**Methods:** The researchers conducted a retrospective analysis of medication-related patient safety incidents reported to the National Reporting and Learning System (NRLS) between 1 January 2015 and 31 December 2019. Incidents were extracted using a free-text search for 'discharge' (including misspellings) within the 'medication' category and 'general practice' care setting. After excluding 203 of 1,324 reports (e.g., non-discharge-related, clinic discharges, duplicates), 1,121 incidents were analyzed. Descriptive analysis examined frequency and nature of incidents. Content analysis of free-text narratives was performed using the Patient Safety Research Group (PISA) classification to identify contributory factors and outcomes. Twenty percent of data (n=237) was independently coded by two researchers for validation.
**Key Results:** Most incidents involved patients aged >65 years (55%, n=626/1,121; 70.4% of reports with age data). More than one in ten incidents (12.6%, n=142/1,121) were associated with actual patient harm: low harm (5.1%, n=58), moderate harm (6.1%, n=69), severe harm (0.7%, n=8), and death (0.6%, n=7). The most common medication process stages were prescribing (42%, n=479/1,121), administration (22.5%, n=253/1,121), and monitoring (12%, n=140/1,121). The monitoring (17%) and administration (15%) stages had higher proportions of harmful incidents than prescribing (12%). The most reported error categories were wrong/unclear dose or strength (19%, n=212/1,121), omitted medicine (13%, n=148/1,121), and wrong drug/medicine (10%, n=118/1,121). Medication omission was associated with the highest proportion of harmful incidents among frequently reported error types (19%, n=28/148). A total of 1,504 medications were involved across incidents. The most common medication classes were cardiovascular (48.8%, n=734/1,504), central nervous system (18%, n=273/1,504), and endocrine (12%, n=183/1,504). Specific high-frequency medications included antiplatelets (n=126), factor Xa inhibitors (n=124), opioids (n=79), insulin (n=76), beta-blockers (n=76), heparins (n=71), vitamin K antagonists (n=67), and diuretics (n=66). Among 408 incidents (36%) with sufficient free-text data to identify contributory factors, 467 factors were coded. Organization factors were most common (82%, n=383/467), with 98% (n=377/383) related to continuity of care—primarily between secondary and primary care (n=308) and between healthcare and pharmacy (n=35). Staff factors accounted for 16% (n=75/467), including cognitive issues (47%, n=35/75) such as mistakes and misreading, and task-related issues (44%, n=33/75) including failure to follow protocol (n=14) and wrong professional carrying out the task (n=14). Patient factors (n=5) and equipment factors (n=4) were rare. Among 1,660 reported outcomes, 34% (n=564) were organizational inconvenience (mostly phone calls/follow-up, n=412), 27% (n=455) were patient inconvenience (including missed doses, n=107; unnecessary treatment, n=90; repeated visits, n=76; hospital admission, n=55), and 13% (n=216) were patient clinical harm.
**Clinical Implications:** This study identifies several actionable targets for improving medication safety after hospital discharge. The high involvement of cardiovascular and central nervous system medications suggests these should be prioritized in prescribing safety indicators and pharmaceutical prioritization tools. The predominance of organizational contributory factors—especially failures in continuity of care and information sharing between secondary care, primary care, and community pharmacies—supports the implementation of electronic interventions like the NHS Discharge Medicine Service (DMS) to improve timely communication. Staff factors, including cognitive errors and task delegation issues (e.g., administrative staff handling prescriptions), highlight the need for appropriate skill mix, such as embedding clinical pharmacists in general practices to triage discharge letters and perform medication reconciliation. The finding that patients or relatives identified errors and prevented harm in 72 incidents supports incorporating patient and family engagement into safety strategies. The association of monitored dosage systems (MDS) with errors indicates a need for further research to improve safety for patients using compliance aids post-discharge.