Validation of the symptom pattern method for analyzing verbal autopsy data
2007

New Method for Analyzing Cause of Death Data

Sample size: 10000 publication Evidence: high

Author Information

Author(s): Chris Murray, Alan D. Lopez, Dennis M. Feehan, Shanon T. Peter, Gonghuan Yang

Primary Institution: Institute for Health Metrics and Evaluation, University of Washington

Hypothesis

Can the symptom pattern method provide more accurate cause-of-death estimates from verbal autopsy data compared to physician-coded verbal autopsy?

Conclusion

The symptom pattern method outperforms physician-coded verbal autopsy in estimating causes of death both at the population and individual levels.

Supporting Evidence

  • The SP method estimates cause-specific mortality fractions with an average relative error of 16%.
  • At the individual level, the SP method assigns the correct cause of death in 83% of cases.
  • PCVA assigns the correct cause of death in only 69% of cases.
  • Without medical record recall, the SP method still performs well with a 14% average relative error.

Takeaway

Researchers created a new way to figure out why people died by asking families about symptoms, and it works better than the old method that needed doctors.

Methodology

The study developed and validated a statistical method using verbal autopsy data to estimate cause-specific mortality fractions and assign causes of death to individuals.

Potential Biases

The reliance on household recall may introduce bias in cause-of-death assignment.

Limitations

The method needs further validation in different cultural contexts and may not work as well for deaths without prior health service contact.

Participant Demographics

Data collected from a community sample in China.

Digital Object Identifier (DOI)

10.1371/journal.pmed.0040327

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication