| Objective. To compare survival among patients enrolled in the Collaborative Ocular Melanoma Study (COMS) large tumor trial, with survival among patients who were eligible but who did not enroll.; Methods. The COMS large tumor trial enrolled 1003 patients, and collected baseline data on 299 additional patients judged to be eligible, but who did not enroll. Vital status of the eligible but not enrolled patients was determined by the local clinical center searching medical records and either sending personal identifying information to the COMS Coordinating Center to search the National Death Index and the Social Security Death Index (SSDI), or by someone at the COMS clinical center searching the SSDI using personal identifying information, but only sending the vital status to the COMS Coordinating Center. Survival of COMS enrolled patients was then compared to survival of COMS eligible but not enrolled patients.; Results. Of the 299 patients eligible but not enrolled in the Collaborative Ocular Melanoma Study (COMS) large tumor trial, 129 were from the clinics with IRB approval for this study. There were no statistically significant differences between the baseline characteristics of the 129 patients in this study and the 170 eligible but not enrolled patients not in this study. Of the patients with survival data, 49 were reported as deceased by at least one method as of July 31, 1997. Cox proportional hazards analysis identified a hazard ratio of 1.20 for enrolling in the trial, i.e. hazard of death was increased by 20% in enrolled patients as compared to those eligible but not enrolled in the trial, but this difference was not statistically significant (95% confidence interval = 0.88--1.64). When age and tumor diameter, covariates previously identified as associated with survival, were entered into the model, they remained statistically significant, whereas trial enrollment status did not.; Conclusions. The COMS large tumor trial results appear to be generalizable to all eligible patients, but it is unknown if they are generalizable to patients who would not be eligible for the trial. |
