Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Medication safety in acute care in Australia: where are we now? Part 2: a review of strategies and activities for improving medication safety 2002-2008

Medication safety in acute care in Australia: where are we now? Part 2: a review of strategies... Background: This paper presents Part 2 of a literature review examining medication safety in the Australian acute care setting. This review was undertaken for the Australian Commission on Safety and Quality in Health Care, updating the 2002 national report on medication safety. Part 2 of the review examined the Australian evidence base for approaches to build safer medication systems in acute care. Methods: A literature search was conducted to identify Australian studies and programs published from 2002 to 2008 which examined strategies and activities for improving medication safety in acute care. Results and conclusion: Since 2002 there has been significant progress in strategies to improve prescription writing in hospitals with the introduction of a National Inpatient Medication Chart. There are also systems in place to ensure a nationally coordinated approach to the ongoing optimisation of the chart. Progress has been made with Australian research examining the implementation of computerised prescribing systems with clinical decision support. These studies have highlighted barriers and facilitators to the introduction of such systems that can inform wider implementation. However, Australian studies assessing outcomes of this strategy on medication incidents or patient outcomes are still lacking. In studies assessing education for reducing medication errors, academic detailing has been demonstrated to reduce errors in prescriptions for Schedule 8 medicines and a program was shown to be effective in reducing error prone prescribing abbreviations. Published studies continue to support the role of clinical pharmacist services in improving medication safety. Studies on strategies to improve communication between different care settings, such as liaison pharmacist services, have focussed on implementation issues now that funding is available for community-based services. Double checking versus single-checking by nurses and patient self-administration in hospital has been assessed in small studies. No new studies were located assessing the impact of individual patient medication supply, adverse drug event alerts or bar coding. There is still limited research assessing the impact of an integrated systems approach on medication safety in Australian acute care. Page 1 of 14 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:24 http://www.anzhealthpolicy.com/content/6/1/24 seeking to blame the individual/s involved. This informa- Background This paper is the second in a two-part literature review of tion can be used to formulate strategies to prevent similar medication safety in the Australian acute care setting [1]. incidents, which in turn can be evaluated as part of a con- It examines the Australian evidence-base for strategies to tinuous quality improvement cycle [10]. In this review, reduce medication errors and updates the data presented root cause analysis strategies were included, as were sys- in the Second National Report on Patient Safety - Improving tems to improve the reporting of medication incidents in Medication Safety [2]. acute care. International evidence has identified a number of tools or Methods practices that may reduce medication incidents in health- Search strategy care. Computerised physician order entry (CPOE) cou- Databases and search terms used in the search strategy pled with clinical decision support systems (CDSS) is have been detailed in Part 1 of the review [1]. Criteria rel- amongst the most studied. These systems allow standard- evant to the general headings in the Second National Report ised online prescribing coupled to computerised advice to on Patient Safety - Improving Medication Safety [2] were support prescribing decisions. The systems may provide used. automated checks or alerts such as drug-drug interaction checking or drug allergy alerts. These systems can reduce The database search was supplemented with review of rel- adverse drug events in hospital settings [3] although there evant reports and resources on the Australian Commis- have been reports of increased error rates when imple- sion on Safety and Quality in Health Care website http:// mentation problems have occurred [4]. Most studies have www.safetyandquality.gov.au/. Additionally, the website examined "home-grown" CPOE/CDSS systems developed of the Australian Government Department Health and and used within an institution; commercially developed Ageing was searched for information and publications systems have been less well studied [3]. relating to government electronic health initiatives including bar coding, shared electronic medical records International studies also support the role of clinical phar- and electronic prescribing. State government websites and macists in hospital wards for preventing adverse drug the website of the New South Wales Therapeutic Advisory events [4-6]. Other strategies examined for improving Group (NSW TAG) were searched for information on medication safety in acute care include individual patient medication safety initiatives in hospitals. supply of medication, robotic systems for medication dis- pensing (including automated bedside systems), bar cod- Selection of studies for review ing of medication packaging and patients, intravenous Letters, study reports and reviews published between devices capable of performing dilutions, computerised 2002 and 2008 were selected if they reported on: medication records, changes to systems to facilitate improved communication and educational interventions - systems to promote improved prescription writing; [4,7,8]. - systems to promote dissemination of information about This review examines the published literature on the evi- medicines and improved prescriber decision-making; dence and implementation of these strategies in the Aus- tralian acute care setting. While these approaches may - systems used to promote accurate dispensing and/or dis- impact particular stages in the medication use process tribution of medicines; (such as prescribing, transcribing of information, dispens- ing, administration), no single strategy can target all - systems to ensure adequate checking; stages of the complex process. While the evidence for these strategies has been reviewed separately the impor- - systems used to promote accurate administration of tance of using a multifaceted approach, incorporating a medicines; number of these strategies, is emphasised to improve medication safety. - systems to improve management of medicines; The analysis of medication incident reports occurring - medicine-specific handling/management strategies; within an institution can allow problems with systems to be identified and corrected. "Root cause analysis" is one - clinical pharmacy services; technique which has been adapted from industries such as aviation and aerospace for use in the health care setting - systems to improve information transfer about medi- [9]. This involves identifying the "root" and contributory cines between hospital and community settings; factors that have led to a medication incident rather than Page 2 of 14 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:24 http://www.anzhealthpolicy.com/content/6/1/24 - medication management services and case conferencing; An implementation strategy for the NIMC in three Victo- rian hospitals involved an interdisciplinary steering group - systems-based approaches to understanding and pre- to support integration across the whole organization, a venting medication errors; dedicated project officer acting as facilitator and coordina- tor and four interdisciplinary working groups addressing - systems to promote reporting of medication incidents supply, communication, education and evaluation issues and adverse drug reactions in hospitals. [14]. Studies undertaken in the community setting were An audit of the NIMC design and performance was also excluded. undertaken at the Royal Perth Hospital (RPH) [15]. Chart design assessment was performed on 15 design features and compared with the previously used RPH chart, four Results and Discussion Evidence for systems to improve prescription writing other WA hospital charts and nine charts from teaching The National Inpatient Medication Chart hospitals in other states and territories. Completion of the A National Inpatient Medication Chart (NIMC) aims to individual fields of the chart was assessed and an audit of improve medication safety through standardisation of NIMC charts from six medical and surgical wards of the medication ordering in all Australian public hospitals, hospital was compared to completion of the previous and a number of private hospitals. The NIMC was RPH chart. Aspects of the design of the NICM considered adopted for national roll-out following an agreement of likely to improve medication safety included a) a section the Australian Health Ministers Council in 2004 [11]. The to complete medication history; b) provision for record- recommendation for a process of pharmaceutical review ing sustained-release dose forms; c) a mechanism to circle of all aspects of medication management in the hospital inpatient drugs intended for provision at discharge; d) was also part of this reform [12]. provision for documenting a medicine's indication; and e) direction to record intended administration times of The chart along with education on safe prescribing and each medicine. Four of these five advantages were poorly administration were piloted in 31 public and private hos- complied with in practice. Overall compliance with the pitals in metropolitan, regional and rural areas in 2004 NIMC was 56% [95% CI 43-67%]. Only 2.3% of charts [13]. Pre and post-implementation audits were under- had the medication history completed. There was a mod- taken by 28 sites. Improvements seen in the combined est, but non-significant increase in overall compliance results included: after the introduction of the NIMC compared to the previ- ous chart (6% [95% CI -0.2 to 13%]). Concerns about - increased documentation of adverse drug reactions design features of the chart included cramped design, lack (21% at baseline to 50% at three months); of colour, lack of provision for variable dosing and the need for, on average, twice as many charts per admission - decreased prescription of drugs to which patient had (and hence increased requirements for rewriting charts an allergy (9% to 6%); and possible transcription errors) [15]. - increased entry of actual administration times (18% Tools have been developed to allow ongoing evaluation to 68%); of the chart [16]. Currently local management of the NIMC is overseen by local jurisdictional bodies at the - increased frequency of providing the indication for a State and Territory level [17]. Where relevant, the local 'prn' (as needed) medication (13% to 26%); body will refer proposed changes to the national NIMC Oversight Committee. Versions of the NIMC for paediat- - increased documentation of the maximum dose for a ric patients and patients requiring long hospital stays are 'prn' medication (24% to 36%); available [18,19]. Prescriber education - increased frequency of the prescriber name being identifiable (41% to 79%); and Education services to individual healthcare professionals are sometimes referred to as 'academic detailing'. This - increased frequency of target international normal- term refers to an educational approach based on princi- ised ratio (INR) documentation (9% to 71%) [13]. ples of communications theory and behaviour change [20]. The study used surrogate, rather than direct, measures of patient harm [13]. A 2001 NSW study examined whether academic detailing could reduce prescription errors for drugs of addiction Page 3 of 14 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:24 http://www.anzhealthpolicy.com/content/6/1/24 (DOA) in the hospital setting [21]. Errors assessed vention (p < 0.001). The rates of abbreviations classified included: as of major significance decreased from 5.8 per 100 pre- scriptions pre-intervention to 2.3 post-intervention (p < - the quantity not written in both words and numbers; 0.001). The study was not controlled and the sustainabil- ity of the effect over time is unknown. - the DOA not written on a separate script; The evidence for systems ensuring better dissemination of - alterations to the script not initialled; knowledge about drugs Clinical decision support systems (CDSS) with computerised - details of the preparation or form of the drug omit- physician order entry (CPOE) ted; In the 2002 national patient safety report, careful imple- mentation and evaluation of electronic prescribing with - liquid preparations without a milligram dose given; decision support was suggested as a priority for improving medication safety in Australia [2]. Implementation - strength omitted or incorrect. research has since been conducted in Australia, however, outcome studies are still lacking. The intervention involved a one-on-one academic detail- ing session for all first and second year post-graduate prac- The limitations of using electronic prescribing alone are titioners. A follow-up session was conducted after two highlighted in the findings of a 2001 study of discharge months. Error rates were assessed two months after the prescriptions at a teaching hospital in Brisbane [23]. A intervention and compared with a control hospital. The computer-generated discharge summary system enabled a baseline levels of errors at the intervention hospital discharge prescription to be generated based on informa- (approximately 40%) were higher than those of the con- tion entered by the medical officer. An observational audit trol hospital (25%), making comparison difficult. Error of 200 discharge prescriptions was conducted; involving rates reduced from 41% of 46 prescriptions pre-interven- 100 handwritten prescriptions (605 medications) and tion to 24% of 128 prescriptions post-intervention (p < 100 computer generated prescriptions (700 medications). 0.001, chi square = 17.3). There was no change in the The same medical staff were responsible for both prescrip- error rate at the control hospital. No assessment was made tion types. There were more errors in the computer gener- on the likely to impact of these errors on patient out- ated prescriptions (81 errors, 11.6% of items) compared comes. to the handwritten ones (30 errors, 5.0% of items, p < 0.001). Errors judged to have the potential to result in Further adequately controlled studies are required to con- patient harm were similar between the groups. Errors that firm whether academic detailing can reduce prescription occurred more frequently when a computer was used error rates in the Australian hospital setting. These should included dosing errors (25 computer generated compared include assessment of errors likely to impact on patient with 5 handwritten errors) and duration of therapy errors outcomes. due to default settings in the computer. The authors con- cluded that electronic prescribing without decision sup- A pre-intervention, post-intervention study in a Mel- port and alerting systems could increase the risk of patient bourne teaching hospital examined an educational inter- harm. vention to reduce the use of "error-prone prescribing abbreviations" in an emergency department (ED) [22]. Recognition that electronic medication management sys- The intervention, involving ED registrars and postgradu- tems can introduce machine-related errors led a Univer- ate course nurses, included small group and one-to-one sity of NSW research group to develop a multilevel tutorials about abbreviations commonly causing medica- "accident model" to examine points in electronic prescrib- tion errors or confusion. Summary cards and posters rein- ing systems where system failures may occur [24]. This forced this information. The intervention ran for six- model used a systematic approach to identify human- months. All medication and fluid charts in the ED depart- computer interaction processes as well as the context in ment were assessed for error-prone abbreviations at a ran- which electronic prescribing systems are used (such as domly selected time each day for one week before the health professional cultures, organizational factors). The intervention and one week following. Abbreviations were aim is to aid the development of electronic prescribing classified as being of major, moderate or minor signifi- systems with features that improve patient safety. The cance by two independent pharmacists. Charts for 166 validity of the model is still to be tested. patients were included in the two assessment phases. The error-prone abbreviation rate per 100 prescriptions Implementation of an inpatient electronic prescribing decreased from 31.8 pre-intervention to 18.7 post-inter- and clinical decision support system in an acute and sub- Page 4 of 14 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:24 http://www.anzhealthpolicy.com/content/6/1/24 acute ward at a metropolitan and rural hospital in Victoria would be willing to adopt electronic prescribing and alert found mixed results [25]. The system allowed a 'point and systems, the implementation would be hampered by click' method of prescribing for physicians' ordering inpa- existing barriers. Barriers included lack of confidence in tient and discharge medications, integration with the hos- the security aspects of the system (e.g. use of electronic sig- pital's pharmacy ordering system and use of clinical natures, failing to log-out of a system), lack of funding decision support tool. The clinical decision support and resources, concern about the time taken to complete allowed checking of interactions, allergies and duplicate and check an electronic prescription (e.g. excessive drug- ordering and access to drug information including the drug interaction checking by some decision support/alert- AusDI drug and therapeutics database and MIMS. The sys- ing systems), lack of compatibility with the existing tem was tested in one acute and one sub-acute 30-bed patient administration system in the hospital and legisla- ward within each hospital. In each ward two clinical com- tive barriers (such as legal requirement for handwritten puters were available, in addition to a wireless laptop as a signatures). The majority of clinicians favoured the idea of point-of-care computer. Intensive training was provided a state-wide EPDS system. Clinicians