BACKGROUND: Audit has been a major part of attempts to improve patient care in Britain, with substantial resources devoted to it since the 1990 National Health Service reforms. Systematic reviews have considered audit to be of variable, but often moderate, effectiveness. However, these have included few studies from British primary care, and as quality improvement activities may be context specific, it is hard to judge how effective audit has been here.
RESULTS: A search for audits published in peer-reviewed journals revealed 48 two-stage projects carried out in British general practice, of which 27 principally concerned chronic disease management and nine prescribing. Most audits showed some improvements in performance, and those using controls showed 27/56 (48%) parameters had changed significantly (P < 0.05).
CONCLUSIONS: This review adds further evidence that audit can often be moderately effective. However, it is frequently used as one of a complex set of interventions making precise evaluation difficult. Those responsible for clinical governance will need to choose carefully the subjects they audit in order to use their limited resources to maximum effect. These projects are illustrative examples but once again do not identify any 'magic bullets' that would be highly likely to improve professional performance.
It is crucial that research findings are implemented in general practice if high-quality care is to be achieved. Multifaceted interventions are usually assumed to be more effective than single interventions, but this hypothesis has yet to be tested for general practice care. This review evaluates the effectiveness of interventions in influencing the implementation of guidelines and adoption of innovations in general practice. A systematic literature study was carried out using MEDLINE searches for the period from January 1980 until June 1994, and 21 medical journals were searched manually. Randomized controlled trials and controlled before and after studies (with pre- and post-intervention measurements in all groups) were selected for the analysis. Clinical area, interventions used, methodological characteristics and effects on clinical behaviour were noted independently by two researchers using a standardized scoring form. Of 143 studies found, 61 were selected for the analysis, covering 86 intervention groups that could be compared with a control group without the intervention. Information transfer alone was effective in two out of 18 groups, whereas combinations of information transfer and learning through social influence or management support were effective in four out of eight and three out of seven groups respectively. Information linked to performance was effective in 10 out of 15 groups, but the combination of information transfer and information linked to performance was effective in only three out of 20 groups. Some, but not all, multifaceted interventions are effective in inducing change in general practice. Social influence and management support can improve the effectiveness of information transfer, but information linked to performance does not necessarily do so. The variation in the effectiveness of interventions needs further analysis.
Doubts have been raised about the effectiveness of traditional types of continuing medical education. Different strategies or combinations of strategies could prove to be more effective for improving the care provided by the general practitioner. For this reason a systematic literature analysis was carried out involving 75 studies of different strategies applied in primary health care. The strategies most often studied were feedback, reminders and group education. Educational material or group education combined with feedback was the combination most frequently studied. One third of all studies could be characterized as randomized controlled trials. Individual instruction, feedback and reminders seem to be the most effective single strategies. The most effective combined strategies appeared to be all combinations with individual instruction and the combination of peer review and feedback.
OBJECTIVE: To establish what is known about the role of feedback of statistical information in changing clinical practice. DESIGN: Review of 36 studies of interventions entailing the use of statistical information for audit or practice review, which used a formal research design. SUBJECTS: Papers identified from computer searches of medical and health service management publications, of which 36 describing studies of interventions designed to influence clinical care and including information feedback from clinical or administrative data systems were reviewed. MAIN OUTCOME MEASURES: Evidence for effect of information feedback on change in clinical practice. RESULTS: Information feedback was most likely to influence clinical practice if it was part of strategy to target decision makers who had already agreed to review their practice. A more direct effect was discernable if the information was presented close to the time of decision making. The questions of the optimum layout and quantity of information were not addressed; the 36 papers were insufficient for defining good formats for information to be used for audit or quality assurance. CONCLUSIONS: Given the cost of information processing and the current emphasis on closing the audit loop in the health services, it is important that the use of information in the audit process should be critically evaluated.
Audit has been a major part of attempts to improve patient care in Britain, with substantial resources devoted to it since the 1990 National Health Service reforms. Systematic reviews have considered audit to be of variable, but often moderate, effectiveness. However, these have included few studies from British primary care, and as quality improvement activities may be context specific, it is hard to judge how effective audit has been here.
RESULTS:
A search for audits published in peer-reviewed journals revealed 48 two-stage projects carried out in British general practice, of which 27 principally concerned chronic disease management and nine prescribing. Most audits showed some improvements in performance, and those using controls showed 27/56 (48%) parameters had changed significantly (P < 0.05).
CONCLUSIONS:
This review adds further evidence that audit can often be moderately effective. However, it is frequently used as one of a complex set of interventions making precise evaluation difficult. Those responsible for clinical governance will need to choose carefully the subjects they audit in order to use their limited resources to maximum effect. These projects are illustrative examples but once again do not identify any 'magic bullets' that would be highly likely to improve professional performance.
