Regional Consumer Guide Methodology

The consumer guides use a 5-star rating system to assess the quality of care across plans. A rating with more stars indicates a better quality. Consumers can view both the overall quality rating and individual domain ratings for a plan. Domains represent the quality of care for a group of people or condition and often include several measures of care relevant to the group or condition. Plan ratings are determined using a four step process:

  1. Prepare Data for Scoring.
    The measure set in this rating system includes fifty three individual measures and two composites (weight assessment and counseling and diabetes screening tests). The composites are constructed on the measures with the same denominators. The composite of weight assessment and counseling is the average of the numerators in BMI, nutrition and physical activity. The composite for diabetes screening tests is the number of members who had all four screening tests (HbA1c, LDL-C, Eye Exam and Medical Attention for Nephropathy).

  2. Standardize Measure Scores.
    Measure results need to be standardized before the calculation of domain scores. Depending on the method of data collection, different statistics are used to create the standardized measure scores.

    1. For sample-based measures - A standard z-score is used based on statewide variation:
      Standardized Score = (plan rate - statewide rate)/SQRT[(statewide rate * (1 - statewide rate))/plan denominator]
    2. For population-based measures - The Nelson's H statistic is used, which is derived from Analysis of Proportions (ANOP) methodology:
      Standardized Score = (plan rate - statewide rate)/(SQRT[statewide rate * (1 - statewide rate)] * SQRT[(statewide denominator - plan denominator)/(statewide denominator * plan denominator)])
    3. For CAHPS measures in the satisfaction domains - The Student's t-statistic is used :
      Standardized Score = (plan rate - statewide rate) / Standard Error

  3. Calculate Domain Scores and Ratings.
    A domain score is the average of the standardized measure scores within the domain. The t-statistic of the domain scores are calculated and converted to the star ratings by the percentile rank inferred from the Student's t-distribution. The cut-point values for the 5-star ratings are shown in the table below.

  4. Percentiles of t-statistic Ratings
       0<= Score Value < 10   1 star
     10<= Score Value < 30   2 stars
     30<= Score Value < 70   3 stars
     70<= Score Value < 90   4 stars
     90<= Score Value   5 stars
  5. Calculate Overall Rating.
    The overall score is the average of the stars in the domain scores. The t-statistic of the overall score is calculated and converted to the 5-star scale by the percentile rank inferred from the student's t-distribution. The method is the same as used in the domain ratings.