stated the need for to medical officers and nursing staff. Use of the system in an integrated system that would overcome the need to log the acute ward setting was discontinued after six weeks in into different systems for different types of results. the rural hospital and eight weeks in the metropolitan hospital. The barriers to effective implementation in the A 2007 project reviewed electronic medication manage- acute ward setting were perceptions of increased clinical ment (EMM) systems in the Australian setting [27]. A risk, workload issues, lack of medical staff commitment, multidisciplinary reference group including medical, insufficient computer access and technical and software pharmacy and nursing representation as well as clinical limitations (including inadequate interaction and allergy information technology experts formulated "key princi- checking, and problems with version control). The experi- ples and core features" for assessing the suitability of an ence in the sub-acute ward setting in both hospitals was EMM systems. Systems from 11 companies were evalu- different, with use becoming accepted practice and the ated with all found to have the majority of the required system being rolled out to other sub-acute wards. Medica- core functions seen as important by the reference group. tion error rates were not assessed. The study highlighted Systems were not ranked relative to each other. It con- the need for document control, understanding clinical cluded that a number of available systems were suitable workflow issues and commitment through the whole for use in Australian hospitals, but that change manage- organization. One of the barriers in the acute care setting ment issues needed to be addressed for implementation was the need to frequently review medication charts and a to occur more widely. ban on handwritten alterations to charts. The authors con- cluded that implementation of electronic prescribing and The Australian Health Information Council (AHIC) pro- clinical decision support systems requires a highly organ- vides advice to the Australian Health Ministers Advisory ised approach at all levels of the institution, giving consid- Council (AHMAC) on information management and eration to the technical and cultural issues and technology in the health care system. A report by the environment in which it is to be used. AHIC in 2008 [28] examined Electronic Decision Support systems, with a focus on medicines, including the current Another qualitative feasibility study for an 'electronic pre- state of their implementation in Australia. This review scribing decision support' (EPDS) system was undertaken found there were some sound evidence-based programs in a NSW public hospital [26]. The Sauer's Triangle of used in hospitals in some States. In Victoria, for example, Dependencies model was used to examine the organisa- the Victorian Clinical Systems provided through Health tional context in which an information system is placed. SMART supported electronic prescribing and decision- The study used interviews and focus groups with medical making by providing patient allergies, adverse reactions staff, pharmacists, nurse managers and clinical informa- and automatic checking for duplication and drug interac- tion technology experts. Questions examined the limita- tions. However, the report highlighted that an agreed set tions of the present paper-based prescribing system, of national standards was lacking. The need to work with technical requirements for an electronic prescribing and professional bodies to examine possible barriers and decision support system, the environment in which the incentives for healthcare professionals in the uptake of system would be used, the political setting, the type of electronic decision support was identified, as was the need ward structure suited to electronic prescribing, perceived for the health care workforce to be adequately supported barriers to implementation and mechanisms for consult- to gain the skills to use the technology [28]. ing with medical staff in the design and implementation of an information system. Only 9% of medical staff and Since the previous medication safety review [2], a number 20% of nurse managers participated. Results found that of new studies have examined the implementation of elec- while nearly all participating clinicians indicated they tronic prescribing in combination with clinical decision Page 5 of 14 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:24 http://www.anzhealthpolicy.com/content/6/1/24 support systems in Australian hospitals, but not the and track administration [34]. No published studies in impact on medication error or patient outcomes. The the Australian setting were identified. potential value of these strategies now appears to be more widely recognised by government, hospitals and health An ongoing project administered by GS1 Australia (for- professionals. Published studies provide useful insights merly EAN Australia) and the National E-Health Transi- into some of the barriers to the introduction of this strat- tion Authority (NEHTA) aims to implement an Australian egy in the acute care setting that need to be considered standard coding system for medicines - the Australian Cat- with wider implementation. alogue of Medicines (ACOM) [35]. This will ensure that all prescription and non-prescription medicines (includ- Evidence for systems promoting better medication ing complementary medicines) have a globally unique distribution code. A national coding system is required to allow the Individual patient based medication distribution electronic transmission, storage and use of medication In the 2002 review of medication safety in Australia [2], information. This has the potential to facilitate the use of individual patient medication supply systems were found bar coding technology. to reduce medication errors. Two Australian studies had compared different medication distribution methods and Computer adverse drug event detection and alerts resultant errors associated with administration in the Aus- A computer system may be used to detect potential tralian hospital setting [29,30], showing a decrease in adverse drug events (such as interactions between differ- error rates. No further Australian studies were located. The ent medications, or abnormal laboratory results for a Society for Hospital Pharmacists of Australia Standards of patient taking a particular medication) and alert a Practice for the Distribution of Medicines in Australian patient's health care professional such as physician or Hospitals published in 2006 [31] state that unit-dose sys- pharmacist. The previous review [2] found limited inter- tems are the preferred method of medicines distribution national evidence but a lack of Australian research to sup- for healthcare facilities in terms of patient safety. How- port the use of such systems for improving medication ever, there is a lack of published data on the uptake of safety. No more recent Australian studies were located. individual patient supply systems for medications in Aus- Single-person versus double-person checking by nurses administering tralian hospitals despite evidence supporting its use in reducing medication errors. medications A study in a Victorian acute care hospital examined the Automated dispensing devices safety of single-checking by a registered nurse of medica- Automated medication dispensing devices are electronic tions that had required double-person checking [36]. storage devices that dispense medications in a controlled These included medications requiring calculations, drugs manner and track the use of medication. The 2002 medi- of addiction, cytotoxics, new drugs, epidurally adminis- cation safety review [2] noted that evidence for automated tered drugs, variable dose insulin, blood products and drug distribution systems in reducing medication inci- high dose potassium chloride. Medication incident dents was limited. Two studies that evaluated automated reports were assessed in the single checking study for a drug distribution in the Australian health care setting seven-month period and compared to those in the same [32,33] did not provide clear evidence of the efficacy of units in the same months of the previous year when dou- automated systems for reducing error. No further Austral- ble-person checking was standard practice. There was no ian studies since 2002 were located. significant difference between the two periods, however the number of reported administration errors was low Evidence for systems ensuring adequate checking (four in the study period and five in the previous year), Bar coding and the required study power to detect a difference was In the previous review [2] there was international evi- not reported. This study analysed medication incidents dence to support the investigation of bar coding as a strat- reported through the hospital's reporting scheme and did egy to reduce medication error. Bar-coding or other not include any independent assessment of errors. Reli- identification systems such as radio frequency identifica- ance on incident reports may have meant errors were tion tags could allow medication packaging supplied for undetected as incidents are known to be under-reported. an individual patient to be cross-checked with patient Further studies are required to provide conclusive evi- identification information (such as a hospital patient dence about the relative safety of single and double-per- identification bracelet) at the point of medication admin- son checking of high risk medications in the Australian istration [34]. Alternatively, more advanced systems could acute care setting. allow electronic prescriptions to interface automated dis- pensing systems to assemble individual medication packs Page 6 of 14 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:24 http://www.anzhealthpolicy.com/content/6/1/24 Evidence for systems to improve medication of the program. A total of 220 patients participated, with administration 45% remaining on Level 1, 26% reaching level 2 and 29% Drug packaging, storage and administration equipment reaching level 3. There were no patient initiated medica- The potential for administration of IV medications by the tion errors in the study period and two errors involving wrong route has been highlighted by cases of inadvertent staff, compared to one error in the previous six-month spinal administration of the anti-cancer medication vinc- period. This study was conducted in a specialized unit. ristine [37]. The Society for Hospital Pharmacists of Aus- The findings, therefore, are not generalisable to other tralia has recommended strategies to reduce the risk of acute care settings but warrant further investigation. error associated with cytotoxic medications, with the abo- lition of syringes for administration of vincristine in Education and training about medication administration errors favour of an infusion bag [37]. An orientation program for newly employed registered nurses at a Queensland teaching hospital examined the A system to prevent infusions being administered by the ability of nurses to identify medication errors and apply incorrect route has been developed at the Women's and strategies to prevent medication incidents [43]. The pro- Children's Hospital, Adelaide [38]. The system, called the gram used simulated medication administration scenar- Adelaide Regional Connector (ARC), was under prototype ios of frequently occurring medication errors with development in 2002. This colour coded luer incompati- potential for harm. After each scenario the nurses were ble system ensures that syringes and other drug adminis- asked whether they detected the errors, whether they tration equipment used to administer epidural and would have modified their practice and whether they were intrathecal doses were not able to be connected to those aware of the error concept. Nurses were presented with used to administer IV infusions. education about concepts of human error and risks, the systems in place in the hospital to prevent medication Further research is needed to examine the current state of errors, roles and responsibilities in detecting errors and implementation of system changes to reduce the risk of preventing harm. The study was conducted over a two- inadvertent administration of IV and intrathecal medica- year period with 591 nurses participating. Results showed tions by the wrong route. that the risk would have been identified and appropriate action taken in a median of 5 and average of 4 of the 6 sce- Incorrect IV administration of potassium chloride can narios. This study did not assess whether this translated potentially cause significant patient harm [39]. The Medi- into improved recognition of medication errors in prac- cation Safety Taskforce of the previous Safety and Quality tice. Further research on the impact of education on inci- Council recommended components to be included in dent rates and medication error detection in the acute care guidelines for potassium chloride [40] and case studies setting is still needed. from two Australian hospitals were developed and made available online. Evidence for systems providing clinical pharmacy services In the previous review [2] some Australian studies were A root cause analysis of an incident in which a bolus dose located which supported the role of clinical pharmacists of IV potassium chloride was inadvertently administered in improving patient safety. Newer studies, published at the Alfred hospital in Melbourne [41] led to the devel- since 2002, further support this role. opment of pre-mixed solutions. These were developed by physician consensus and in collaboration with the prod- A pre-test, post-test study examined the impact of an uct manufacturer. This allowed all concentrated potas- emergency department (ED) clinical pharmacist on pre- sium chloride preparations to be removed from all scribing errors in a Victorian metropolitan teaching hospi- general wards. A policy for prescribing potassium chloride tal [44]. Prescription error rates for patients during a 5 day in millimoles rather than grams was also implemented. control period were compared with error rates in the fol- Outcomes were not evaluated. lowing week when a pharmacist ED service was provided. In the intervention period a dedicated ED pharmacist Patient self-administration in the acute care setting interviewed patients using a structured medication recon- A pilot study in a Nursing Convalescent Unit of a large ciliation form to obtain a medication history and recon- metropolitan teaching hospital examined the effective- ciled the history with the ED medication chart where ness of an inpatient self-medication program [42]. The possible or passed the information to the ward pharma- six-month study examined three levels of administration: cist. At 24 hours post-admission a senior clinical pharma- 1) registered nurse (RN) administration; 2) patient medi- cist reviewed the medication history and medication chart cation with direct supervision from an RN; and 3) self- and recorded and resolved any prescribing errors. Error medication with indirect RN supervision. Patient educa- types were classified using an in-house classification sys- tion and a medication record card were key components tem and the risk rating was assessed by a blinded, inde- Page 7 of 14 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:24 http://www.anzhealthpolicy.com/content/6/1/24 pendent physician using standard risk assessment criteria. munity pharmacist. This included a medication transfer There were 56 patients in the control period and 55 in the summary (which supplemented normal discharge infor- intervention period with patient characteristics and the mation), coordination of a medication review by the number of drugs ordered per patient similar between pharmacist contracted to the facility and a case conference groups. There were 88 prescription errors detected in the including the pharmacist coordinator, family physician, control period (1.6 errors/patient) and 25 in the interven- community pharmacist and registered nurse from the tion period (0.5/patient) (p < 0.0001). There was a rela- facility. The main outcome measure (the Medication tive reduction of errors rated as high/extreme (64% Appropriateness Index -MAI) was assessed at discharge reduction), moderate (71% reduction) and minor (90% (baseline) and at 8 weeks post-discharge by independent reduction). This study supports the role of an ED pharma- pharmacists blinded to patient group allocation. The MAI cist in reducing prescription errors, however further con- was not significantly different between the groups at base- trolled studies in other acute care hospitals in Australia are line (intervention group 3.2 [95%CI 1.8-4.6]; control needed to determine the generalisability of these findings. group 3.7 [95% CI 2.2-5.2]), while at 8 weeks the MAI was unchanged in the intervention group (2.5 [95%CI 1.4- Less rigorous evidence for the effectiveness of clinical 3.7]), but significantly higher (worse) in the control group pharmacy interventions is provided by studies in which (6.5 [95% CI 3.9-9.1]). When patients who were alive at interventions undertaken by clinical pharmacists have the 8 week follow-up were included in the analysis, there been independently reviewed in order to assess their clin- were significantly fewer hospital admissions and ical significance or impact on patient outcomes or medi- unplanned emergency department attendances in the cation error rates. A number of studies of this type have intervention group (RR 0.38 [95% CI 0.15-0.99]). How- been undertaken in the Australian setting as summarised ever, there was no significant difference if all patients were in the previous review [2]. The largest study conducted in included in the analysis. The intervention group reported Australian acute care was published in 2004 [45]. Clinical less "worsening pain" compared to the control group (RR pharmacist interventions in eight major public hospitals 0.55 [95% CI 0.32-0.94]). No difference in adverse drug over an average of 21 days were reviewed by an independ- events was detected (RR 1.05 [95% CI 0.66-1.68]). There ent multidisciplinary panel. There were 1,399 interven- were no significant differences for falls, worsening mobil- tions during 24,866 patient separations. Of these, 96 ity, worsening behaviour or increased confusion. This interventions (7%) were judged to have reduced the study suggests a transition coordinator can improve patient's length of stay in hospital and 156 (11%) were aspects of medication management during the