| The potential for using the case-control design with only hospital separation records was explored in a study of ectopic pregnancy.; The Saskatchewan Hospital Services Commission hospital morbidity database was used. Cases were women hospitalized for ectopic pregnancy in the years 1984 through 1986. Controls were women hospitalized for any other pregnancy condition.; In a case-control study using pregnant controls, the odds ratio is an appropriate estimator of relative risk, conditional on conception. Among the potential risk factors (exposures) represented in the data were tubal surgery, prior ectopic pregnancy, pelvic inflammatory disease, and infertility. Of these, the first two are completely ascertained during the time a woman was resident in Saskatchewan. Under representation of exposures should be non-differential, leading to odds ratios which are biased towards the null. Multiple logistic regression analysis yielded estimates of the odds ratios for acute PID, chronic PID or pelvic adhesions, prior ectopic pregnancy, tubal sterilization, and tubal repair. Of particular interest was the association between ectopic pregnancy and tubal surgery performed in the hope of maintaining or restoring normal tubal function. The attributable risk proportion of ectopic pregnancy due to such surgery was 8.1%. This study design can be used to follow trends in surgical causes of ectopic pregnancy in Canada. (Abstract shortened by UMI.)... |
