cards, we N % N % N % tabulated its absence and presence, respectively, with Sex respect to the following variables: ‘survey’, sex, age, lan- Female 448 56 528 50 976 53 guage region, anamnestic vaccination status, prior measles Male 355 44 520 50 875 47 history and estimated a linear unweighted logistic regres- Total 803 100 1048 100 1851 100 \ 0.05 sion model. Age group We performed all analyses using R for Windows 3.1.3, \ 25 years 403 50 687 66 1090 59 respectively, 3.3.1 under Windows 7, made available by [ 24 years 400 50 361 34 761 41 the R Core Team (2015). Total 803 100 1048 100 1851 100 \ 0.05 Birth cohort Results C 1990 226 28 768 73 994 54 \ 1990 576 72 280 27 856 46 Response rate and basic characteristics Missing 1 0 0 0 1 0 of participants Total 803 100 1048 100 1851 100 \ 0.05 Nationality The response rate for completed interviews was 50% (2012 Swiss 751 94 982 94 1733 94 survey 50%, 2015 survey 49%) and for interviews with Foreign 52 6 66 6 118 6 matching vaccination card 31% (2012 survey 28%, 2015 Total 803 100 1048 100 1851 100 0.87 survey 34%). These proportions are calculated as follows: Language region the proportion of observed eligible subjects was 2132 out German 430 54 557 53 987 53 of 3054 (70%) in 2012 and 2294 out of 2827 (81%) in French 373 46 491 47 864 47 2015; these proportions were multiplied by the number of Total 803 100 1048 100 1851 100 0.86 subjects with whom no contact could be established Educational level (N = 886 in 2012 and N = 934 in 2015) or with whom no Secondary 565 70 776 74 1341 72 communication was possible (N = 155 in 2012 and N =37 Tertiary 236 29 271 26 507 27 in 2015), e.g. 1041 in 2012 and 971 in 2015; these esti- Missing 2 0 1 0 3 0 mates were added to the number of observed eligible per- Total 803 100 1048 100 1851 100 0.08 sons, leading to the estimated number of eligible subjects: Having own (1041 times 2132/3054) plus 2132 in 2012 and (971 times children 2294/2827) plus 2294 in 2015, i.e. 100% of the response Yes 125 16 64 6 189 10 rate (Fig. 1). The complete combined dataset includes 2942 No 678 84 984 94 1662 90 observations (2011 survey N = 1421, 2015 survey Total 803 100 1048 100 1851 100 \ 0.05 N = 1521). We included 1851 observations with valid vaccination documents in the analysis (2012 survey: N = 803, 43%; 2015 survey: N = 1048, 57%).The two of being vaccinated with more than one dose (Table 3). survey populations were slightly different with respect to Most participants received their first measles vaccination birth year, age, sex and having children (Table 1). according to the recommendations: the median age at We validated the question about vaccination by com- uptake was 1 year, both in 2012 (interquartile range paring the information given by memory and written on the 1–2 years, N = 729) and in 2015 (interquartile range vaccination card. The agreement was poor (see Table 2). 1–1 year, N = 1006). The median age at the second vac- cination was 13 years in 2012 (interquartile range Vaccination coverage 10–15 years, N = 610) and 10 years in 2015 (interquartile range 7–13 years, N = 923). Very few second doses and Vaccination coverage with at least two doses increased even fewer first doses were received between the ages of 20 from 77% (95% CI 73–81%) in 2012 to 88% (95% CI and 29 years (Fig. 2) in both surveys. According to the 85–90%) in 2015. Only a minority was vaccinated with just 2015 survey, only 4% of respondents (37/932) received one dose and the proportion decreased from 16% (95% CI their second dose after 2011, the year a national measles 13–19%) in 2012 to 9% in 2015 (95% CI 7–12%) in favour vaccination campaign was launched, compared to 8% of 123 Marked increase in measles vaccination coverage among young adults in Switzerland: a… 593 Table 2 Agreement between anamnesis and documented measles vaccination by use of the vaccination document at interview, measles survey 2012 and 2015, Switzerland Documented number of doses 0 1 2 3 Total Vaccination document used during interview Have you been vaccinated against measles? Yes 5 (1.2%) 48 (11.8%) 304 (74.5%) 34 (8.3%) 391 (95.8%) No 5 (1.2%) 2 (0.5%) 6 (1.5%) 1 (0.2%) 14 (3.4%) I do not know 1 (0.2%) 0 (0%) 2 (0.5%) 0 (0%) 3 (0.7%) Total 408 (100%) Vaccination document not used during interview Have you been vaccinated against measles? Yes 26 (1.8%) 122 (8.4%) 886 (61.3%) 91 (6.3%) 1125 (77.8%) No 34 (2.4%) 19 (1.3%) 56 (3.9%) 2 (0.1%) 111 (7.7%) I do not know 13 (0.9%) 21 (1.5%) 159 (11.0%) 17 (1.2%) 210 (14.5%) Total 1446 (100.0%) respondents (46/610) during an equally long time span educational level remained statistically significant in the after 2008 for the 2012 survey (OR = 0.51, 95% CI multivariate analysis. Thus, the major determinants of 0.32–0.81). However, the catch-up needs for second doses complete measles vaccination were the year of the survey, 3 years before the start of the respective survey were lower i.e. respondents surveyed in 2015 were more likely to be in 2015 (151/932 = 16%) than in 2012 (228/610 = 37%) fully vaccinated than those surveyed in 2012; birth cohort, (OR = 0.32, 95% CI 0.25–0.41). Accordingly, these catch- i.e. persons born in 1990 or after were more likely to be up vaccinations with a second dose during the campaign fully vaccinated than those born before 1990; sex, i.e. covered 25% of the needs in 2015 respondents compared to females were more likely to be fully vaccinated than males; 20% in 2012 (OR = 1.28, 95% CI 0.76–2.16). There is a and prior measles, i.e. those who were sure that they had strong cohort effect on the complete vaccination coverage not had measles were more likely to be fully vaccinated (supplementary material Tables 3 and 4). than those with measles in their medical history (Table 4). Knowledge about vaccination Sensitivity analyses Awareness of measles and measles vaccination improved The sensitivity analysis either neglecting the weights, between 2012 and 2015 (Table 3). First, significantly more assuming ‘missing completely at random’ or ‘missing at participants knew that adults can get measles: 32% in 2012 random’ and their combination confirmed the results of the (95% CI 28–36%) were sure that adults can get measles weighted univariate and multivariate analysis (supple- compared to 41% in 2015 (95% CI 37–45%). Second, there mentary material Tables 1 and 2). was a trend, although not statistically significant, for more There is a strong cohort effect on the primary endpoint participants to be sure that they could get missing vacci- complete vaccination coverage. If the analysis is restricted nations as adults: from 18% in 2012 (95% CI 14–21%) to to the birth cohorts 1986–1992, the odds ratio between 23% in 2015 (95% CI 19–26%). There was practically no 2012 and 2015 is 1.40 (95% CI 0.87–2.10). In the multi- change with respect to the other questions on awareness. In variate analysis, this estimate drops to 1.29 (95% CI both surveys, a similarly low proportion had previously 0.82–2.03) (supplementary material Tables 3 and 4). contracted measles. About a quarter of participants had The participants with and without vaccination cards are professional contacts to risk groups (Table 3). different with respect to the factors ‘survey’, age, language region and anamnestic vaccination status (supplementary Determinants of vaccination material Table 5). There were more missing vaccination cards in 2012 compared to 2015. In both surveys, 82% of In the weighted univariate analysis, we identified the fol- respondents reported during the telephone interview to be lowing factors as significantly associated with complete completely vaccinated, which contrasted with information vaccination:‘survey’, birth cohort, sex, educational level gained from the vaccination cards. and prior measles (Table 4). All these factors except 123 594 E. Altpeter et al. Table 3 Vaccination coverage and knowledge about measles, measles surveys 2012 and 2015, Switzerland Survey 2012 Survey 2015 N % 95% CI N % 95% CI Documented measles vaccination Not vaccinated 51 7 5 9 30 3 2 4 1 dose 130 16 13 19 82 9 7 12 2 doses 577 71 67 75 836 78 75 81 3 doses 45 6 4 8 100 10 7 12 Vaccinated with at least two doses No 181 23 19 27 112 12 10 15 Yes 622 77 73 81 936 88 85 90 Prior