transition deemed to have reduced the potential for the patient to be from hospital to residential aged care, however, no impact readmitted to hospital. The clinical significance of the on adverse drug events was demonstrated. This study is intervention was deemed to be life saving in 15 (1.1%), limited by its small sample size and larger studies may be major in 351 (25%), moderate in 535 (38%) and minor required to determine whether the endpoints are sensitive in 425 (30%). to this type of intervention. Evidence for systems improving information transfer Qualitative research has been published exploring the Information transfer at the hospital-community interface potential role of a liaison pharmacist between hospital In the previous review [2] it was noted that controlled and community health care settings in Australia [47]. This studies undertaken in Australia to assess the impact of dis- study involved semi-structured interviews and a focus charge medication management services implemented by group examining the discharge process, liaison between pharmacists or by pharmacists and nurses showed hospital and community settings and the possible role of improvements on patient outcomes and reductions in a community liaison pharmacist. Participants included undesirable medication events. Further research on this medical practitioners, community nurses, community type of service has been subsequently published. pharmacists, hospital pharmacists, consumers and hospi- tal administrators from a division of general practice in A randomised, single blind, controlled trial conducted in Victoria. In general, participants felt that a community South Australia examined whether the addition of a phar- liaison service should be targeted to those most a risk of macist transition coordinator could impact on medica- medication misadventure. Potential roles for the service tion management and health outcomes in older people included providing advice and reassurance about medica- undergoing transition from a hospital to a long-term aged tions, assessment of a patient's medication understanding care facility [46]. The study included 110 older adults dis- and ability to manage their medicines at home, education charged from three metropolitan hospitals to long-term and reinforcement of instructions about medicines and care. The transition coordinator focused on the transfer of communication of patient progress with service providers. medicines information to care providers in the long-term In general, domiciliary visits were considered the most care facility and the patient's family physician and com- appropriate mechanism, however telephone calls were Page 8 of 14 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:24 http://www.anzhealthpolicy.com/content/6/1/24 also suggested. The logistics of providing the service - requesting the community pharmacist arrange an included the need for domiciliary visits to be conducted accredited pharmacist to undertake the HMR or order- within one week of discharge, but preferably within 24-48 ing a hospital-funded review if the GP and community hours. pharmacist could not be involved; Other models for development of medication liaison serv- - sending the HMR report to the hospital outpatient ices between hospital and community settings have been clinic, the GP and community pharmacist. examined in major hospitals in NSW and South Australia. Of the 50 eligible patients, 38 gave consent and 21 In NSW a 'Heartlink' medication management pathway patients received the full service. The mean time for the for patients with chronic heart failure was developed HMR was 18 ± 7 days post-discharge. Barriers to the serv- involving a community liaison pharmacist and medica- ice included GP time constraints and unwillingness to tion management review facilitator [48]. On hospital learn how to make an HMR referral. Community pharma- admission, patient consent was obtained for the hospital cist barriers included time constraints and a lack of remu- pharmacist to communicate with the patient's preferred neration. community pharmacy to obtain a complete medication history. The community pharmacist inserted an alert onto Information transfer between hospitals and general practitioners the patient's computer file. On discharge from hospital An Australian study to promote medication information the patient was given a medication list by the hospital transfer at the hospital community interface focused on pharmacist, and the community pharmacist was sent the the transfer of information between general practitioners discharge prescription, information about medication and hospitals at both hospital admission and discharge changes, any relevant information to assist long-term [50]. This quasi-pre-test post-test design study aimed to patient monitoring and risk factors for medication misad- improve communication between general practitioners venture. A community liaison pharmacist requested the and hospital staff in an Area Health Service in NSW. Stage patient's GP refer the patient for a home medication one of the project [51], indicated that compliance with review (HMR) and provided the GP with information the Australian Pharmaceutical Advisory Council (APAC) about specific medication risk factors or recommenda- National guidelines to achieve the continuum of quality use of tions from the hospital team. The liaison pharmacist medicines between hospital and community was poor, and accompanied the accredited pharmacist on the HMR visit that a number of barriers to effective communication to provide written resources and other information to the existed. A series of workshops were conducted which patient. An HMR facilitator worked to provide informa- identified changes that could be made to overcome these tion to GPs and community pharmacists about the serv- communication barriers. In stage two, progress was ice. A retrospective survey of GPs, community pharmacists assessed using specific indicators. GPs (n = 122) com- and accredited pharmacists found most agreed that the pleted questionnaires for two of their elderly patients dis- service improved the link between the hospital and com- charged from a hospital in the area health service. munity setting. Time factors and lack of patient interest Subsequently, a forum was held to review results and reas- was identified as the major barriers to the HMR process. sess action plans. Three months later another survey was Patients receiving the service who responded to the ques- conducted. In comparison with stage one, there were sub- tionnaire (n = 27) reported they were more confident tak- stantial and maintained improvements in faxing of dis- ing medications regularly after the HMR and felt they charge summaries from hospitals to GPs (p < 0.001) and learnt something from the HMR (24 respondents, 89%). provision of medication information to hospitals by GPs for patients at risk (p < 0.05). Some problems, however, In South Australia a pilot study examined a service from had changed little including a poor rate of hospital notifi- within the hospital (before discharge) to organize an cation to GPs of a patient's admission to hospital. This HMR for patients at high risk of medication misadventure study did not use adverse drug events or medication error [49]. Standard care involved mailing a discharge summary as an outcome measure. to the patient's GP. The added service involved a liaison pharmacist: Shared electronic medication records Initiatives to develop systems to improve the sharing of - sending a medication discharge summary to the medication information between patients and various patient's GP and community pharmacist; healthcare providers through a shared electronic medical record have been funded through the Australian Govern- - organising an appointment for the patient with their ment. Since the last medication safety review [2] a Medi- GP two days after discharge to order a HMR; Connect program (formerly the Better Medication Management System) began development. This program Page 9 of 14 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:24 http://www.anzhealthpolicy.com/content/6/1/24 aimed to develop a system to allow consumers to consent data management process involved including a category to various different healthcare professionals accessing on the reporting form for "adverse medication effect" and and, where necessary, using and recording information in a process to ensure that reports where medications were a a shared medication record. A secure national electronic contributing factor were communicated monthly to the system was trialled successfully in two sites in Victoria and pharmacy department. This allowed review and inclusion Tasmania in 2003. In 2004 the MediConnect program in the hospital continuous quality improvement program. was incorporated within the wider HealthConnect pro- The intervention enabled detection of adverse drug events gram [52]. HealthConnect is an ongoing partnership that would have been missed in the existing system. between the Commonwealth and State and Territory Gov- ernments to facilitate sharing of health information elec- Other studies have examined factors that influence tronically. Systems for electronic sharing of patient health adverse drug reaction (ADR) reporting [61] and incident information, including medication information between reporting [62] in the acute care setting in Australia. Sug- different settings are currently under development and gested strategies to improve reporting included improving evaluation in various states [53]. the accessibility of report forms, encouraging computer- based reporting and implementing educational initiatives Medication record cards for nurses and junior medical staff [61]. In response to The previous medication safety review [2] identified one barriers to incident reporting [62] suggestions included randomised controlled study in the hospital outpatient development of time-efficient reporting systems and setting which assessed the impact of using a medication resources to provide feedback and action in relation to card in conjunction with medication counselling for reported incidents and near-misses. Personal digital assist- improving knowledge about medications and compliance ants (PDAs) were successfully utilized by anaesthetists to [54,55]. No further Australian studies were located. facilitate incident reporting [63]. There is a need for fur- ther research in the Australian acute care setting to evalu- Evidence for systems promoting multidisciplinary care ate whether these suggested strategies could increase Medication Management Review and Case Conferencing Services reporting rates. The only controlled Australian studies assessing medica- tion management and case conferencing services have Evidence for systems-based approaches to understanding been undertaken as part of hospital-community discharge and preventing medication errors liaison studies [46,56-59]. No further studies in the acute Systems to allow health services to assess medication systems and care setting in Australia were located. performance The National Medication Safety Breakthrough Collabora- Evidence for systems to promote reporting of medication tive was a key initiative of the former Australian Council incidents and adverse drug reactions for Safety and Quality, which aimed to reduce harm from The previous medication safety review did not include medications. The collaborative of 100 health service analysis of the evidence for improving adverse drug reac- teams worked towards a goal of reducing medication- tion reporting or incident reporting [2], however it high- related harm to patients by 50% [64]. Reported achieve- lighted that routine data collection about undesirable ments included more than halving the percentage of medication events is important for understanding why patients experiencing a high-risk adverse drug event (by medication incidents occur and how they might be pre- the top 8 hospital teams) and an increase in the percent- vented. While there are well established mechanisms to age of hospitalised patients who had medicines informa- collect data on adverse drug reactions and medication tion communicated to their primary health care providers incidents in Australia, medication incidents and adverse in a timely manner - from less than 30% to over 90% (by drug reactions are still under-recognised and under- the top 8 hospital teams) [64]. The collaborative devel- reported. Continued and increased participation needs to oped "toolkits" to improve medication safety, including be encouraged. Some strategies described in the recent lit- alert cards, incident report forms, education tools for staff erature have been developed to promote participation in and patients, communication tools and guidelines for reporting. high-risk medications [65,66]. A strategy to increase the reporting of adverse drug events The NSW Therapeutic Advisory Group (NSW TAG) and in the Alfred Hospital in Melbourne involved ensuring the Clinical Excellence Commission have adapted medication errors identified via calls for the medical resources developed by the North American Institute for emergency team (MET) were included in the hospital Safe Medication Practices (ISMP). The "Medication Safety quality programs [60]. The MET provides early interven- Self Assessment for Australian Hospitals" and "Medication tion when a patient's condition deteriorates and causes of Safety Self Assessment for Antithrombotic Therapy in Austral- the deterioration are recorded. A modification of the MET ian Hospitals" are available through the NSW Government Page 10 of 14 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:24 http://www.anzhealthpolicy.com/content/6/1/24 Clinical Excellence Commission (CEC) website [67]. lar medications. The impact of the program on medica- Completed assessments can be submitted to the CEC tion infusion error rates in the hospital was not reported. through a secure site which provides a confidential online report back to the hospital. The resources are designed to These studies describe initiatives to design system-based allow hospital administrators, in conjunction with a approaches to reduce drug administration errors. There is multidisciplinary team, to self assess their hospital's per- a need for further research in the Australian acute care set- formance on elements that have been shown to improve ting examining the actual impact of these approaches on safe use of medicines generally and anti-thrombotic medication errors and adverse drug events. agents specifically. The system is currently used, but has not been evaluated for its impact on improving medica- Conclusion tion safety [68]. In 2002, the former Australian Council for Safety and Quality in Health Care, in its national report on medica- System-based approaches to drug administration errors tion safety, highlighted a number of systems solutions A study conducted at a NSW tertiary hospital examined a known to be effective in improving medication safety [2]. systems-based approach to reporting, review and feedback These included clinical decision support systems; adverse of data obtained on prescribing incidents in the hospital drug event alerts; systems that provide adequate checking, [69]. A database of prevented prescribing incidents (near- such as bar coding; individual patient medication supply miss incidents) detected by hospital pharmacists was systems; as well as provision of clinical pharmacy services developed. A pharmacist classified incidents by type and and discharge medication management services. Since potential severity, and recorded descriptive data and a this review, further Australian studies addressing the evi- brief narrative of the incident. Systems failures were iden- dence base and experiences in their implementation have tified from the data and fedback to specific clinical areas now been published. However, gaps still exist. and specialist medical officers. Clinical pharmacists spe- cialising in the clinical area were involved in multidisci- There has been significant progress in strategies to plinary forums with junior and senior medical officers in improve prescription writing for medications in hospitals which intervention reports were discussed. A survey was with the introduction of a NIMC now used in all Austral- sent to 21 senior clinicians who received reports through ian public hospitals and many private hospitals. Pub- the program, of which 10 (47%) responded. All indicated lished research has highlighted limitations with the chart, that the feedback was of value in improving prescribing however there are systems in place to allow ongoing eval- practice and was incorporated into clinical quality pro- uation and nationally coordinated strategies for making grams. Most respondents (80%) indicated they found the changes to the chart design. comparative data between departments useful. Studies have now assessed the implementation of compu- A program to reduce the potential for medication infu- terised prescribing and clinical decision support in Aus- sion-related error was undertaken at an acute care hospital tralian hospitals, suggesting computerised prescribing in Melbourne [70]. A multidisciplinary team examined alone without decision support, may lead to increased medication administration errors over a 29-month period error. Studies have also highlighted that implementation identifying root causes and contributing systems failures. of these systems in acute care must include appropriate Identified systems failures included design flaws in tech- education and training for staff, a change management nology currently in use, deviations from safe practice that strategy and a highly organised approach at all levels of had become culturally accepted, complex and variable the institution, giving consideration to the technical issues medication prescribing, unnecessary administration prac- and culture and environment in which it is to be used. tices, lack of accessible medication calculation resources Similarly, the need for standardisation of systems and an and limited accessible drug information. Improvement agreed set of national standards has been highlighted. initiatives included a medication safety education pro- There remains a lack of published research on the impact gram incorporating medication calculations initiatives, a of electronic prescribing in combination with CDSS on campaign to increase reporting of near-miss incidents, medication errors or adverse drug events when used by strategies to address unsafe practices that had become health care practitioners in the acute care setting in Aus- accepted