Population characteristics in cervical cancer trials: Search for external validity

Objective The purpose of this study was to compare the characteristics of patients with cervical cancer who were enrolled in cooperative group trials with characteristics of the cervical cancer population of the United States and to determine the generalizability of the results of those trials to the disease population in the United States. Study design Phase III trials in cervical cancer that were conducted by cooperative groups in the United States from 1981 through 1997 were identified. These groups were placed into 4 categories on the basis of disease stage and primary treatment modality: Stage IB, negative pelvic nodes that were treated with radical hysterectomy (n = 277 patients); Stage IB-IIA, positive pelvic nodes that were treated with radical hysterectomy (n = 239 patients); Stage IB2, negative pelvic and para-aortic nodes that were treated with radiation therapy (n = 369 patients); and Stage IIB-IVA, negative para-aortic nodes that were treated with radiation therapy (n = 1190 patients). For each category, comparable patients from the Surveillance, Epidemiology, and End Results (SEER) database were identified. The age and ethnic distributions of each study population and the distributions of the SEER program were compared. Results The age distributions were equivalent, except for patients with IB2 disease that was treated with radiation therapy where cooperative group subjects were more likely to be younger than 50 years, (odds ratio, 0.17; 95% CI, 0.11-0.26). A statistically significant higher proportion of black and Hispanic women enrolled in cooperative group studies in comparison with surveillance, epidemiology, and end results. Conclusion Hispanic and black women were recruited proportionately to cooperative group randomized cervical cancer trials in comparison to the United States population. The age distribution of the clinical trial population is also comparable to that of the general population.