measles disease Surely yes 55 7 5 10 82 8 6 10 Rather yes 31 5 3 7 32 3 2 5 Rather no 113 12 9 15 176 19 16 23 Surely no 570 70 65 74 723 65 61 69 I do not know 34 6 3 8 35 4 2 6 Can adults get measles? Surely yes 280 32 28 36 374 41 37 45 Rather yes 402 53 48 57 529 48 44 52 Rather no 56 7 5 9 72 6 4 8 Surely no 8 0 0 1 11 1 0 1 I do not know 57 8 5 10 62 5 3 6 Can adults catch up missing vaccinations? Surely yes 147 18 14 21 200 23 19 26 Rather yes 408 51 47 56 566 51 47 55 Rather no 100 14 11 17 132 14 11 17 Surely no 28 3 1 5 17 1 0 2 I do not know 120 14 11 17 133 11 9 13 I can endanger others if I am not vaccinated Complete agreement 345 42 38 47 451 44 41 48 Slight agreement 287 36 31 40 405 37 34 41 Slight disagreement 94 13 10 16 148 15 12 18 Complete disagreement 46 5 3 7 26 2 1 3 I do not know 31 4 2 6 18 1 0 2 Is vaccination against measles mandatory in Switzerland? Surely yes 41 5 3 7 60 5 3 6 Yes, I believe so 106 13 10 16 159 13 11 16 No, I do not believe so 281 32 28 36 410 37 34 41 Surely no 317 44 39 48 379 42 38 46 I do not know 58 6 4 7 40 3 2 4 123 Marked increase in measles vaccination coverage among young adults in Switzerland: a… 595 Table 3 (continued) Survey 2012 Survey 2015 N % 95% CI N % 95% CI Did you know about the collaboration of Switzerland with the WHO in eliminating measles? Yes 287 38 33 42 438 44 40 48 No 516 62 58 67 610 56 52 60 Is measles elimination necessary in Switzerland? Surely yes 332 38 34 43 361 34 30 38 Rather yes 285 35 31 39 393 37 33 41 Rather no 102 15 12 18 177 18 15 21 Surely no 40 5 3 7 44 5 3 6 I do not know 44 7 4 9 73 7 5 9 Do you have professional contact to children, pregnant women or sick persons? Yes 259 28 24 32 302 25 22 29 No 544 72 68 76 745 75 71 78 Missing 0 1 Percent taking into account the weighting. Deviations from the total of 100 per cent is due to rounding errors N total: 803 for measles survey 2012, 1048 for measles survey 2015 this study provide the first reliable estimates of measles 2 doses, survey 2012, N= 610 vaccination coverage in young adults in Switzerland. We assessed vaccination coverage and determinants of 1 dosis, survey 2012, N= 729 vaccination among young adults by comparing the results 2 doses, survey 2015, N= 932 of two cross-sectional surveys conducted in 2012 and 2015. 1 dosis, survey 2015, N= 1006 Documented complete vaccination increased from 77 to 88%. However, there is no evidence for an association between vaccination and the recent national awareness campaign or increased awareness. Knowing that measles can also affect adults was the only factor that increased between the two surveys, but this knowledge was not related to the likelihood of vaccination. Strengths and limitations 0 5 10 15 20 25 30 The careful sampling mechanism stratified by sex and age language region, the adequately high sample size of valid vaccination card-based data, the thorough statistical anal- Fig. 2 Age distribution at first and second dose of vaccination, 2012 ysis taking into account various sensitivity analyses for and 2015 measles surveys, Switzerland missing data and weighting, as well as application of the Discussion identical questionnaire tools make us confident of the quality of the estimate as well as the significant increase in Main results vaccination coverage between 2012 and 2015. However, the marked increase in vaccination coverage is very likely While routine monitoring of vaccination coverage is almost due to a cohort effect. This cohort effect, unexpected at the universally established for measles in children and ado- design stage of the studies, is so strong, that all multivariate lescents, it generally does not include adults. The results of analyses have to be interpreted with care. The observed percent 020 406080 596 E. Altpeter et al. Table 4 Logistic regression models of documented vaccination coverage with at least two doses, weighted and assuming missing completely at random, measles survey 2012 and 2015, Switzerland Univariate analysis Multivariate analysis OR 95% CI Adj. OR 95% CI Survey 2012 1 1 2015 2.15 1.54 3.01 1.67 1.15 2.42 Birth cohort C 1990 1 1 \ 1990 0.32 0.23 0.45 0.41 0.28 0.61 Sex Female 1 1 Male 0.59 0.42 0.81 0.58 0.42 0.82 Nationality Swiss 1 n.i. Foreign 1.15 0.59 2.23 n.i. Language region German 1 n.i. French 1.11 0.83 1.48 n.i. Educational level Secondary 1 1 Tertiary 0.60 0.43 0.86 0.77 0.53 1.12 Having own children Yes 1 n.i. No 1.48 0.96 2.31 n.i. Prior measles Surely yes 1 1 Yes, I believe so 1.43 0.62 3.28 1.69 0.71 4.05 No, I do not believe so 1.79 0.97 3.31 1.83 0.94 3.54 Surely not 2.33 1.40 3.89 2.46 1.43 4.23 I do not know 1.19 0.49 2.91 1.51 0.61 3.77 Can adults get measles? Surely yes 1 n.i. Rather yes 1.13 0.79 1.62 n.i. Rather no 0.98 0.52 1.86 n.i. Surely no 0.85 0.25 2.89 n.i. I do not know 0.93 0.50 1.74 n.i. Can adults catch up missing vaccinations? Surely yes 1 n.i. Rather yes 1.00 0.66 1.53 n.i. Rather no 0.77 0.44 1.36 n.i. Surely no 1.65 0.63 4.37 n.i. I do not know 0.89 0.51 1.55 n.i. I can endanger others if I am not vaccinated Complete agreement 1 n.i. Slight agreement 0.76 0.52 1.10 n.i. Slight disagreement 0.78 0.48 1.27 n.i. Complete disagreement 0.74 0.38 1.41 n.i. I do not know 1.60 0.40 6.45 n.i. 123 Marked increase in measles vaccination coverage among young adults in Switzerland: a… 597 Table 4 (continued) Univariate analysis Multivariate analysis OR 95% CI Adj. OR 95% CI Is vaccination against measles mandatory in Switzerland? Surely yes 1 n.i. Yes, I believe so 2.06 0.73 5.83 n.i. No, I do not believe so 0.96 0.38 2.45 n.i. Surely no 0.90 0.35 2.27 n.i. I do not know 1.33 0.43 4.09 n.i. Did you know about the collaboration of Switzerland with the WHO in eliminating measles? Yes 1 n.i. No 0.99 0.71 1.38 n.i. Is measles elimination necessary in Switzerland? Surely yes 1 n.i. Rather yes 1.06 0.72 1.57 n.i. Rather no 0.73 0.46 1.17 n.i. Surely no 1.43 0.67 3.04 n.i. I do not know 1.03 0.55 1.91 n.i. Do you have professional contact to children, pregnant women or sick persons? Yes 1 n.i. No 0.95 0.65 1.38 n.i. OR odds ratio, CI confidence interval, n.i. not included into the multivariate model cohort effect is confirmed by the comparison with the information about their vaccination status, provided by values from the external reference the SNVCS, which also participants not consulting their vaccination card, was often shows the increasing values over the study periods of the uncertain (see Table 2). We therefore restricted our anal- SNVCS. yses to individuals who provided a vaccination card, The high number of missing vaccination cards—an thereby reducing response rate from 50 to 31%. This could exclusion criterion of our study—could lead to a selection have introduced selection bias and led us to overestimate bias. The contrast between self-reported vaccination status the effect, because we assume that people without a vac- and documented vaccination history confirms our study cination card are less likely to be vaccinated. However, this design, which focused on the collection of vaccination was partially addressed by a complex weighting procedure, cards and restricted the analysis on these written records. a common method of dealing with the unit non-response. The return rate of vaccination cards (63%) was in the range After weighting, the study population had a similar distri- of the expectation during the design stage and was antici- bution to the Swiss general population in this age range in pated in the sample size calculation for interviews with regard to age, sex and cantonal distribution. matching vaccination cards. However, there is evidence Our results are in line with the vaccination coverage in that subjects with missing vaccination cards are different the 16-year-old in the SNVCS (Bundesamt fu ¨ r Gesundheit between the two surveys, with respect to age, language 2016b). They are as well comparable to findings from region, and history of measles vaccination. The process of Germany and France. In a study, conducted 2012 in Ger- differential missing data between 2012 and 2015 may lead many, vaccination coverage for two doses was found to be to an overestimation of the marked increase in vaccination 56% in adults aged 20–34 years (Schuster et al. 2015). 