K Furu1, D S Thelle. 1. Institute of General Practice and Community Medicine, University of Oslo, P.O. Box 1130 Blindem, N-0318 OSLO, Norway. kari.furu@samfunnsmed.uio.no
Abstract
OBJECTIVE: The aim of this study was to investigate whether morbidity in the general population could be assessed by questions on drug use in the Norwegian Health Survey 1995. MATERIAL AND METHOD: A sample of 6,702 persons, aged 20-79 years was interviewed in their homes using computer-assisted personal interviewing (CAPI). MEAN OUTCOME MEASURE: The validity of questions on use of analgesics and drugs against dyspepsia/peptic ulcer has been assessed according to categories of self-evaluated health. RESULTS: There was a difference between sporadic and daily users of the drugs to what extent they rated their health as poor. The validity of the drug questions assessed by sensitivity and specificity, showed that only using a dichotomous outcome variable, is too low to give a sufficiently valid measure of the morbidity in the population. CONCLUSIONS: Using "yes" or "no" as the only outcome of drug questions has the unfortunate effect of putting together chronic users of drugs with infrequent users for all of the subsequent analyses, which results in a considerable measurement error. This implies a need for improved methods to determine the optimal recall period for different drugs and it is crucial to include more details in questions on drug use to increase the validity of this information.
OBJECTIVE: The aim of this study was to investigate whether morbidity in the general population could be assessed by questions on drug use in the Norwegian Health Survey 1995. MATERIAL AND METHOD: A sample of 6,702 persons, aged 20-79 years was interviewed in their homes using computer-assisted personal interviewing (CAPI). MEAN OUTCOME MEASURE: The validity of questions on use of analgesics and drugs against dyspepsia/peptic ulcer has been assessed according to categories of self-evaluated health. RESULTS: There was a difference between sporadic and daily users of the drugs to what extent they rated their health as poor. The validity of the drug questions assessed by sensitivity and specificity, showed that only using a dichotomous outcome variable, is too low to give a sufficiently valid measure of the morbidity in the population. CONCLUSIONS: Using "yes" or "no" as the only outcome of drug questions has the unfortunate effect of putting together chronic users of drugs with infrequent users for all of the subsequent analyses, which results in a considerable measurement error. This implies a need for improved methods to determine the optimal recall period for different drugs and it is crucial to include more details in questions on drug use to increase the validity of this information.