EVALUATION OF A POPULATION BASED INTERVENTION TO IMPROVE THE VIETNAMESE MORTALITY REPORTING SYSTEM

Abstract

Background Objectives were to examine the benefits of a populationbased intervention to improve mortality reporting system for NCDs in Central Viet Nam.
Methods An intervention was performed at Dien Chau district Health Center for 85 health officers in applying Verbal Autopsy (VA) to determine the causes of death and ICD-10 coding. Data were repeatedly obtained by three times (before, after the intervention, and final VA) for all deaths occurred in 2014. Verbal autopsy was the reference group to estimate diagnostic tests for all NCDs and its sub-groups. We compared these indicators between the second and first obtained data to see the improvements. Results Among 1,581 deaths, 1,572 cases were determined by VAs. Sensitivity, specificity, positive predictive value, and negative predictive value was 0.95, 0.54, 0.83, and 0.80, respectively for before and 0.93, 0.76, 0.90, and 0.81, respectively for after the intervention. The significant changes for all NCDs were not seen for sensitivity and negative predictive value, P > 0.05, but that were observed improving for specificity and positive predictive value, P < 0.05. After the intervention, sensitivity was increased from 0.28 to 0.76 for chronic respiratory disease, from 0.23 to 0.50 for diabetes, from 0.79 to 0.87 for cancer; and 0.87 (not changed) for cardiovascular disease. Conclusions Data quality for NCDs and its sub-groups was improved after the intervention, which would be suggested to perform nationwide to enhance mortality data for identifying and planning preventive strategy’s targeting the leading-causes of deaths in Viet Nam.