80% coverage. of these participants relied on their vaccination card to There are some critical points with respect to our anal- answer. Furthermore, the fact that in our study younger age ysis. The participants were asked about vaccinations that groups and females were more likely to be vaccinated is often occurred during their early infancy. Anamnestic consistent with the finding of Schuster et al. (2015). In 123 598 E. Altpeter et al. France, measles vaccination coverage (based on vaccina- causal pathway between knowledge, attitude and actual tion cards of first-year health care students) was 78% in vaccination; this leads to an inconclusive effect, a problem 2011 (Faure et al. 2013). In another French study in 2015, that was encountered by Larson et al. (2014). coverage was 93% in 16–18-year-olds (Buscail and Gag- niere 2016). Conclusion Our results show that the sharp increase of vaccination coverage in just 3 years was very likely due to a ‘‘cohort An 11% increase in measles vaccination coverage within a effect’’ resulting from a change in recommendations made 3-year period in a population where vaccination coverage in 1996: a second dose of measles vaccine was introduced is already high is a historic finding. This marked increase in to the national vaccination plan for 4–7-year-old children, measles vaccination coverage is, however, due to a cohort together with the recommendation to catch up at any age. effect, owing to the introduction of the second dose of Only four out of ten birth cohorts (1989–1992) among the vaccine in 1996 in the national vaccination plan. This study participants in the first survey had the opportunity to provides evidence of an improvement in the awareness receive a second dose as routine vaccination, compared to about measles and measles vaccination in young adults, seven cohorts (1989–1995) of participants in the second which may result in an impact on measles vaccination survey in 2015. In addition, it took a few years for the coverage in the future. The achievement of this marked second dose to become a well-established practice. The gap increase in measles vaccination coverage merits acknowl- of catch-up vaccinations with a second dose 3 years before edgment of the efforts made by general practitioners and the surveys was therefore larger in the group surveyed in paediatricians in Switzerland, who remain the main and the 2012 than that in 2015. most important promoters for measles vaccination. Most of these catch-up vaccinations were administered Governments at all levels are challenged to use their lee- before the age of 16 years. They were received before the way to support any efforts made by Swiss physicians to recent promotion campaign and should be credited to promote vaccination in general. paediatricians, school health services and parents. Although one quarter of the pre-campaign vaccination gaps Funding This study was fully funded by the Swiss government (Swiss were filled during the campaign, this proportion was not Federal Office of Public Health) based on the federal epidemic law significantly higher than during the 3 years before the (EpG, SR 818.101 of December 18, 1970). campaign suggesting that the campaign had little to no effect on the number of catch-up doses administered. Compliance with ethical standards We estimated the number of catch-up vaccinations administered by primary care physicians in a study within Conflict of interest The authors declare that they have no conflict of the Swiss Sentinella network (Bundesamt fu ¨ r Gesundheit interest. 2016c). In 2014 and 2015, estimated totals of 33,000 and 37,000 doses of measles catch-up vaccinations were Ethical approval We conducted this survey as an evaluation of ongoing surveillance of vaccination coverage according to the Swiss administered by general practitioners and paediatricians. law on communicable diseases in 2012 and 2015 (EpG, SR 818.101 This corresponds to 37% of all doses needed in 2015 to of December 18, 1970). close the estimated gap (with a 95% objective) in measles vaccination for 2–51-year-olds in Switzerland. Two-thirds Informed consent Informed consent was obtained from all individual participants included in the study. of these catch-up vaccinations were given to adults aged 20 years and older. Eighty-eight percent of all vaccinations Open Access This article is distributed under the terms of the Creative were initiated by the physicians, underlining their crucial Commons Attribution 4.0 International License (http://creative role in promoting vaccination and filling the gaps. commons.org/licenses/by/4.0/), which permits unrestricted use, dis- Awareness of measles has increased slightly in our study tribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a and can almost be considered as ‘‘common sense’’, with link to the Creative Commons license, and indicate if changes were 89% of all people knowing that measles can also affect made. adults. There mere fact that the study topic was measles elimination and the interviewees were primed by an invi- tation letter sent out by the FOPH may lead to overesti- References mating results about knowledge and attitude towards measles elimination. Knowledge and attitude questions Bundesamt fu ¨ r Gesundheit (2013) Nationale Strategie zur Maserne- limination 2011–2015. BAG Bull 17:269–271 were highly correlated in our survey. But the data were too Bundesamt fu ¨ r Gesundheit (2016a) Schweiz ohne Masern. BAG Bull scarce to be stratified by all variables and evaluated in an 5:93–95 integrated way. In addition, there is no simple and direct 123 Marked increase in measles vaccination coverage among young adults in Switzerland: a… 599 Bundesamt fu ¨ r Gesundheit (2016b) Durchimpfung von 2-, 8- und Lumley T (2004) Analysis of complex survey samples. J Stat Softw 16-jahrigen Kindern in der Schweiz, 1999–2016. https://www. 9(1):1–19 bag.admin.ch/bag/de/home/themen/mensch-gesundheit/ueber Lumley T (2014a) Survey: analysis of complex survey samples. R tragbare-krankheiten/impfungen-prophylaxe/informationen- package version 3.30 fachleute-gesundheitspersonal/durchimpfung.html. Accessed 22 Lumley T (2014b) mitools: Tools for multiple imputation of missing Nov 2017 data. R package version 2.3. http://CRAN.R-project.org/pack Bundesamt fu ¨ r Gesundheit (2016c) Nachholimpfung gegen Masern age=mitools. Accessed 22 Nov 2017 und Auffrischimpfung gegen Keuchhusten: wesentlicher Beitrag Quartagno M and Carpenter J (2016) jomo: a package for multilevel zur Masernelimination und zum Schutz von Sa ¨uglingen. BAG joint modelling multiple imputation. http://CRAN.R-project.org/ Bull 35:533–541 package=jomo. Accessed 22 Nov 2017 Buscail C, Gagnie `re B (2016) Vaccination coverage of adolescents: Richard JL, Masserey Spicher V (2009) Large measles epidemic in results of a Defense and Citizenship Day-based survey. Med Mal Switzerland from 2006 to 2009: consequences for the elimina- Infect 46(1):25–31. https://doi.org/10.1016/j.medmal.2015.11. tion of measles in Europe. Eur Surveill 14(50):pii: 19443 010 R Core Team (2015) R: a language and environment for statistical Faure E, Cortot C, Gosset D, Cordonnier A, Deruelle P, Guery B computing. R Foundation for Statistical Computing, Vienna, (2013) Vaccinal status of healthcare students in Lille. Med Mal Austria. http://www.R-project.org/. Accessed 22 Nov 2017 Infect 43(3):114–117. https://doi.org/10.1016/j.medmal.2013.02. Schuster M, Stelzer T, Burckhardt F (2015) Why are young adults 002 affected? Estimating measles vaccination coverage in 20–34 year Larson HJ, Jarrett C, Eckersberger E, Smith DMD, Paterson P (2014) old Germans in order to verify progress towards measles elimina- Understanding vaccine hesitancy around vaccines and vaccina- tion. PLoS Curr 7:pii: ecurrents.outbreaks.0a2d3e9 tion from a global perspective: a systematic review of published 465f067a0b2933d598d504d2e. https://doi.org/10.1371/currents. literature, 2007–2012. Vaccine 32(19):2150–2159 outbreaks.0a2d3e9465f067a0b2933d598d504d2e http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png International Journal of Public Health Springer Journals

Marked increase in measles vaccination coverage among young adults in Switzerland: a campaign or cohort effect?

