Uncategorized

Massachusetts and Hawaii Offer the Most Cost-Effective Health Insurance Coverage

What makes a state’s health insurance successful for its citizens? It should be affordable, it should cover a lot of people, and it should manage its members well, keeping people healthy as measured both by preventive care as well as actual health outcomes.

It turns out that, using those criteria, the state with the highest Health Insurance Success Score (HISS) is Massachusetts. One would expect high quality, good outcomes and of course close to 100% coverage in the Bay State, but it also — quite surprisingly — ranks 5th in affordability, as described below.

Hawaii is a very close second. (One could also argue that Hawaii’s circumstances are unique and non-comparable because that state differentially attracts and retains healthy residents, but the analysis eschewed all subjectivity and second-guessing of the data.) Texas is last, one point behind Arkansas. In both the best and worst listings, there is a noticeable gap between the two states at the extremes and their respective runner-up pelotons.

Out of a potential overall score of 4 to 204 (4 would be a #1 ranking in all 4 categories), the top ten states would be:

The bottom ten states would be:

The 51-state (including DC) ranking may be obtained gratis from the author.

Affordability is measured using the Commonwealth Fund’s recent report on family health plan cost and annual deductibles by state, and is compared to the family income of that state. For instance, health insurance costs about the same in Maine as in Massachusetts…but Massachusetts median family income is about 40% higher. Therefore (and also because the annual deductible is lower in Massachusetts) Massachusetts ranks 5th in affordability while Maine ranks 46th, with health insurance taking a 34.8% bite of a median family’s income. (Massachusetts policies also pay for more services, like infertility treatments, but no attempt was made to capture and control for differences in covered services among states.)

Coverage is tallied using a recent Gallup-Healthways poll, which correlates with other sources.

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.

Outcomes were trickier to measure. Though the criteria were equally weighted to avoid any element of subjectivity (and those who would like to weight the criteria can obtain the analysis from me and apply their own weightings), in my personal opinion, chronic disease outcomes matter a great deal. Health plans expend tremendous resources trying to improve them, as do employers through wellness programs and the states themselves through public health initiatives. Chronic disease outcomes are a combination of avoiding the onset of chronic disease plus keeping people who have chronic disease out of the hospital. The combination of those two endeavors can be summed into the rate of chronic disease inpatient events vs. the state’s population. Of course there was an adjustment for age in the commercially insured population using census data, as people 45-65 are much more likely to have chronic disease events than younger people.

Chronic disease events paid for with private insurance were gleaned from the AHRQ’s Healthcare Cost and Utilization Project (HCUP) database.

Principal diagnosis event rates would be the basket of ICD9s used by DMPC for its comparisons across health plans and employers in the five so-called “common chronic” conditions: asthma, coronary artery disease, congestive heart failure, chronic obstructive pulmonary disease and diabetes. (15 states still don’t report to HCUP. Their event rates were assumed to be average for their age distribution.)

HCUP provides total numbers of events. That figure is then divided by the Kaiser Family Foundation’s estimate of commercially insured people by state, in order to produce an event rate.

As mentioned, rankings were equally weighted to avoid subjectivity, but another stealth subjective factor is the choice of variables to rank. For instance, the current rankings reveal that, while high per capita income correlates loosely with top performance, low income correlates tightly with poor performance. And health status does correlate with people’s wealth, above and beyond anything that a state’s health insurers or medical system might do. If the adverse event rates are considered to be a function of both age and income instead of age alone and therefore to score high a state must “outperform” the expectation created by its own wealth, the rankings change as follows, with some movement among and in and out of the best ten but mostly compression of the top scores:

The lowest-ranked states are still the lowest-ranked states, even accounting for the more limited financial resources of their populations:

This HISS analysis is fully replicable. All the variables are completely objective and available to everyone using the links. The analysis itself – which is free for journalists, academic researchers, bloggers, federal employees and retainer-level members of the Disease Management Purchasing Consortium — may be obtained from the author. Rather than defend this analysis as definitive (and as many limitations as I could think of are listed on my website), I instead encourage people to improve upon this analysis and will be of assistance to anyone who would like to refine it.

Al Lewis, called “the country’s leading outcomes measurement guru” in the 9th Annual Report on the Disease Management and Wellness Industries, is president of the Disease Management Purchasing Consortium, www.dismgmt.com. He was founder and first president of the Care Continuum Alliance. An excerpt from his forthcoming book Why Nobody Believes the Numbers: The Outcomes Measurement Guide for Grown-Ups may be downloaded here. You can email Lewis for more info at diseasmgmt@aol.com.

