BACKGROUND: Geriatric hip fracture is a common condition, and there are many open questions regarding patient management. Among the various types of medical evidence, the prospective randomized controlled trial (RCT) is considered the best. Our primary hypothesis was that small sample size would be seen frequently among RCTs involving geriatric patients with hip fracture. A related hypothesis was that studies from the United States would have particularly large deficits in sample size. Therefore, we asked the following research questions: (1) What is the mean sample size of RCTs involving geriatric patients with hip fracture? (2) How do sample sizes for studies from the U.S. differ from those performed elsewhere? METHODS: Following the PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses) guidelines, a systematic review of hip fracture RCTs was conducted. The Embase and MEDLINE databases were searched. Additional data included the country of origin, the power of the study, and whether sample size calculations were performed. One hundred and forty-seven RCTs were identified. RESULTS: The mean sample size of the 147 RCTs was 134.9. The mean sample size for the 7 American trials was 110.3, and the mean sample size for all trials conducted outside of the United States was 136.1. A sample size that was sufficient to ensure 80% power was used in only 31.3% of the RCTs. CONCLUSIONS: RCTs for hip fracture are small and underpowered. Moreover, <5% of the RCT studies have been conducted in the U.S., and they were smaller than those conducted elsewhere. The shortage of American trials may be a feature of the dispersion of geriatric hip fracture care across many hospitals in the United States. If so, better clinical research might require more centralized care (e.g., in specialized geriatric hip fracture centers) or greater collaboration among the many hospitals that provide care.
BACKGROUND: Geriatric hip fracture is a common condition, and there are many open questions regarding patient management. Among the various types of medical evidence, the prospective randomized controlled trial (RCT) is considered the best. Our primary hypothesis was that small sample size would be seen frequently among RCTs involving geriatric patients with hip fracture. A related hypothesis was that studies from the United States would have particularly large deficits in sample size. Therefore, we asked the following research questions: (1) What is the mean sample size of RCTs involving geriatric patients with hip fracture? (2) How do sample sizes for studies from the U.S. differ from those performed elsewhere? METHODS: Following the PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses) guidelines, a systematic review of hip fracture RCTs was conducted. The Embase and MEDLINE databases were searched. Additional data included the country of origin, the power of the study, and whether sample size calculations were performed. One hundred and forty-seven RCTs were identified. RESULTS: The mean sample size of the 147 RCTs was 134.9. The mean sample size for the 7 American trials was 110.3, and the mean sample size for all trials conducted outside of the United States was 136.1. A sample size that was sufficient to ensure 80% power was used in only 31.3% of the RCTs. CONCLUSIONS: RCTs for hip fracture are small and underpowered. Moreover, <5% of the RCT studies have been conducted in the U.S., and they were smaller than those conducted elsewhere. The shortage of American trials may be a feature of the dispersion of geriatric hip fracture care across many hospitals in the United States. If so, better clinical research might require more centralized care (e.g., in specialized geriatric hip fracture centers) or greater collaboration among the many hospitals that provide care.