Methodology of studies evaluating death certificate accuracy were flawed

Area of Science:

• Epidemiologic research
• Public health surveillance
• Biostatistics

Background:

• Accurate statistics on causes of death are vital for epidemiologic research.
• Existing studies evaluating death data sources yield conflicting results, questioning method comparability and validity.

Purpose of the Study:

• To examine the methods and assess the reproducibility of 44 recent studies evaluating causes of death data.
• To identify inconsistencies and areas for improvement in the methodology of death statistics evaluation.

Main Methods:

• Systematic review of 44 evaluation studies on causes of death data.
• Assessment of described methods, reproducibility, reliability tests, and handling of multiple conditions.
• Comparison of study criteria with international standards, such as World Health Organization guidelines.

Main Results:

• Thirty studies specified data reviewers; six reported reliability tests.
• Twelve studies included all causes of death, but lacked criteria for identifying the underlying cause in complex cases, rendering them not reproducible.
• Of studies focusing on specific conditions, 21 had reproducible diagnostic criteria; however, three of eight studies addressing competing causes conflicted with international standards.

Conclusions:

• Methods and criteria for evaluating causes of death are frequently underspecified.
• When described, methods sometimes deviate from international standards, impacting data comparability.
• Clearer methodological descriptions are needed to enhance study reproducibility and allow readers to assess generalizability.