28 replies »

  1. Fibromyalgia is a common syndrome in which a person has long-term, body-wide pain and tenderness in the joints, muscles, tendons, and other soft tissues.Fibromyalgia has also been linked to fatigue, sleep problems, headaches, depression, and anxiety.’

    Most current post on our personal web site
    <.http://www.foodsupplementcenter.com/l-lysine-for-cats/

  2. Let’s see, Cape Cod or Honolulu? Hard to decide where to move to. Maybe my wife will finally agree to move when I show her we’ll save a ton on insurance.

  3. http://www.dol.gov/ebsa/pdf/ACASelfFundedHealthPlansReport032811.pdf

    Dafny, Ho, and Varela (2010) estimated a hedonic pricing model, using the LEHID, to conclude that employees prefered self-insured plans over fully-insured plans. The authors found the self-insurance preference to be above and beyond the appeal of lower premium payments (which are controlled for in the model). Given that self-insurance allows an employer to choose not to offer state-mandated benefits, this result suggests that employees valued the other attributes of self-insured plans more highly than they valued the state-mandated benefits that would be available under a fully-insured plan.

  4. “That may in fact be because in their desire to attract businesses that could locate anywhere, some poorer states overlook the hidden costs created by low-wage businesses.”

    What would these be? Retail and fast food is pretty consistant across all states. Your remaining low wage jobs like Timber, Coal, Construction, etc are located where the trees, Coal, building need is. What location flexible business do they target? Auto, its comming from Blue states and is still more dominate in Blue States. Your not going to grow cotton in Mass no matter how much you pay per hour. Some other industries like trucking they choose to live in Red States because the cost of living is so much lower.

    The study, from what I have seen ranks Red States lower becuase the measures are bias in favor of blue states. A few slight changes to the way you score your study and things flip. Is a loaf of bread selling for $2.00 in MA more affordable then the same load of bread selling for $1.50 in TX? Is an HMO operating the exact same way better in CA then OK based upon a poll? For the same reason CA is Blue and OK Red people have different opinions, polls aren’t a very accurate way of measuring services across disparet demographics.

    Whats better writing out checks and mailing them or sending ACH? If you ask old poeple like my dad checks, younger people would say ACH. In CA Kaiser is very popular. In OH they can’t get any traction. TX is at the forefront of self funding, next to that no carrier plan is going to rank very high. Red States tend to be more favorable to self funding while Blue States go so far as to try and forbid it. If all you ever know is carrier plans then they wont seem that bad. Not till your covered by a self funded plan do you see how different it can be.

  5. Whoa, guys…

    You are both raising points that can be argued, and that’s precisely why I am offering all the backup to folks to do further analysis.

    I personally like the idea of adjusting for some demographic factors, but I don’t know that one should “adjust” for diet and smoking rates etc. since if you ask most insurors and some employers, improvements in those factors are a focus of their DM and wellness efforts. And smoking rates are driven in large part by taxes on cigarettes, which tend to be lower in the South and border states.

    So unlike demographics that factor would be a controllable one — an outcome rather than something that gets adjusted for.

    One place I might have been a bit sloppy is calling this a Health Insurance Success Score. It is a combination of that plus health policy at the state level. So it’s not just the insurors that drive the score. It’s the climate created for them by the states. Someone pointed out that “red states” dominate the low-ranked list. That may in fact be because in their desire to attract businesses that could locate anywhere, some poorer states overlook the hidden costs created by low-wage businesses.

  6. “diet, genetic disease, poverty are not symptoms or curable by a healthcare system. If your trying to compare the effectivness of a system you must measure from the same base or an adjusted base that equalizes the populations.”

    No, but they’re all treatable by an affordable, accessible health care system, and the south’s failures are all inter-connected. You can’t judge a health care system in a vacuum of other social factors. The south continues to tell it’s citizens that unions are evil but low wages , poor access to health care, and poverty are the just alternatives.

    “are junk science”

    Sure Nate, all your prejudices are explained by that phrase.

  7. Which is why all these studies getting press about american life expectancy are junk science written by quack academics who should be dragged out of their ivory tower and have some common sense beaten into them. Any system would look good if it only cared for Japanese or higher income white europeans. Take those exact same systems and plan them in Africa or South America they wont accomplish anything.Ugonda with the NHS isn’t going to see their life expectancy jumo 50%.

    You must control for what is a result of the system and what is the result out non measured outside forces.

  8. diet, genetic disease, poverty are not symptoms or curable by a healthcare system. If your trying to compare the effectivness of a system you must measure from the same base or an adjusted base that equalizes the populations. Otherwise a poor system starting with better inputs would appear better then a great system starting with a terrible input.

