Ms Elena Gerstman1,2,3, Dr Jennifer Jones1,2,3, Ms Chris Michaels3, Prof Sue Berney1,2, Prof Karin Thursky2, Prof David Berlowitz1,2,3
1Austin Health, Heidelberg, Melbourne, Australia, 2University of Melbourne, Parkville, Australia, 3Institute for Breathing and Sleep, Heidelberg, Australia
Biography:
Elena is a motivated clinician-researcher and experienced physiotherapist working towards becoming a leader in health services research. Elena has extensive clinical experience working in acute care with a focus on general medicine, frailty, big data and novel models of care. She is currently pursuing a research higher degree on “using health data analytics to identify complex patients in general medicine”. Career highlights include the MacHSR Future Fellows leadership and research program. Elena is committed to using research to improve health care for older Australians.
Abstract:
Background:
Patients who are “complex” experience poorer outcomes during and after inpatient care. At our health service, patients identified as complex are referred to a specialist transdisciplinary allied health pathway, but this process is subjective and predominantly based on clinical expertise only.
Aims:
To characterise the patients referred to the complex pathway by developing a list of words clinicians associate with complexity, and by describing the characteristics and outcomes of this cohort.
Methods:
A cross-sectional survey of clinicians (allied health, medical, nursing) and a retrospective observational cohort study of all patients admitted to General Medicine at a quaternary metropolitan hospital in Melbourne over a 10-month period. We compared the demographics, clinical features and outcomes of the complex patients to their non-complex peers. The survey data scored the likelihood of complexity-suggestive words from a clinician's perspective. Cohort outcomes included length of stay, readmissions, discharge destination, mortality and adverse event rates.
Results:
Eighty clinicians (allied health (50%), medical (31%) and nursing (19%)) generated a dictionary of 18 words that described a complex patient. In the general medicine cohort (n=3061), 328 (11%) were complex. Complex patients were older, frail and more multimorbid. Complex patients stayed longer in hospital and more required rehabilitation, with increased mortality and readmissions (p<0.01).
Conclusion:
Frailty, age and high hospital utilisation were associated with complexity across both studies. Combining clinical and demographic data with natural language processing of complexity words may allow prospective digital prediction of patients likely to benefit from complex care pathways.