Healthcare Risk Adjustment and Predictive Modeling

By Ian Duncan · ACTEX Learning · 2018

A guide for healthcare actuaries and other professionals interested in healthcare data and predictive modeling, in its second edition of 2018. Its publisher describes three parts: health risk, available data, identification algorithms and grouper models; data mining and common modeling approaches; and case studies from Medicaid, Medicare Advantage, ACA exchanges and other settings, including use outside the United States.

On the syllabus

Exam GH-301 — Risk Adjustment; Ch. 13: Medicaid Risk Adjustment (3a, 3b); Ch. 14: Risk Adjustment in Medicare, excluding Appendix 14.1 & 14.2 (3a, 3b); Ch. 21: Risk Adjustment on the ACA Exchanges (3a).

Foreword to the Second Edition

Foreword to the First Edition

Preface

Section One Introduction and Background

1 Introduction to Health Risk

1.1 What is Health Risk?
1.2 Health Risk Management
1.3 What Is the Difference between Risk Adjustment and Risk Prediction?
Appendix: Clinical Risk – The Framingham Heart Study

2 Models for Predicting Health Costs

2.1 Introduction
2.2 Non-Condition Risk-Based Models
2.3 Age/Sex and Prior Cost Models: Summary
2.4 Risk Factor-Based Risk Models
2.5 Medical Condition-Based Risk
2.6 Lifestyle-Related Risk Factors
2.7 Exogenous Risk Factors

3 Data: The Raw Material of Modeling

3.1 Overview
3.2 Physician Referral, Medical Charts, and Electronic Medical Records
3.3 Enrollment (Eligibility) Data
3.4 Medical Claims
3.5 Claims Costs
3.6 Paid Claims vs. Allowed Charges
3.7 Facility and Physician Claim Forms
3.8 Interpretation of Claims Codes
3.9 Pharmacy Claims
3.10 The Anatomical Therapeutic Chemical Classification System
3.11 Clinical, Laboratory and EMR Data
3.12 Other Data Available from Non-Claims Sources
3.13 Survey Data
3.14 Conclusion
Appendix 3.1: HCPCS Codes
Appendix 3.2: CMS: Place of Service Codes
Appendix 3.3 Type of Bill (TOB) Code
Appendix 3.4 Revenue Code Categories
Appendix 3.5 Type and Source of Admission Codes
Appendix 3.6: Patient Status (at Discharge) Codes
Appendix 3.7: Physician Specialty Codes
Appendix 3.8: Mapping MS-Diagnosis Related Groups
Appendix 3.9: Legislation Affecting the Use of Data
Appendix 3.10: R Code for Grouping 2017 Medical and Surgical MS-DRG Codes

4 Clinical Identification Algorithms

4.1 Clinical Identification Algorithms
4.2 Sensitivity-Specificity Trade-Off
4.3 Constructing an Identification Algorithm
4.4 Sources of Algorithms

5 Grouper Models

5.1 Use of Grouper Models
5.2 Features of Grouper Models
5.3 Episode Grouper Models
5.4 Therapeutic Class Grouper Models
5.5 Drug-Based Risk Adjustment Models

6 Development and Construction of DRGs, DCGs, and ETGs

6.1 History, Development, and Use of Diagnosis-Related Groups (DRGs)
6.2 DxCG's Clinical Classification System
6.3 Development of Relative Risk Scores
6.4 Concurrent and Prospective Models
6.5 Episode Groupers: Symmetry's Episode Treatment Groups (ETGS)
Appendix: Examples of DRGs

Section Two Model Construction

7 Introduction to Modeling

7.1 What is a Model?
7.2 Example of a Model Application
7.3 Modeling Tools
7.4 The Sample Database
7.5 Modeling
7.6 Testing a Model's Reproducibility
7.7 Selecting Between Models
7.8 Variable Transformation
7.9 Implementation and Computation
Appendix: List of Variables Available in the Test Dataset

8 Linear Regression Models

8.1 Model Specification and Assumptions
8.2 Least Squares Estimation
8.3 ANOVA
8.4 Statistical Inference
8.5 Prediction
8.6 Model Diagnostics
8.7 Example of Linear Regression
8.8 Flexibility of Linearity
8.9 Summary
Appendix 1: Sample Data Set
Appendix 2: R Code for Linear Regression Model

9 The Generalized Linear Model

9.1 Components of GLM
9.2 Maximum Likelihood Estimation (MLE)
9.3 Statistical Inference in GLM: LRT, Wald, and Score Tests
9.4 Model Selection in GLM
9.5 Summary

10 Logistic Regression Models

10.1 Odds and Odds Ratio
10.2 Estimation
10.3 Example of Logistic Regression
10.4 Predictive Accuracy
10.5 Optimal Cutoff Point
10.6 ROC Curve
10.7 Summary
Appendix: R Code for Logistic Regression Models

11 Tree-Based Methods

11.1 Introduction
11.2 Tree Terminology
11.3 Finding the Best Split
11.4 Pruning and Tree Size Determination
11.5 Use and Advantages of Tree-Based Methods
11.6 Decision Tree: Example
11.7 Summary
Appendix: R Code

