Early Detection, Containment, and Management of COVID-19 in Dialysis Facilities Using Multi-Modal Data Sources

Background: Dialysis patients were at high risk for severe complications and death from COVID-19 given older ages and multiple comorbidities. Minorities, especially Blacks and Hispanics, were disproportionately represented among this population. Over 85% of hemodialysis patients must travel three times per week to dialysis facilities for life-sustaining treatments and could not shelter in place. This study aimed to characterize COVID-19 transmission pathways in dialysis patients and clinics, identify potential carriers, and develop procedures to curb spread.

Materials/Methods: This study leveraged demographic, clinical, treatment, laboratory, socioeconomic, serological, metabolomic, wearable, machine-integrated sensor, and COVID-19 surveillance data to develop mathematical and statistical models. These models were implemented across a large number of dialysis clinics to better understand how COVID-19 spread in these settings, identify patients before symptoms appeared, and detect asymptomatic carriers. Novel modeling approaches were designed to fully utilize the high-dimensional, multimodal data available.

Outcome/Impact: The study capitalized on the intrinsic advantages of dialysis clinics to implement and validate prediction models. The intended outcome was to improve patient and staff safety, enhance early identification of infection risk, and support delivery of high-quality, individualized care to this high-risk population.