Shared Decision Making, Patient Satisfaction and Medication Use Behavior

Gediwon Negash Milky, Purdue University

Abstract

The objectives for this study were to assess association between demographic variables and clinical variables with shared decision making, to assess association between shared decision-making and patient satisfaction with care, and to assess association between shared decision-making and adherence to anti-diabetic medication among patients with diabetes. An observational retrospective cohort study design was employed to achieve the objectives using data from the 2011 to 2014 Medical Expenditure Panel Survey (MEPS). To be included in the sample, MEPS respondents had to have a diagnosis for diabetes, be 18 years old or older, have identified a usual source of care, and made a visit to that usual source of care in the prior 12 months. Patients who had missing responses on survey items used to assess shared decision making or satisfaction with care, those who had less than two prescription claims for non-insulin anti-diabetic medication, those who had reported cognitive limitation, and those who had proxy responders to survey items were excluded. A total of 797 respondents met the sample selection criteria and were included in the study sample. A majority (52 percent) of the sample was 45 to 64 years old, 57 percent was female, and 54 percent was non-white. A shared decision-making score was calculated for each patient based on responses to seven survey items. Satisfaction with care was measured using a global health care rating survey item, approximately 12 months following assessment of shared decision making. Medication adherence was assessed for approximately 12 months following assessment of shared decision-making. Medication Possession Ratio (MPR) was calculated for each patient, and patients with an MPR score of 85 percent were considered as adherent while patients with MPR less than 85 percent were considered as non-adherent. Multivariate logistic regression was used to examine association between demographic variables and clinical variables with shared decision-making. Variables associated with high shared decision making in the adjusted model include younger age, female sex, having public insurance, and having an excellent perceived health status. Multivariate linear regression was used to assess association between shared decision-making and satisfaction with care. After adjusting for demographic and clinical variables, high shared decision-making was associated with approximately 40 percent higher satisfaction with care than low shared decision making (effect estimate=0.396, p<0.0001). Multivariate logistic regression was used to assess association between shared decision-making and adherence to non-insulin anti-diabetic medications. No association was found between shared decision making and medication adherence as reflected in medication possession ratio. In conclusion, patients with diabetes who had high shared decision-making were found to be more satisfied than those who had low shared decision-making. Shared decision-making among patients with diabetes was not associated with medication adherence as reflected in medication possession ratio.

Degree

M.S.

Advisors

Thomas, Purdue University.

Subject Area

Pharmaceutical sciences

COinS