Algorithm-generated educational alerts improved cardiology guideline adherence from 45.8% to 48.7% (RR 1.06, p=0.0002) in hospitalized patients.
Do algorithm-generated educational alerts improve adherence to clinical practice guidelines in cardiology in-patients?
Automated patient-specific educational alerts integrated into the hospital information system significantly improve physician adherence to clinical practice guidelines for cardiology inpatients.
Absolute Event Rate: 0% vs 0%
Abstract Background Compliance with clinical practice guideline recommendations is suboptimal in real-world clinical settings. Purpose The hypothesis of this trial was that patient-specific educational alerts on guideline recommendations for cardiology in-patients would improve adherence to guidelines compared to standard of care. Methods This dual center, superiority, blinded, randomized trial randomly assigned patients admitted as regular cardiology in-patients to an intervention group or standard of care. In the intervention group, 143 algorithm-generated educational alerts were automatically displayed in the hospital information system to the care team during the entire hospitalization, in case a deviation from clinical practice guideline recommendations was detected. The primary endpoint was guideline adherence, defined as the ratio of resolved to unresolved guideline recommendations at discharge. By contrast, the safety endpoint analyzed the number of educational alerts interpreted as incorrect by the treating physician. All patients were included in the final analysis as intention to treat. Results Between Jan 10, 2023 and Jul 9, 2023, 1467 patients were randomized to the intervention and 1471 patients to the control group. The mean age was 69 years and 64% of patients were male. The primary outcome indicating adherence to guidelines was higher in the intervention group with 4100 resolved versus 4325 non-resolved recommendations (48.7%) compared to the control group with 3883 resolved versus 4596 (45.8%) non-resolved recommendations (relative risk: 1.06, 95% CI: 1.03-1.10, p=0.0002). This corresponds to a "number needed to show" of 34.5 guideline recommendations and a "number needed to treat" of 5.9 patients. Conclusion The trial results suggest that algorithm-generated, automatically displayed educational alerts based on electronic health data can increase compliance with clinical practice guideline recommendations and should therefore be implemented in hospital information systems by default.Consort diagram of patient flow Primary endpoint
Scherer et al. (Sat,) reported a other. Algorithm-generated educational alerts improved cardiology guideline adherence from 45.8% to 48.7% (RR 1.06, p=0.0002) in hospitalized patients.