Persistence of HMO performance measures.

Area of Science:

• Health Services Research
• Quality Improvement in Healthcare
• Health Plan Performance Analysis

Background:

• Assessing health plan performance is complex due to multiple quality indicators.
• Decision-making processes for employers and consumers rely on accurate, stable performance data.
• Understanding the temporal stability of health plan ratings is crucial for informed choices.

Purpose of the Study:

• To develop and apply a method for evaluating the stability of health plan performance over time.
• To simplify decision-making by providing insights into performance predictability.
• To analyze the persistence of quality ratings across multiple indicators.

Main Methods:

• Utilized longitudinal data from the National Committee for Quality Assurance Health Care Effectiveness Data and Information Set (HEDIS) for childhood immunization (1998-2002).
• Employed a parametric Multiple Indicator Multiple Cause (MIMIC) Model to aggregate multiple performance measures.
• Controlled for plan and market characteristics while accounting for unmeasured heterogeneity.

Main Results:

• Found moderate persistence in health plan performance, with a 0.47 probability of a top-tier plan remaining in that tier two years later.
• Demonstrated that multiple years of good performance enhance future performance probability (63% of top-tier plans in 1999 remained in 2000).
• Indicated that recent performance is a stronger predictor of future performance than distant past performance.

Conclusions:

• Health plan ratings exhibit some persistence but are subject to change, impacting decision-making timelines.
• Employers and consumers should be aware of potential rating fluctuations between data generation and decision implementation.
• Further development of methods is needed to mitigate the consequences of performance rating instability.