Exploring the characteristics, outcomes and clinician perspectives of general medicine patients with complex allied health needs in the digital age – an observational cohort and cross-sectional survey

Abstract

Elena Gerstman 1,2,3, Jennifer Jones1,2,3, Chris Michael4, Sue Berney1,2 Karin Thursky2, David J Berlowitz1,2,3 Exploring the characteristics, outcomes and clinician perspectives of general medicine patients with complex allied health needs in the digital age – an observational cohort and cross-sectional survey 1. Physiotherapy Department, Division of Allied Health, Austin Health, Heidelberg, Australia 2. University of Melbourne, Parkville, Australia 3. Institute of Breathing and Sleep, Heidelberg, Australia 4. Business Intelligence Unit, Austin Health, Heidelberg, Australia Aim: Patients who are “complex” experience poorer outcomes for inpatient care, yet there is no objective means to identify this group. To address complexity at our health service, patients deemed to have complex allied health needs in general medicine are referred to a specialist transdisciplinary allied health pathway. We aimed to build an integrated clinical and digital picture of those patients clinically identified as complex by examining their characteristics and outcomes, and by developing a list of words that clinicians associate with allied health complexity. Methods: Two studies were performed in general medicine at a quaternary hospital in Melbourne, Australia: 1) a cross-sectional survey of clinicians; 2) a retrospective observational cohort study of all patients admitted over a 10-month period. We compared the demographics, clinical features and outcomes of a complex and non-complex cohort of patients using a routinely-collected dataset. Results: In the general medicine cohort (n=3061), 328 (11%) were identified as complex by treating clinicians and referred for the complex pathway. The complex cohort were frail, significantly older, more multimorbid and more likely to have cognitive impairment and multiple previous admissions than non-complex patients. Complex patients stayed longer in hospital and had increased mortality and re-admissions (p<0.01). Survey participants (n=80) including allied health (50%), medical (31%) and nursing (19%), generated a dictionary of 18/37 words that describe a complex patient in the progress notes, providing additional information for the digital picture. Conclusion: Frailty, cognitive impairment, age and high hospital utilisation were associated with allied health complexity across both studies. Impact: These results may inform digital strategies to identify complex patients at risk of poor outcomes. Combining clinical and demographic data with natural language processing of “complexity descriptors” may provide for early algorithmic prediction of patients likely to benefit from complex care pathways.