Prognosis of Cardiac Arrest in Patients Receiving Home Care

Abstract

Background: The home care population is a cohort of medically complex older adults at risk for cardiac arrest and poor post-cardiac arrest health outcomes. Research Question: What is the prognosis of cardiac arrest among patients receiving home care, and what pre-arrest features and geriatric syndromes (e.g., frailty) are prognostic of survival and post-cardiac arrest health? Methods: Following a systematic review and meta-analysis that evaluated the prognostic association between frailty and post-cardiac arrest outcomes, a population-based retrospective cohort was created of adults (≥18 years) who received cardiac arrest care at a hospital in Ontario, Canada, between 2006 and 2018. Patients receiving home care and nursing home residents were identified using the Home Care Dataset and the Continuing Care Reporting System. Arrests were analyzed overall and within distinct sub-groups of in-hospital (IHCA) and out-of-hospital cardiac arrests (OHCA). The primary outcome for this thesis was 30-day survival post-cardiac arrest. Frailty was measured using the Clinical Frailty Scale and a valid frailty index. The odds of survival from cardiac arrest were estimated using multivariable logistic regression. Prognostic models were internally validated using bootstrap resampling (n= 2000). Results: We found high certainty evidence for an association between the Clinical Frailty Scale and death prior to hospital discharge after IHCA (OR = 2.93; 95% CI = 2.43 – 3.53) after adjusting for age as a minimum confounder. Our retrospective cohort contained 86,836 unique adult cardiac arrests, of which 39,610 were OHCA and 47,226 were IHCA. Patients receiving home care represented 10.7% of the cohort and were less likely to survive to hospital discharge (RD = -6.4; 95%CI = -7.4– -5.2) and one-year (RD = -12.8; 95%CI = -14.6 – -10.9) post-cardiac arrest compared to community-dwelling individuals receiving no support in the community. Frail patients receiving home care had worse odds of 30-day survival when measured with the CFS (OR=0.78; 95%CI = 0.61-0.98) and a frailty index (OR=0.89; 95%CI = 0.85-0.95), after adjusting for age, sex, and arrest setting. My prognostic model out-performed the two valid frailty measures and demonstrated fair discriminative accuracy (AUROC = 0.66; 95%CI=0.65-0.65) and good calibration (Slope = 0.95) for group-level prognostication when internally validated among patients receiving home care. Conclusion:Patients receiving home care have a worse absolute risk of death when compared to community-dwelling individuals receiving no community-based support services. Frailty is associated with survival and post-cardiac arrest declines in cognition and function when evaluated in patients receiving home care. The prognostic model developed within my thesis outperformed the ability of frailty to predict 30-day survival and is suitable for group-level prognostication.