Natural history, reasons for, and impact of low/non-adherence to medications for osteoporosis in a cohort of community-dwelling older women already established on medication: a 2-year follow-up study.

Abstract

Approximately 15 % of older women on oral medications for osteoporosis could be considered for alternatives including parenteral therapies. Collection of data on socio-demographic/clinical variables is unlikely to be helpful in predicting low/non-adherence. Alternative approaches are needed to identify individuals at risk of low/non-adherence. This study aims to identify individual patient reasons for stopping medications for osteoporosis, and to investigate whether this can be predicted from knowledge about socio-demographic/clinical data, or whether alternative approaches need to be used. The Cohort for Skeletal Health in Bristol and Avon (COSHIBA) recruited 3200 older women from South West UK, of whom a proportion were on medications for osteoporosis at baseline. Information on self-reported adherence and reasons for low/non-adherence were collected at 6-monthly intervals over a 2-year period. Data was also collected on potential predictors of and impact of low/non-adherence. Two hundred thirty-three of 3200 (7.3 %) women were on medications for osteoporosis at baseline. Mean length of time on treatment prior to enrolment was 46 months. Of those on osteoporosis medications, 94.9 % were on bisphosphonates; 8.5 % reported low adherence and 21.6 % stopped their medication completely over the 2-year follow-up period. Length of time on medication at baseline did not influence rates of low/non-adherence. Reasons for low/non-adherence to bisphosphonates included side effects (53.9 %), practical reasons such as forgetting to take them (18.0 %) and beliefs about medications (20.5 %). No convincing predictors of low/non-adherence were identified. Approximately 15 % of older women on oral medications for osteoporosis could be considered for alternatives including parenteral therapies. This has important implications for healthcare provision. Collection of data on socio-demographic/clinical variables is unlikely to be helpful in predicting low/non-adherence. Alternative approaches are needed to identify individuals at risk of low/non-adherence to osteoporosis medications.