Automatic Feedback Indicator to Enhance the Hospital Discharge Communication Between Acute Care and Primary Care.

Detailed Description Effective communication at hospital discharge is vital for continuity of care and patient safety. Discharge letters summarize the hospital stay, outlining diagnoses, treatments, and follow-up care. Despite national guidelines mandating that discharge letters be validated and provided to patients on the day of discharge, compliance remains suboptimal in France, with average performance scores well below targets.

This study seeks to address this gap through an automated feedback mechanism. Using the hospital's electronic health record (EHR) system, the study will generate monthly dashboards for each participating service in the intervention group. These dashboards will provide a real-time view of performance metrics, including the proportion of discharge letters validated on the day of discharge and the completeness of required content fields.

The trial employs a cluster-randomized controlled design with 40 hospital services as the unit of randomization. Services are stratified by activity type (medicine, surgery/obstetrics) and baseline performance. The study is divided into two phases:

Pre-implementation Phase (January 2024 - January 2025): A 12-month observational period to collect baseline data and stratify services for randomization.

Implementation Phase (February 2025 - February 2026): Intervention services receive monthly performance feedback, while control services continue with standard care practices.

The primary endpoint is the proportion of hospital stays where discharge letters are validated on the day of discharge. Secondary outcomes include:

Patient satisfaction, measured through the national "e-Satis" survey. Rates of unplanned readmissions within 30 days of discharge. Completeness of discharge letters, evaluated across mandated content fields (e.g., patient identification, discharge summary, follow-up plan).

This study will enroll all eligible patient stays within the 40 participating services, excluding stays of less than 24 hours or cases where the patient died during hospitalization. The anticipated sample size is 132,000 stays.

Data collection will rely on routine administrative data from the EHR system, anonymized at the patient level. Statistical analyses will adopt a "difference-in-differences" approach, comparing changes in outcomes between the intervention and control groups over time. A mixed-effects logistic regression model will account for intra-cluster correlations.

The results of this study aim to demonstrate the effectiveness of automated feedback in driving quality improvements in hospital discharge processes. If successful, the approach could be scaled across other hospitals in France, contributing to better continuity of care and patient outcomes.

• Proportion of Discharge Letters Validated on the Day of Discharge [Срок оценки: Measured monthly over the study period (January 2024 to February 2026), comparing a 12-month pre-implementation period to a 12-month intervention period.]
• Median Time to Generate Discharge Letters [Срок оценки: Measured monthly over the study period (January 2024 to February 2026), comparing a 12-month pre-implementation period to a 12-month intervention period.]
• Median Time to Validate Discharge Letters [Срок оценки: Measured monthly over the study period (January 2024 to February 2026), comparing a 12-month pre-implementation period to a 12-month intervention period.]
• Time from Patient Discharge to Electronic Submission of Discharge Letter to Primary Care Physicians [Срок оценки: Measured monthly during the study period (January 2024 to February 2026), comparing the 12-month pre-implementation period to the 12-month intervention period.]
• Patient Satisfaction with Discharge Process (e-Satis Survey) [Срок оценки: Collected monthly during the 12-month intervention period (February 2025 to February 2026).]
• Rate of Unplanned 30-Day Readmissions [Срок оценки: Measured monthly over the study period (January 2024 to February 2026), comparing a 12-month pre-implementation period to a 12-month intervention period.]