Free
11 pages
Loading next page...
 
/lp/springer_journal/marked-increase-in-measles-vaccination-coverage-among-young-adults-in-DK1YpOBBOK
Publisher
Springer International Publishing
Copyright
Copyright © 2018 by The Author(s)
Subject
Medicine & Public Health; Public Health; Occupational Medicine/Industrial Medicine; Environmental Health
ISSN
1661-8556
eISSN
1661-8564
D.O.I.
10.1007/s00038-018-1102-x
Publisher site
See Article on Publisher Site

Abstract

Objectives To evaluate the impact of the Swiss measles elimination strategy—including a mass media campaign—on vaccination coverage and awareness among young adults aged 20–29 years. Methods Comparison of the results of two cross-sectional population surveys in 2012 and 2015. Results Documented vaccination coverage increased from 77 to 88% for two doses of measles vaccine. Major determinants of complete vaccination were survey year, birth cohort, sex and the absence of prior measles disease. If birth cohort and prior history of measles disease are included as factors in a multivariate model, the difference between 2012 and 2015 vanishes. Conclusions The marked increase in complete measles vaccination coverage is due to a cohort effect, owing to the introduction of the second dose of vaccine in 1996. Most of the vaccinations were administered before the national strategy was implemented and vaccination catch-ups did not increase during the campaign in young adults. Nevertheless, this study provides evidence of an improvement in the awareness of measles and measles vaccination in young adults, which may result in an impact on measles vaccination coverage in the near future. Keywords Measles  Vaccination  Vaccination coverage  Epidemiology  Survey Introduction Despite a relatively high vaccination coverage of 93% with two doses during the years 2014–2016 in the 16-year- Measles is a highly transmissible infectious disease. Its olds, measles is still an endemic disease in Switzerland. complications are particularly frequent and severe in The annual baseline incidence is low at 3–10 cases per infants and adults. However, it is possible to prevent the million people. However, a large outbreak between 2007 disease effectively with vaccination. The Swiss Federal and 2009 led to over 4000 cases (Richard and Masserey Office of Public Health (FOPH) has recommended the Spicher 2009). On this occasion, Switzerland also exported vaccination of 12-month-old children against measles since measles, causing outbreaks in various European countries 1976, and has recommended the combined vaccine against and the United States. In response to this national outbreak measles, rubella and mumps since 1985. In 1996, a second and to achieve the WHO’s goal of eliminating measles in dose was added to the vaccination schedule for 4–7-year- the European region by 2015, the FOPH together with the old children; in 2001, the recommended age for this second cantons (states) and other public health partners launched a dose was reduced to 15–24 months. national strategy for 2011–2015 to eliminate measles (Bundesamt fu ¨ r Gesundheit 2013, 2016a). One of the major objectives of this strategy was to raise public awareness of measles and vaccination against this disease to sustainably Electronic supplementary material The online version of this article increase two-dose vaccination coverage to at least 95% in (https://doi.org/10.1007/s00038-018-1102-x) contains supplementary material, which is available to authorized users. children and adolescents, including through catch-up vac- cinations for individuals born after 1963. The strategy & Ekkehardt Altpeter consisted of six areas of intervention, including facilitating ekkehardt.altpeter@bag.admin.ch and encouraging catch-up vaccinations in the over 2-year- Division of Communicable Diseases, Federal Office of Public olds to fill vaccination gaps, as well as communication and Health, Berne, Switzerland 123 590 E. Altpeter et al. promotion to raise public awareness of measles and vac- Participants, recruitment and sample size cination against this disease. Although vaccination coverage for measles has contin- Inclusion criteria for participation in the study were: age ually increased in children of all age groups since 2000, the 20–29 years, address within Switzerland recorded in a national target of 95% of people vaccinated with two doses commercial directory, and ability to answer questions in has not yet been reached. According to the last cantonal German, French or Italian. The survey company used a surveys of 2-, 8- and 16-year-olds during the period commercial household directory, which covers about 95% 2014–2016 (Swiss National Vaccination Coverage Sur- of all Swiss private households. Stratified random sampling vey—SNVCS), national coverage for a single dose was based on two (baseline) or three (endline) language regions 94–96% depending on age group, but only 87–93% for two and municipality size was applied. Each selected household doses (Bundesamt fu ¨ r Gesundheit 2016b). However, a few received an invitation letter from the FOPH to announce cantons did reach 95% for two doses. Vaccination coverage the study and its background. A total of 8645 invitation of adults is not routinely monitored in Switzerland and is letters were sent. For households without a registered unknown. telephone number, the letter requested for it. A free tele- The aim of this study was to assess the effect of the phone hotline for participants who wished to know more measles campaign in regard to improved vaccination cov- about the study was in place during both study periods. erage in young adults by catch-up vaccinations and The telephone contact started with a screening for increased awareness. With two consecutive surveys, we willingness to participate and to assess eligibility by age. If assessed the vaccination coverage as well as awareness of several persons within a household fulfilled the age criteria, measles in 20–29-year-olds in 2012 before the campaign one was selected at random. At the end of the telephone (baseline) and in 2015 afterwards (endline). interview, respondents were asked to send their vaccination card or a copy of it to the FOPH. The originals were returned after data collection. Those consenting to send in Methods their vaccination card received an instruction letter with a stamped and addressed envelope and an incentive of CHF Study design 10. If the FOPH did not receive the vaccination card within 3 weeks, a reminder was sent. To get the vaccination cards, We designed two identical nationally representative cross- a total of 2717 letters were sent followed by 1373 sectional telephone surveys before and after the campaign reminders. in 2012 and 2015, respectively, conducted by the same We anticipated a ratio of one vaccination card received commercial telephone survey company, using identical per two completed telephone interviews. The number of sampling methods and assessment tools for both baseline attempted interviews was calculated on this basis. The and endline survey. The same professional telephone sur- target sample size was 350 interviews with matching vac- vey company conducted recruitment and data collection for cination cards per language region. We aimed at a total both surveys. The interview included the same questions sample size for interviews with matching vaccination cards regarding: participants’ self-reported measles vaccination of 700 in the baseline with two language regions and 1050 status, reasons for receiving measles catch-up vaccination, in the endline with three language regions. willingness to be vaccinated if vaccinations were missing and measles disease history, as well as knowledge of the Statistical analyses disease and the elimination campaign (for the wording of the questions see Table 3). All participants were asked to We calculated sampling weights according to the same send a copy of their vaccination card to the Federal Office protocol for both surveys: The sampling frame was a of Public Health. The average interview time was about database that covers 95% of all households in Switzerland. 