Benzodiazepines (BZDs) are important, effective and widely used medications for treating anxiety, panic, sleep, seizure, and bipolar disorders. Epidemiological studies to date have produced conflicting data about hip fracture risk associated with BZD use by the elderly. This study provides a unique opportunity to compare an epidemiological assessment of the BZD-hip fracture relationship with results based on the effects of a statewide natural experiment that resulted in a sudden, sustained reduction in BZD use. The overall aim of this study is to determine whether BZDs are associated with increased hip fracture incidence in the elderly and, if so, which characteristics of BZD use are most likely to increase the risk. The proposed investigation will use a classical epidemiological design (Phase I) and a unique longitudinal, controlled, intervention-based (quasi-experimental) design (Phase II). The study populations will consist of elderly New York (NY) and New Jersey (NJ) Medicaid enrollees between October 1987 and December 1990. Phase I, an open cohort study of more than 50,000 elderly NJ Medicaid recipients, will estimate the magnitude and significance of the BZD-hip fracture association overall and in specific risk groups. Most importantly, Phase II will test whether a NY policy that dramatically reduced BZD use also decreased hip fracture incidence (overall and in specific risk groups). On January 1, 1989, New York implemented a triplicate prescription program, a potent barrier to BZD prescribing, that resulted in a sudden, sustained decrease of over 50 percent in BZD use in most patient groups, including the elderly. The Phase II survival analyses of this natural experiment will compare changes in hip fracture incidence associated with the dramatic decline in BZD use in NY (intervention state) with changes in NJ (control state) where BZD use remained constant during the study period (total N=14,000). Medicaid and Medicare data provide reliable, valid, person-specific measures of demographics, medication and health services use over time, including hip fracture occurrence. The proposed study will provide important data on the association between BZDs, used by more than 20 percent of the elderly, and hip fractures, a major cause of morbidity and mortality. Phase II analyses will also contribute a novel method for evaluating health outcomes of changes in medication use in large populations.
| Zhang, Fang; Wagner, Anita K; Soumerai, Stephen B et al. (2009) Methods for estimating confidence intervals in interrupted time series analyses of health interventions. J Clin Epidemiol 62:143-8 |
| Wagner, Anita K; Ross-Degnan, Dennis; Gurwitz, Jerry H et al. (2007) Effect of New York State regulatory action on benzodiazepine prescribing and hip fracture rates. Ann Intern Med 146:96-103 |
| Wagner, Anita K; Zhang, Fang; Soumerai, Stephen B et al. (2004) Benzodiazepine use and hip fractures in the elderly: who is at greatest risk? Arch Intern Med 164:1567-72 |