12 Artificial Neural Networks

12.1 Multilayer Perceptron (MLP)
12.2 Radial Basis Function (RBF) Networks
12.3 Fitting Neural Networks
12.4 Example of Artificial Neural Networks
12.5 Summary

Section Three Applications

13 Medicaid Risk Adjustment

13.1 Background
13.2 Methodology
13.3 Newborn Model
13.4 Application of the Methodology to Capitation Rate Adjustment
Appendix: Acronyms and Other Regulatory References Used in this Chapter

14 Risk Adjustment in Medicare

14.1 Introduction
14.2 Medicare Risk Adjuster Methodology
14.3 Illustrative Examples of 2017 Medicare Risk Score Calculations
14.4 Use of the Projected Medicare Risk Score in the Part C Medicare Bid
14.5 Medicare vs. ACA Health Insurance Exchange Risk Adjustment
14.6 Part D Risk Adjustment
14.7 Conclusion
Appendix 14.1: Medicare's STARS Rating System
Appendix 14.2: Medicare Advantage Bid Process
Appendix 14.3: Revisions and Updates to HCC Models

15 Risk Adjustment and Health Care Reform: The Example of Massachusetts

15.1 Background
15.2 Risk Adjustment in the Commonwealth Care Program
15.3 Choice and Calibration of the Risk Adjustment Model
15.4 Practical Application of Adjustment in the Exchange
15.5 Postscript

16 Developing and Using Predictive Models for Care Management Programs

16.1 The Risk Management Economic Model
16.2 Identification, Stratification, and Prediction
16.3 What Is Opportunity Analysis?
16.4 Member Stratification for Opportunity Analysis
16.5 Designing a Program
16.6 Implementing a Program
16.7 The Value of Opportunity Analysis
16.8 Conclusion

17 Provider Efficiency Assessment and Reimbursement Using Risk Adjustment

17.1 Provider Efficiency Assessment
17.2 Impact of Provider Practice and Care Management on Risk Adjustment
17.3 An Empirical Study of Provider Reimbursement and Independence of Outcomes
17.4 Risk Adjustment Issues in Provider Contracting
17.5 Conclusions

18 Developing a Predictive Model for Depression Severity

18.1 Background
18.2 The Study
18.3 Conclusion

19 A Predictive Model for Disability Underwriting Profitability

19.1 Background
19.2 The Model
19.3 Choice of Decision Tree
19.4 Choice of Variables and Development of Final Model
19.5 Model Evaluation
19.6 Using the Model for Underwriting

20 Predicting Hospital Readmissions

20.1 Why Are Hospital Readmissions Important?
20.2 Readmission Adjustment Penalty Calculation
20.3 Existing Readmission Predictive Models
20.4 Data
20.5 The Model
20.6 Practical Application
20.7 An Economic Model for Readmission Intervention
Appendix: R Code for Running the Model

21 Risk Adjustment on the ACA Exchanges

21.1 The Affordable Care Act and the State Exchanges
21.2 History and Operation of Exchanges
21.3 Risk Adjustment in Exchanges
21.4 Revenue Transfer Process
21.5 Underlying Theory of Risk Adjustment
21.6 A Problem with Non-Linear Cost Relationships
21.7 Empirical Analysis of Cost-Risk Relationship
21.8 Some Other Practical Risk Adjustment Issues
21.9 Conclusion

22 Accountable Care Organizations

22.1 Accountable Care Organizations
22.2 Structure and Organization of Medicare ACOs
22.3 ACO Assignment Process
22.4 ACO Per Capita Costs
22.5 ACO Risk Adjustment
22.6 ACO Savings Calculation
Appendix 22.1: Providers of Primary Care Services (for Determination of Assignment)

23 Predictive Modeling and Risk Adjustment outside the United States

23.1 Predictive Modeling in the United Kingdom
23.2 Risk Adjustment in the Netherlands
23.3 Risk Adjustment in Israel
23.4 Risk Adjustment in Colombia
23.5 Risk Adjustment in South Africa
23.6 Conclusion

24 Communicating Results

24.1 Background
24.2 Persuasion: The Purpose of Communication
24.3 What Constitutes Good Writing?
24.4 Communicating Technical Topics
24.5 A Word about Powerpoint Presentations
24.6 Report Structure
24.7 Actuarial Reporting Requirements
24.8 Conclusion

A Statistical Appendix

A.1 Random Variables and Distributions
A.2 Confidence Intervals and Hypothesis Testing
A.3 Two-Sample Inference
A.4 Analysis of Variance (ANOVA)
A.5 Appendix: Statistical Tables

Bibliography

Index

Sources

Healthcare Risk Adjustment and Predictive Modeling, 2nd Edition, sample pages (ACTEX Learning) — the book itself: title page, copyright page (ACTEX Learning, 2018, ISBN 978-1-63588-413-5) and the table of contents by section, chapter and section (read from the PDF; sha256 f29a3f9a292b8ef7…)
Healthcare Risk Adjustment & Predictive Modeling (ACTEX Learning) — the publisher's record: author, edition, the ISBNs and the description of the three parts for the lead
GH-301 Syllabus, Nov 2026/Mar 2027/Jul 2027 (SOA) — the citation (2nd Edition, 2018) and the chapters and outcomes assigned