Predictors of adverse outcomes and factors associated with a risk-stratification based community virtual ward model for older persons with complex health and social care needs

<p><b>Introduction</b></p><p>Population ageing is occurring rapidly worldwide, particularly in more developed<br></p><p>countries. As a result, a greater proportion of frail older adults are expected to be living</p><p>in the community but at increased risk of emergency department (ED) presentation and</p><p>hospital admission. While risk-stratification is useful in allocating limited resources, few</p><p>instruments or models are available to support healthcare professionals to manage the</p><p>complex health and social care needs of these patients. The recent introduction of a</p><p>Community Virtual Ward (CVW) for older persons with complex healthcare needs,</p><p>reduced ED presentations and unplanned hospital admissions. The research presented</p><p>in this thesis expands on previous studies examining predictors of adverse healthcare</p><p>outcomes in community-dwellers. It also investigates the impact of health trajectories</p><p>(transient health states defined as ‘unstable’, ‘deteriorated’ and ‘stable’) and factors</p><p>associated with the risk-stratification model including individual screening and</p><p>assessment scale scores on the risk of hospitalisation, institutionalisation and death</p><p>among patients admitted to a CVW.</p><p><b><br></b></p><p><b>Methods</b></p><p>This study is a longitudinal non-experimental descriptive and correlational design</p><p>investigating patients selected for admission to a CVW in a single centre in Ireland. The</p><p>theoretical principles of a Markov model informed the study design. Correlations and</p><p>associations were examined using Chi-square, and Spearman’s rho tests. To examine</p><p>relationships between risk scores, health states and adverse outcomes a univariate,</p><p>multinomial and multivariable regression analysis was performed.</p><p><br></p><p><b>Results</b></p><p>In total, 88 community-dwellers, with a mean age (+/- standard deviation) of 82.8 +/-6.4</p><p>years were included. Most were severely frail on the Rockwood Clinical Frailty Scale</p><p>(mean 6.8/9 +/-1.33) and highly dependent on the Barthel Index (2.7/20 +/-0.8). Health</p><p>states were predictive of adverse outcomes 60 observed up to 90 days. Reaching a</p><p>level of stability (‘stable’ state) within 60 days of admission to the CVW was a predictor</p><p>for stability at 90 days and remaining at home. Stability was also associated with less</p><p>episodes of care (<2), lower risk of experiencing delirium, and fewer numbers of health</p><p>care professional direct involvement in care (<7). In contrast an ‘unstable’ health state</p><p>observed at 60 days was predictive of institutionalisation or death with an increase in</p><p>the number (>2) of episodes of care, and higher numbers of healthcare professionals</p><p>(>7) direct involvement in care. Several screening and assessment instruments scores</p><p>covering pressure ulcer risk, functional and frailty status, mobility, nutrition and cognition</p><p>were predictive of reaching stability and risk of institutionalisation.</p><p><br></p><p><b>Conclusion</b></p><p>The CVW provides a framework for case management as a conceptual model of risk to</p><p>support older people at home or identify those at risk of institutional care. The use of</p><p>defined health states assisted in stratifying those at lower or higher risk in a high-risk</p><p>frail population. The selection of brief, targeted screening and assessment instruments</p><p>used to monitor risk in this CVW model were also helpful in stratifying participants.</p><p>Further research is needed to confirm these findings and refine this model.</p><p><br></p>