Hospitals and community-facing networks of care need tools to anticipate the rising and falling rates of patients with SARS-COV-2 to build resilient and available healthcare for patients. The risk to patients of uncertain and strained capacity is evident. Hospital systems that fail to anticipate local increases in patients infected with SARS-COV-2 are likely to experience shortages of providers and equipment. In the absence of reliable, hospital-specific forecasting, hospitals may be forced to increase patients' risk in seeking non-COVID related care. At the same time, they are underprepared to care for COVID-infected patients.
To gain insights into the number of COVID-infected patients who may come to their hospital, many institutions rely on nationally-published models. These national models do not consider local infection patterns, hospital-level policies, clinical decision-making, or shifts in available capacity.
The Beth Israel Deaconess Medical Center's Center for Healthcare Delivery Science used methods derived from machine learning and epidemiology with data from a single hospital. These methods identified that local patterns of infection and social distancing are critical parameters in predicting hospital capacity demands from patients with SARS-COV-2. The Center anticipates that its approach will be generalizable to different healthcare regions and hospital sizes.
Steven Horng, MD, is the Clinical Lead for Machine Learning at the Center for Healthcare Delivery Science at the Beth Israel Deaconess Medical Center and an Instructor of Emergency Medicine at Harvard Medical School.
Ashley O’Donoghue, PhD, is an economist at the Center for Healthcare Delivery Science at the Beth Israel Deaconess Medical Center.
Massachusetts Health Data Consortium
460 Totten Pond Road | Suite 690
Waltham, Massachusetts 02451
For more information,please contact Arleen Colettiat firstname.lastname@example.org
STAY CONNECTEDjoin our mailing list
© Massachusetts Health Data Consortium