The rural dispensing practice – does it achieve better medication adherence and clinical outcomes compared to non-dispensing practices? A cross-sectional analysis of routine data

Problem

Rights to dispense medications, as opposed to issuing prescriptions that may not be dispensed, distinguish rural dispensing practices from their urban counterparts. The clinical implications of this difference are unknown. We hypothesised that dispensing status may be associated with better medication adherence than prescribing alone. This could impact intermediate clinical outcomes dependent on medication adherence in, for example, hypertension or diabetes.

Approach

We analysed Quality and Outcome Framework (QOF) data from 2016/17 linked to NHS Business Services Authority dispensing status for England. QOF performance indicators were classified into three groups, namely; 1) indicators dependent on medication adherence (for example, HYP006: the percentage of patients with hypertension whose last blood pressure reading is ≤150/90 mmHg), 2) indicators related to prescribing but independent of medication adherence (e.g. CHD005: percentage of patients with coronary heart disease and a record of anti-platelet or an anti-coagulant prescribed in the last year) and 3) indicators unrelated to prescribing (e.g. AF006: percentage of patients with atrial fibrillation in whom stroke risk has been assessed). Mixed-effects logistic regression was used to estimate differences between dispensing and non-dispensing practices. Adjustment was made for percentage of practice population aged ≥65, percentage of practice population who were male, practice deprivation, practice list size, single-handed status and rurality. Analyses were restricted to practices with list sizes ≥1000.

Findings

Complete data existed for 6,596 practices with over 1,000 patients, of which 976 (14.8%) had dispensing status. Dispensing practices typically had more patients over the age of 65, served less deprived populations, were more likely to be single-handed practices and were much more likely to be in rural areas. We found evidence (p<0.05) that achievement on seven out of nine QOF indicators dependant on medication adherence was higher in dispensing practices than non-dispensing practices. We found greater achievement of blood pressure targets in hypertension (e.g. for HYP006 Odds ratio (OR) 1.07; 95%CI 1.04-1.11, p < 0.0001), coronary heart disease, peripheral arterial disease and diabetes, and greater lowering of total cholesterol to 5mmol/L or less in diabetes. One of the three HBA1c targets in diabetes showed greater achievement in dispensing practices whereas two others showed no difference. We found evidence of differential performance between dispensing and non-dispensing practices for one out of nine indicators related to prescribing but independent of medication adherence; only one of six indicators unrelated to prescribing or dispensing showed any difference.

Consequences

On conservative estimates, between 3% and 10% of prescriptions are not dispensed. Dispensing to patients presents one less barrier to obtaining medication than prescribing alone. Practice dispensing may be associated with better clinical outcomes, possibly through improved medication adherence. Further work is required to clarify the possible underlying mechanisms for, and significance of, these observations.