Purpose: Low accrual to clinical trials can result in selective samples compromising generalizability and increasing time and costs. Therefore, accurate accrual estimates are essential. This project explored the number and characteristics of subjects who would, or would not, enter an RCT of bracing vs. observation in scoliosis. In addition, this study derived an estimate of the patient-desired effect size for use in power analysis and trial logistics. Methods: Two groups participated: children eligible for the RCT and their parents (scoliosis group), and children at high risk of developing scoliosis and their parents (surrogates). All read a Web-based module describing RCT procedures and the protocols. We surveyed subjects' willingness to participate, decision-making preferences, and desired reduction in surgical rate due to treatment. This risk reduction reflects the patient-desired effect size. Results: 19 scoliosis pairs and 59 surrogate pairs completed the survey. Overall, 56% of parents and 48% of children agreed to participate, but agreement rates were 50% lower in the scoliosis group (39 and 26% respectively). Overall, 32% of parents and children in the surrogate group jointly agreed to participate vs. 11% of the scoliosis group. The majority of all subjects would participate determine the best treatment. The majority who refused did so in order to share in decision-making. 33% of the children wanted to make their own decision. Children with a preference for one treatment arm over the other were less likely to participate than those without a preference. The patient-desired effect size was 50%, indicating that patients' required a 50% reduction in surgical rates before they would consider bracing a viable treatment option. Conclusions: Surrogates and patients apparently differ in their willingness to participate in this RCT, therefore, the use of surrogates in planning studies can be misleading. Additionally, children and parents differ in their treatment preferences, potentially leading to low enrollment and compliance and crossover problems during the study itself. The effect size of 50% corresponded with that proposed by our clinical trial protocol committee and will be used in the power analysis for the grant submission.