in the hospital such as storing potassium chlo- tralia. ride ampoules in bedside drawers with other medications and poor labelling of drug infusions. Other initiatives New strategies that have been assessed include double included the implementation of an auditing program, checking versus single checking by nurses for safe medica- changes to infusion pump equipment and methods to tion administration and patient self-administration in standardize the prescribing and administration of particu- hospital. The small studies found no significant differ- ences between groups, however, the studies were only Page 11 of 14 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:24 http://www.anzhealthpolicy.com/content/6/1/24 located in single centres and possibly had insufficient Authors' contributions power to detect differences. While these studies do not SJS was the main author of Part 2 of this review and was provide sound evidence of effectiveness, they warrant fur- involved in reviewing the literature, summarising study ther research in these areas. findings and synthesis of the findings with those from the previous medication safety review. EER made a substan- Other strategies that have been implemented but the tial contribution to drafting and editing of this paper and impact on medication error rates not reported include provided direction and guidance in the review of the rele- leur incompatible systems to avoid incorrect route of vant literature. administration for intravenous and intrathecal injections, as well as the removal of concentrated potassium chloride Authors' informations from wards with replacement by pre-mixed solutions. EER is an Associate Professor and co-director in the Qual- ity Use of Medicines and Pharmacy Research Centre Academic detailing has been demonstrated to reduce (QUMPRC), Sansom Institute, University of South Aus- errors in prescriptions for Schedule 8 medicines where tralia. SJS is a Research Fellow in the QUMPRC. EER and error rates were high and an uncontrolled study suggested SJS were the primary authors of the Second National Report an education program was effective in reducing the use of on Patient Safety Report - Improving Medication Safety for error prone prescribing abbreviations in the emergency theAustralian Council for Safety and Quality in Health department setting. Care in 2002. Studies have continued to assess hospital discharge plan- Acknowledgements The authors wish to acknowledge staff of the New South Wales (NSW) ning or liaison pharmacy services primarily focusing on Medicines Information Centre, St Vincent's Hospital for conducting the implementation issues. Barriers to home medication database search for the literature review. The review was conducted with reviews after hospital discharge include workload factors financial support from the Australian Commission on Safety and Quality in for both general practitioners and pharmacists and lack of Health Care. The Commission initiated the decision to submit the manu- patient interest, as well as the ability to engage an accred- script for publication. ited pharmacist in a timely manner. One new model included a transition co-ordinator to assist transfer of References medication information for patients discharged from hos- 1. Roughead EE, Semple SJ: Medication safety in acute care in Aus- tralia: where are we now? Part 1: a review of the extent and pital to residential aged-care facilities. The model demon- causes of medication problems 2002-2008. Aust New Zealand strated an improvement in medication appropriateness, Health Policy 2009, 6(1):18. however there was no significant impact on adverse drug 2. Safety and Quality Council: Second National Report on Patient Safety, Improving Medication Safety. Canberra: Australian events, but the study may have had insufficient power to Council for Safety and Quality in Health Care; 2002. detect this endpoint. 3. Wolfstadt J, Gurwitz J, Field T, Lee M, Kalkar S, Wu W, Rochon P: The Effect of Computerized Physician Order Entry with Clinical Decision Support on the Rates of Adverse Drug No new studies were located that have assessed the impact Events: A Systematic Review. Journal of General Internal Medicine of individual patient medication supply, adverse drug 2008, 23(4):451-458. 4. Rommers M, Teepe-Twiss I, Guchelaar H-J: Preventing adverse event alerts or bar coding. The extent of their implemen- drug events in hospital practice: an overview. Pharmacoepide- tation across Australia is unknown. miology and Drug Safety 2007, 16:1129-1135. 5. Kaboli PJ, Hoth AB, McClimon BJ, Schnipper JL: Clinical Pharma- cists and Inpatient Medical Care: A Systematic Review. Arch While there is now a stronger evidence base demonstrat- Intern Med 2006, 166(9):955-964. ing that systems factors are major contributors to medica- 6. Kaushal R, Bates DW, Abramson EL, Soukup JR, Goldmann DA: Unit-based clinical pharmacists' prevention of serious medi- tion errors, there is still very limited research assessing the cation errors in pediatric inpatients. American Journal of Health- impact of an integrated set of activities on medication System Pharmacy 2008, 65(13):1254-1260. safety. 7. Shojania KG, Duncan BW, KM M, Wachter R: A Critical Analysis of Patient Safety Practices. Evidence Report/Technology Assessment No.43 (Prepared by the University of California List of abbreviations used at San Francisco-Stanford Evidence-based Practice Center), AHRQ Publication No.01-E058. Rockville, MD: Agency for ADR: (adverse drug reaction); CDSS: (clinical decision Healthcare Research and Quality; 2001. support systems); CPOE: (computerised physician order 8. Wong ICK, Wong LYL, Cranswick NE: Minimising medication entry); DOA: (drugs of addiction); ED: (emergency errors in children. Archives of Disease in Childhood 2009, 94:161-164. department); MAI: (medication appropriateness index); 9. Rex JH, Turnbull JE, Allen SJ, Voorde KV, Luther K: Systematic NIMC: (National Inpatient Medication Chart); NSW: Root Cause Analysis of Adverse Drug Events in a Tertiary (New South Wales); RN: (registered nurse). Referral Hospital. Joint Commission Journal on Quality and Patient Safety 2000, 26:563-575. 10. Hoffman C, Beard P, Greenall J, U D, White J: Canadian Root Competing interests Cause Analysis Framework. A tool for identifying and addressing the root causes of critical incidents in healthcare. The authors declare that they have no competing interests. Page 12 of 14 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:24 http://www.anzhealthpolicy.com/content/6/1/24 [http://www.patientsafetyinstitute.ca/English/toolsResources/rca/ 33. Coombes I, Maclean C, Rigby D, Petrie G: Medication administra- Documents/March%202006%20RCA%20Workbook.pdf]. tion observational study. (Abstract). Proc Soc Hosp Pharm Aust 11. National Inpatient Medication Chart [http://www.safetyand 1999, 1:119. quality.gov.au/internet/safety/publishing.nsf/content/NIMC_001] 34. Technology Solutions to Patient Misidentification. Report of 12. Leach H: National inpatient medication chart implementa- Review [http://www.safetyandquality.gov.au/internet/safety/publish tion. Journal of Pharmacy Practice & Research 2006, 36(1):6-7. ing.nsf/Content/CE1F60BEAF285FF7CA2574E400219A9F/$File/ 13. Australian Council for Safety and Quality in Health Care 19794-TechnologyReview.PDF] National Inpatient Medication Chart Pilot Aggregate Data 35. Fact Sheet Australian Catalogue of Medicines [http:// Report [http://www.health.gov.au/internet/safety/publishing.nsf/ www.nehta.gov.au/dmdocuments/ACOM%20FACT%20SHEET.pdf] Content/national-inpatient-medication-chart] 36. Jarman H, Jacobs E, Zielinski V: Medication study supports regis- 14. Bucknall TK: Implementing guidelines to improve medication tered nurses' competence for single checking. International safety for hospitalised patients: experiences from Western Journal of Nursing Practice 2002, 8(6):330-335. Health, Australia. Worldviews on Evidence-Based Nursing 2007, 37. Gilbar PJ, Carrington CV: Preventing intrathecal administration 4(1):51-53. of vincristine. Medical Journal of Australia 2004, 181(8):464. 15. Millar JA, Silla RC, Lee GE, Berwick A: The national inpatient 38. Cyna A, Simmons S, Osborn K, Andrew M: Fatal epidural infusion- medication chart: critical audit of design and performance at call for a system-wide change. Anaesthesia & Intensive Care 2002, a tertiary hospital. Medical Journal of Australia 2008, 188(2):95-99. 30(1):99-100. 16. National Inpatient Medication Chart Audit Tool - User 39. Reeve JF, Allinson YM, Stevens A: High-risk medication alert: Guide 2008 [http://www.safetyandquality.gov.au/internet/safetpub- Intravenous potassium chloride. Australian Prescriber 2008, lishing.nsf/Content/com-pubs_NIMC/$File/20186-User 28(1):14-16. Guide%202009.pdf] 40. Medication Alert. Intravenous potassium chloride can be 17. Jurisdictional Guidelines for Local Management of the fatal if given inappropriately [http://www.health.gov.au/inter- National Inpatient Medication Chart [http://www.safetyand nesafety/publishing.nsf/Content/former-pubs-archive-potassium-chlo quality.gov.au/internet/safety/publishing.nsf/Content/ ride] 80A0EF37F281A8D7CA25718F000CCC2F/$File/ 41. Lubliner M, Vreede M Van de: Preventing potassium problems nimc%20Guidelines%20for%20LM.pdf] permanently: Eliminating concentrated potassium chloride 18. National Inpatient Medication Chart Paediatric Version errors. Journal of Pharmacy Practice & Research 2006, 36(3):181-184. [http://www.safetyandquality.gov.au/internet/safety/publishing.nsf/ 42. Grantham G, McMillan V, Dunn SV, Gassner L-A, Woodcock P: content/NIMC_010_PM-Charts] Patient self-medication--a change in hospital practice. Journal 19. National Impatient Medication Chart. Long Stay Version of Clinical Nursing 2006, 15(8):962-970. (for acute care use) [http://www.safetyandquality.gov.au/internet/ 43. Coombes ID, Heel AC, Stowasser DA, Reid CM, Henderson A, safety/publishing.nsf/content/NIMC_001] Mitchell CA: Identification of medication errors by nurses dur- 20. Soumerai S, Avorn J: Principles of educational outreach ('aca- ing a simulated ward, medication safety orientation pro- demic detailing') to improve decision making. Journal of the gram. Journal of Pharmacy Practice & Research 2005, 35(3):190-194. American Medical Association 1990, 263:549-556. 44. deClifford JM, Caplygin FM, Lam SS, Leung BK: Impact of an emer- 21. Shaw J, Harris P, Keogh G, Graudins L, Perks E, Thomas PS: Error gency department pharmacist on prescribing errors in an reduction: academic detailing as a method to reduce incor- Australian hospital. Journal of Pharmacy Practice & Research 2007, rect prescriptions. European Journal of Clinical Pharmacology 2003, 37(4):284-286. 59(8-9):697-699. 45. Dooley MJ, Allen KM, Doecke CJ, Galbraith KJ, Taylor GR, Bright J, 22. Taylor SE, Chu MT, Haack LA, McGrath A, To TP: An intervention Carey DL: A prospective multicentre study of pharmacist ini- to reduce the use of error-prone prescribing abbreviations in tiated changes to drug therapy and patient management in the emergency department. Journal of Pharmacy Practice & acute care government funded hospitals. British Journal of Clini- Research 2007, 37(3):214-216. cal Pharmacology 2004, 57(4):513-521. 23. Coombes ID, Stowasser DA, Mitchell CA, Varghese P: Effect of 46. Crotty M, Rowett D, Spurling L, Giles LC, Phillips PA: Does the computerised prescribing on use of antibiotics. Medical Journal addition of a pharmacist transition coordinator improve evi- of Australia 2004, 180(3):140-141. dence-based medication management and health outcomes 24. Magrabi F, McDonnell G, Westbrook JI, Coiera E: Using an acci- in older adults moving from the hospital to a long-term care dent model to design safe electronic medication manage- facility? Results of a randomized, controlled trial. American ment systems. Medinfo 2007, 12(Pt 2):948-952. Journal Geriatric Pharmacotherapy 2004, 2(4):257-264. 25. Ribbons R, Milne L, Allan K: A difficult pill to swallow: Inpatient 47. Vuong T, Marriott JL: Potential role of the community liaison e-prescribing, rural and metropolitan experiences com- pharmacist: stakeholder views. International Journal of Pharmacy pared. Journal on Information Technology in Healthcare 2005, Practice 2006, 14(2):135-148. 3(2):79-88. 48. Blennerhassett JD, Cusack BM, Smith CD, Green L, Tribe KL: A 26. Bomba D, Land T: The feasibility of implementing an electronic novel medicines management pathway. Journal of Pharmacy prescribing decision support system: a case study of an Aus- Practice & Research 2006, 36(3):175-179. tralian public hospital. Australian Health Review 2006, 49. Yu K, Nguyen A, Shakib S, Doecke CJ, Boyce M, March G, Anderson 30(3):380-388. BA, Gilbert AL, Angley MT: Enhancing continuity of care in ther- 27. Ryan MJ: Australian electronic medication management pol- apeutics: Development of a post-discharge home medicines icy and systems. Journal of Pharmacy Practice & Research 2007, review model. Journal of Pharmacy Practice & Research 2007, 37(1):49-55. 37(1):22-26. 28. Electronic Decision Support Systems Report - 2008 [http:// 50. Mant A, Kehoe L, Cockayne NL, Kaye KI, Rotem WC: A quality use www.health.gov.au/internet/main/publishing.nsf/Content/ of medicines program for continuity of care in therapeutics 9B50F6B541401BBACA25753F007C283A/$File/ from hospital to community. Medical Journal of Australia 2002, AHIC%20EDSS%20web%20version%2012%20January%202009.pdf] 177(1):32-34. 29. McNally KM, Page MA, Sunderland VB: Failure-mode and effects 51. Mant A, Rotem WC, Kehoe L, Kaye KI: Compliance with guide- analysis in improving a drug distribution system. Am J Health lines for continuity of care in therapeutics from hospital to Syst Pharm 1997, 54(2):171-177. community. Medical Journal of Australia 2001, 174(6):277-280. 30. Boyle CM, Maxwell DJ, Meguid SE, Ouaida D, Krass I: Missed doses: 52. MediConnect [http://www.health.gov.au/internet/hconnect/pub evaluation in two drug distribution systems. Australian Journal lishing.nsf/Content/mediconnect-1lp] of Hospital Pharmacy 1998, 28(6):413-416. 53. eHealth. Health Connect [http://www.health.gov.au/healthcon 31. O'Leary K, Burke R, Kirsa S: SHPA standards of practice for the nect] distribution of medicines in Australian hospitals. Journal of 54. Lourens H: Impact of a medication card on compliance. Hei- Pharmacy Practice & Research 2006, 36(2):143-149. delberg West, Victoria: Pharmacy Department, Heidelberg Repatria- 32. Martin ED, Burgess NG, Doecke CJ: Evaluation of an automated tion Hospital; 1991. drug distribution system in an Australian teaching hospital. 55. Lourens H, Woodward MC: Impact of a medication card on Australian Journal of Hospital Pharmacy 2000, 30(4):141-145. compliance in older people. Aust J Ageing 1994, 13:72-76. Page 13 of 14 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:24 http://www.anzhealthpolicy.com/content/6/1/24 56. Collins DM, Stowasser DA: Liaison pharmacy project. the development of pharmaceutical care liaison models in the community and hospital setting. Final Report to the Pharmaceuti- cal Benefits Scheme Education Program 1998. 57. Spurling L, Gilbert A, Phillips P, Odgers C: A hospital medication discharge service: a novel approach utilising a community- based medication liaison pharmacist. (Abstract). Proc Soc Hosp Pharm Aust 2001, 2(174):. 58. Stewart S, Pearson S, Horowitz JD: Effects of a home-based inter- vention among patients with congestive heart failure dis- charged from acute hospital care. Archives of Internal Medicine 1998, 158(10):1067-1072. 59. Stewart S, Pearson S, Luke CG, Horowitz JD: Effects of home- based intervention on unplanned readmissions and out-of- hospital deaths. Journal of the American Geriatrics Society 1998, 46(2):174-180. 60. Vreede M Van De, Leong T: Using the medical emergency team to supplement an existing adverse drug event reporting framework. Journal of Pharmacy Practice & Research 2007, 37(3):197-199. 61. Kelly M, Kaye KI, David SR, Shenfield GM: Factors influencing adverse drug reaction reporting in New South Wales teach- ing hospitals. Journal of Pharmacy Practice & Research 2004, 34(1):32-35. 62. Evans SM, Berry JG, Smith BJ, Esterman A, Selim P, O'Shaughnessy J, DeWit M: Attitudes and barriers to incident reporting: a col- laborative hospital study. Quality & Safety in Health Care 2006, 15(1):39-43. 63. Freestone L, Bolsin SN, Colson M, Patrick A, Creati B: Voluntary incident reporting by anaesthetic trainees in an Australian hospital. Int J Qual Health Care 2006, 18(6):452-457. 64. Safety and Quality Council: Medication Safety Breakthrough Collaborative. Project Chronicle. Canberra: Commonwealth of Australia; 2005. 65. Safety and Quality Council: National Medication Safety Break- through Collaborative. Improvement Toolkit. Wave 1. Can- berra: Commonwealth of Australia; 2005. 66. Safety and Quality Council: National Medication Safety Break- through Collaborative. Improvement Toolkit. Wave 2. Can- berra: Commonwealth of Australia; 2005. 67. Medication Safety Self Assessment for Australian Hospitals [http://mssa.cec.health.nsw.gov.au/MSSA_introduction.html] 68. Hughes CF: Medication errors in hospitals: what can be done? [comment]. Medical Journal of Australia 2008, 188(5):267-268. 69. D'Souza DC, Koller LJ, Ng K, Thornton PD: Reporting, review and application of near-miss prescribing medication incident data. Journal of Pharmacy Practice & Research 2004, 34(3):190-193. 70. Burdeu G, Crawford R, Vreede M van de, McCann J: Taking aim at infusion confusion. Journal of Nursing Care Quality 2006, 21(2):151-159. Publish with Bio Med Central and every scientist can read your work free of charge "BioMed Central will be the most significant development for disseminating the results of biomedical researc h in our lifetime." Sir Paul Nurse, Cancer Research UK Your research papers will be: available free of charge to the entire biomedical community peer reviewed and published immediately upon acceptance cited in PubMed and archived on PubMed Central yours — you keep the copyright BioMedcentral Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv.asp Page 14 of 14 (page number not for citation purposes) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Australia and New Zealand Health Policy Springer Journals

Medication safety in acute care in Australia: where are we now? Part 2: a review of strategies and activities for improving medication safety 2002-2008

Loading next page...
 
/lp/springer-journals/medication-safety-in-acute-care-in-australia-where-are-we-now-part-2-a-YnRm9H60pa
Publisher
Springer Journals
Copyright
Copyright © 2009 by Semple and Roughead; licensee BioMed Central Ltd.
