Performance of the WHO Rose angina questionnaire in post-menopausal women: are all of the questions necessary?

OBJECTIVE To assess the performance of a shortened version of the Rose angina questionnaire focusing on exertional chest pain. METHODS Cross sectional analysis of 3987 women aged 60 to 79 years from 23 British towns. The performances of definite Rose angina (using data from the full Rose angina questionnaire) and exertional chest pain (using data from a subset of three questions from the Rose angina questionnaire) were assessed against a medical record of angina. RESULTS The sensitivity (the proportion with a medical record of angina who were identified as having angina by the questionnaire) was 29.9% (95% confidence intervals 25.7% to 34.4%) comparing definite Rose angina to any medical record of angina since 1978 and 50.7% (45.9% to 55.5%) comparing exertional chest pain to any medical record diagnosis of angina. The positive predictive values of both questionnaires were similar. When the two questionnaires were compared with a gold standard of a primary care consultation for angina symptoms within the past five years the sensitivity of definite Rose angina was 33.0% (26.9% to 39.6%) and that of exertional chest pain was 51.8% (45.1% to 58.5%). Although the sensitivity of both versions of the questionnaire was greater in those aged 60-69 years compared with those aged 70-79 years, it remained higher in the exertional chest pain version of the questionnaire than for definite Rose angina based on the full version of the questionnaire in both age groups. Performance of either version of the questionnaire was not affected by occupational social class. CONCLUSIONS With respect to identifying women with a medical diagnosis of angina or those presenting to primary care with anginal symptoms, these results suggest that a shortened version of the Rose angina questionnaire focusing on exertional chest pain performs better than the full version. Other studies suggest that exertional chest pain is the crucial element of the Rose angina questionnaire with respect to predicting future coronary events. It is concluded that using a shortened version of the Rose angina questionnaire is adequate in epidemiological studies.