MRC MPH students

Quality of cause of natural death certification at Groote Schuur Hospital in Cape Town

Cause of death information (mortality statistics) is reported on death certificates and is amongst the mostly widely used source of epidemiological and clinical and health investigations and for mortality statistics at national and local levels. Its quality is greatly dependent on the physician and his or her understanding of the guidelines for reporting. Data input personnel are limited by the certificate data and the accuracy of the software product. The quality of mortality data in our nation reflects the importance we attach to the structure and process of the data gathering. In the public health field, mortality statistics are an important comparison measure dependent on data quality. When state or national mortality statistics are compared, inconsistent data could lead to false conclusions.

Numerous studies have shown that coding accuracy affects the quality of data. Yet for many developing countries data on cause of death are often missing, and/or inaccurate and/or incomplete.  Complete and accurate reporting of cause of death is thus extremely important for establishing meaningful statistics for public health policies and is also required for the patient’s families for emotional and legal purposes. Poor quality death certificate data have been noted in South Africa, but accuracy and completion of cause of death certification has not been assessed countrywide. In addition there is indirect evidence of misclassification or underreporting of HIV and AIDS deaths. Given the growing numbers of HIV related deaths and the concerns of stigma, the extent of incomplete information on the HIV status on the death certificate might be large.  Therefore assessment of the extent of cause of death errors and the extent of misclassification of AIDS mortality would be useful.

This will be done at a local teaching hospital. All deaths certificates in a particular year will be reviewed to examine the quality of information provided about the cause of death, the completeness and whether they comply with the International Statistical Classification of Diseases and related Health problems - Tenth Revision (ICD-10) guidelines. 

The specific objectives of the study are:

1. To determine what proportion of causes of natural death listed on death certificates were adequate for coding to ICD-10
2. To determine what proportion of death certificates had sufficiently logically consistent information to identify an underlying cause
3. To determine the extent of major and minor errors in the cause of death certification
4. To assess whether factors such as age of deceased, gender of deceased, category of health service and category of cause of death are associated with the adequacy of cause of death information for coding to ICD 10.
5. To assess whether factors such as age of deceased, gender of deceased, category of health service and category of cause of death are associated with the adequacy of cause of death information for identification of an underlying cause of death.
6. To assess the completeness of reporting of HIV/AIDS data on the death certificates

Supervisor: Dr Pam Groenewald, MRC Burden of Disease Research Unit