Subject
Medicine & Public Health; Public Health; Social Policy
eISSN
1743-8462
DOI
10.1186/1743-8462-6-24
pmid
19772663
Publisher site
See Article on Publisher Site

Abstract

Background: This paper presents Part 2 of a literature review examining medication safety in the Australian acute care setting. This review was undertaken for the Australian Commission on Safety and Quality in Health Care, updating the 2002 national report on medication safety. Part 2 of the review examined the Australian evidence base for approaches to build safer medication systems in acute care. Methods: A literature search was conducted to identify Australian studies and programs published from 2002 to 2008 which examined strategies and activities for improving medication safety in acute care. Results and conclusion: Since 2002 there has been significant progress in strategies to improve prescription writing in hospitals with the introduction of a National Inpatient Medication Chart. There are also systems in place to ensure a nationally coordinated approach to the ongoing optimisation of the chart. Progress has been made with Australian research examining the implementation of computerised prescribing systems with clinical decision support. These studies have highlighted barriers and facilitators to the introduction of such systems that can inform wider implementation. However, Australian studies assessing outcomes of this strategy on medication incidents or patient outcomes are still lacking. In studies assessing education for reducing medication errors, academic detailing has been demonstrated to reduce errors in prescriptions for Schedule 8 medicines and a program was shown to be effective in reducing error prone prescribing abbreviations. Published studies continue to support the role of clinical pharmacist services in improving medication safety. Studies on strategies to improve communication between different care settings, such as liaison pharmacist services, have focussed on implementation issues now that funding is available for community-based services. Double checking versus single-checking by nurses and patient self-administration in hospital has been assessed in small studies. No new studies were located assessing the impact of individual patient medication supply, adverse drug event alerts or bar coding. There is still limited research assessing the impact of an integrated systems approach on medication safety in Australian acute care. Page 1 of 14 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:24 http://www.anzhealthpolicy.com/content/6/1/24 seeking to blame the individual/s involved. This informa- Background This paper is the second in a two-part literature review of tion can be used to formulate strategies to prevent similar medication safety in the Australian acute care setting [1]. incidents, which in turn can be evaluated as part of a con- It examines the Australian evidence-base for strategies to tinuous quality improvement cycle [10]. In this review, reduce medication errors and updates the data presented root cause analysis strategies were included, as were sys- in the Second National Report on Patient Safety - Improving tems to improve the reporting of medication incidents in Medication Safety [2]. acute care. International evidence has identified a number of tools or Methods practices that may reduce medication incidents in health- Search strategy care. Computerised physician order entry (CPOE) cou- Databases and search terms used in the search strategy pled with clinical decision support systems (CDSS) is have been detailed in Part 1 of the review [1]. Criteria rel- amongst the most studied. These systems allow standard- evant to the general headings in the Second National Report ised online prescribing coupled to computerised advice to on Patient Safety - Improving Medication Safety [2] were support prescribing decisions. The systems may provide used. automated checks or alerts such as drug-drug interaction checking or drug allergy alerts. These systems can reduce The database search was supplemented with review of rel- adverse drug events in hospital settings [3] although there evant reports and resources on the Australian Commis- have been reports of increased error rates when imple- sion on Safety and Quality in Health Care website http:// mentation problems have occurred [4]. Most studies have www.safetyandquality.gov.au/. Additionally, the website examined "home-grown" CPOE/CDSS systems developed of the Australian Government Department Health and and used within an institution; commercially developed Ageing was searched for information and publications systems have been less well studied [3]. relating to government electronic health initiatives including bar coding, shared electronic medical records International studies also support the role of clinical phar- and electronic prescribing. State government websites and macists in hospital wards for preventing adverse drug the website of the New South Wales Therapeutic Advisory events [4-6]. Other strategies examined for improving Group (NSW TAG) were searched for information on medication safety in acute care include individual patient medication safety initiatives in hospitals. supply of medication, robotic systems for medication dis- pensing (including automated bedside systems), bar cod- Selection of studies for review ing of medication packaging and patients, intravenous Letters, study reports and reviews published between devices capable of performing dilutions, computerised 2002 and 2008 were selected if they reported on: medication records, changes to systems to facilitate improved communication and educational interventions - systems to promote improved prescription writing; [4,7,8]. - systems to promote dissemination of information about This review examines the published literature on the evi- medicines and improved prescriber decision-making; dence and implementation of these strategies in the Aus- tralian acute care setting. While these approaches may - systems used to promote accurate dispensing and/or dis- impact particular stages in the medication use process tribution of medicines; (such as prescribing, transcribing of information, dispens- ing, administration), no single strategy can target all - systems to ensure adequate checking; stages of the complex process. While the evidence for these strategies has been reviewed separately the impor- - systems used to promote accurate administration of tance of using a multifaceted approach, incorporating a medicines; number of these strategies, is emphasised to improve medication safety. - systems to improve management of medicines; The analysis of medication incident reports occurring - medicine-specific handling/management strategies; within an institution can allow problems with systems to be identified and corrected. "Root cause analysis" is one - clinical pharmacy services; technique which has been adapted from industries such as aviation and aerospace for use in the health care setting - systems to improve information transfer about medi- [9]. This involves identifying the "root" and contributory cines between hospital and community settings; factors that have led to a medication incident rather than Page 2 of 14 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:24 http://www.anzhealthpolicy.com/content/6/1/24 - medication management services and case conferencing; An implementation strategy for the NIMC in three Victo- rian hospitals involved an interdisciplinary steering group - systems-based approaches to understanding and pre- to support integration across the whole organization, a venting medication errors; dedicated project officer acting as facilitator and coordina- tor and four interdisciplinary working groups addressing - systems to promote reporting of medication incidents supply, communication, education and evaluation issues and adverse drug reactions in hospitals. [14]. Studies undertaken in the community setting were An audit of the NIMC design and performance was also excluded. undertaken at the Royal Perth Hospital (RPH) [15]. Chart design assessment was performed on 15 design features and compared with the previously used RPH chart, four Results and Discussion Evidence for systems to improve prescription writing other WA hospital charts and nine charts from teaching The National Inpatient Medication Chart hospitals in other states and territories. Completion of the A National Inpatient Medication Chart (NIMC) aims to individual fields of the chart was assessed and an audit of improve medication safety through standardisation of NIMC charts from six medical and surgical wards of the medication ordering in all Australian public hospitals, hospital was compared to completion of the previous and a number of private hospitals. The NIMC was RPH chart. Aspects of the design of the NICM considered adopted for national roll-out following an agreement of likely to improve medication safety included a) a section the Australian Health Ministers Council in 2004 [11]. The to complete medication history; b) provision for record- recommendation for a process of pharmaceutical review ing sustained-release dose forms; c) a mechanism to circle of all aspects of medication management in the hospital inpatient drugs intended for provision at discharge; d) was also part of this reform [12]. provision for documenting a medicine's indication; and e) direction to record intended administration times of The chart along with education on safe prescribing and each medicine. Four of these five advantages were poorly administration were piloted in 31 public and private hos- complied with in practice. Overall compliance with the pitals in metropolitan, regional and rural areas in 2004 NIMC was 56% [95% CI 43-67%]. Only 2.3% of charts [13]. Pre and post-implementation audits were under- had the medication history completed. There was a mod- taken by 28 sites. Improvements seen in the combined est, but non-significant increase in overall compliance results included: after the introduction of the NIMC compared to the previ- ous chart (6% [95% CI -0.2 to 13%]). Concerns about - increased documentation of adverse drug reactions design features of the chart included cramped design, lack (21% at baseline to 50% at three months); of colour, lack of provision for variable dosing and the need for, on average, twice as many charts per admission - decreased prescription of drugs to which patient had (and hence increased requirements for rewriting charts an allergy (9% to 6%); and possible transcription errors) [15]. - increased entry of actual administration times (18% Tools have been developed to allow ongoing evaluation to 68%); of the chart [16]. Currently local management of the NIMC is overseen by local jurisdictional bodies at the - increased frequency of providing the indication for a State and Territory level [17]. Where relevant, the local 'prn' (as needed) medication (13% to 26%); body will refer proposed changes to the national NIMC Oversight Committee. Versions of the NIMC for paediat- - increased documentation of the maximum dose for a ric patients and patients requiring long hospital stays are 'prn' medication (24% to 36%); available [18,19]. Prescriber education - increased frequency of the prescriber name being identifiable (41% to 79%); and Education services to individual healthcare professionals are sometimes referred to as 'academic detailing'. This - increased frequency of target international normal- term refers to an educational approach based on princi- ised ratio (INR) documentation (9% to 71%) [13]. ples of communications theory and behaviour change [20]. The study used surrogate, rather than direct, measures of patient harm [13]. A 2001 NSW study examined whether academic detailing could reduce prescription errors for drugs of addiction Page 3 of 14 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:24 http://www.anzhealthpolicy.com/content/6/1/24 (DOA) in the hospital setting [21]. Errors assessed vention (p < 0.001). The rates of abbreviations classified included: as of major significance decreased from 5.8 per 100 pre- scriptions pre-intervention to 2.3 post-intervention (p < - the quantity not written in both words and numbers; 0.001). The study was not controlled and the sustainabil- ity of the effect over time is unknown. - the DOA not written on a separate script; The evidence for systems ensuring better dissemination of - alterations to the script not initialled; knowledge about drugs Clinical decision support systems (CDSS) with computerised - details of the preparation or form of the drug omit- physician order entry (CPOE) ted; In the 2002 national patient safety report, careful imple- mentation and evaluation of electronic prescribing with - liquid preparations without a milligram dose given; decision support was suggested as a priority for improving medication safety in Australia [2]. Implementation - strength omitted or incorrect. research has since been conducted in Australia, however, outcome studies are still lacking. The intervention involved a one-on-one academic detail- ing session for all first and second year post-graduate prac- The limitations of using electronic prescribing alone are titioners. A follow-up session was conducted after two highlighted in the findings of a 2001 study of discharge months. Error rates were assessed two months after the prescriptions at a teaching hospital in Brisbane [23]. A intervention and compared with a control hospital. The computer-generated discharge summary system enabled a baseline levels of errors at the intervention hospital discharge prescription to be generated based on informa- (approximately 40%) were higher than those of the con- tion entered by the medical officer. An observational audit trol hospital (25%), making comparison difficult. Error of 200 discharge prescriptions was conducted; involving rates reduced from 41% of 46 prescriptions pre-interven- 100 handwritten prescriptions (605 medications) and tion to 24% of 128 prescriptions post-intervention (p < 100 computer generated prescriptions (700 medications). 0.001, chi square = 17.3). There was no change in the The same medical staff were responsible for both prescrip- error rate at the control hospital. No assessment was made tion types. There were more errors in the computer gener- on the likely to impact of these errors on patient out- ated prescriptions (81 errors, 11.6% of items) compared comes. to the handwritten ones (30 errors, 5.0% of items, p < 0.001). Errors judged to have the potential to result in Further adequately controlled studies are required to con- patient harm were similar between the groups. Errors that firm whether academic detailing can reduce prescription occurred more frequently when a computer was used error rates in the Australian hospital setting. These should included dosing errors (25 computer generated compared include assessment of errors likely to impact on patient with 5 handwritten errors) and duration of therapy errors outcomes. due to default settings in the computer. The authors con- cluded that electronic prescribing without decision sup- A pre-intervention, post-intervention study in a Mel- port and alerting systems could increase the risk of patient bourne teaching hospital examined an educational inter- harm. vention to reduce the use of "error-prone prescribing abbreviations" in an emergency department (ED) [22]. Recognition that electronic medication management sys- The intervention, involving ED registrars and postgradu- tems can introduce machine-related errors led a Univer- ate course nurses, included small group and one-to-one sity of NSW research group to develop a multilevel tutorials about abbreviations commonly causing medica- "accident model" to examine points in electronic prescrib- tion errors or confusion. Summary cards and posters rein- ing systems where system failures may occur [24]. This forced this information. The intervention ran for six- model used a systematic approach to identify human- months. All medication and fluid charts in the ED depart- computer interaction processes as well as the context in ment were assessed for error-prone abbreviations at a ran- which electronic prescribing systems are used (such as domly selected time each day for one week before the health professional cultures, organizational factors). The intervention and one week following. Abbreviations were aim is to aid the development of electronic prescribing classified as being of major, moderate or minor signifi- systems with features that improve patient safety. The cance by two independent pharmacists. Charts for 166 validity of the model is still to be tested. patients were included in the two assessment phases. The error-prone abbreviation rate per 100 prescriptions Implementation of an inpatient electronic prescribing decreased from 31.8 pre-intervention to 18.7 post-inter- and clinical decision support system in an acute and sub- Page 4 of 14 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:24 http://www.anzhealthpolicy.com/content/6/1/24 acute ward at a metropolitan and rural hospital in Victoria would be willing to adopt electronic prescribing and alert found mixed results [25]. The system allowed a 'point and systems, the implementation would be hampered by click' method of prescribing for physicians' ordering inpa- existing barriers. Barriers included lack of confidence in tient and discharge medications, integration with the hos- the security aspects of the system (e.g. use of electronic sig- pital's pharmacy ordering system and use of clinical natures, failing to log-out of a system), lack of funding decision support tool. The clinical decision support and resources, concern about the time taken to complete allowed checking of interactions, allergies and duplicate and check an electronic prescription (e.g. excessive drug- ordering and access to drug information including the drug interaction checking by some decision support/alert- AusDI drug and therapeutics database and MIMS. The sys- ing systems), lack of compatibility with the existing tem was tested in one acute and one sub-acute 30-bed patient administration system in the hospital and legisla- ward within each hospital. In each ward two clinical com- tive barriers (such as legal requirement for handwritten puters were available, in addition to a wireless laptop as a signatures). The majority of clinicians favoured the idea of point-of-care computer. Intensive training was provided a state-wide EPDS system. Clinicians stated