Accuracy of Cause-of-Death Classification in Ghana: Evidence from the 2023 District Health Information Management System (DHIMS) II

Résumé

Abstract

Background and Purpose:

Accurate and reliable cause-of-death (COD) data are crucial for informing public health policy, tracking epidemiological trends, and allocating health resources effectively. In Ghana, institutional mortality data captured through the District Health Information Management System (DHIMS) II provide critical insights into disease burden and health system performance. However, the use of these data is limited by inconsistencies in medical certification, coding practices, and data completeness. This study aims to identify the leading causes of institutional deaths in Ghana in 2023 and assess the quality of ICD-11 coding. The findings are expected to help strengthen cause-of-death reporting systems and improve the accuracy of health data in Ghana.

Methods:
The study analysed the 2023 mortality data recorded in the DHIMS II, covering 30,397 institutional deaths after excluding records with missing age data. The dataset was cleaned, retaining cases with estimated ages to preserve demographic patterns. The ANACOD3 tool was used to assess the completeness, specificity, and quality of causes of the data, with a focus on identifying garbage codes and misclassified entries. Ethical approval was obtained from the Ghana Health Service.

Results:
Non-communicable diseases accounted for 57.4% of institutional deaths, followed by communicable conditions (36.6%) and injuries (3.8%). Gender-specific patterns revealed differences in the leading causes of death. The quality assessment showed a high frequency of ill-defined underlying causes and invalid ICD-11 codes, with over 30% of the records classified as garbage codes, and incomplete records were also prevalent.

Conclusion:
Significant gaps in COD data quality compromise its utility for health planning in Ghana. Addressing deficiencies in certification, coder training, and diagnostic infrastructure is crucial for improving mortality data accuracy and supporting evidence-based health interventions.

Keywords: Cause-of-death data; Institutional mortality; DHIMS II; ICD-11 coding; ANACOD3 tool; Garbage codes.