On the estimation of population cause-specific mortality fractions from in-hospital deaths

Improving methods to measure comparable mortality cause (IMMCMC) gold standard verbal autopsy dataset

Objectives

Gold standard cause of death data is critically important to improve verbal autopsy (VA) methods in diagnosing cause of death where civil and vital registration systems are inadequate or poor. As part of a three-country research study—Improving Methods to Measure Comparable Mortality by Cause (IMMCMC) study—data were collected on clinicopathological criteria-based gold standard cause of death from hospital record reviews with matched VAs. The purpose of this data note is to make accessible a de-identified format of these gold standard VAs for interested researchers to improve the diagnostic accuracy of VA methods.

Data description

The study was conducted between 2011 and 2014 in the Philippines, Bangladesh, and Papua New Guinea. Gold standard diagnoses of underlying causes of death for deaths occurring in hospital were matched to VAs conducted using a standardized VA questionnaire developed by the Population Health Metrics Consortium. 3512 deaths were collected in total, comprised of 2491 adults (12 years and older), 320 children (28 days to 12 years), and 702 neonates (0–27 days).

Who dies where? Estimating the percentage of deaths that occur at home

Introduction

The majority of low-income and middle-income countries (LMICs) have incomplete death registration systems and so the proportion of deaths that occur at home (ie, home death percentage) is generally unknown. However, home death percentage is important to estimate population-level causes of death from integration of data of deaths at home (verbal autopsies) and in hospitals (medical certification), and to monitor completeness of death notification and verbal autopsy data collection systems. This study proposes a method to estimate home death percentage using data readily available at the national and subnational level.

Methods

Data on place of death from 152 country-years in 49 countries from 2005 to 2019, predominantly from vital registration systems, were used to model home death percentage standardised for population age and cause distribution. A national-level model was developed using Bayesian model averaging to estimate national, regional and global home death percentage. A subnational-level model was also developed and assessed in populations where alternative data on home death percentage were available.

Results

Globally, it is estimated that 53.4% (95% uncertainty interval (UI) 50.8%–55.9%) of deaths occur at home, slightly higher (59.7%, 95% UI 56.5%–62.7%) in LMICs, substantially higher in low-income countries (79.5%, 95% UI 77.3%–81.5%) and much lower (27.3%, 95% UI 25.2%–29.6%) in high-income countries. Countries with the highest home death percentage are mostly found in South, East and Southeast Asia and sub-Saharan Africa (above 90% in Ethiopia, Chad and South Sudan). As expected, the national model has smaller error than the subnational model.

Conclusion

The study demonstrates substantial diversity in the location of deaths in LMICs and fills a significant gap in knowledge about where people die, given its importance for health systems and policies. The high proportion of deaths in LMICs that occur at home reinforces the need for routine verbal autopsy to determine the causes of death.

Monitoring progress with national and subnational health goals by integrating verbal autopsy and medically certified cause of death data

Introduction

The measurement of progress towards many Sustainable Development Goals (SDG) and other health goals requires accurate and timely all-cause and cause of death (COD) data. However, existing guidance to countries to calculate these indicators is inadequate for populations with incomplete death registration and poor-quality COD data. We introduce a replicable method to estimate national and subnational cause-specific mortality rates (and hence many such indicators) where death registration is incomplete by integrating data from Medical Certificates of Cause of Death (MCCOD) for hospital deaths with routine verbal autopsy (VA) for community deaths.

Methods

The integration method calculates population-level cause-specific mortality fractions (CSMFs) from the CSMFs of MCCODs and VAs weighted by estimated deaths in hospitals and the community. Estimated deaths are calculated by applying the empirical completeness method to incomplete death registration/reporting. The resultant cause-specific mortality rates are used to estimate SDG Indicator 23: mortality between ages 30 and 70 years from cardiovascular diseases, cancers, chronic respiratory diseases and diabetes. We demonstrate the method using nationally representative data in Myanmar, comprising over 42 000 VAs and 7600 MCCODs.

Results

In Myanmar in 2019, 89% of deaths were estimated to occur in the community. VAs comprised an estimated 70% of community deaths. Both the proportion of deaths in the community and CSMFs for the four causes increased with older age. We estimated that the probability of dying from any of the four causes between 30 and 70 years was 0.265 for men and 0.216 for women. This indicator is 50% higher if based on CSMFs from the integration of data sources than on MCCOD data from hospitals.

Conclusion

This integration method facilitates country authorities to use their data to monitor progress with national and subnational health goals, rather than rely on estimates made by external organisations. The method is particularly relevant given the increasing application of routine VA in country Civil Registration and Vital Statistics systems.

Using health facility deaths to estimate population causes of neonatal and child mortality in four African countries

BACKGROUND

Verbal autopsy is the main method used in countries with weak civil registration systems for estimating community causes of neonatal and 1-59-month-old deaths. However, validation studies of verbal autopsy methods are limited and assessment has been dependent on hospital-based studies, with uncertain implications for its validity in community settings. If the distribution of community deaths by cause was similar to that of facility deaths, or could be adjusted according to related demographic factors, then the causes of facility deaths could be used to estimate population causes.

METHODS

Causes of neonatal and 1-59-month-old deaths from verbal/social autopsy (VASA) surveys in four African countries were estimated using expert algorithms (EAVA) and physician coding (PCVA). Differences between facility and community deaths in individual causes and cause distributions were examined using chi-square and cause-specific mortality fractions (CSMF) accuracy, respectively. Multinomial logistic regression and random forest models including factors from the VASA studies that are commonly available in Demographic and Health Surveys were built to predict population causes from facility deaths.

RESULTS

Levels of facility and community deaths in the four countries differed for one to four of 10 EAVA or PCVA neonatal causes and zero to three of 12 child causes. CSMF accuracy for facility compared to community deaths in the four countries ranged from 0.74 to 0.87 for neonates and 0.85 to 0.95 for 1-59-month-olds. Crude CSMF accuracy in the prediction models averaged 0.86 to 0.88 for neonates and 0.93 for 1-59-month-olds. Adjusted random forest prediction models increased average CSMF accuracy for neonates to, at most, 0.90, based on small increases in all countries.

CONCLUSIONS

There were few differences in facility and community causes of neonatal and 1-59-month-old deaths in the four countries, and it was possible to project the population CSMF from facility deaths with accuracy greater than the validity of verbal autopsy diagnoses. Confirmation of these findings in additional settings would warrant research into how medical causes of deaths in a representative sample of health facilities can be utilized to estimate the population causes of child death.