Pharmacoepidemiology of Antihypertensive Medication Adherence in Older Adults and Feasibility of Intervening in Community Pharmacy
This thesis aimed to identify the evidence base regarding determinants, outcomes, and measures of antihypertensive medication adherence, and to assess the feasibility of monitoring adherence in community pharmacy. The overarching aim is to inform the development of pharmacy-led adherence interventions. Specific aims were to evaluate 1) patient and health-system determinants of antihypertensive adherence, 2) adverse consequences of poor antihypertensive adherence, 3) validity of pharmacy adherence measures, and 4) to assess the feasibility of identifying poor adherence during repeat dispensing.
A 12-month prospective cohort study of community-dwelling hypertensive older adults (≥65 years) recruited from 106 community pharmacies was undertaken from 2014 to 2015. Participants completed structured questionnaires at baseline (n=1,564) and follow-up (n=1,232), which were linked to their pharmacy dispensing records. At follow-up, high-levels of antihypertensive adherence were observed with 52.2% reporting high adherence using the Morisky Medication Adherence Scale (MMAS-8); 89.9% were estimated to be adherent from pharmacy records using the Proportion of Days Covered (PDC). However, a large proportion of participants (39%) had uncontrolled blood pressure.
The Beliefs about Medication Questionnaire (BMQ) was used to evaluate necessity-beliefs and concerns about antihypertensive medication, a patient determinant of medication adherence. Polynomial regression demonstrated the commonly employed necessity-concerns differential model to be suboptimal and identified a more complex 3-D relationship between concerns, necessity-beliefs and adherence. This 3-D model may facilitate stratifying patients for adherence interventions that target medication beliefs.
For patients eligible for the General Medical Services (GMS) scheme, 30% reported financial burden due to the cost of a €2.50 prescription co-payment charge. Financially burdened participants were less likely to have higher education attainment and private health insurance, which are indicators of lower socioeconomic status. This health-system determinant of adherence was associated with 3.6% lower self-reported adherence at 12-months, although the 2.8% lower PDC was not statistically significant. Alternative policies, which remove this cost-barrier to adherence, should be considered.
Injurious falls are an adverse event of particular interest with antihypertensive medication use in older adults. Gaps in antihypertensive adherence specifically may result in fluctuations in blood pressure that could precipitate falls. In adjusted regression models, each 5-day gap in antihypertensive refill-adherence was associated with an 18% increased risk of an injurious fall. However, this finding is hypothesis generating and should be replicated using methods that are more precise to measure gaps in adherence and injurious falls.
Group-based Trajectory Modelling (GBTM) is a novel refill-adherence measure, which could stratify patients for adherence interventions. However, in this cohort study GBTM failed to demonstrate predictive validity with blood pressure and number of hospital visits, although participants grouped into a low adherence trajectory had a 37% higher rate of General Practitioner (GP) visits. In contrast, each 10% increase in PDC was associated with a 16% and 8% lower rate of hospital and GP visits respectively. Statistically PDC was the superior metric in associations with outcomes. However, contextual reimbursement and workflow practices likely overestimate pharmacy refill adherence in this setting.
Finally, in a factorial survey design, the feasibility of monitoring adherence within the current workflow of community pharmacy was evaluated. Responses from 258 Irish community pharmacists indicate that examining dispensing records is a more feasible method to evaluate medication adherence than questioning patients regarding adherence behaviour or beliefs. Contextual factors such as time-pressures negatively influenced pharmacists’ reported intentions to monitor adherence, while pharmacists with more positive attitudes towards medication monitoring, reported higher intentions to monitor adherence.
Together, the findings of this thesis are useful to inform the future development of community pharmacy-led adherence interventions with particular findings relating to stratification of patients and specific intervention targets relating to medication beliefs underpinning adherence behaviours.