The accuracy of clinical diagnosis for pelvic inflammatory disease was determined in 95 women who presented with pelvic pain to primary care physicians and then were referred to gynecologists. Laparoscopy or laparotomy with endometrial biopsy and fimbrial minibiopsy revealed that prevalence of pelvic inflammatory was 46% (44/95) and positive and negative predictive values of gynecologists were 74% (23/31) and 67% (43/64) (p = 0.0002). If histopathologic diagnosis was the standard, clinical accuracies of the gynecologists were no better than chance (p = 0.43), suggesting an expectation bias for visual diagnosis. Laparoscopy had a sensitivity of 50% (12/24) and a specificity of 80% (40/50) for salpingitis if the standard was fimbrial histopathologic diagnosis (p = 0.01). These results support the routine use of laparoscopy, supplemented when negative by endometrial and fimbrial minibiopsy, to accurately diagnose pelvic inflammatory disease.