7 min. The questionnaire was available in German, French Weights accounted for each language region and were then and Italian. The sampling procedure and questionnaire calibrated to the population total considering sex and were piloted with 38 completed interviews. The baseline canton (Lumley 2004, 2014a). survey took place between April 23 and July 22, 2012 and We combined the data of the two surveys and defined an the second survey took place between November 9 and indicator variable identifying the 2012 survey as zero and December 18, 2015. the 2015 survey as one. We will call this variable ‘survey’ The two surveys were conducted under the Swiss Epi- in the following. Since we did not sample data from the demics Act of December 18, 1970, thus ethical committee Italian-speaking part of Switzerland in the 2012 survey, we approval was not required. discarded these data from the 2015 survey for this analysis. 123 Marked increase in measles vaccination coverage among young adults in Switzerland: a… 591 Only datasets with a vaccination card were considered for We evaluated the primary endpoint in a univariate the analysis, resulting in an overall total N of 1851 (Fig. 1). analysis and then included all variables with a significant The primary endpoint was defined as documented vac- effect in a multivariate logistic regression model consid- cination with at least two doses of measles vaccine. We ering the survey weights and calculating odds ratios (OR) defined two cohorts based on the median birth year of the and their 95% CI. This analysis assumes ‘missing com- complete study population: one born in 1990 or later pletely at random’ (MCAR). formed the reference group, while the other cohort included All logistic regression models (weighted as well as respondents born before 1990. In a sensitivity analysis, we unweighted) were estimated with and without multiple included annual birth cohort as a factor into the analyses. imputations assuming a joint model (Quartagno and Car- We recoded nationality for 2015 to the standard of 2012, penter 2016). The model included all categorical variables since in 2012 we had two groups (Swiss and foreign as well as age and weight as continuous variables. We nationality) and in 2015 three groups (Swiss, foreign and estimated five imputed datasets and combined the results both). We defined people with both nationalities as Swiss. by means of Rubin’s rule (Lumley 2014b). We analysed the data excluding and including the We tabulated the documented vaccine coverage by sampling weights. For unweighted analysis, we used cohort and the variable ‘survey’. We determined the vac- standard tabulations and standard inference. The level of cination coverage at age 16-years by cohort and compared significance was set to alpha \ 0.05. We calculated 95% these estimates with coverage of the SNVCS for that age. confidence intervals (CI) where applicable. For the We further derive the catch-up vaccination by subtracting weighted analysis, we described the data by means of the overall cohort-specific vaccine coverage at the age of tabulation and calculated proportions (Thomas Lumley 16 years from the overall cohort-specific vaccine coverage 2004, 2014a). at the time of the interview. We restricted our analysis to the birth cohorts 1986–1992 which overlap between the randomly selected households based on an address directory with sex and age attributes 2012 N = 4519 2015 N = 4126 phone number invalid (out of phone number valid service, fax) 2012 N = 4095 2012 N = 424 2015 N = 3798 2015 N = 328 contact with household no contact established (no answer, line established occupied, answering machine) 2012 N = 886 2012 N = 3209 2015 N = 2864 2015 N = 934 communication not possible (language communication possible barrier) 2012 N = 3054 2012 N = 155 2015 N = 2827 2015 N = 37 eligible for age not eligible for age 2012 N = 2132 2012 N = 922 2015 N = 2294 2015 N = 533 interview not completed interview completed 2012 N = 711 2012 N = 1421 2015 N = 771 2015 N = 1521 vaccinaon card vaccinaon card exploitable not exploitable 2012 N = 803 2012 N = 618 2015 N = 1048 2015 N = 473 Fig. 1 Data flow chart, measles surveys 2012 and 2015, Switzerland 123 592 E. Altpeter et al. two surveys and compared the vaccination coverage for at Table 1 Demographic characteristics, measles surveys 2012 and 2015, Switzerland least two doses at the time of interview between the survey 2012 and 2015. We included annual birth cohort as a factor Survey Survey Total p value in a logistic regression model. 2012 2015 To evaluate the effect of absent vaccination cards, we N % N % N % tabulated its absence and presence, respectively, with Sex respect to the following variables: ‘survey’, sex, age, lan- Female 448 56 528 50 976 53 guage region, anamnestic vaccination status, prior measles Male 355 44 520 50 875 47 history and estimated a linear unweighted logistic regres- Total 803 100 1048 100 1851 100 \ 0.05 sion model. Age group We performed all analyses using R for Windows 3.1.3, \ 25 years 403 50 687 66 1090 59 respectively, 3.3.1 under Windows 7, made available by [ 24 years 400 50 361 34 761 41 the R Core Team (2015). Total 803 100 1048 100 1851 100 \ 0.05 Birth cohort Results C 1990 226 28 768 73 994 54 \ 1990 576 72 280 27 856 46 Response rate and basic characteristics Missing 1 0 0 0 1 0 of participants Total 803 100 1048 100 1851 100 \ 0.05 Nationality The response rate for completed interviews was 50% (2012 Swiss 751 94 982 94 1733 94 survey 50%, 2015 survey 49%) and for interviews with Foreign 52 6 66 6 118 6 matching vaccination card 31% (2012 survey 28%, 2015 Total 803 100 1048 100 1851 100 0.87 survey 34%). These proportions are calculated as follows: Language region the proportion of observed eligible subjects was 2132 out German 430 54 557 53 987 53 of 3054 (70%) in 2012 and 2294 out of 2827 (81%) in French 373 46 491 47 864 47 2015; these proportions were multiplied by the number of Total 803 100 1048 100 1851 100 0.86 subjects with whom no contact could be established Educational level (N = 886 in 2012 and N = 934 in 2015) or with whom no Secondary 565 70 776 74 1341 72 communication was possible (N = 155 in 2012 and N =37 Tertiary 236 29 271 26 507 27 in 2015), e.g. 1041 in 2012 and 971 in 2015; these esti- Missing 2 0 1 0 3 0 mates were added to the number of observed eligible per- Total 803 100 1048 100 1851 100 0.08 sons, leading to the estimated number of eligible subjects: Having own (1041 times 2132/3054) plus 2132 in 2012 and (971 times children 2294/2827) plus 2294 in 2015, i.e. 100% of the response Yes 125 16 64 6 189 10 rate (Fig. 1). The complete combined dataset includes 2942 No 678 84 984 94 1662 90 observations (2011 survey N = 1421, 2015 survey Total 803 100 1048 100 1851 100 \ 0.05 N = 1521). We included 1851 observations with valid vaccination documents in the analysis (2012 survey: N = 803, 43%; 2015 survey: N = 1048, 57%).The two of being vaccinated with more than one dose (Table 3). survey populations were slightly different with respect to Most participants received their first measles vaccination birth year, age, sex and having children (Table 1). according to the recommendations: the median age at We validated the question about vaccination by com- uptake was 1 year, both in 2012 (interquartile range paring the information given by memory and written on the 1–2 years, N = 729) and in 2015 (interquartile range vaccination card. The agreement was poor (see Table 2). 