Heart Failure Prediction

Patients often receive hospital treatment for heart failure and can safely return home, with proper monitoring. Dr. Katherine Kim at UC Davis is tackling a fundamental challenge: determining from limited patient data whether a need for readmission has arisen.  Inaccuracies in this assessment are harmful to patients. Dr. Kim’s approach, applying AI to patient data analysis, is yielding very encouraging accuracy improvements.

The HIBAR elements of this project:

Dual motivations:  Kim is deeply concerned with the problem of improving patient well-being.  Simultaneously, as a scientist, she is fascinated by the prospects of AI in evaluating data provided by patients, and in particular, she has made a fundamental discovery about the objective accuracy of subjective patient assessments concerning their own well-being.

Dual methods: Kim’s work includes data that are provided by patients in their homes, so it has many of the features of traditional patient care, as opposed to research carried out in a highly controlled scientific setting. Nevertheless, her research methods have all the required checks and balances of traditional scientific research.

Dual partners: Kim’s research partners span the gamut from leading computer scientists and clinicians to patient participants. These external partners are not just participating in the research – their views are having a critical impact on it.

Dual time frame: Ultimately, there is a very good chance that Kim’s research discoveries in this project will have profound positive impact on patients. Naturally, this is a long-term process.  If Dr. Kim had restricted her efforts only to short-term improvements, this project would not have taken place. Having said that, her project management efforts have focused urgently on enabling beneficial impact as quickly as possible.