Integrated multisource estimates of mortality for Thailand in 2005

Peter Byass

doi:10.1186/1478-7954-8-10

Abstract

Estimates of mortality in Thailand during 2005 have been published, integrating multiple data sources including national vital registration and a national follow-up cluster sample, covering both deaths in health facilities (approximately one-third) and elsewhere. The methodological challenge is to make the best use of the existing data, supplemented by additional data that are feasible to obtain, in order to arrive at the best possible overall estimates of mortality. In this case, information from the national vital registration database was supplemented by a verbal autopsy survey of approximately 2.5% of deaths, the latter being used to validate routine cause-of-death data and information from medical records. This led to a revised national cause-specific mortality envelope for Thailand in 2005, amounting to 447,104 deaths. However, difficulties over standardizing verbal autopsy interpretation may mean that there are still some uncertainties in these revised estimates.

Highlights

This commentary relates to a set of four papers by Rao and colleagues that relate to a detailed investigation of cause-specific mortality in Thailand during 2005, integrating a number of different data sources
About one-third of registered deaths in Thailand occur in hospitals, and paper II looks into the validity of the registered causes of these deaths, using additional information from medical records and verbal autopsies, in order to attribute misclassification errors [2]
For the majority of deaths, which occur outside hospitals, registered causes can only be validated by carrying out verbal autopsies, and this process is covered in paper III [3], which provides part of the validation comparisons for the hospital deaths in paper II

Summary

Introduction

This commentary relates to a set of four papers by Rao and colleagues that relate to a detailed investigation of cause-specific mortality in Thailand during 2005, integrating a number of different data sources. About one-third of registered deaths in Thailand occur in hospitals, and paper II looks into the validity of the registered causes of these deaths, using additional information from medical records and verbal autopsies, in order to attribute misclassification errors [2]. For the majority of deaths, which occur outside hospitals, registered causes can only be validated by carrying out verbal autopsies, and this process is covered in paper III [3], which provides part of the validation comparisons for the hospital deaths in paper II. Paper IV integrates the findings from papers II and III into a complete overview and estimate of mortality for Thailand in 2005, with a discussion of implications for practice and policy [4]. The somewhat complex inter-relationships between these four papers are illustrated conceptually in Figure 1, with paper I being an overview of the whole

Methods

Findings

Discussion

Conclusion