Contextualizing functional status with comorbidities in prognosticating post-hospital discharge outcomes: a retrospective cohort study

Abstract

Background Discharges from acute episodes of hospitalization are well recognized as periods where individuals are at high risk of poor outcomes including hospital and emergency re-presentation, death, and long-term care facility (LTCF) admission or wait listing for older adults. As the population of Ontario ages and is at higher risk of these outcomes, efforts to prognosticate these risks to optimize care delivery have recognized demographics and medical morbidity have been well-associated with these events using large datasets. The impact of individual patient function and capacity to carry out their activities of daily living (ADLs), however, is not as well understood despite being central to an individual’s ability to remain successfully within the community. Recent efforts have allowed for the administrative collection of data concerning functional capacity to both demonstrate the functional characteristics of the acutely hospitalized adult and understand its impact on post-discharge outcomes.

Methods Using databases housed at the Institute for Clinical Evaluative Sciences (ICES), a retrospective cohort of all adults 65 years of age or older who underwent assessment of their ADLs in the Healthcare Outcomes for Better Information and Care (HOBIC) during acute hospital admissions between 2008 and 2016. Individuals who were expected to return to hospital (e.g. due to chemotherapy or dialysis) or were receiving palliative care, as well as individuals assessed at hospitals who did not complete enough HOBIC assessments to provide reliable information were excluded. Descriptive analysis was performed to understand the functional and medical characteristics of the cohort. Multivariable logistic regression analysis was used to determine how demographics, comorbidities, previous health service use, and functional status were associated with hospital re-admission, emergency re-presentation, death, and LTCF admission or wait-listing at 180 days post-discharge. Subgroup analysis was also performed across common comorbidities including chronic obstructive pulmonary disease, coronary artery disease, congestive heart failure, falls, delirium, and dementia.

Results An analytic cohort of 139 798 was constructed for whom 67.4% of admission HOBIC assessments and 56.7% of discharge assessments were complete. With increasing age there was increasing functional disability as well as increasing burdens of congestive heart failure, delirium, dementia, and falls. The majority of individuals who were admitted with some degree of functional impairment experienced an improvement during their admission. Function and comorbidities were found to be equally and moderately associated with increased odds of experiencing ED re-presentation and re-hospitalization, though there was poor discriminability (AUROC 0.62, 95% confidence interval 0.62-0.63 and 0.64, 0.63 – 0.65 respectively). For death and LTCF admission or wait listing, however, functional impairment had greater association than comorbidities, and there higher discriminability was seen (AUROC 0.84, 0.83-0.85 and 0.79, 0.79-0.80). Functional impairment was also more highly associated with poor outcomes in those with chronic obstructive pulmonary disease and dementia.

Discussion This analysis demonstrates that administrative data can be used to assess function, and that function is of comparable importance to the studied outcomes as medical comorbidity. It demonstrates that meaningful information concerning function can be derived from secondary data, though information gathered concerning function at admission appears to be of questionable reliability, and comorbidities are likely under-reported. Overall, however, it suggests that future efforts to assess individual risks of post-discharge outcomes should consider functional status as equally important to other factors more extensively historically studied.