Improving the Performance of Risk Adjustment Systems

McGuire, Thomas G.; Zink, Anna; Rose, Sherri

doi:10.1086/716199

Cited by 14 publications

(3 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In a recent paper on Switzerland, for example, Beck et al (2020) find that when the fit effect of Swiss risk sharing is taken into account, the overall fit of the payment system rises from just above 20% as measured by the R 2 from a regression to 57%. Substantial improvements in fit have also been observed in work by McGuire et al (2020), andMcGuire et al (2021). The second reason for the big jump in fit in our case is the extreme skewness of Chilean health care spending, which may be more skewed than in other countries, implying that a very large portion of the total variation in spending is concentrated among the high spenders.…”

Section: Performance Measures (Fit) Across Risk-equalization Formulassupporting

confidence: 59%

Designing feasible and effective health plan payments in countries with data availability constraints

Henríquez

Iommi

McGuire

et al. 2022

J of Risk & Insurance

View full text Add to dashboard Cite

Risk equalization is a fundamental tool in health plan payment in many countries. Data availability often constrains the feasible models. This paper proposes, implements and quantifies the gains of a risk equalization scheme which incorporates risk sharing in a data poor context. Risk sharing relies on total spending data likely available for purposes of payment, potentially increasing feasibility of an effective payment design. To examine incentives for risk selection, alternative models are evaluated in terms of fit at individual, insurer, and group level. Using Chile's private health insurance market as case study, we show that modest amount of risk sharing greatly improves fit even in simple demographic-based risk equalization.Expanding the model's formula to include morbidity-based adjustors and risk sharing redirects compensations at insurer level and reduces opportunity to engage in profitable risk selection at group level. Our emphasis on feasibility may make alternatives proposed attractive to countries facing data-availability constraints.

show abstract

Section: Performance Measures (Fit) Across Risk-equalization Formulassupporting

confidence: 59%

Designing feasible and effective health plan payments in countries with data availability constraints

Henríquez

Iommi

McGuire

et al. 2022

J of Risk & Insurance

View full text Add to dashboard Cite

show abstract

“…The DXIs reduced average errors by 80% to 90% relative to the HHS-HCC model for enrollees with rare (1-in-1000 to 1-in-1 000 000) diagnoses, as shown in Figure 2. Modeling with DXI categories thus fixes a concerning selection problem that remains even when the global fit of payments to expected costs is improved by other means, such as constrained regression, reinsurance, mixed payment, and outlier adjustments that have recently been proposed …”

Section: Discussionmentioning

confidence: 99%

Development and Assessment of a New Framework for Disease Surveillance, Prediction, and Risk Adjustment

et al. 2022

View full text Add to dashboard Cite

show abstract

“…The aim was to improve model performance for a particular ethnic group in order to understand fairness in disease prediction. Our measure is somewhat similar to the group fit measurement employed by McGuire et al (31) However, while these authors used group fit for the total payment ratio received for groups by health condition (cancer, heart health, diabetes and mental health), our group fit was AUC by ethnicity. Furthermore, McGuire et al set up a constraint on the group fit measurement (ie, the total payment ratio equals one reflecting a balance between budgeted and actual health expenditures), but we optimized our group fit level by ethnicity through parameter tuning.…”

Section: Model Optimizationmentioning

confidence: 99%

Predicting the risk of diabetes complications using machine learning and social administrative data in a country with ethnic inequities in health: Aotearoa New Zealand

Nghiem,

Wilson,

Krebs

et al. 2023

Preprint

View full text Add to dashboard Cite

BackgroundIn the age of big data, linked social and administrative health data in combination with machine learning (ML) is being increasingly used to improve prediction in cardiovascular diseases (CVD). We aimed to apply ML methods on extensive national-level health and social administrative datasets to predict future diabetes complications by ethnicity.MethodsFive ML models were used to predict CVD events among all people with known diabetes in the population of New Zealand, utilizing national-level administrative data at the individual level.ResultsThe Xgboost ML model had the best predictive power for predicting CVD events three years into the future among the population with diabetes. The optimization procedure also found limited improvement in AUC by ethnicity. The results indicated no trade-off between model predictive performance and equity gap of prediction by ethnicity. The list of variables of importance was different among different models/ethnic groups, for examples: age, deprivation, having had a hospitalization event, and the number of years living with diabetes.Discussion and conclusionsWe provide further evidence that ML with administrative health data can be used for meaningful future prediction of health outcomes. As such it could be utilized to inform health planning and healthcare resource allocation for diabetes management and the prevention of CVD events. Our results may suggest limited scope for developing prediction models by ethnic group and that the major ways to reduce inequitable health outcomes is probably via improved delivery of prevention and management to those groups with diabetes at highest need.

show abstract

Improving the Performance of Risk Adjustment Systems

Cited by 14 publications

References 30 publications

Designing feasible and effective health plan payments in countries with data availability constraints

Designing feasible and effective health plan payments in countries with data availability constraints

Development and Assessment of a New Framework for Disease Surveillance, Prediction, and Risk Adjustment

Predicting the risk of diabetes complications using machine learning and social administrative data in a country with ethnic inequities in health: Aotearoa New Zealand

Contact Info

Product

Resources

About