Digital Primary Care Tools Improve Patient Safety Measures
Key Takeaways
- Across randomized trials of digital patient safety interventions in primary care, medication-related and non-medication safety process measures improved significantly.
- Adverse events were also reduced, but the signal was more modest than the gains seen in safety process measures.
- Multicomponent digital interventions showed larger pooled effects than single-component approaches across medication safety, non-medication process measures, and adverse events.
- In the narrative synthesis, better-performing interventions often featured workflow integration, relational support, simplicity, and sustained engagement, while poor workflow fit and low use were common barriers.
In a systematic review and meta-analysis, investigators evaluated English-language randomized controlled trials (RCTs) in primary care, reported in line with Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) and registered in PROSPERO as CRD42024620160. Searches of MEDLINE, EMBASE, and CENTRAL covered January 2001 through June 2025. Eligible interventions had to include at least one digital element, and outcomes were grouped as medication safety process measures, non-medication process measures, and adverse events. The review included 74 RCTs from multiple countries, most using cluster designs; most interventions incorporated clinical decision support systems (CDSS), and nearly three-quarters were multicomponent.
Medication safety process measures improved across 42 randomized trials contributing 54 comparisons with OR 1.46 (95% CI 1.30–1.64; I² 98.6%), non-medication process measures improved across 30 randomized trials and 40 comparisons with OR 1.77 (95% CI 1.56–2.01; I² 96.4%), and adverse events were reduced across 12 randomized trials and 12 comparisons with OR 1.17 (95% CI 1.06–1.30; I² 82.6%). Overall, the pooled gains were larger for safety processes than for downstream harm outcomes.
Multicomponent interventions showed larger pooled effects than single-component approaches across medication safety, non-medication process measures, and adverse events. For medication safety, the largest reported subgroup effect came from interventions combining CDSS with audit and feedback (A&F), OR 1.75 (95% CI 1.35–2.25). Patient-facing components were also associated with larger effects for non-medication process measures, and sensitivity analyses restricted to low-risk-of-bias trials remained directionally consistent.
Interpretation of the pooled estimates was tempered by substantial heterogeneity across countries, intervention formats, clinical domains, and implementation settings. Egger testing also suggested small-study effects in all three outcome categories. Only 21 studies were judged low risk of bias, while most had some concerns and 10 were rated high risk. In the narrative synthesis, workflow integration, relational support, intervention simplicity, and sustained engagement tracked with better performance, whereas poor workflow fit, limited time or resources, low use, and mismatches between CDSS prompts and real-world clinical complexity repeatedly limited uptake.
The authors concluded that digital interventions in primary care were associated with moderate improvements in medication and non-medication safety process measures and with modest reductions in adverse events, with multicomponent approaches appearing most promising. They also called for more work on the sociotechnical mechanisms behind those effects and on whether benefits are sustained over longer-term follow-up.
Clinician Questions
What outcomes counted as medication safety, non-medication process measures, and adverse events in digital primary care safety research?
In this review, medication safety process measures included inappropriate prescribing, appropriate prescribing for high-risk groups, and medication errors; non-medication process measures included diagnostic errors or delays, timely investigations or laboratory monitoring, and optimized management of high-risk groups; and adverse events included unplanned and cause-specific hospitalizations, cardiovascular events, asthma exacerbations, and adverse drug events.
Which primary care digital intervention features were linked to larger effects in the pooled analyses?
Multicomponent interventions generally showed larger pooled effects than single-component interventions across medication safety, non-medication process measures, and adverse events. Among component comparisons, CDSS combined with A&F showed the largest medication-safety signal, while patient-facing components were associated with larger effects for non-medication process measures.
What implementation barriers limited digital patient safety interventions in primary care?
Common barriers included low engagement or low use, difficulty integrating tools into existing systems and workflows, limited time and resources, and mismatches between CDSS recommendations and the complexity of real-world clinical decisions. Some patient-facing interventions also struggled to obtain accurate and timely self-reported data, while relational supports such as academic detailing or pharmacist-led meetings were described as helpful for engagement and sustainability.
Recommended Reading
- For more on e-prescribing workflow integration: Navarre e-Prescribing Channel Improved Pharmacy–PCP Coordination
- For more on medication management in primary care: ADHD Trajectories in Children Linked to Social Context