the need for to medical officers and nursing staff. Use of the system in an integrated system that would overcome the need to log the acute ward setting was discontinued after six weeks in into different systems for different types of results. the rural hospital and eight weeks in the metropolitan hospital. The barriers to effective implementation in the A 2007 project reviewed electronic medication manage- acute ward setting were perceptions of increased clinical ment (EMM) systems in the Australian setting [27]. A risk, workload issues, lack of medical staff commitment, multidisciplinary reference group including medical, insufficient computer access and technical and software pharmacy and nursing representation as well as clinical limitations (including inadequate interaction and allergy information technology experts formulated "key princi- checking, and problems with version control). The experi- ples and core features" for assessing the suitability of an ence in the sub-acute ward setting in both hospitals was EMM systems. Systems from 11 companies were evalu- different, with use becoming accepted practice and the ated with all found to have the majority of the required system being rolled out to other sub-acute wards. Medica- core functions seen as important by the reference group. tion error rates were not assessed. The study highlighted Systems were not ranked relative to each other. It con- the need for document control, understanding clinical cluded that a number of available systems were suitable workflow issues and commitment through the whole for use in Australian hospitals, but that change manage- organization. One of the barriers in the acute care setting ment issues needed to be addressed for implementation was the need to frequently review medication charts and a to occur more widely. ban on handwritten alterations to charts. The authors con- cluded that implementation of electronic prescribing and The Australian Health Information Council (AHIC) pro- clinical decision support systems requires a highly organ- vides advice to the Australian Health Ministers Advisory ised approach at all levels of the institution, giving consid- Council (AHMAC) on information management and eration to the technical and cultural issues and technology in the health care system. A report by the environment in which it is to be used. AHIC in 2008 [28] examined Electronic Decision Support systems, with a focus on medicines, including the current Another qualitative feasibility study for an 'electronic pre- state of their implementation in Australia. This review scribing decision support' (EPDS) system was undertaken found there were some sound evidence-based programs in a NSW public hospital [26]. The Sauer's Triangle of used in hospitals in some States. In Victoria, for example, Dependencies model was used to examine the organisa- the Victorian Clinical Systems provided through Health tional context in which an information system is placed. SMART supported electronic prescribing and decision- The study used interviews and focus groups with medical making by providing patient allergies, adverse reactions staff, pharmacists, nurse managers and clinical informa- and automatic checking for duplication and drug interac- tion technology experts. Questions examined the limita- tions. However, the report highlighted that an agreed set tions of the present paper-based prescribing system, of national standards was lacking. The need to work with technical requirements for an electronic prescribing and professional bodies to examine possible barriers and decision support system, the environment in which the incentives for healthcare professionals in the uptake of system would be used, the political setting, the type of electronic decision support was identified, as was the need ward structure suited to electronic prescribing, perceived for the health care workforce to be adequately supported barriers to implementation and mechanisms for consult- to gain the skills to use the technology [28]. ing with medical staff in the design and implementation of an information system. Only 9% of medical staff and Since the previous medication safety review [2], a number 20% of nurse managers participated. Results found that of new studies have examined the implementation of elec- while nearly all participating clinicians indicated they tronic prescribing in combination with clinical decision Page 5 of 14 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:24 http://www.anzhealthpolicy.com/content/6/1/24 support systems in Australian hospitals, but not the and track administration [34]. No published studies in impact on medication error or patient outcomes. The the Australian setting were identified. potential value of these strategies now appears to be more widely recognised by government, hospitals and health An ongoing project administered by GS1 Australia (for- professionals. Published studies provide useful insights merly EAN Australia) and the National E-Health Transi- into some of the barriers to the introduction of this strat- tion Authority (NEHTA) aims to implement an Australian egy in the acute care setting that need to be considered standard coding system for medicines - the Australian Cat- with wider implementation. alogue of Medicines (ACOM) [35]. This will ensure that all prescription and non-prescription medicines (includ- Evidence for systems promoting better medication ing complementary medicines) have a globally unique distribution code. A national coding system is required to allow the Individual patient based medication distribution electronic transmission, storage and use of medication In the 2002 review of medication safety in Australia [2], information. This has the potential to facilitate the use of individual patient medication supply systems were found bar coding technology. to reduce medication errors. Two Australian studies had compared different medication distribution methods and Computer adverse drug event detection and alerts resultant errors associated with administration in the Aus- A computer system may be used to detect potential tralian hospital setting [29,30], showing a decrease in adverse drug events (such as interactions between differ- error rates. No further Australian studies were located. The ent medications, or abnormal laboratory results for a Society for Hospital Pharmacists of Australia Standards of patient taking a particular medication) and alert a Practice for the Distribution of Medicines in Australian patient's health care professional such as physician or Hospitals published in 2006 [31] state that unit-dose sys- pharmacist. The previous review [2] found limited inter- tems are the preferred method of medicines distribution national evidence but a lack of Australian research to sup- for healthcare facilities in terms of patient safety. How- port the use of such systems for improving medication ever, there is a lack of published data on the uptake of safety. No more recent Australian studies were located. individual patient supply systems for medications in Aus- Single-person versus double-person checking by nurses administering tralian hospitals despite evidence supporting its use in reducing medication errors. medications A study in a Victorian acute care hospital examined the Automated dispensing devices safety of single-checking by a registered nurse of medica- Automated medication dispensing devices are electronic tions that had required double-person checking [36]. storage devices that dispense medications in a controlled These included medications requiring calculations, drugs manner and track the use of medication. The 2002 medi- of addiction, cytotoxics, new drugs, epidurally adminis- cation safety review [2] noted that evidence for automated tered drugs, variable dose insulin, blood products and drug distribution systems in reducing medication inci- high dose potassium chloride. Medication incident dents was limited. Two studies that evaluated automated reports were assessed in the single checking study for a drug distribution in the Australian health care setting seven-month period and compared to those in the same [32,33] did not provide clear evidence of the efficacy of units in the same months of the previous year when dou- automated systems for reducing error. No further Austral- ble-person checking was standard practice. There was no ian studies since 2002 were located. significant difference between the two periods, however the number of reported administration errors was low Evidence for systems ensuring adequate checking (four in the study period and five in the previous year), Bar coding and the required study power to detect a difference was In the previous review [2] there was international evi- not reported. This study analysed medication incidents dence to support the investigation of bar coding as a strat- reported through the hospital's reporting scheme and did egy to reduce medication error. Bar-coding or other not include any independent assessment of errors. Reli- identification systems such as radio frequency identifica- ance on incident reports may have meant errors were tion tags could allow medication packaging supplied for undetected as incidents are known to be under-reported. an individual patient to be cross-checked with patient Further studies are required to provide conclusive evi- identification information (such as a hospital patient dence about the relative safety of single and double-per- identification bracelet) at the point of medication admin- son checking of high risk medications in the Australian istration [34]. Alternatively, more advanced systems could acute care setting. allow electronic prescriptions to interface automated dis- pensing systems to assemble individual medication packs Page 6 of 14 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:24 http://www.anzhealthpolicy.com/content/6/1/24 Evidence for systems to improve medication of the program. A total of 220 patients participated, with administration 45% remaining on Level 1, 26% reaching level 2 and 29% Drug packaging, storage and administration equipment reaching level 3. There were no patient initiated medica- The potential for administration of IV medications by the tion errors in the study period and two errors involving wrong route has been highlighted by cases of inadvertent staff, compared to one error in the previous six-month spinal administration of the anti-cancer medication vinc- period. This study was conducted in a specialized unit. ristine [37]. The Society for Hospital Pharmacists of Aus- The findings, therefore, are not generalisable to other tralia has recommended strategies to reduce the risk of acute care settings but warrant further investigation. error associated with cytotoxic medications, with the abo- lition of syringes for administration of vincristine in Education and training about medication administration errors favour of an infusion bag [37]. An orientation program for newly employed registered nurses at a Queensland teaching hospital examined the A system to prevent infusions being administered by the ability of nurses to identify medication errors and apply incorrect route has been developed at the Women's and strategies to prevent medication incidents [43]. The pro- Children's Hospital, Adelaide [38]. The system, called the gram used simulated medication administration scenar- Adelaide Regional Connector (ARC), was under prototype ios of frequently occurring medication errors with development in 2002. This colour coded luer incompati- potential for harm. After each scenario the nurses were ble system ensures that syringes and other drug adminis- asked whether they detected the errors, whether they tration equipment used to administer epidural and would have modified their practice and whether they were intrathecal doses were not able to be connected to those aware of the error concept. Nurses were presented with used to administer IV infusions. education about concepts of human error and risks, the systems in place in the hospital to prevent medication Further research is needed to examine the current state of errors, roles and responsibilities in detecting errors and implementation of system changes to reduce the risk of preventing harm. The study was conducted over a two- inadvertent administration of IV and intrathecal medica- year period with 591 nurses participating. Results showed tions by the wrong route. that the risk would have been identified and appropriate action taken in a median of 5 and average of 4 of the 6 sce- Incorrect IV administration of potassium chloride can narios. This study did not assess whether this translated potentially cause significant patient harm [39]. The Medi- into improved recognition of medication errors in prac- cation Safety Taskforce of the previous Safety and Quality tice. Further research on the impact of education on inci- Council recommended components to be included in dent rates and medication error detection in the acute care guidelines for potassium chloride [40] and case studies setting is still needed. from two Australian hospitals were developed and made available online. Evidence for systems providing clinical pharmacy services In the previous review [2] some Australian studies were A root cause analysis of an incident in which a bolus dose located which supported the role of clinical pharmacists of IV potassium chloride was inadvertently administered in improving patient safety. Newer studies, published at the Alfred hospital in Melbourne [41] led to the devel- since 2002, further support this role. opment of pre-mixed solutions. These were developed by physician consensus and in collaboration with the prod- A pre-test, post-test study examined the impact of an uct manufacturer. This allowed all concentrated potas- emergency department (ED) clinical pharmacist on pre- sium chloride preparations to be removed from all scribing errors in a Victorian metropolitan teaching hospi- general wards. A policy for prescribing potassium chloride tal [44]. Prescription error rates for patients during a 5 day in millimoles rather than grams was also implemented. control period were compared with error rates in the fol- Outcomes were not evaluated. lowing week when a pharmacist ED service was provided. In the intervention period a dedicated ED pharmacist Patient self-administration in the acute care setting interviewed patients using a structured medication recon- A pilot study in a Nursing Convalescent Unit of a large ciliation form to obtain a medication history and recon- metropolitan teaching hospital examined the effective- ciled the history with the ED medication chart where ness of an inpatient self-medication program [42]. The possible or passed the information to the ward pharma- six-month study examined three levels of administration: cist. At 24 hours post-admission a senior clinical pharma- 1) registered nurse (RN) administration; 2) patient medi- cist reviewed the medication history and medication chart cation with direct supervision from an RN; and 3) self- and recorded and resolved any prescribing errors. Error medication with indirect RN supervision. Patient educa- types were classified using an in-house classification sys- tion and a medication record card were key components tem and the risk rating was assessed by a blinded, inde- Page 7 of 14 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:24 http://www.anzhealthpolicy.com/content/6/1/24 pendent physician using standard risk assessment criteria. munity pharmacist. This included a medication transfer There were 56 patients in the control period and 55 in the summary (which supplemented normal discharge infor- intervention period with patient characteristics and the mation), coordination of a medication review by the number of drugs ordered per patient similar between pharmacist contracted to the facility and a case conference groups. There were 88 prescription errors detected in the including the pharmacist coordinator, family physician, control period (1.6 errors/patient) and 25 in the interven- community pharmacist and registered nurse from the tion period (0.5/patient) (p < 0.0001). There was a rela- facility. The main outcome measure (the Medication tive reduction of errors rated as high/extreme (64% Appropriateness Index -MAI) was assessed at discharge reduction), moderate (71% reduction) and minor (90% (baseline) and at 8 weeks post-discharge by independent reduction). This study supports the role of an ED pharma- pharmacists blinded to patient group allocation. The MAI cist in reducing prescription errors, however further con- was not significantly different between the groups at base- trolled studies in other acute care hospitals in Australia are line (intervention group 3.2 [95%CI 1.8-4.6]; control needed to determine the generalisability of these findings. group 3.7 [95% CI 2.2-5.2]), while at 8 weeks the MAI was unchanged in the intervention group (2.5 [95%CI 1.4- Less rigorous evidence for the effectiveness of clinical 3.7]), but significantly higher (worse) in the control group pharmacy interventions is provided by studies in which (6.5 [95% CI 3.9-9.1]). When patients who were alive at interventions undertaken by clinical pharmacists have the 8 week follow-up were included in the analysis, there been independently reviewed in order to assess their clin- were significantly fewer hospital admissions and ical significance or impact on patient outcomes or medi- unplanned emergency department attendances in the cation error rates. A number of studies of this type have intervention group (RR 0.38 [95% CI 0.15-0.99]). How- been undertaken in the Australian setting as summarised ever, there was no significant difference if all patients were in the previous review [2]. The largest study conducted in included in the analysis. The intervention group reported Australian acute care was published in 2004 [45]. Clinical less "worsening pain" compared to the control group (RR pharmacist interventions in eight major public hospitals 0.55 [95% CI 0.32-0.94]). No difference in adverse drug over an average of 21 days were reviewed by an independ- events was detected (RR 1.05 [95% CI 0.66-1.68]). There ent multidisciplinary panel. There were 1,399 interven- were no significant differences for falls, worsening mobil- tions during 24,866 patient separations. Of these, 96 ity, worsening behaviour or increased confusion. This interventions (7%) were judged to have reduced the study suggests a transition coordinator can