1–1 year, N = 1006). The median age at the second vac- cination was 13 years in 2012 (interquartile range Vaccination coverage 10–15 years, N = 610) and 10 years in 2015 (interquartile range 7–13 years, N = 923). Very few second doses and Vaccination coverage with at least two doses increased even fewer first doses were received between the ages of 20 from 77% (95% CI 73–81%) in 2012 to 88% (95% CI and 29 years (Fig. 2) in both surveys. According to the 85–90%) in 2015. Only a minority was vaccinated with just 2015 survey, only 4% of respondents (37/932) received one dose and the proportion decreased from 16% (95% CI their second dose after 2011, the year a national measles 13–19%) in 2012 to 9% in 2015 (95% CI 7–12%) in favour vaccination campaign was launched, compared to 8% of 123 Marked increase in measles vaccination coverage among young adults in Switzerland: a… 593 Table 2 Agreement between anamnesis and documented measles vaccination by use of the vaccination document at interview, measles survey 2012 and 2015, Switzerland Documented number of doses 0 1 2 3 Total Vaccination document used during interview Have you been vaccinated against measles? Yes 5 (1.2%) 48 (11.8%) 304 (74.5%) 34 (8.3%) 391 (95.8%) No 5 (1.2%) 2 (0.5%) 6 (1.5%) 1 (0.2%) 14 (3.4%) I do not know 1 (0.2%) 0 (0%) 2 (0.5%) 0 (0%) 3 (0.7%) Total 408 (100%) Vaccination document not used during interview Have you been vaccinated against measles? Yes 26 (1.8%) 122 (8.4%) 886 (61.3%) 91 (6.3%) 1125 (77.8%) No 34 (2.4%) 19 (1.3%) 56 (3.9%) 2 (0.1%) 111 (7.7%) I do not know 13 (0.9%) 21 (1.5%) 159 (11.0%) 17 (1.2%) 210 (14.5%) Total 1446 (100.0%) respondents (46/610) during an equally long time span educational level remained statistically significant in the after 2008 for the 2012 survey (OR = 0.51, 95% CI multivariate analysis. Thus, the major determinants of 0.32–0.81). However, the catch-up needs for second doses complete measles vaccination were the year of the survey, 3 years before the start of the respective survey were lower i.e. respondents surveyed in 2015 were more likely to be in 2015 (151/932 = 16%) than in 2012 (228/610 = 37%) fully vaccinated than those surveyed in 2012; birth cohort, (OR = 0.32, 95% CI 0.25–0.41). Accordingly, these catch- i.e. persons born in 1990 or after were more likely to be up vaccinations with a second dose during the campaign fully vaccinated than those born before 1990; sex, i.e. covered 25% of the needs in 2015 respondents compared to females were more likely to be fully vaccinated than males; 20% in 2012 (OR = 1.28, 95% CI 0.76–2.16). There is a and prior measles, i.e. those who were sure that they had strong cohort effect on the complete vaccination coverage not had measles were more likely to be fully vaccinated (supplementary material Tables 3 and 4). than those with measles in their medical history (Table 4). Knowledge about vaccination Sensitivity analyses Awareness of measles and measles vaccination improved The sensitivity analysis either neglecting the weights, between 2012 and 2015 (Table 3). First, significantly more assuming ‘missing completely at random’ or ‘missing at participants knew that adults can get measles: 32% in 2012 random’ and their combination confirmed the results of the (95% CI 28–36%) were sure that adults can get measles weighted univariate and multivariate analysis (supple- compared to 41% in 2015 (95% CI 37–45%). Second, there mentary material Tables 1 and 2). was a trend, although not statistically significant, for more There is a strong cohort effect on the primary endpoint participants to be sure that they could get missing vacci- complete vaccination coverage. If the analysis is restricted nations as adults: from 18% in 2012 (95% CI 14–21%) to to the birth cohorts 1986–1992, the odds ratio between 23% in 2015 (95% CI 19–26%). There was practically no 2012 and 2015 is 1.40 (95% CI 0.87–2.10). In the multi- change with respect to the other questions on awareness. In variate analysis, this estimate drops to 1.29 (95% CI both surveys, a similarly low proportion had previously 0.82–2.03) (supplementary material Tables 3 and 4). contracted measles. About a quarter of participants had The participants with and without vaccination cards are professional contacts to risk groups (Table 3). different with respect to the factors ‘survey’, age, language region and anamnestic vaccination status (supplementary Determinants of vaccination material Table 5). There were more missing vaccination cards in 2012 compared to 2015. In both surveys, 82% of In the weighted univariate analysis, we identified the fol- respondents reported during the telephone interview to be lowing factors as significantly associated with complete completely vaccinated, which contrasted with information vaccination:‘survey’, birth cohort, sex, educational level gained from the vaccination cards. and prior measles (Table 4). All these factors except 123 594 E. Altpeter et al. Table 3 Vaccination coverage and knowledge about measles, measles surveys 2012 and 2015, Switzerland Survey 2012 Survey 2015 N % 95% CI N % 95% CI Documented measles vaccination Not vaccinated 51 7 5 9 30 3 2 4 1 dose 130 16 13 19 82 9 7 12 2 doses 577 71 67 75 836 78 75 81 3 doses 45 6 4 8 100 10 7 12 Vaccinated with at least two doses No 181 23 19 27 112 12 10 15 Yes 622 77 73 81 936 88 85 90 Prior measles disease Surely yes 55 7 5 10 82 8 6 10 Rather yes 31 5 3 7 32 3 2 5 Rather no 113 12 9 15 176 19 16 23 Surely no 570 70 65 74 723 65 61 69 I do not know 34 6 3 8 35 4 2 6 Can adults get measles? Surely yes 280 32 28 36 374 41 37 45 Rather yes 402 53 48 57 529 48 44 52 Rather no 56 7 5 9 72 6 4 8 Surely no 8 0 0 1 11 1 0 1 I do not know 57 8 5 10 62 5 3 6 Can adults catch up missing vaccinations? Surely yes 147 18 14 21 200 23 19 26 Rather yes 408 51 47 56 566 51 47 55 Rather no 100 14 11 17 132 14 11 17 Surely no 28 3 1 5 17 1 0 2 I do not know 120 14 11 17 133 11 9 13 I can endanger others if I am not vaccinated Complete agreement 345 42 38 47 451 44 41 48 Slight agreement 287 36 31 40 405 37 34 41 Slight disagreement 94 13 10 16 148 15 12 18 Complete disagreement 46 5 3 7 26 2 1 3 I do not know 31 4 2 6 18 1 0 2 Is vaccination against measles mandatory in Switzerland? Surely yes 41 5 3 7 60 5 3 6 Yes, I believe so 106 13 10 16 159 13 11 16 No, I do not believe so 281 32 28 36 410 37 34 41 Surely no 317 44 39 48 379 42 38 46 I do not know 58 6 4 7 40 3 2 4 123 Marked increase in measles vaccination coverage among young adults in Switzerland: a… 595 Table 3 (continued) Survey 2012 Survey 2015 N % 95% CI N % 95% CI Did you know about the collaboration of Switzerland with the WHO in eliminating measles? Yes 287 38 33 42 438 44 40 48 No 516 62 58 67 610 56 52 60 Is measles elimination necessary in Switzerland? Surely yes 332 38 34 43 361 34 30 38 Rather yes 285 35 31 39 393 37 33 41 Rather no 102 15 12 18 177 18 15 21 Surely no 40 5 3 7 44 5 3 6 I do not know 44 7 4 9 73 7 5 9 Do you have professional contact to children, pregnant women or sick persons? Yes 259 28 24 32 302 25 22 29 No 544 72 68 76 745 75 71 78 Missing 0 1 Percent taking into account the weighting. Deviations from the total of 100 per cent is due to rounding errors N total: 803 for measles survey 2012, 1048 for measles survey 2015 this study provide the first reliable estimates of measles 2 doses, survey 2012, N= 610 vaccination coverage in young adults in Switzerland. We assessed vaccination coverage and determinants of 1 dosis, survey 2012, N= 729 vaccination among young adults by comparing the results 2 doses, survey 2015, N= 932 of two cross-sectional surveys conducted in 2012 and 2015. 