Falls prevention in community-dwelling older people with co-morbidity : a targeted approach

This thesis examines the prevalence and patterns of comorbidity in community-dwelling older people hospitalised for fall-related injury, and assesses the impact of comorbidity on inpatient service use by this population. This information is important for identifying target groups for falls prevention. The thesis is motivated by a desire to increase potential gains in avoidable inpatient service use due to falls in community-dwelling older people. Five research questions were answered using hospital discharge data from Victoria and New Zealand between 01 July 2005 and 30 June 2008, inclusive. Two of the research questions were methodological and integral to the rigour of the thesis given limitations in the Victorian data. The first methodological question addressed the accuracy of both incidence and comorbidity estimates derived from unlinked (episode-based) Victorian data, compared to those estimated from internally linked (person-based) Victorian data. The second methodological question addressed the accuracy of evidence-based criteria for identifying fall-related incident injury from internally linked data in the absence of the date of injury. New Zealand data containing the date of injury were used for this purpose. Findings from the methodological research were used to optimise case selection from Victorian data for empirical research. The empirical research examined the burden of hospitalised fall-related injury in community-dwelling older people in Victoria; and assessed the prevalence, patterns and impact of comorbidity on inpatient service use by those hospitalised. On the methodological front, the thesis concluded that methods of estimating fall-related hip fracture incidence in community-dwelling older people using unlinked hospital discharge data overestimated fall-related hip fracture incidence, underestimated the prevalence of some comorbid conditions and may alter the relative ranking of these conditions. A further conclusion was that in the absence of the date of injury, validated evidence-based criteria may be used to reliably estimate the incidence of fall-related hip fractures in community-dwelling older people from good quality internally linked data. On the empirical front, the thesis found that the total burden of hospitalised fall-related injury in community-dwelling older people in Victoria was 284,781 hospital bed days in 2005–06, rising to 310,031 hospital bed days in 2007–08. This burden was considerably higher than previously estimated. Hospitalisation rates were highest in women and people aged 85+ years. The thesis further found that more than one in four patients had at least one comorbid condition and among these, one in three had multi-comorbidity. Five distinct, biologically plausible clusters of comorbidity were identified. Comorbidity was shown to be associated with significant increases in inpatient service use and the overall pattern was one of highly prevalent comorbidities having lower impact. These findings suggest that, on the basis of avoidable fall-related inpatient service use, a targeted approach in which community-dwelling older people with comorbidity are identified and provided with cost-effective falls prevention interventions may potentially be cost-saving to the health system, compared to a universal population-based approach. The targeted approach may be complementary to a more universal population-based approach offering falls prevention programs to women and those aged 85+ years.