improve patient's length of stay in hospital and 156 (11%) were aspects of medication management during the transition deemed to have reduced the potential for the patient to be from hospital to residential aged care, however, no impact readmitted to hospital. The clinical significance of the on adverse drug events was demonstrated. This study is intervention was deemed to be life saving in 15 (1.1%), limited by its small sample size and larger studies may be major in 351 (25%), moderate in 535 (38%) and minor required to determine whether the endpoints are sensitive in 425 (30%). to this type of intervention. Evidence for systems improving information transfer Qualitative research has been published exploring the Information transfer at the hospital-community interface potential role of a liaison pharmacist between hospital In the previous review [2] it was noted that controlled and community health care settings in Australia [47]. This studies undertaken in Australia to assess the impact of dis- study involved semi-structured interviews and a focus charge medication management services implemented by group examining the discharge process, liaison between pharmacists or by pharmacists and nurses showed hospital and community settings and the possible role of improvements on patient outcomes and reductions in a community liaison pharmacist. Participants included undesirable medication events. Further research on this medical practitioners, community nurses, community type of service has been subsequently published. pharmacists, hospital pharmacists, consumers and hospi- tal administrators from a division of general practice in A randomised, single blind, controlled trial conducted in Victoria. In general, participants felt that a community South Australia examined whether the addition of a phar- liaison service should be targeted to those most a risk of macist transition coordinator could impact on medica- medication misadventure. Potential roles for the service tion management and health outcomes in older people included providing advice and reassurance about medica- undergoing transition from a hospital to a long-term aged tions, assessment of a patient's medication understanding care facility [46]. The study included 110 older adults dis- and ability to manage their medicines at home, education charged from three metropolitan hospitals to long-term and reinforcement of instructions about medicines and care. The transition coordinator focused on the transfer of communication of patient progress with service providers. medicines information to care providers in the long-term In general, domiciliary visits were considered the most care facility and the patient's family physician and com- appropriate mechanism, however telephone calls were Page 8 of 14 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:24 http://www.anzhealthpolicy.com/content/6/1/24 also suggested. The logistics of providing the service - requesting the community pharmacist arrange an included the need for domiciliary visits to be conducted accredited pharmacist to undertake the HMR or order- within one week of discharge, but preferably within 24-48 ing a hospital-funded review if the GP and community hours. pharmacist could not be involved; Other models for development of medication liaison serv- - sending the HMR report to the hospital outpatient ices between hospital and community settings have been clinic, the GP and community pharmacist. examined in major hospitals in NSW and South Australia. Of the 50 eligible patients, 38 gave consent and 21 In NSW a 'Heartlink' medication management pathway patients received the full service. The mean time for the for patients with chronic heart failure was developed HMR was 18 ± 7 days post-discharge. Barriers to the serv- involving a community liaison pharmacist and medica- ice included GP time constraints and unwillingness to tion management review facilitator [48]. On hospital learn how to make an HMR referral. Community pharma- admission, patient consent was obtained for the hospital cist barriers included time constraints and a lack of remu- pharmacist to communicate with the patient's preferred neration. community pharmacy to obtain a complete medication history. The community pharmacist inserted an alert onto Information transfer between hospitals and general practitioners the patient's computer file. On discharge from hospital An Australian study to promote medication information the patient was given a medication list by the hospital transfer at the hospital community interface focused on pharmacist, and the community pharmacist was sent the the transfer of information between general practitioners discharge prescription, information about medication and hospitals at both hospital admission and discharge changes, any relevant information to assist long-term [50]. This quasi-pre-test post-test design study aimed to patient monitoring and risk factors for medication misad- improve communication between general practitioners venture. A community liaison pharmacist requested the and hospital staff in an Area Health Service in NSW. Stage patient's GP refer the patient for a home medication one of the project [51], indicated that compliance with review (HMR) and provided the GP with information the Australian Pharmaceutical Advisory Council (APAC) about specific medication risk factors or recommenda- National guidelines to achieve the continuum of quality use of tions from the hospital team. The liaison pharmacist medicines between hospital and community was poor, and accompanied the accredited pharmacist on the HMR visit that a number of barriers to effective communication to provide written resources and other information to the existed. A series of workshops were conducted which patient. An HMR facilitator worked to provide informa- identified changes that could be made to overcome these tion to GPs and community pharmacists about the serv- communication barriers. In stage two, progress was ice. A retrospective survey of GPs, community pharmacists assessed using specific indicators. GPs (n = 122) com- and accredited pharmacists found most agreed that the pleted questionnaires for two of their elderly patients dis- service improved the link between the hospital and com- charged from a hospital in the area health service. munity setting. Time factors and lack of patient interest Subsequently, a forum was held to review results and reas- was identified as the major barriers to the HMR process. sess action plans. Three months later another survey was Patients receiving the service who responded to the ques- conducted. In comparison with stage one, there were sub- tionnaire (n = 27) reported they were more confident tak- stantial and maintained improvements in faxing of dis- ing medications regularly after the HMR and felt they charge summaries from hospitals to GPs (p < 0.001) and learnt something from the HMR (24 respondents, 89%). provision of medication information to hospitals by GPs for patients at risk (p < 0.05). Some problems, however, In South Australia a pilot study examined a service from had changed little including a poor rate of hospital notifi- within the hospital (before discharge) to organize an cation to GPs of a patient's admission to hospital. This HMR for patients at high risk of medication misadventure study did not use adverse drug events or medication error [49]. Standard care involved mailing a discharge summary as an outcome measure. to the patient's GP. The added service involved a liaison pharmacist: Shared electronic medication records Initiatives to develop systems to improve the sharing of - sending a medication discharge summary to the medication information between patients and various patient's GP and community pharmacist; healthcare providers through a shared electronic medical record have been funded through the Australian Govern- - organising an appointment for the patient with their ment. Since the last medication safety review [2] a Medi- GP two days after discharge to order a HMR; Connect program (formerly the Better Medication Management System) began development. This program Page 9 of 14 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:24 http://www.anzhealthpolicy.com/content/6/1/24 aimed to develop a system to allow consumers to consent data management process involved including a category to various different healthcare professionals accessing on the reporting form for "adverse medication effect" and and, where necessary, using and recording information in a process to ensure that reports where medications were a a shared medication record. A secure national electronic contributing factor were communicated monthly to the system was trialled successfully in two sites in Victoria and pharmacy department. This allowed review and inclusion Tasmania in 2003. In 2004 the MediConnect program in the hospital continuous quality improvement program. was incorporated within the wider HealthConnect pro- The intervention enabled detection of adverse drug events gram [52]. HealthConnect is an ongoing partnership that would have been missed in the existing system. between the Commonwealth and State and Territory Gov- ernments to facilitate sharing of health information elec- Other studies have examined factors that influence tronically. Systems for electronic sharing of patient health adverse drug reaction (ADR) reporting [61] and incident information, including medication information between reporting [62] in the acute care setting in Australia. Sug- different settings are currently under development and gested strategies to improve reporting included improving evaluation in various states [53]. the accessibility of report forms, encouraging computer- based reporting and implementing educational initiatives Medication record cards for nurses and junior medical staff [61]. In response to The previous medication safety review [2] identified one barriers to incident reporting [62] suggestions included randomised controlled study in the hospital outpatient development of time-efficient reporting systems and setting which assessed the impact of using a medication resources to provide feedback and action in relation to card in conjunction with medication counselling for reported incidents and near-misses. Personal digital assist- improving knowledge about medications and compliance ants (PDAs) were successfully utilized by anaesthetists to [54,55]. No further Australian studies were located. facilitate incident reporting [63]. There is a need for fur- ther research in the Australian acute care setting to evalu- Evidence for systems promoting multidisciplinary care ate whether these suggested strategies could increase Medication Management Review and Case Conferencing Services reporting rates. The only controlled Australian studies assessing medica- tion management and case conferencing services have Evidence for systems-based approaches to understanding been undertaken as part of hospital-community discharge and preventing medication errors liaison studies [46,56-59]. No further studies in the acute Systems to allow health services to assess medication systems and care setting in Australia were located. performance The National Medication Safety Breakthrough Collabora- Evidence for systems to promote reporting of medication tive was a key initiative of the former Australian Council incidents and adverse drug reactions for Safety and Quality, which aimed to reduce harm from The previous medication safety review did not include medications. The collaborative of 100 health service analysis of the evidence for improving adverse drug reac- teams worked towards a goal of reducing medication- tion reporting or incident reporting [2], however it high- related harm to patients by 50% [64]. Reported achieve- lighted that routine data collection about undesirable ments included more than halving the percentage of medication events is important for understanding why patients experiencing a high-risk adverse drug event (by medication incidents occur and how they might be pre- the top 8 hospital teams) and an increase in the percent- vented. While there are well established mechanisms to age of hospitalised patients who had medicines informa- collect data on adverse drug reactions and medication tion communicated to their primary health care providers incidents in Australia, medication incidents and adverse in a timely manner - from less than 30% to over 90% (by drug reactions are still under-recognised and under- the top 8 hospital teams) [64]. The collaborative devel- reported. Continued and increased participation needs to oped "toolkits" to improve medication safety, including be encouraged. Some strategies described in the recent lit- alert cards, incident report forms, education tools for staff erature have been developed to promote participation in and patients, communication tools and guidelines for reporting. high-risk medications [65,66]. A strategy to increase the reporting of adverse drug events The NSW Therapeutic Advisory Group (NSW TAG) and in the Alfred Hospital in Melbourne involved ensuring the Clinical Excellence Commission have adapted medication errors identified via calls for the medical resources developed by the North American Institute for emergency team (MET) were included in the hospital Safe Medication Practices (ISMP). The "Medication Safety quality programs [60]. The MET provides early interven- Self Assessment for Australian Hospitals" and "Medication tion when a patient's condition deteriorates and causes of Safety Self Assessment for Antithrombotic Therapy in Austral- the deterioration are recorded. A modification of the MET ian Hospitals" are available through the NSW Government Page 10 of 14 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:24 http://www.anzhealthpolicy.com/content/6/1/24 Clinical Excellence Commission (CEC) website [67]. lar medications. The impact of the program on medica- Completed assessments can be submitted to the CEC tion infusion error rates in the hospital was not reported. through a secure site which provides a confidential online report back to the hospital. The resources are designed to These studies describe initiatives to design system-based allow hospital administrators, in conjunction with a approaches to reduce drug administration errors. There is multidisciplinary team, to self assess their hospital's per- a need for further research in the Australian acute care set- formance on elements that have been shown to improve ting examining the actual impact of these approaches on safe use of medicines generally and anti-thrombotic medication errors and adverse drug events. agents specifically. The system is currently used, but has not been evaluated for its impact on improving medica- Conclusion tion safety [68]. In 2002, the former Australian Council for Safety and Quality in Health Care, in its national report on medica- System-based approaches to drug administration errors tion safety, highlighted a number of systems solutions A study conducted at a NSW tertiary hospital examined a known to be effective in improving medication safety [2]. systems-based approach to reporting, review and feedback These included clinical decision support systems; adverse of data obtained on prescribing incidents in the hospital drug event alerts; systems that provide adequate checking, [69]. A database of prevented prescribing incidents (near- such as bar coding; individual patient medication supply miss incidents) detected by hospital pharmacists was systems; as well as provision of clinical pharmacy services developed. A pharmacist classified incidents by type and and discharge medication management services. Since potential severity, and recorded descriptive data and a this review, further Australian studies addressing the evi- brief narrative of the incident. Systems failures were iden- dence base and experiences in their implementation have tified from the data and fedback to specific clinical areas now been published. However, gaps still exist. and specialist medical officers. Clinical pharmacists spe- cialising in the clinical area were involved in multidisci- There has been significant progress in strategies to plinary forums with junior and senior medical officers in improve prescription writing for medications in hospitals which intervention reports were discussed. A survey was with the introduction of a NIMC now used in all Austral- sent to 21 senior clinicians who received reports through ian public hospitals and many private hospitals. Pub- the program, of which 10 (47%) responded. All indicated lished research has highlighted limitations with the chart, that the feedback was of value in improving prescribing however there are systems in place to allow ongoing eval- practice and was incorporated into clinical quality pro- uation and nationally coordinated strategies for making grams. Most respondents (80%) indicated they found the changes to the chart design. comparative data between departments useful. Studies have now assessed the implementation of compu- A program to reduce the potential for medication infu- terised prescribing and clinical decision support in Aus- sion-related error was undertaken at an acute care hospital tralian hospitals, suggesting computerised prescribing in Melbourne [70]. A multidisciplinary team examined alone without decision support, may lead to increased medication administration errors over a 29-month period error. Studies have also highlighted that implementation identifying root causes and contributing systems failures. of these systems in acute care must include appropriate Identified systems failures included design flaws in tech- education and training for staff, a change management nology currently in use, deviations from safe practice that strategy and a highly organised approach at all levels of had become culturally accepted, complex and variable the institution, giving consideration to the technical issues medication prescribing, unnecessary administration prac- and culture and environment in which it is to be used. tices, lack of accessible medication calculation resources Similarly, the need for standardisation of systems and an and limited accessible drug information. Improvement agreed set of national standards has been highlighted. initiatives included a medication safety education pro- There remains a lack of published research on the impact gram incorporating medication calculations initiatives, a of electronic prescribing in combination with CDSS on campaign