Recommendation Model
Once we have the projected score, measure level cut-offs, and the final STAR ratings at a measure level and contract level, we move to the prescriptive section. Improving the projected STAR ratings lead to monetary gains and improved quality of care. A simple way to improve ratings would be to focus on all measures at once. But this can be time consuming, and infeasible in some cases. The Optimization approach that has been proposed here, gives a list of focus measures that need to be worked upon to improve the projected STAR rating. Thus, it aims to achieve maximum gain, with minimum effort. The selection of these focus measures is done based on the following factors -
Difficulty associated with per increment in measure level rating, Possible discount gained that can be a result of focusing on correlated measures, Maximizing final rating, and Minimizing number of measures to be focused on.
METHODOLOGY:
Problem statement- Given the projected ratings at a measure-level (mi), the contract’s own assessment of the difficulty of gap closure across different measures (di), and total expected difficulty (D), what are the minimum number of measures whose projected score needs to improve in order to cause a positive change in the contract’s final star rating. The mathematical formulation chosen for this problem was that of mixed integer programming.
Input variables : Data source (CMS technical notes) and potential rating
Output Variables : Recommended Score
Metrics to Monitor
|
Statistical
|
:
|
Mallow's CP |
R Squared |
Mean Square Error |
Adjusted R Squared |
Mean Absolute Error |
Huber Loss
|
|
Infrastructure
|
:
|
Log Bytes |
Logging/User/IAMPolicy |
Logging/User/VPN |
CPU Utilization |
Memory Usage |
Error Count |
Prediction Count |
Prediction Latencies |
Private Endpoint Prediction Latencies |
Private Endpoint Response Count
|
Visit Model :
medictiv.citiustech.com