Calibration Model – Clinical

• CitiusTech

All other consider year level data as inputs. As the measure level performance differs throughout the year, considering these changes can lead to a significant improvement in the projection accuracy. The Calibration model for predicting end-of-year star rating uses the periodic measure-level progress data and updates the initial compliance rate projections. This will help the contracts have a realistic picture of their expected EOY rating at every reporting period. It is executed Weekly/Monthly based on the data available at clients end, Works on HEDIS and PDE datasource. Contract-measure level projections. METHODOLOGY: Contracts have their own method for tracking progress made along different measures relevant for stars goals. This data can come from HEDIS engines like BI-Clinical or tables generated as part of the health plan’s reporting provided they match the schema required for ingestion. To calibrate the initial EOY measure-level projection using evidence from reporting, a flexible framework is created that takes in historic measure-level data at a granularity of month or week. The Time Series approach can be validated using forecast for previous years. Example - • For 2021 forecast, build model on data from Jan2019 till Jun 2020 • Forecast for Jul 2020 till Dec 2020 • Validate the forecast against actual values, select the model with lowest error • Rebuild the model with the best algorithm Input data: Jan2019 till Dec 2020; Forecast: Jan 2021 onwards

Input variables : Compliance per measure (BIC) and projected compliance rates
Output Variables : Projected score & Projected rating

Visit Model : medictiv.citiustech.com