to increase reporting of near-miss incidents, medication errors or adverse drug events when used by strategies to address unsafe practices that had become health care practitioners in the acute care setting in Aus- accepted in the hospital such as storing potassium chlo- tralia. ride ampoules in bedside drawers with other medications and poor labelling of drug infusions. Other initiatives New strategies that have been assessed include double included the implementation of an auditing program, checking versus single checking by nurses for safe medica- changes to infusion pump equipment and methods to tion administration and patient self-administration in standardize the prescribing and administration of particu- hospital. The small studies found no significant differ- ences between groups, however, the studies were only Page 11 of 14 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:24 http://www.anzhealthpolicy.com/content/6/1/24 located in single centres and possibly had insufficient Authors' contributions power to detect differences. While these studies do not SJS was the main author of Part 2 of this review and was provide sound evidence of effectiveness, they warrant fur- involved in reviewing the literature, summarising study ther research in these areas. findings and synthesis of the findings with those from the previous medication safety review. EER made a substan- Other strategies that have been implemented but the tial contribution to drafting and editing of this paper and impact on medication error rates not reported include provided direction and guidance in the review of the rele- leur incompatible systems to avoid incorrect route of vant literature. administration for intravenous and intrathecal injections, as well as the removal of concentrated potassium chloride Authors' informations from wards with replacement by pre-mixed solutions. EER is an Associate Professor and co-director in the Qual- ity Use of Medicines and Pharmacy Research Centre Academic detailing has been demonstrated to reduce (QUMPRC), Sansom Institute, University of South Aus- errors in prescriptions for Schedule 8 medicines where tralia. SJS is a Research Fellow in the QUMPRC. EER and error rates were high and an uncontrolled study suggested SJS were the primary authors of the Second National Report an education program was effective in reducing the use of on Patient Safety Report - Improving Medication Safety for error prone prescribing abbreviations in the emergency theAustralian Council for Safety and Quality in Health department setting. Care in 2002. Studies have continued to assess hospital discharge plan- Acknowledgements The authors wish to acknowledge staff of the New South Wales (NSW) ning or liaison pharmacy services primarily focusing on Medicines Information Centre, St Vincent's Hospital for conducting the implementation issues. Barriers to home medication database search for the literature review. The review was conducted with reviews after hospital discharge include workload factors financial support from the Australian Commission on Safety and Quality in for both general practitioners and pharmacists and lack of Health Care. The Commission initiated the decision to submit the manu- patient interest, as well as the ability to engage an accred- script for publication. ited pharmacist in a timely manner. One new model included a transition co-ordinator to assist transfer of References medication information for patients discharged from hos- 1. Roughead EE, Semple SJ: Medication safety in acute care in Aus- tralia: where are we now? Part 1: a review of the extent and pital to residential aged-care facilities. The model demon- causes of medication problems 2002-2008. Aust New Zealand strated an improvement in medication appropriateness, Health Policy 2009, 6(1):18. however there was no significant impact on adverse drug 2. Safety and Quality Council: Second National Report on Patient Safety, Improving Medication Safety. Canberra: Australian events, but the study may have had insufficient power to Council for Safety and Quality in Health Care; 2002. detect this endpoint. 3. Wolfstadt J, Gurwitz J, Field T, Lee M, Kalkar S, Wu W, Rochon P: The Effect of Computerized Physician Order Entry with Clinical Decision Support on the Rates of Adverse Drug No new studies were located that have assessed the impact Events: A Systematic Review. Journal of General Internal Medicine of individual patient medication supply, adverse drug 2008, 23(4):451-458. 4. Rommers M, Teepe-Twiss I, Guchelaar H-J: Preventing adverse event alerts or bar coding. The extent of their implemen- drug events in hospital practice: an overview. Pharmacoepide- tation across Australia is unknown. miology and Drug Safety 2007, 16:1129-1135. 5. Kaboli PJ, Hoth AB, McClimon BJ, Schnipper JL: Clinical Pharma- cists and Inpatient Medical Care: A Systematic Review. Arch While there is now a stronger evidence base demonstrat- Intern Med 2006, 166(9):955-964. ing that systems factors are major contributors to medica- 6. Kaushal R, Bates DW, Abramson EL, Soukup JR, Goldmann DA: Unit-based clinical pharmacists' prevention of serious medi- tion errors, there is still very limited research assessing the cation errors in pediatric inpatients. American Journal of Health- impact of an integrated set of activities on medication System Pharmacy 2008, 65(13):1254-1260. safety. 7. Shojania KG, Duncan BW, KM M, Wachter R: A Critical Analysis of Patient Safety Practices. Evidence Report/Technology Assessment No.43 (Prepared by the University of California List of abbreviations used at San Francisco-Stanford Evidence-based Practice Center), AHRQ Publication No.01-E058. Rockville, MD: Agency for ADR: (adverse drug reaction); CDSS: (clinical decision Healthcare Research and Quality; 2001. support systems); CPOE: (computerised physician order 8. Wong ICK, Wong LYL, Cranswick NE: Minimising medication entry); DOA: (drugs of addiction); ED: (emergency errors in children. Archives of Disease in Childhood 2009, 94:161-164. department); MAI: (medication appropriateness index); 9. Rex JH, Turnbull JE, Allen SJ, Voorde KV, Luther K: Systematic NIMC: (National Inpatient Medication Chart); NSW: Root Cause Analysis of Adverse Drug Events in a Tertiary (New South Wales); RN: (registered nurse). Referral Hospital. Joint Commission Journal on Quality and Patient Safety 2000, 26:563-575. 10. Hoffman C, Beard P, Greenall J, U D, White J: Canadian Root Competing interests Cause Analysis Framework. A tool for identifying and addressing the root causes of critical incidents in healthcare. The authors declare that they have no competing interests. Page 12 of 14 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:24 http://www.anzhealthpolicy.com/content/6/1/24 [http://www.patientsafetyinstitute.ca/English/toolsResources/rca/ 33. Coombes I, Maclean C, Rigby D, Petrie G: Medication administra- Documents/March%202006%20RCA%20Workbook.pdf]. tion observational study. (Abstract). Proc Soc Hosp Pharm Aust 11. National Inpatient Medication Chart [http://www.safetyand 1999, 1:119. quality.gov.au/internet/safety/publishing.nsf/content/NIMC_001] 34. Technology Solutions to Patient Misidentification. Report of 12. Leach H: National inpatient medication chart implementa- Review [http://www.safetyandquality.gov.au/internet/safety/publish tion. Journal of Pharmacy Practice & Research 2006, 36(1):6-7. ing.nsf/Content/CE1F60BEAF285FF7CA2574E400219A9F/$File/ 13. Australian Council for Safety and Quality in Health Care 19794-TechnologyReview.PDF] National Inpatient Medication Chart Pilot Aggregate Data 35. Fact Sheet Australian Catalogue of Medicines [http:// Report [http://www.health.gov.au/internet/safety/publishing.nsf/ www.nehta.gov.au/dmdocuments/ACOM%20FACT%20SHEET.pdf] Content/national-inpatient-medication-chart] 36. Jarman H, Jacobs E, Zielinski V: Medication study supports regis- 14. Bucknall TK: Implementing guidelines to improve medication tered nurses' competence for single checking. International safety for hospitalised patients: experiences from Western Journal of Nursing Practice 2002, 8(6):330-335. Health, Australia. Worldviews on Evidence-Based Nursing 2007, 37. Gilbar PJ, Carrington CV: Preventing intrathecal administration 4(1):51-53. of vincristine. Medical Journal of Australia 2004, 181(8):464. 15. Millar JA, Silla RC, Lee GE, Berwick A: The national inpatient 38. Cyna A, Simmons S, Osborn K, Andrew M: Fatal epidural infusion- medication chart: critical audit of design and performance at call for a system-wide change. Anaesthesia & Intensive Care 2002, a tertiary hospital. Medical Journal of Australia 2008, 188(2):95-99. 30(1):99-100. 16. National Inpatient Medication Chart Audit Tool - User 39. Reeve JF, Allinson YM, Stevens A: High-risk medication alert: Guide 2008 [http://www.safetyandquality.gov.au/internet/safetpub- Intravenous potassium chloride. Australian Prescriber 2008, lishing.nsf/Content/com-pubs_NIMC/$File/20186-User 28(1):14-16. Guide%202009.pdf] 40. Medication Alert. Intravenous potassium chloride can be 17. Jurisdictional Guidelines for Local Management of the fatal if given inappropriately [http://www.health.gov.au/inter- National Inpatient Medication Chart [http://www.safetyand nesafety/publishing.nsf/Content/former-pubs-archive-potassium-chlo quality.gov.au/internet/safety/publishing.nsf/Content/ ride] 80A0EF37F281A8D7CA25718F000CCC2F/$File/ 41. Lubliner M, Vreede M Van de: Preventing potassium problems nimc%20Guidelines%20for%20LM.pdf] permanently: Eliminating concentrated potassium chloride 18. National Inpatient Medication Chart Paediatric Version errors. Journal of Pharmacy Practice & Research 2006, 36(3):181-184. [http://www.safetyandquality.gov.au/internet/safety/publishing.nsf/ 42. Grantham G, McMillan V, Dunn SV, Gassner L-A, Woodcock P: content/NIMC_010_PM-Charts] Patient self-medication--a change in hospital practice. Journal 19. National Impatient Medication Chart. Long Stay Version of Clinical Nursing 2006, 15(8):962-970. (for acute care use) [http://www.safetyandquality.gov.au/internet/ 43. Coombes ID, Heel AC, Stowasser DA, Reid CM, Henderson A, safety/publishing.nsf/content/NIMC_001] Mitchell CA: Identification of medication errors by nurses dur- 20. Soumerai S, Avorn J: Principles of educational outreach ('aca- ing a simulated ward, medication safety orientation pro- demic detailing') to improve decision making. Journal of the gram. Journal of Pharmacy Practice & Research 2005, 35(3):190-194. American Medical Association 1990, 263:549-556. 44. deClifford JM, Caplygin FM, Lam SS, Leung BK: Impact of an emer- 21. Shaw J, Harris P, Keogh G, Graudins L, Perks E, Thomas PS: Error gency department pharmacist on prescribing errors in an reduction: academic detailing as a method to reduce incor- Australian hospital. Journal of Pharmacy Practice & Research 2007, rect prescriptions. European Journal of Clinical Pharmacology 2003, 37(4):284-286. 59(8-9):697-699. 45. Dooley MJ, Allen KM, Doecke CJ, Galbraith KJ, Taylor GR, Bright J, 22. Taylor SE, Chu MT, Haack LA, McGrath A, To TP: An intervention Carey DL: A prospective multicentre study of pharmacist ini- to reduce the use of error-prone prescribing abbreviations in tiated changes to drug therapy and patient management in the emergency department. Journal of Pharmacy Practice & acute care government funded hospitals. British Journal of Clini- Research 2007, 37(3):214-216. cal Pharmacology 2004, 57(4):513-521. 23. Coombes ID, Stowasser DA, Mitchell CA, Varghese P: Effect of 46. Crotty M, Rowett D, Spurling L, Giles LC, Phillips PA: Does the computerised prescribing on use of antibiotics. Medical Journal addition of a pharmacist transition coordinator improve evi- of Australia 2004, 180(3):140-141. dence-based medication management and health outcomes 24. Magrabi F, McDonnell G, Westbrook JI, Coiera E: Using an acci- in older adults moving from the hospital to a long-term care dent model to design safe electronic medication manage- facility? Results of a randomized, controlled trial. American ment systems. Medinfo 2007, 12(Pt 2):948-952. Journal Geriatric Pharmacotherapy 2004, 2(4):257-264. 25. Ribbons R, Milne L, Allan K: A difficult pill to swallow: Inpatient 47. Vuong T, Marriott JL: Potential role of the community liaison e-prescribing, rural and metropolitan experiences com- pharmacist: stakeholder views. International Journal of Pharmacy pared. Journal on Information Technology in Healthcare 2005, Practice 2006, 14(2):135-148. 3(2):79-88. 48. Blennerhassett JD, Cusack BM, Smith CD, Green L, Tribe KL: A 26. Bomba D, Land T: The feasibility of implementing an electronic novel medicines management pathway. Journal of Pharmacy prescribing decision support system: a case study of an Aus- Practice & Research 2006, 36(3):175-179. tralian public hospital. Australian Health Review 2006, 49. Yu K, Nguyen A, Shakib S, Doecke CJ, Boyce M, March G, Anderson 30(3):380-388. BA, Gilbert AL, Angley MT: Enhancing continuity of care in ther- 27. Ryan MJ: Australian electronic medication management pol- apeutics: Development of a post-discharge home medicines icy and systems. Journal of Pharmacy Practice & Research 2007, review model. Journal of Pharmacy Practice & Research 2007, 37(1):49-55. 37(1):22-26. 28. Electronic Decision Support Systems Report - 2008 [http:// 50. Mant A, Kehoe L, Cockayne NL, Kaye KI, Rotem WC: A quality use www.health.gov.au/internet/main/publishing.nsf/Content/ of medicines program for continuity of care in therapeutics 9B50F6B541401BBACA25753F007C283A/$File/ from hospital to community. Medical Journal of Australia 2002, AHIC%20EDSS%20web%20version%2012%20January%202009.pdf] 177(1):32-34. 29. McNally KM, Page MA, Sunderland VB: Failure-mode and effects 51. Mant A, Rotem WC, Kehoe L, Kaye KI: Compliance with guide- analysis in improving a drug distribution system. Am J Health lines for continuity of care in therapeutics from hospital to Syst Pharm 1997, 54(2):171-177. community. Medical Journal of Australia 2001, 174(6):277-280. 30. Boyle CM, Maxwell DJ, Meguid SE, Ouaida D, Krass I: Missed doses: 52. MediConnect [http://www.health.gov.au/internet/hconnect/pub evaluation in two drug distribution systems. Australian Journal lishing.nsf/Content/mediconnect-1lp] of Hospital Pharmacy 1998, 28(6):413-416. 53. eHealth. Health Connect [http://www.health.gov.au/healthcon 31. O'Leary K, Burke R, Kirsa S: SHPA standards of practice for the nect] distribution of medicines in Australian hospitals. Journal of 54. Lourens H: Impact of a medication card on compliance. Hei- Pharmacy Practice & Research 2006, 36(2):143-149. delberg West, Victoria: Pharmacy Department, Heidelberg Repatria- 32. Martin ED, Burgess NG, Doecke CJ: Evaluation of an automated tion Hospital; 1991. drug distribution system in an Australian teaching hospital. 55. Lourens H, Woodward MC: Impact of a medication card on Australian Journal of Hospital Pharmacy 2000, 30(4):141-145. compliance in older people. Aust J Ageing 1994, 13:72-76. Page 13 of 14 (page number not for citation purposes) Australia and New Zealand Health Policy 2009, 6:24 http://www.anzhealthpolicy.com/content/6/1/24 56. Collins DM, Stowasser DA: Liaison pharmacy project. the development of pharmaceutical care liaison models in the community and hospital setting. Final Report to the Pharmaceuti- cal Benefits Scheme Education Program 1998. 57. Spurling L, Gilbert A, Phillips P, Odgers C: A hospital medication discharge service: a novel approach utilising a community- based medication liaison pharmacist. (Abstract). Proc Soc Hosp Pharm Aust 2001, 2(174):. 58. Stewart S, Pearson S, Horowitz JD: Effects of a home-based inter- vention among patients with congestive heart failure dis- charged from acute hospital care. Archives of Internal Medicine 1998, 158(10):1067-1072. 59. Stewart S, Pearson S, Luke CG, Horowitz JD: Effects of home- based intervention on unplanned readmissions and out-of- hospital deaths. Journal of the American Geriatrics Society 1998, 46(2):174-180. 60. Vreede M Van De, Leong T: Using the medical emergency team to supplement an existing adverse drug event reporting framework. Journal of Pharmacy Practice & Research 2007, 37(3):197-199. 61. Kelly M, Kaye KI, David SR, Shenfield GM: Factors influencing adverse drug reaction reporting in New South Wales teach- ing hospitals. Journal of Pharmacy Practice & Research 2004, 34(1):32-35. 62. Evans SM, Berry JG, Smith BJ, Esterman A, Selim P, O'Shaughnessy J, DeWit M: Attitudes and barriers to incident reporting: a col- laborative hospital study. Quality & Safety in Health Care 2006, 15(1):39-43. 63. Freestone L, Bolsin SN, Colson M, Patrick A, Creati B: Voluntary incident reporting by anaesthetic trainees in an Australian hospital. Int J Qual Health Care 2006, 18(6):452-457. 64. Safety and Quality Council: Medication Safety Breakthrough Collaborative. Project Chronicle. Canberra: Commonwealth of Australia; 2005. 65. Safety and Quality Council: National Medication Safety Break- through Collaborative. Improvement Toolkit. Wave 1. Can- berra: Commonwealth of Australia; 2005. 66. Safety and Quality Council: National Medication Safety Break- through Collaborative. Improvement Toolkit. Wave 2. Can- berra: Commonwealth of Australia; 2005. 67. Medication Safety Self Assessment for Australian Hospitals [http://mssa.cec.health.nsw.gov.au/MSSA_introduction.html] 68. Hughes CF: Medication errors in hospitals: what can be done? [comment]. Medical Journal of Australia 2008, 188(5):267-268. 69. D'Souza DC, Koller LJ, Ng K, Thornton PD: Reporting, review and application of near-miss prescribing medication incident data. Journal of Pharmacy Practice & Research 2004, 34(3):190-193. 70. Burdeu G, Crawford R, Vreede M van de, McCann J: Taking aim at infusion confusion. Journal of Nursing Care Quality 2006, 21(2):151-159. Publish with Bio Med Central and every scientist can read your work free of charge "BioMed Central will be the most significant development for disseminating the results of biomedical researc h in our lifetime." Sir Paul Nurse, Cancer Research UK Your research papers will be: available free of charge to the entire biomedical community peer reviewed and published immediately upon acceptance cited in PubMed and archived on PubMed Central yours — you keep the copyright BioMedcentral Submit your manuscript here: http://www.biomedcentral.com/info/publishing_adv.asp Page 14 of 14 (page number not for citation purposes)

Journal

Australia and New Zealand Health PolicySpringer Journals

Published: Sep 22, 2009

References