This thesis sets out to demonstrate by way of four sub-studies that elderly people who live in the community do not constitute, as might be imagined, a homogeneous group, but rather make up identifiable 'at-risk' subsets or niche populations, all of which have higher needs for health care services and are at higher risk of medication misadventure, including adverse drug reactions, than the general public.
Both qualitative and quantitative research methods were employed, and a grounded theory approach adopted, with the findings of each sub-study impacting in turn on the design and direction of subsequent work.
Data were gathered from two separate focus group studies with community nurses, and medication chart data was collected from 1591 community nursing patients. For the purpose of assessing risk, a unique Risk Assessment Instrument (RAI) was designed and developed across six dimensions of risk - frailty, symptoms, current prescription medications, hoarding of currant and non-current prescriptions medications, medication compliance and medication management. This RAI was administered to 200 community-dwelling elderly, including community nursing patients.
Well-recognised, statistical analysis methods, such as principle components analysis and regression modelling, were employed in risk factor reduction and explaining variance within each of the risk dimensions. Careful examination was made of the various models to accurately predict outcomes and additional analyses performed to verify the ability of the models to differentiate 'at-risk' patients within those risk dimensions.
There are many positive findings from this work. The first is the positive conclusion that community nursing patients represent a subset of community dwelling elderly more 'at-risk' across all frailty and medication-related indicators examined, regardless of the use of more health care services and multiple health-care practitioners, with all models being highly predictive.
Secondly, this thesis has identified an unmet need or 'gap' in the home health-care infrastructure arising from practice change that has reduced the involvement of community nurses with patients' medications management, resulting in less opportunity to assess patients' medication knowledge and compliance. Simply put, there is a lack of pharmaceutical assessment and medication optimisation in these patients.
Thirdly, this thesis presents data clearly indicating that community nursing patients (living in their own homes) have similar health and medication characteristics to hostel and/or nursing home residents, along with all the associated risks.
Fourthly, thesis data clearly identify that no significant progress has been made towards building community-based health-care teamwork, particularly with respect to community-based health-care professionals sharing clinical information to optimise patients' health care and medication outcomes in the home-care environment.
This work also describes a new and effective way to model medication-related risk, presenting a battery of such risk indicators highly predictive of the medication care needs across community-dwelling patients, stratifying them into three categories along a continuum of increasing risk - independent, semi-dependent and dependent elderly.
It also strongly argues that the models of examined risk factors, developed within the study, should be utilised to facilitate the preferential selection of community-dwelling elderly, locked in a continuum of increasing frailty and risk of medication misadventure (including adverse drug reactions), for Homes Medicines Review.