1 dosis, survey 2015, N= 1006 Documented complete vaccination increased from 77 to 88%. However, there is no evidence for an association between vaccination and the recent national awareness campaign or increased awareness. Knowing that measles can also affect adults was the only factor that increased between the two surveys, but this knowledge was not related to the likelihood of vaccination. Strengths and limitations 0 5 10 15 20 25 30 The careful sampling mechanism stratified by sex and age language region, the adequately high sample size of valid vaccination card-based data, the thorough statistical anal- Fig. 2 Age distribution at first and second dose of vaccination, 2012 ysis taking into account various sensitivity analyses for and 2015 measles surveys, Switzerland missing data and weighting, as well as application of the Discussion identical questionnaire tools make us confident of the quality of the estimate as well as the significant increase in Main results vaccination coverage between 2012 and 2015. However, the marked increase in vaccination coverage is very likely While routine monitoring of vaccination coverage is almost due to a cohort effect. This cohort effect, unexpected at the universally established for measles in children and ado- design stage of the studies, is so strong, that all multivariate lescents, it generally does not include adults. The results of analyses have to be interpreted with care. The observed percent 020 406080 596 E. Altpeter et al. Table 4 Logistic regression models of documented vaccination coverage with at least two doses, weighted and assuming missing completely at random, measles survey 2012 and 2015, Switzerland Univariate analysis Multivariate analysis OR 95% CI Adj. OR 95% CI Survey 2012 1 1 2015 2.15 1.54 3.01 1.67 1.15 2.42 Birth cohort C 1990 1 1 \ 1990 0.32 0.23 0.45 0.41 0.28 0.61 Sex Female 1 1 Male 0.59 0.42 0.81 0.58 0.42 0.82 Nationality Swiss 1 n.i. Foreign 1.15 0.59 2.23 n.i. Language region German 1 n.i. French 1.11 0.83 1.48 n.i. Educational level Secondary 1 1 Tertiary 0.60 0.43 0.86 0.77 0.53 1.12 Having own children Yes 1 n.i. No 1.48 0.96 2.31 n.i. Prior measles Surely yes 1 1 Yes, I believe so 1.43 0.62 3.28 1.69 0.71 4.05 No, I do not believe so 1.79 0.97 3.31 1.83 0.94 3.54 Surely not 2.33 1.40 3.89 2.46 1.43 4.23 I do not know 1.19 0.49 2.91 1.51 0.61 3.77 Can adults get measles? Surely yes 1 n.i. Rather yes 1.13 0.79 1.62 n.i. Rather no 0.98 0.52 1.86 n.i. Surely no 0.85 0.25 2.89 n.i. I do not know 0.93 0.50 1.74 n.i. Can adults catch up missing vaccinations? Surely yes 1 n.i. Rather yes 1.00 0.66 1.53 n.i. Rather no 0.77 0.44 1.36 n.i. Surely no 1.65 0.63 4.37 n.i. I do not know 0.89 0.51 1.55 n.i. I can endanger others if I am not vaccinated Complete agreement 1 n.i. Slight agreement 0.76 0.52 1.10 n.i. Slight disagreement 0.78 0.48 1.27 n.i. Complete disagreement 0.74 0.38 1.41 n.i. I do not know 1.60 0.40 6.45 n.i. 123 Marked increase in measles vaccination coverage among young adults in Switzerland: a… 597 Table 4 (continued) Univariate analysis Multivariate analysis OR 95% CI Adj. OR 95% CI Is vaccination against measles mandatory in Switzerland? Surely yes 1 n.i. Yes, I believe so 2.06 0.73 5.83 n.i. No, I do not believe so 0.96 0.38 2.45 n.i. Surely no 0.90 0.35 2.27 n.i. I do not know 1.33 0.43 4.09 n.i. Did you know about the collaboration of Switzerland with the WHO in eliminating measles? Yes 1 n.i. No 0.99 0.71 1.38 n.i. Is measles elimination necessary in Switzerland? Surely yes 1 n.i. Rather yes 1.06 0.72 1.57 n.i. Rather no 0.73 0.46 1.17 n.i. Surely no 1.43 0.67 3.04 n.i. I do not know 1.03 0.55 1.91 n.i. Do you have professional contact to children, pregnant women or sick persons? Yes 1 n.i. No 0.95 0.65 1.38 n.i. OR odds ratio, CI confidence interval, n.i. not included into the multivariate model cohort effect is confirmed by the comparison with the information about their vaccination status, provided by values from the external reference the SNVCS, which also participants not consulting their vaccination card, was often shows the increasing values over the study periods of the uncertain (see Table 2). We therefore restricted our anal- SNVCS. yses to individuals who provided a vaccination card, The high number of missing vaccination cards—an thereby reducing response rate from 50 to 31%. This could exclusion criterion of our study—could lead to a selection have introduced selection bias and led us to overestimate bias. The contrast between self-reported vaccination status the effect, because we assume that people without a vac- and documented vaccination history confirms our study cination card are less likely to be vaccinated. However, this design, which focused on the collection of vaccination was partially addressed by a complex weighting procedure, cards and restricted the analysis on these written records. a common method of dealing with the unit non-response. The return rate of vaccination cards (63%) was in the range After weighting, the study population had a similar distri- of the expectation during the design stage and was antici- bution to the Swiss general population in this age range in pated in the sample size calculation for interviews with regard to age, sex and cantonal distribution. matching vaccination cards. However, there is evidence Our results are in line with the vaccination coverage in that subjects with missing vaccination cards are different the 16-year-old in the SNVCS (Bundesamt fu ¨ r Gesundheit between the two surveys, with respect to age, language 2016b). They are as well comparable to findings from region, and history of measles vaccination. The process of Germany and France. In a study, conducted 2012 in Ger- differential missing data between 2012 and 2015 may lead many, vaccination coverage for two doses was found to be to an overestimation of the marked increase in vaccination 56% in adults aged 20–34 years (Schuster et al. 2015). 