The reliability of medication and death recording in electronic health records among people with schizophrenia in Hong Kong

Objective: This study aimed to examine the reliability of death and medication records among individuals diagnosed with schizophrenia between two widely-used electronic health record (EHR) databases in Hong Kong: the Clinical Management System (CMS) and the Clinical Data Analysis and Reporting System (CDARS). Methods: A cohort of patients with schizophrenia-spectrum disorders enrolled in public psychiatric services in Hong Kong between 1998 and 2003 was identified from the CMS. The unique IDs, vital status, and clozapine prescription information of these patients were extracted from both the CMS and CDARS. The positive predictive value (PPV) was computed to evaluate the concordance of records between the two databases. Results: Of the 1234 patients identified in the CMS, 1119 (90.7 %) had a record of schizophrenia in the CDARS. All deaths recorded in the CDARS matched those in the CMS, but the PPV for the exact date of death was 66.4 %, improving to 86.0 % when allowing a 30-day margin of error. The PPV for clozapine prescribing was 96.1 %, and 91.8 % for the initiation date when allowing a 30-day margin of error. Conclusion: This study provides evidence on the reliability of mortality and medication data for patients with schizophrenia in two widely-used EHR databases in Hong Kong. While the record on whether an event has occurred had excellent concordance, the exact timing of events showed lower reliability, highlighting the need to consider data resolution when utilizing these databases for research.