|
Do computerised clinical decision support systems for prescribing change practice? A systematic review of the literature (1990-2007) |
Pearson SA, Moxey A, Robertson J, Hains I, Williamson M, Reeve J, Newby D |
|
|
CRD summary This review found that computer-based clinical decision support systems appeared to be effective for supporting some areas of the prescribing process, but there was little evidence on their impact in specific clinical domains and settings. The authors’ cautious conclusions appear to be reasonable, but an unrestricted literature search might have retrieved more evidence. Authors' objectives To evaluate the impact of computer-based clinical decision support systems (CDSSs) on specific aspects of prescribing. Searching MEDLINE, PREMEDLINE, EMBASE, CINAHL, PsycINFO, Inspec, and the Cochrane Database of Systematic Reviews were searched for articles from 1990 to November 2007. Search terms were reported and the reference lists of retrieved articles and reviews were checked. The search was restricted to studies published in English. Study selection Randomised or non-randomised controlled trials and controlled or uncontrolled interrupted time series of computerised CDSSs compared with usual care or paper-based clinical decision support were eligible for inclusion. The CDSS had to supply the provider, at the time of prescribing, with patient-specific information that was generated and delivered electronically. Studies had to report the changes in prescribing that were consistent with the intention of the CDSS, such as initiating, monitoring, or stopping treatment. Studies were excluded if they targeted only medical students, pharmacists, and/or nurses, had non-clinical settings, were audits, or lacked statistical analysis.
The included studies were conducted in out-patient care (including medical centres) or in-patient care or both and mostly in North America. The clinical focus included the management of cardiovascular disease, anticoagulation, vaccination, and respiratory conditions. The systems were used by physicians alone or with other health professionals and they delivered a variety of system- or user-initiated interventions, including rule-based alerts, reminders, suggestions, and links to evidence. About half of them were delivered with other interventions (e.g. audit). Studies of treatment initiation used the CDSS either before or after drug selection. Outcomes included the rates of appropriate drug use, lab monitoring, guideline adherence, and the number of follow-up visits.
Two reviewers independently selected the studies. Assessment of study quality A published 10-point scale (Garg, et al. 2005, see Other Publications of Related Interest) was used to assess the randomisation method, unit of allocation, group equivalence, objectivity of outcome measures, and follow-up rate. It was also noted whether statistical analyses were adjusted for clustering. Two reviewers assessed study validity. Data extraction Data were extracted on whether the CDSS was favoured for at least one prescribing outcome, whether the results were statistically significant, and whether 50% or more of the outcomes significantly favoured the intervention (based on the methods of Garg, et al. 2005).
Two reviewers independently extracted the data and disagreements were resolved by consensus. Methods of synthesis Studies were combined in a narrative synthesis and the findings were grouped by prescribing domain (initiating, monitoring, or stopping therapy). The proportion of relevant studies reporting positive findings for each outcome was calculated. Studies were subgrouped to explore the impact of the initiation method (system or user), care setting (institutional or ambulatory), type of intervention (single or multifaceted), and clinical area (eight options). Results of the review Fifty-six articles were included in the review; 50 were randomised controlled trials (RCTs), five were interrupted time series, and one was a quasi-experimental study; no further details were given. Where reported, sample sizes ranged from nine to 778 providers and from 25 to over 26,000 patients. The mean quality score was 7.9 points out of 10 (range four to 10). Forty-four RCTs used cluster randomisation or statistically adjusted for clustering; 36 studies adjusted for baseline group differences (if required), 53 used objective or blinded outcome measures, and 50 followed up at least 90% of participants.
In 19 of 38 studies of treatment initiation, the CDSS was associated with statistically significant improvement for most of the prescribing outcomes. The CDSS appeared to be more effective if used after drug selection (significant in seven of 12 studies) rather than before (significant in 12 of 26 studies). In eight of 23 studies of monitoring therapy, the CDSS was associated with statistically significant improvement for most of the prescribing outcomes. No studies of stopping treatment reported effects that significantly favoured the CDSS, but the evidence was sparse (three studies).
Positive findings for CDSSs appeared to be more common in studies in institutional settings and where the CDSS was system initiated. Stand-alone CDSSs did not appear to be inferior to multifaceted interventions. There was limited evidence on the impact of CDSSs in specific clinical settings. Authors' conclusions Computerised CDSSs appeared to be effective for supporting some areas of the prescribing process, but there was little evidence for specific clinical domains and settings. CRD commentary The objectives and inclusion criteria were clear and relevant sources were searched for studies. The restrictions for publication status and language meant that some studies might have been missed. Steps were taken to minimise the risk of reviewer bias and error by having more than one reviewer independently select studies, assess validity, and extract the data. Adequate details were provided about most of the characteristics of the included studies, but the quality scores were reported only as totals. The use of a narrative synthesis to combine the studies appears to have been appropriate, given their heterogeneity. There was potential for bias in the summary measure of effect (significant improvement in most of the outcomes) for each study as this was influenced by the number and relevance of study outcomes, as the authors pointed out. The vote-counting method of combining the studies might not have been reliable due to the differences between the studies, but it was a reasonable way of summarising the evidence available, given its range and variety.
The authors’ cautious conclusions appear to be reasonable, but an unrestricted literature search might have retrieved more evidence. Implications of the review for practice and research Practice: The authors stated that further implementation of CDSSs, by policy makers and system developers, should include safety alerts, reminders for patients on long-term therapy, and suggestions for high-risk patients.
Research: The authors stated that more research on the short-, medium- and long-term economic and clinical impacts of CDSSs was required, especially for stopping therapy. Different types of system and different implementation strategies should be tested and the outcome measures should be clinical. Future studies should provide details of the individual features of the systems, their development methods, and the settings in which they were used. Funding National Prescribing Service (NPS) Limited. Bibliographic details Pearson SA, Moxey A, Robertson J, Hains I, Williamson M, Reeve J, Newby D. Do computerised clinical decision support systems for prescribing change practice? A systematic review of the literature (1990-2007) BMC Health Services Research 2009; 9:154 Other publications of related interest Garg AX, Adhikari NK, McDonald H, Rosas-Arellano MP, Devereaux PJ, Beyene J, Sam J, Haynes RB. Effects of computerized clinical decision support systems on practitioner performance and patient outcomes: a systematic review. JAMA 2005; 293: 1223-1238. Indexing Status Subject indexing assigned by NLM MeSH Decision Support Systems, Clinical; Drug Prescriptions; Humans; Practice Patterns, Physicians'; User-Computer Interface AccessionNumber 12009110260 Date bibliographic record published 17/03/2010 Date abstract record published 16/06/2010 Record Status This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn. |
|
|
|