Domain

Analytic Methods

Type

Empirical Study

Theme

population; operations

Start Date

7-6-2014 2:55 PM

End Date

7-6-2014 4:15 PM

Structured Abstract

Introduction:

Certain populations require special attention to assist them in managing their care. Individuals with persistent high costs are especially challenging. Identifying individuals with high costs, whose cost of care can be influenced, will help care managers better allocate their limited resources. Unfortunately, most models for predicting persistent high costs focus solely on factors related to the individual. This approach ignores decades of research highlighting the multiple determinants of health. The public health social ecological model exemplifies a more robust view of the determinants of health. The importance of this model in explaining health care utilization is illustrated in this presentation where we present applied predictive analytic models developed for commercial payers targeting care management at "impactable" persistent high cost members.

Methods and Findings:

Three years of total cost of care claims data in a large commercial payer data base along with community level and provider level data were analyzed to identify persistent high cost members and the factors that contributed to “persistence” year over year. Community, health care system, direct provider, and patient level factors were assessed in logistic regression models. C statistics approaching .80 were achieved with the full complement of variables and substantially lower statistics were achieved with only person level characteristics. Including multiple determinants of health in predictive models increases their predictive accuracy.

Discussion and Conclusion:

As predictive analytics matures and narrows in targeting populations and outcomes, it is important to recognize that not all outcomes are the result of individual behaviors or disease processes. Establishing the context of care from multiple data sources will focus attention on system interventions as well as care management for individuals attempting to navigate complex care patterns.

Acknowledgements

N/A

Creative Commons License

Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 3.0 License.

Share

COinS
 
Jun 7th, 2:55 PM Jun 7th, 4:15 PM

Recognizing Multiple Determinants of Health in Predictive Analytics

Introduction:

Certain populations require special attention to assist them in managing their care. Individuals with persistent high costs are especially challenging. Identifying individuals with high costs, whose cost of care can be influenced, will help care managers better allocate their limited resources. Unfortunately, most models for predicting persistent high costs focus solely on factors related to the individual. This approach ignores decades of research highlighting the multiple determinants of health. The public health social ecological model exemplifies a more robust view of the determinants of health. The importance of this model in explaining health care utilization is illustrated in this presentation where we present applied predictive analytic models developed for commercial payers targeting care management at "impactable" persistent high cost members.

Methods and Findings:

Three years of total cost of care claims data in a large commercial payer data base along with community level and provider level data were analyzed to identify persistent high cost members and the factors that contributed to “persistence” year over year. Community, health care system, direct provider, and patient level factors were assessed in logistic regression models. C statistics approaching .80 were achieved with the full complement of variables and substantially lower statistics were achieved with only person level characteristics. Including multiple determinants of health in predictive models increases their predictive accuracy.

Discussion and Conclusion:

As predictive analytics matures and narrows in targeting populations and outcomes, it is important to recognize that not all outcomes are the result of individual behaviors or disease processes. Establishing the context of care from multiple data sources will focus attention on system interventions as well as care management for individuals attempting to navigate complex care patterns.