OBSERVER VARIABILITY IN THE DIAGNOSIS AND MANAGEMENT OF THE HYSTEROSALPINGOGRAM

Objective: To determine the reproducibility of hysterosalpingogram (HS G) interpretation and clinical management recommendations among traine d observers. Design: Fifty HSG films were distributed to five fertilit y practitioners with a mean of 20 years clinical experience. Each obse rver evaluated components of uterine and tubal status and provided cli nical recommendations for hysteroscopy and laparoscopy. Setting: Unive rsity hospital-affiliated reproductive endocrine practice. Interventio n(s): None Main Outcome Measure(s): Tile level of agreement among obse rvers for each uterine and tubal category as determined by the kappa ( kappa) statistic. Determinants of clinical recommendation for further diagnostic studies were assessed. Result(s): The level of agreement be tween observers as determined by kappa ranged from 0.645 in the hydros alpinx category, indicating fair reliability, to 0.111 for pelvic adhe sions, indicating poor reliability. The composite kappa for uterine st atus was 0.345 whereas the composite kappa for tubal status was 0.430. Agreement among observers concerning management showed marginal repro ducibility with a kappa of 0.261. Overall, more than one abnormality o f either the cavity or the fallopian tubes led to a diagnostic recomme ndation for further workup in greater than or equal to 90% of cases. C onclusion(s): In a group of five experienced clinicians, there was con siderable variability in the interpretation as well as the clinical ma nagement of the HSG. Physicians caring for infertile couples should be aware of this discrepancy and should, if possible, review carefully b oth the original films as well as the report of the attending radiolog ist in formulating their diagnostic evaluation and management plan.