Evaluating Triple Aim in integrated care through claims data

• Main Authors: Elias Van Deun, Walter Sermeus, Geert Goderis
• Format: Article
• Language: English
• Published: Ubiquity Press 2019-08-01
• Series: International Journal of Integrated Care
• Subjects: triple aim; claims data; integrated care; measures
• Online Access: https://www.ijic.org/articles/5269

Introduction: In Belgium, 12 integrated care pilot projects (Integreo) have been started in beginning 2018. Triple Aim is used as evaluation framework: improving the health of populations, patient experience of care and reducing the per capita cost of healthcare (Stiefel & Nolan, 2012). To make evaluation sustainable in the long term and limit registration workload, we will evaluate if relevant Triple Aim measures can be calculated based on claims data, which are already systematically collected. One of the most important routine health databases in Belgium is the Inter Mutualistic Agency (IMA) database. This database contains all the claims for the compulsory health insurance in Belgium including doctor and hospital visits, technical interventions and drug deliveries. As such the IMA database is very useful to picture healthcare related processes and trajectories, but also to create accurate proxy parameters for a range of patient conditions (Vaes et al., 2018). Theory/Methods: In a first stage, we created a list of “Triple Aim Claims Indicators” (TACI). To do this, we used the list of Triple Aim measures from the systematic review of Hendrikx et al. (2016). In parallel, we analysed the list of available data in the IMA-database. We combined the two lists to identify the Triple Aim measures that can be calculated based on the IMA-data. This exercise resulted in the TACI-list. Selection criteria were applied. The selection process was double blinded. Differences were resolved by consensus. In a second stage, we validated this list. Hypothesis was that the TACI’s were able to discriminate care with high from care with low added value. As such these TACI’s should be useful to evaluate healthcare for subpopulations of patients and for comparing different regions and settings. To validate the list, we evaluated the TACI-list in the population with at least one chronic condition, compared to the total Belgian population. Results: Out of the list of 865 Triple Aim measures identified by Hendrikx et al. (2016), we were able to identify a list of 662 measures, including hospital admissions, disease prevalence and cost. This list need to be further checked, for ea. project specific measures and incompatibilities with the Belgian health system. Discussion: A rough check of what is available in the IMA-database for measuring Triple Aim, provides the following pre-analysis: IMA has health outcome information through proxies, although these need to be handled with care (dimension 1) There is very limited information on patient experiences (dimension 2)  IMA provides an extensive range of cost information (dimension 3) Conclusions: Triple Aim can partially be measured through claims data, although not all dimensions to the same extent, especially the patient experience of care. Cost is the most easy to track. Lessons learned: New data collections are not always necessary. Making good use of already available data can create important efficiency gains. Limitations: This research is limited to the Belgian IMA-database. Suggestions for future research: Checking the TACI-list against claims data in other countries.  Checking the list of Triple Aim measures with other (routine) databases.