Patient Coded Severity and Payment Penalties Under the Hospital Readmissions Reduction Program: A Machine Learning...

Area of Science:

• Health Services Research
• Health Economics
• Medical Informatics

Background:

• The Centers for Medicare and Medicaid expanded allowable secondary diagnoses in January 2011.
• This policy change aimed to capture patient complexity more accurately.
• Hospitals' responses to this change varied, potentially affecting financial performance.

Purpose of the Study:

• To examine hospital variation in response to the expansion of allowable secondary diagnoses.
• To assess the association between changes in coded severity and financial penalties under the Hospital Readmission Reduction Program (HRRP).

Main Methods:

• Analysis of Medicare administrative claims for 3102 hospitals (July 2008-June 2011).
• Description of changes in comorbidity coding before and after the policy change.
• Use of random forest machine learning regression and a 2-part model to assess associations.

Main Results:

• Significant variation observed in hospital severity coding changes.
• Hospital characteristics like baseline condition categories, case-mix index, and size were associated with coding changes.
• Coding an additional condition category showed a nonsignificant association with increased HRRP penalties.

Conclusions:

• Changes in patient coded severity did not significantly affect readmission penalties under the HRRP.
• The policy expansion's impact on financial penalties appears minimal despite coding variations.