Are all tuberculosis deaths correctly attributed due to tuberculosis? Analysis of causes of death (COD) using verbal autopsy: A cross-sectional study in Puducherry.

Abstract

Background: State Task Force National Tuberculosis Elimination Programme (STF-NTEP) wanted to ascertain the cause of death of tuberculosis (TB) patients was due to TB. Hence, this study was done to determine the cause of death among TB patients using verbal autopsy.

Material and method: The community-based cross-sectional study was conducted among the closest caregivers of the reported TB death cases using a verbal autopsy questionnaire. A list of TB deaths was extracted from Puducherry's State Tuberculosis Office (STO). The VA questionnaires were reviewed by two trained faculties and experts independently who were also asked to code the cause of death according to the International Classification of Diseases - 10^th Revision. Data were analysed using the SPSS software receiver operator curve (ROC) curve and Cohen's kappa statistics with 95% confidence interval (CI) were applied to compare agreement between the raters.

Results: Out of 307 TB deaths, 143 were contacted and the remaining were migrated, missing or not available. Among them 34.9% of the declared TB deaths are not due to TB. Cohen's kappa = 0.4; 95% CI: (0.858-0.985) showed moderate agreement between the raters. Experts' review was used as a reference standard to compare rater agreements. The Predictive accuracy of Rater 2 (area under the curve [AUC] -0.953; P value <0.001) was slightly higher than Rater 1 (AUC -0.942; P value <0.001).

Conclusion: The findings of the study have clearly indicated that 34.9% of the declared TB deaths are not due to TB.