Code used in the 2021-2022 ASPE supported frailty validation using EHR, titled: "Developing and Assessing the Validity of Claims-based Indicators of Frailty & Functional Disabilities and Testing their Use in Other Data (EHRs and linked EHR-claims)
To address the operational gap between CFIs and EFIs, this project focused on validating an established CFI using linked claims-EHR databases of multiple large health systems. Additionally, the project aimed to refine the EFI for potential use across both older (65+) and younger adults (18-64). This project also included an assessment of the validated CFI versus its EFI version as a tool for concurrent and prospective prediction of healthcare utilization among older and younger adults.
In this project, each task included multiple subtasks. Task-1 included administrative subtasks such as acquiring the required legal contracts and data use agreements to access the linked EHR-claims data sources; gathering information on various CFIs and EFIs; and, soliciting feedback from EHR Learning Network 30 on operationalizing EFI in an EHR environment. Task-2 included data management subtasks focusing on linking EHRs with claims; formulating and identifying the patient denominators (e.g., 18-64 vs. 65+); measuring the completeness of frailty-specific codes across EHRs and claims data; and comparing patient-level comorbidity scores across EHRs and claims data. Task-3 included risk prediction subtasks such as validating and refining CFI and EFI against a set of functional disability outcomes; assessing the distribution and concordance of CFI vs. EFI; comparing CFI vs. EFI across multiple health systems; measuring potential disparities using CFI vs. EFI; and, comparing the predictive value of CFI vs. EFI in forecasting healthcare utilization. Task-4 included dissemination subtasks such as presentation at scientific venues; sharing the underlying code in a standardized format; and publishing the study findings in scientific journals.
- Code related to completion of tasks 2 and 3
- Application of Kim et al CFI
- Assessing concordance of CFI-EFI
- Prediction of healthcare utilization with FI scores
- Code related to completion of task 3.3
- Applies to OLDW data resources only
- Assesses counts and averages of FI across claims and EHR for multiple sourceids