Published in the November issue of Medical Care, Moore Foundation grantee, Gabriel Escobar, M.D., and a team of researchers at Kaiser Permanente Hospital, show how the use of predictive models can support planning for when people are leaving the hospital, adjusting for patient risks and enhancing existing quality assurance and case management efforts.
The ability to predict patient outcomes by identifying risk factors and patterns may help to save lives and reduce preventable hospital readmissions. Researchers were able to leverage data from existing systems at two Kaiser hospitals undertaking an early-warning system pilot project that provides emergency and surgical medical staff with a severity of illness score and longitudinal co-morbidity score, as well as an in-hospital deterioration risk estimate, in real time. Using these exiting systems and data, Escobar and team tracked the rate of non-elective rehospitalizations (patients who needed to be readmitted within seven to 30 days of having been discharged) and the rate of mortality among those same patients. Data from 360,036 patients who were hospitalized and discharged from June 2010 to December 2013 were tracked and analyzed.
Results for non-elective rehospitalizations were 5.8 percent and 12.4 percent, respectively; 1.3 percent and 3.7 percent, respectively, for mortality; and composite outcome rates were 6.3% and 14.9%, respectively.
Read the full abstract here.
Message sent
Thank you for sharing.