80% coverage. of these participants relied on their vaccination card to There are some critical points with respect to our anal- answer. Furthermore, the fact that in our study younger age ysis. The participants were asked about vaccinations that groups and females were more likely to be vaccinated is often occurred during their early infancy. Anamnestic consistent with the finding of Schuster et al. (2015). In 123 598 E. Altpeter et al. France, measles vaccination coverage (based on vaccina- causal pathway between knowledge, attitude and actual tion cards of first-year health care students) was 78% in vaccination; this leads to an inconclusive effect, a problem 2011 (Faure et al. 2013). In another French study in 2015, that was encountered by Larson et al. (2014). coverage was 93% in 16–18-year-olds (Buscail and Gag- niere 2016). Conclusion Our results show that the sharp increase of vaccination coverage in just 3 years was very likely due to a ‘‘cohort An 11% increase in measles vaccination coverage within a effect’’ resulting from a change in recommendations made 3-year period in a population where vaccination coverage in 1996: a second dose of measles vaccine was introduced is already high is a historic finding. This marked increase in to the national vaccination plan for 4–7-year-old children, measles vaccination coverage is, however, due to a cohort together with the recommendation to catch up at any age. effect, owing to the introduction of the second dose of Only four out of ten birth cohorts (1989–1992) among the vaccine in 1996 in the national vaccination plan. This study participants in the first survey had the opportunity to provides evidence of an improvement in the awareness receive a second dose as routine vaccination, compared to about measles and measles vaccination in young adults, seven cohorts (1989–1995) of participants in the second which may result in an impact on measles vaccination survey in 2015. In addition, it took a few years for the coverage in the future. The achievement of this marked second dose to become a well-established practice. The gap increase in measles vaccination coverage merits acknowl- of catch-up vaccinations with a second dose 3 years before edgment of the efforts made by general practitioners and the surveys was therefore larger in the group surveyed in paediatricians in Switzerland, who remain the main and the 2012 than that in 2015. most important promoters for measles vaccination. Most of these catch-up vaccinations were administered Governments at all levels are challenged to use their lee- before the age of 16 years. They were received before the way to support any efforts made by Swiss physicians to recent promotion campaign and should be credited to promote vaccination in general. paediatricians, school health services and parents. Although one quarter of the pre-campaign vaccination gaps Funding This study was fully funded by the Swiss government (Swiss were filled during the campaign, this proportion was not Federal Office of Public Health) based on the federal epidemic law significantly higher than during the 3 years before the (EpG, SR 818.101 of December 18, 1970). campaign suggesting that the campaign had little to no effect on the number of catch-up doses administered. Compliance with ethical standards We estimated the number of catch-up vaccinations administered by primary care physicians in a study within Conflict of interest The authors declare that they have no conflict of the Swiss Sentinella network (Bundesamt fu ¨ r Gesundheit interest. 2016c). In 2014 and 2015, estimated totals of 33,000 and 37,000 doses of measles catch-up vaccinations were Ethical approval We conducted this survey as an evaluation of ongoing surveillance of vaccination coverage according to the Swiss administered by general practitioners and paediatricians. law on communicable diseases in 2012 and 2015 (EpG, SR 818.101 This corresponds to 37% of all doses needed in 2015 to of December 18, 1970). close the estimated gap (with a 95% objective) in measles vaccination for 2–51-year-olds in Switzerland. Two-thirds Informed consent Informed consent was obtained from all individual participants included in the study. of these catch-up vaccinations were given to adults aged 20 years and older. Eighty-eight percent of all vaccinations Open Access This article is distributed under the terms of the Creative were initiated by the physicians, underlining their crucial Commons Attribution 4.0 International License (http://creative role in promoting vaccination and filling the gaps. commons.org/licenses/by/4.0/), which permits unrestricted use, dis- Awareness of measles has increased slightly in our study tribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a and can almost be considered as ‘‘common sense’’, with link to the Creative Commons license, and indicate if changes were 89% of all people knowing that measles can also affect made. adults. There mere fact that the study topic was measles elimination and the interviewees were primed by an invi- tation letter sent out by the FOPH may lead to overesti- References mating results about knowledge and attitude towards measles elimination. Knowledge and attitude questions Bundesamt fu ¨ r Gesundheit (2013) Nationale Strategie zur Maserne- limination 2011–2015. BAG Bull 17:269–271 were highly correlated in our survey. But the data were too Bundesamt fu ¨ r Gesundheit (2016a) Schweiz ohne Masern. BAG Bull scarce to be stratified by all variables and evaluated in an 5:93–95 integrated way. In addition, there is no simple and direct 123 Marked increase in measles vaccination coverage among young adults in Switzerland: a… 599 Bundesamt fu ¨ r Gesundheit (2016b) Durchimpfung von 2-, 8- und Lumley T (2004) Analysis of complex survey samples. J Stat Softw 16-jahrigen Kindern in der Schweiz, 1999–2016. https://www. 9(1):1–19 bag.admin.ch/bag/de/home/themen/mensch-gesundheit/ueber Lumley T (2014a) Survey: analysis of complex survey samples. R tragbare-krankheiten/impfungen-prophylaxe/informationen- package version 3.30 fachleute-gesundheitspersonal/durchimpfung.html. Accessed 22 Lumley T (2014b) mitools: Tools for multiple imputation of missing Nov 2017 data. R package version 2.3. http://CRAN.R-project.org/pack Bundesamt fu ¨ r Gesundheit (2016c) Nachholimpfung gegen Masern age=mitools. Accessed 22 Nov 2017 und Auffrischimpfung gegen Keuchhusten: wesentlicher Beitrag Quartagno M and Carpenter J (2016) jomo: a package for multilevel zur Masernelimination und zum Schutz von Sa ¨uglingen. BAG joint modelling multiple imputation. http://CRAN.R-project.org/ Bull 35:533–541 package=jomo. Accessed 22 Nov 2017 Buscail C, Gagnie `re B (2016) Vaccination coverage of adolescents: Richard JL, Masserey Spicher V (2009) Large measles epidemic in results of a Defense and Citizenship Day-based survey. Med Mal Switzerland from 2006 to 2009: consequences for the elimina- Infect 46(1):25–31. https://doi.org/10.1016/j.medmal.2015.11. tion of measles in Europe. Eur Surveill 14(50):pii: 19443 010 R Core Team (2015) R: a language and environment for statistical Faure E, Cortot C, Gosset D, Cordonnier A, Deruelle P, Guery B computing. R Foundation for Statistical Computing, Vienna, (2013) Vaccinal status of healthcare students in Lille. Med Mal Austria. http://www.R-project.org/. Accessed 22 Nov 2017 Infect 43(3):114–117. https://doi.org/10.1016/j.medmal.2013.02. Schuster M, Stelzer T, Burckhardt F (2015) Why are young adults 002 affected? Estimating measles vaccination coverage in 20–34 year Larson HJ, Jarrett C, Eckersberger E, Smith DMD, Paterson P (2014) old Germans in order to verify progress towards measles elimina- Understanding vaccine hesitancy around vaccines and vaccina- tion. PLoS Curr 7:pii: ecurrents.outbreaks.0a2d3e9 tion from a global perspective: a systematic review of published 465f067a0b2933d598d504d2e. https://doi.org/10.1371/currents. literature, 2007–2012. Vaccine 32(19):2150–2159 outbreaks.0a2d3e9465f067a0b2933d598d504d2e

Journal

International Journal of Public HealthSpringer Journals

Published: Apr 19, 2018

References

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