OBJECTIVE: To estimate the effect of a triplicate prescription policy on administration of benzodiazepines and possible substitute drugs in nursing home residents during 1994-1995. DESIGN: Retrospective cross-sectional study SETTING: Medicare- and Medicaid-certified nursing homes in five states: New York, which has a policy requiring triplicate prescription forms for certain drugs, and four states without such a policy. SUBJECTS: A total of 211,177 nursing home residents. MEASUREMENTS AND MAIN RESULTS: Data were collected using the Minimum Data Set; logistic regression provided an estimated effect of the policy on prescribing benzodiazepines. We also estimated possible substitute drugs for benzodiazepines (e.g., barbiturates, antidepressants, antipsychotics, and miscellaneous sedative-hypnotics) while adjusting simultaneously for potential confounders. New York nursing home residents were less likely to be receiving benzodiazepine therapy than those in nonpolicy states (odds ratio 0.42; 95% confidence interval 0.41-0.44). No increases occurred in administration of benzodiazepine substitute drugs in New York versus states without a triplicate prescription policy. CONCLUSION: Policy did affect prescription and administration of benzodiazepines in nursing homes in states with versus without a triplicate prescription policy Lack of substitution with other drugs makes it unclear whether overprescribing occurred in states without a policy, or undertreatment occurred in the state with a policy.
OBJECTIVE: To estimate the effect of a triplicate prescription policy on administration of benzodiazepines and possible substitute drugs in nursing home residents during 1994-1995. DESIGN: Retrospective cross-sectional study SETTING: Medicare- and Medicaid-certified nursing homes in five states: New York, which has a policy requiring triplicate prescription forms for certain drugs, and four states without such a policy. SUBJECTS: A total of 211,177 nursing home residents. MEASUREMENTS AND MAIN RESULTS: Data were collected using the Minimum Data Set; logistic regression provided an estimated effect of the policy on prescribing benzodiazepines. We also estimated possible substitute drugs for benzodiazepines (e.g., barbiturates, antidepressants, antipsychotics, and miscellaneous sedative-hypnotics) while adjusting simultaneously for potential confounders. New York nursing home residents were less likely to be receiving benzodiazepine therapy than those in nonpolicy states (odds ratio 0.42; 95% confidence interval 0.41-0.44). No increases occurred in administration of benzodiazepine substitute drugs in New York versus states without a triplicate prescription policy. CONCLUSION: Policy did affect prescription and administration of benzodiazepines in nursing homes in states with versus without a triplicate prescription policy Lack of substitution with other drugs makes it unclear whether overprescribing occurred in states without a policy, or undertreatment occurred in the state with a policy.