Validity of self reported utilisation of primary health care services in an urban population in Spain

Study objective-To assess the validity and factors related with the validity of self reported numbers of visits to a primary health care centre, in comparison with the recorded number. Design-Cross sectional study. Setting-The urban area served by the Zaidin-Sur Primary Health Care Centre (Granada, Spain). Participants-Two population samples (236 high users and 420 normal users) who were seen at the centre from 1985 to 1991 were interviewed in 1993. Main results-A net tendency to overreport the actual number of visits was observed. Absolute concordance between self reported and recorded utilisation decreased as time interval lengthened, although this mainly reflected the increase in maximum variability both with time interval length and with the number of recorded visits. Corrected Spearman rho coefficients obtained between the number of self reported and recorded visits ranged from 0.602 for the two weeks before the interview to 0.678 for the year before. Regression slopes of self reported utilisation upon recorded utilisation did not change between periods. In multiple regression analyses the actual number of visits was the main factor associated with both underreporting and overreporting. Older age was also significantly associated with underreporting. Poor health status and high satisfaction with health care were significantly associated with overreporting. Conclusions-There was a substantial degree of inaccuracy in self reported utilisation, with a net tendency to overreport the number of visits, in relative terms, however, accuracy of self reports did not seem to decrease appreciably as the recall time lengthened. To compare the accuracy of different measures, it is important to take into account the maximum variability of each one. Otherwise, contradictory results may be obtained.