Scores are tabulated using the (unweighted) rankings for affordability, quality, coverage, and outcomes.

Affordability tallies the cost of family health coverage to an employer and employee combined, in total annual policy cost plus annual deductible. (Co-pays are not included due to lack of data.)

Limitations:

1. Some states have mandates for greater coverage than others. For instance, Massachusetts covers infertility treatments. This is not controlled for.
2. The annual deductible is for individual policies. (Family deductibles were not available.) This is presumably correlated with family deductibles, though.

Quality was derived by assigning the HEDIS health plan ranks to the states in which the health plans are listed as doing business. For instance, Harvard-Pilgrim Health Care, the country’s highest-rated plan, does business in Massachusetts and Maine, so each of those states would score a “1”for that plan and then lower scores for the other plans doing business in those states. Using the example of Hawaii. NCQA ranks four health plans doing business there. Those plans are ranked 37, 80, 104, and 286. The average of those four rankings is 127, good enough for second place behind Massachusetts, where the average plan licensed to sell health insurance is ranked 57.

Limitations: These are average rankings, not weight-average rankings. So the rank of a health plan with 20,000 members counts the same in the average as the rank of a health plan with 200,000 members that also does business in that state.

Coverage is the percentage of people who report having health insurance.

Limitation: This is self-reported data.

Outcomes, more formally Adverse Event Avoidance vs. Age-Adjusted Expectation, is the ranking of each state, in the commercially insured population, in avoiding inpatient events in the categories covered by DMPC: asthma, coronary artery disease, congestive heart failure, chronic obstructive pulmonary disease, and diabetes. “Avoidance” is defined as the inverse of event incidence in those categories.

The ranking of these events was compared to the percentage of the state’s population 45-64.9 as most events happen in that age group. The age group rank is the proxy for the expected rank of these events if insurers and providers (and other variables) were average.

Limitations:

1. “Commercially insured population” figures likewise also vary by source, though not much
2. Event rates are for 2009 as of this writing. 2010 not yet available.
3. Others would differ in their selection of a pool of the most likely avoidable events. Readers are encouraged to substitute their own ICD9s for the ones used in this analysis
4. For the 15 states with no HCUP data, the event rates were assumed to correlate identically with what would be predicted by age alone
5. The ranking methodology magnifies very small differences in performance. This would not affect the states at either end but might jiggle some rankings in the middle
6. It could be easily argued that per capita income should also be included as a separate variable since avoidance of adverse events correlates to some degree with income. We encourage others to take this analysis forward using that variable and others we haven’t thought of

The analysis is provided on two spreadsheets, making it possible to substitute your own assumptions or databases for one or more of those used and add weightings. All databases are sourced and hyperlinked for your convenience.

Order Now

Please use the PayPal button below to order the 2011 DMPC Health Insurance Success Scores (HISS).