Predicting death from coronary heart disease using a questionnaire

Area of Science:

• Cardiovascular epidemiology
• Public health screening

Background:

• Coronary heart disease (CHD) remains a leading cause of mortality.
• Effective screening tools are crucial for population health management.
• The Whitehall study provides a valuable dataset for epidemiological research.

Purpose of the Study:

• To evaluate the predictive power of a questionnaire for ten-year CHD mortality.
• To assess the utility of self-reported symptoms as a screening tool for CHD risk.
• To identify optimal combinations of questions for population-based screening.

Main Methods:

• Analysis of questionnaire responses from 18,322 male civil servants (aged 40-64).
• Examination of ten-year CHD mortality data (1714 deaths, 723 from CHD).
• Evaluation of different algorithms based on reported symptoms (angina, chest pain, dyspnea, diabetes, physician visits).

Main Results:

• Angina plus possible myocardial infarction (PMI) yielded 99% specificity but only 7% sensitivity.
• Angina and/or PMI improved sensitivity to 29% with 90% specificity.
• An 11-question algorithm including dyspnea, diabetes, and physician visits achieved 85% specificity and 44% sensitivity.
• Excluding 'yes-set' responders slightly increased specificity but reduced sensitivity.

Conclusions:

• A multi-symptom questionnaire, particularly an 11-question combination, is recommended for CHD screening.
• This screening approach balances high specificity with improved sensitivity for identifying at-risk individuals.
• Further refinement of screening questions may enhance predictive accuracy for coronary heart disease mortality.