Effectiveness of Clinical Audit and Feedback in Improving Healthcare Professionals’ Practice in Primary Healthcare Settings: A Systematic Review

Abstract

Clinical audit and feedback (A&F) is one of the most widely applied quality-improvement strategies for changing healthcare professional behaviour. It entails methodical appraisal of clinical practice versus a clear standard, benchmark or peer comparator, and then feedback that is aimed to provoke reflection, objective and corrective action. It is particularly relevant in primary healthcare where the bulk of prescribing, investigation, preventive and chronic disease management is practised and where relatively small changes in professional behaviour have major population impacts. This systematic review critically assesses the modern evidence of effectiveness of clinical audit and feedback on the practice of healthcare professionals in the primary healthcare setting with special attention to the articles published in the period between 2022 and 2025 considered in the corpus of references used in this review. The review was designed following PRISMA 2020 guidelines. Bibliographic records were initially availed (n=40); ten of them were replicated creating 30 reports. The primary narrative synthesis included sixteen key effectiveness studies and six systematic reviews or meta-analyses and eight qualitative, implementation, design or consensus studies were left as the contextual evidence. Another pooled meta-analysis was not methodologically justified because of the large differences between the results, in terms of clinical domain, unit of analysis, intensity of intervention, and statistical measurement. The data confirm that A&F can most likely lead to small-to-moderate alterations in professional practice but is not consistent or automatic. In a primary care-based trial involving 5,046 primary care physicians in Ontario, the feedback to the physicians decreased overall antibiotic prescribing by about 5 percent compared to control, unnecessary antibiotic prescribing by 11 percent, prescriptions greater than seven days by 15 percent and broad-spectrum prescribing by about 6 percent (Schwartz et al., 2024). Personalised peer-comparison feedback with therapeutic recommendations in British Columbia with appropriate first-line nitrofurantoin treatment of uncomplicated lower urinary tract infection resulted in an odds ratio of 1.28 (95% CI 1.07-1.52) in 5,073 family physicians (Carney et al., 2025). A trial with high-imaging general practitioners in Australia reported an adjusted decrease in musculoskeletal imaging requests of 2.66 requests per 1,000 visits at 12 months (O’Connor et al., 2022). Whereas in South African community primary-care centres, peer audit and feedback boosted adherence to antimicrobial prescribing quality indicators (19% to 47%), it also resulted in a 19.3% reduction in antibiotic use (de Vries et al., 2022). On the other hand, a Swiss multi-centre randomised trial of 3,426 physicians reported no essential change in total antibiotic prescribing with quarterly feedback on personalised rates and a British Columbia trial reported no significant impact of feedback on initial opioid prescribing (Aghlmandi et al., 2023; McCracken et al., 2023). Similarly, an A&F trial of patient-safety in Spanish, 59 primary healthcare centres and 1,053 professionals did not demonstrate any significant change in safety culture in which the implementation was mostly disrupted by the COVID-19 pandemic (Fiol-deRoque et al., 2025). These results demonstrate that the effectiveness is determined by the design of interventions, baseline performance, actionability, clinician engagement, implementation fidelity, peer comparison, contextual barriers and the amount of control clinicians have over the target behaviour. These findings are in line with the wider evidence. In Cochrane review by Ivers et al., the median of the desired professional practice was observed to improve by 2.7 percent and the model-based mean of the desired improvement by 6.2 percent but in the primary-care antibiotic meta-analysis by Xu et al., the percentage change in total antibiotic prescribing, unnecessary initiation, excessive duration and broad-spectrum antibiotic choice were all reduced by 11, 23, 13 and 17 percent respectively. However, the heterogeneity was high and that implies that there is no average effect that can be applied to explain performance in various settings. Clinical A&F is not in general to be regarded as a stand-alone information-delivery endeavour but as a behaviour-change system. Its most powerful impacts are those in which feedback is centred around a significant performance gap, where it contains plausible individualised data, provides an actionable comparator, fosters interpretation, allows action planning and is situated in an environment that has the capacity to make change. The evidence appears to support the idea of A&F used regularly and thoughtfully in primary care and caution against the assumption that dashboards, mailed reports or repeated measurement will only improve care.