Random digit dialing (RDD) has served for decades as the control selection method of choice but, due to social and technological change, is becoming increasingly unfeasible. Thus the COG Epidemiology Committee is actively seeking strategies alternative to RDD for case-control studies of pediatric cancers. United States (US) birth registries are an appealing source of young control children since they are population-based, provide demographic and pregnancy data for comparison of responders to the study base, and are able to match on birth characteristics. Not all registries can participate in epidemiological studies, however, and cases born in such states should technically be excluded. In studies of rare pediatric cancers, for which cases must be recruited nationwide, this could result in the exclusion of an unacceptable number of subjects. If participating states are representative of the US then a relaxation of orthodoxy may be warranted. The purpose of this report is 1) to describe the ability of US birth registries to release information sufficient to locate and contact control subjects and 2) to determine if participating states adequately represent the US as a whole with respect to characteristics reported in the US Census 2004 American Household Survey. We contacted 52 US birth registries (50 states plus the District of Columbia and New York City) between 12/2004 and 6/2005 to obtain rosters of controls subjects for COG case-control studies of hepatoblastoma and infant leukemia. Thirty-two states (n = 32), in which 79.9% of children 0-5 years of age reside, approved the recruitment of controls through their birth registries. Reasons for nonparticipation of the remaining birth registries were statute (n = 8), departmental policy (n = 4), lack of resources (n = 4), and nonresponse (n = 4). We calculated mean values of selected Census characteristics of participating states, weighted for population 0-5 years of age, and compared them to US values using the one-sample t-test. Participating states did not differ significantly from the US as whole with respect to the mean percentage (% age) of white residents (75.7 vs. 75.6; p = 0.97), mean % age of black residents (11.7 vs. 12.2; p = 0.69), mean % age of children living below the poverty level (18.5 vs. 18.4; p = 0.91), mean % age of adults with at least a high school education (83.4 vs. 83.9; p = 0.48), and median household income ($45,510 vs. $44,684; p = 0.46). Mean number of people per square mile differed significantly (226 vs. 79.6; p = 0.001) due to the nonparticipation of several sparsely populated states. We conclude that is possible to select controls through a preponderance of state birth registries, which account for a large majority of young children and which adequately represent the characteristics of the US as a whole. This work was supported by National Institutes of Health grants R01CA111355 and R01CA079940.

[Proc Amer Assoc Cancer Res, Volume 47, 2006]