Visit our website: http://bit.ly/2GtXaiw
Romy Hussai, Director, Data Science and Healthcare Economics, Johns Hopkins Healthcare
Abstract:
I will presents a tripartite model that we nicknamed Callisto. Using a stepwise accounting of individual heterogeneity and stochasticity, we developed three intertwined models to more accurately route patients into complex care management programs. The three models predict interdependent processes in the patient care pathway, small changes to which can vastly change the trajectory and outcomes of a patient’s care.
The first model is the prediction of future cost and utilization: the idea here is that patients who are high cost today may or may not be high cost tomorrow, and today’s low utilizers may not stay that way forever.
The second model is regression to the mean, a major confounder that is not usually fully accounted for in healthcare economic modeling. Two members who have the same future cost structure could revert to the population average at different rates: our research shows that optimum results are achieved by placing the patients who are resistant to regression in care management programs.
The last model is impactability, and it posits that some patients – even those who are resistant to regression and future high cost – simply may not be impactable by standard care management programs. There may be good programs out there for them, but complex care management isn’t it. Using Callisto, Johns Hopkins Healthcare has been able to simulate vastly superior ROI on our care management programs. Implementation is in its early phases, and results are being measured presently.
Add comment