Validity of self-reports of fractures in perimenopausal women

Abstract

The validity of self-report of fractures in postal inquiry among perimenopausal women was evaluated. Self-reports of fractures in the 1989 baseline postal inquiry data of the Kuopio Osteoporosis Risk Factor and Prevention Study (OSTPRE) were compared with information in patient records. The study population consisted of 373 women who reported fractures sustained during the last 10 years and 200 randomly selected women who did not report fractures from a population base of 2,007 women aged 47-56 years. Self-report as a screening test for fracture was evaluated in the total sample of 2,007 women by estimating the number of false negative reports in all the women who did not report a fracture with the information on these 200 women. Of the self-reports of fractures, 84% proved to be true fractures, 12% soft tissue injuries, and the rest either self-diagnoses or misnomers. Self-report of wrist fracture was more accurate (95%). The sensitivity of self-report to detect fracture was 78% for all fractures and 95% for wrist fracture, while the respective specificities were 96 and 99%. Self-report is a relatively accurate way to obtain information about past major fractures in perimenopausal women. However, it is rather insensitive in the detection of minor fractures, if the reporting period is several years.