The Analytic Foundation of a Robust Population Health Management Program
Most risk-bearing organizations, including health plans, accountable care organizations (ACOs), and self-funded employers deploy some form of analytical strategy to inform their approach to population health management. But what that means could be as wide-ranging as:
- Using registries to track compliance with quality measures
- Generating static reports on medical cost trend for the chronically ill
- Creating predictive models using artificial intelligence techniques to forecast who will need increased services in the future
The primary goal of a population health management program is to measurably improve health outcomes. To be successful, your analytical strategy should support the key building blocks of the population health value chain, which is defined by the Population Health Alliance (PHA) and theHealth Enhancement Research Organization (HERO). We have illustrated the concept as follows:
To design an effective program that achieves improvement in clinical, utilization, and financial outcomes, there are three analytical strategies that risk-bearing organizations should not overlook:
- Identification and Stratification (aka “ID and Strat”)—the obvious.
Finding the right people to receive the right intervention is key to ensuring the remainder of the population health value chain will work effectively. This is done through data mining, clinical algorithms, and/or predictive modeling. Identification and stratification could be based on financial risk, gaps in care, disease stage, likelihood of high costs, likelihood of non-compliance with quality measures, and more.
- Evaluation–the hygiene factor.
Part of an analytic team’s role should be in unbiased evaluation of your intervention. A review of the PHA/HERO’s publication “Core Metrics for Employee Health Management” is worth your time here – you will find a comprehensive review of the various methodologies for evaluating the performance of a population health management program and how to interpret changes in the health status of a population. Examples of evaluation analytics include:
- Financial outcomes
- Health impact
- Participation rates
- Satisfaction rates
- Quality Assurance – the best practice.
Analytics should be a part of your ongoing QA process. You may have heard this referred to as a “Logical Model” or a “Logical Framework”. It simply means that in order to believe the outcomes from your evaluation, there will be a set of leading indicators that will give you confidence in this result. Analytics should be involved in each step of your intervention method to ensure that these leading indicators are present. Examples of quality assurance analytics should help you address the following:
- Healthy Behaviors
- Do we see improvements in modifiable risk factors?
- Do we see improvements in adherence to prescribed treatments?
- Perceived Health
- Do we see improvements in self-perceived health status?
- Do we see improvements in self-efficacy?
- Are we seeing improvements in productivity?
- Biometric Values
- Are leading biometric indicators improving?
- Healthy Behaviors