Data-Quality Audit of Bacteriology Electronic Medical Records at a Nigerian Tertiary Hospital

In low- and middle-income countries (LMICs), routine microbiology data from clinical laboratories make up most of the data used for antimicrobial resistance (AMR) surveillance. Most of these data are stored in the electronic medical records (EMRs) of clinical and public health laboratories. However, these data are rarely evaluated for quality. We audited EMR data for all cultures and susceptibility tests from the Department of Medical Microbiology and Parasitology, University of Calabar Teaching Hospital (UCTH), Nigeria, from 2021 to 2025. The authors assessed the data for conformance, completeness, plausibility, uniqueness, and timeliness. They also checked antimicrobial susceptibility test results against CLSI and EUCAST intrinsic resistance and reporting rules. The change from year to year was analyzed using the Cochran–Armitage test. A total of 222,620 rows (including missing data) of information were analyzed in the bacteriology export for the five years under study. The bacteriology export included 2,250 culture-positive requests and 2,704 organism records. Of the 19,476 reported susceptibility test results for bacteria, 2,427 (12.5%) were either invalid or not reported. Of the reported susceptibility information for bacteria tested against antibiotics, 180 susceptible results were reported for intrinsically resistant bacteria. Also, 16 disk diffusion tests for vancomycin susceptibility were reported as resistant for S. aureus. Clinical diagnosis information was missing for 61% of the bacteriology requests analyzed in this study. Information on the pathologist‘s verification of identity for organisms in the bacteriology requests analyzed in this study increased from 0% to 26% for three consecutive years (2019 to 2023) and increased to 99% for the two years (2024 to 2025) analyzed in this study. The inbuilt demographic and date fields of the UCTH EMR were found to be captured appropriately; however, the manner in which susceptibility test results were stored in the system for various organisms allowed for the entry of inappropriate values. Organism-specific antibiotic panels, rule-based entry validation, and mandatory coded clinical fields are required to enable the use of the data for AMR surveillance.

Keywords: data quality; electronic medical records; antimicrobial resistance; laboratory surveillance