Prevalence and data transparency of national clinical registries in the United States

Objective: To determine the prevalence and characteristics of national clinical registries. Methods: Review of clinical registries through the following: (1) PubMed search using MeSH term "registries," (2) clinical trials database search using the term "registry," (3) review of the American Medical Association (AMA) recognized specialty societies for registry affiliation, and (4) consultation with a panel representing the American Board of Medical Specialties (ABMS). Main Outcome Measures: Outcomes that characterize registries (type, participants, specialty affiliation, funding), reflect data quality (risk adjustment, auditing practices), and indicate transparency (public reporting). Results: We identified 153 clinical registries of which 47.7% (73) were health services registries, 43.1% (66) were disease registries, and 9.2% (14) were combination registries. The mean number of hospitals per registry was 1,693 (interquartile range [IQR] = 45-230), and the mean number of patients per registry was 1,160,492 (IQR = 2,150-10,045). Among the 117 AMA specialty societies, 16.2% (19) were affiliated with a registry. Government funding was associated with 26.1% (40/153) of registries. Of the 153 registries, 23.5% (36) risk adjusted outcomes and 18.3% (23) audited data. Mandatory public reporting of hospital outcomes for all participating hospitals was associated with 2.0% (3/153) of registries. Conclusion: There is substantial opportunity to develop more specialty-specific clinical registries with publicly available data.