Prescribing of psychotropic medications to the elderly population of a Canadian province: a retrospective study using administrative databases

Background. Psychotropic medications, in particular second-generation antipsychotics (SGAs) and benzodiazepines, have been associated with harm in elderly populations. Health agencies around the world have issued warnings about the risks of prescribing such medications to frail individuals affected by dementia and current guidelines recommend their use only in cases where the benefits clearly outweigh the risks. This study documents the use of psychotropic medications in the entire elderly population of a Canadian province in the context of current clinical guidelines for the treatment of behavioural disturbances. Methods. Prevalent and incident utilization of antipsychotics, benzodiazepines and related medications (zopiclone and zaleplon) were determined in the population of Manitobans over age 65 in the time period 1997/98 to 2008/09 fiscal years. Comparisons between patients living in the community and those living in personal care (nursing) homes (PCH) were conducted. Influence of sociodemographic characteristics on prescribing was assessed by generalized estimating equations. Non-optimal use was defined as the prescribing of high dose of antipsychotic medications and the use of combination therapy of a benzodiazepine (or zopiclone/zaleplon) with an antipsychotic. A decrease in intensity of use over time and lower proportions of patients treated with antipsychotics at high dose or in combination with benzodiazepines (or zopiclone/zaleplon) was considered a trend toward better prescribing. Multiple regression analysis determined predictors of non-optimal use in the elderly population. Results. A 20-fold greater prevalent utilization of SGAs was observed in PCH-dwelling elderly persons compared to those living in the community. In 2008/09, 27% of PCH-dwelling individuals received a prescription for an SGA. Patient characteristics, such as younger age, male gender, diagnoses of dementia (or use of an acetylcholinesterase inhibitor) or psychosis in the year prior the prescription, were predictors of non-optimal prescribing (e.g., high dose antipsychotics). During the period 2002/3 and 2007/8, amongst new users of SGAs, 10.2% received high doses. Those receiving high dose antipsychotics did not show high levels of polypharmacy. Conclusions. Despite encouraging trends, the use of psychotropic medications remains high in elderly individuals, especially in residents of nursing homes. Clinicians caring for such patients need to carefully assess risks and benefits.