Approaches to informing policy and planning in residential aged care: new knowledge from existing data

Abstract

Background In public health policy, where decisions are made that affect the health and wellbeing of thousands or millions of people, relevant information to aid decision-making may be sparse or misleading. With rapidly ageing populations, decisions are increasingly needed to inform service provision for late-life care of older people. In New Zealand, residential aged care (RAC, equivalent to nursing homes and/or care homes elsewhere) is part of that service, but is not well described. Data such as administrative or transactional data are frequently used for official reports but much is unknown about usage patterns or future demand. If analysed differently, data currently held in information systems have potential to reduce information gaps. Objectives To answer fundamental questions about the use of RAC and hospital services by older people in New Zealand (NZ) using analytic methods that are often used in population surveys but seldom used in epidemiology or health services research. Methods Data from many administrative, survey and research sources are employed, using ratio estimation and survey reweighting when needed to adjust for study design. Analytical methods include logistic regression, generalised linear regression, point-process renewal methods, and proportional hazards models. Findings In recent decades, reports of the percentage of those aged 65 years and over living in RAC varied markedly for a range of reasons. Since 2008 reports were more consistent, stabilising between 4%-6%. When using payments data to describe residents, bias is introduced because the 25% who are not subsidised differ systematically from the subsidised. At least 47% of individuals reaching 65 years of age enter RAC for late-life care. After the age of 85 years that likelihood reaches over 58% for men and 70% for women. Using survey reweighting techniques, the median length of completed stay in RAC was estimated at 2.0 years, with 17% dying within 3 months and 23% surviving more than 5 years. Over a 12 month period, an estimated 64 new admissions to RAC occur for every 100 occupied beds, 14 being transfers from other facilities. Half the new residents enter directly from an acute hospital stay and these are likely to die sooner than others. Over the period of a year, for every 100 residents, an estimated 41 emergency department presentations and 52 hospitalisations occur, of which 38 are acute. When seeking to identify facilities with higher use of acute hospitals, for example to offer support to facility staff, studies must choose between several measures of level of use. Different measures produce very different rankings of RAC facilities and the choice is best determined by the reason for making the selection. Conclusions Planning for long-term care needs of older people deserves high priority given the costs of such care and expected growth in demand. Identified weaknesses and information gaps in information systems may hamper good debate and policy. The novel use of existing data has answered questions that were previously unanswerable. Survey and other research sources, in conjunction with administrative data and with adjustments for length-biased sampling, are useful in this and other settings. The results demonstrate the entrenched role RAC plays in NZ. For some, that signals the importance of strategies to better support people living at home, to avoid or delay RAC entry. For others, it emphasises the need for better planning for the provision and funding of care for older people. Yet others will be interested in developing and testing options that provide both support and choice to individuals and their families. It is important for the ageing population of the future to create and maintain information systems that measure service utilisation consistently and that monitor and project trends across time in order to inform policy and practice.