  9. So what are we trying to adjust for? If the south is poorer, their diet worse, their education less, their environment less safe, their health plan cost to income worse, their rates of chronic disease higher, and their ethic and cultural mix skewed – let’s factor out the negatives so that we can cover up the facts and say it’s a great place to live if health care were not a factor?

  10. ethnic and racial make up and those trends should all be adjusted for. A state with a large population of recent immigrants would have a much different health make up then HI which has a much moe stable population. Like the terrible life expectancy studies going around again factors that can’t be controlled by the health system should not be counted against the health system.

    allowance for plan type, while some states have 20%+ HMO pentration others have next to none. Its seems a lot of your measures would tend to favor HMOs or more managed plans

    In regards to affordability I have seen recent studies that adjust for cost of living, if you make 15% more in State A then B but the cost of living is 30% higher a $100 premium would be less affordable in the state with the higher income

  11. It sounds like I should do an adjustment for a state’s ethnic and racial makeup in the next round? As mentioned above I would do another round when the new HCUP data comes in. I could add this adjustment as well.

    Expect something in January -February. Expect Alabama to shine. It was the best-performing state in its region in this analysis, so any adjustment that elevates southern states in general will elevate Alabama the most.

    Blue Cross is probably a more dominant carrier in Alabama than in almost any other state, and BCBSAL has won many awards for its excellent and improving disease management outcomes.

  12. The south has some core differences that will skew the results. I have always heard that the south has more medical problems than the north. Some of this is due to diet but also due to the environment. Also, the south has a higher black and Hispanic population that is both underinsured and has some differences in terms of health care, lifestyle, treatments and results, Shouldn’t the analysis be adjusted for regional bias since it isn’t the system that causes or prevents this. I don’t know how to normalize the data for diet and environment and cultural norms. I don’t think income is a suitable adjustment.

  13. Cedric and Steve, I can’t access you via the links in your comments, which are not to email addresses. (Your emails are not published.) Bill, you didn’t leave an email. If you are interested just contact me directly at alewis@dismgmt.com

    Nate, I would totally agree that my state-average HEDIS quality ratings are a blunt instrument to measure quality. As mentioned in the article, you are welcome to the HISS analysis if you would like to refine it further. (There was, however, enough clustering of quality ratings by state that I doubt adding self-insured non-NCQA plans would change the relative ratings much, if there were a way to add them.)

    As I mentioned in the article using the infertility example, I also didn’t attempt to adjust for coverage mandates but my strong suspicion is that they correlate with state average income as richer states can afford more mandates. if there were some way of removing those mandates to compare the affordability of bare-bones programs, probably the difference in affordability by state would be magnified since richer states “spend” some of their money on mandates. (In other words, Massachusetts has more mandates than Mississippi.)

    PS I am guessing you don’t live in Massachusetts or Hawaii 🙂

  14. How does the quality measure adjust for the fact that 30-40% of the insured population is not in a plan rated by NCQA? Your inserting a bias against States that have a higher population of self funded members when in fact self funded plans perform considerably better on average then the NCQA plans.

    I also don’t see how you can measure affordability on the state level. A large portion of insured premium is determined by state insurance laws, coverage mandates etc. The income in Boston or the rich areas is many magnatude higher then the poor areas compared to the difference in the cost of insurance.

    The quality of care in Boston also shares little resembalnce to many other parts of the state. In a state like NV the centralized population would make any assumptions on the remainder of the state worthless.

    I also see a number of healthplans that don’t cover the entire state or have so little membership in parts of the state equating their value to the entire state would be inaccurate. If your in Philadelphia what does it matter to you how great UPMC is?
    The top two NCQA insurers in Ohio serve roughly they same metro area which isn’t even in the top 5 largest of the state. You can’t even buy their plans in over 75% of the state. Good plans and people love them but they say nothing about healthcare at the state level.

  15. My apologies! I was wondering why download requests were no longer coming in. I will take care of this for you as soon as I can find the kid who takes care of this for me (usually quite excellent, by the way, in case anyone is looking for someone who knows how to do this stuff, but is having a family emergency right now). Meantime hit “contact us” from the website and I’ll fulfill tonight or tomorrow morning. And I’ll take care of you three as well.

    By the way, I tried doing the thing where I “controlled” more generally for income as suggested above. Great idea but the problem is that the states that come out best and worst are disproportionately among the 15 for whom I have no outcomes-specific data (a study limitation noted on my site). I didn’t consider the lack of that particular data point to be an issue when those states landed in the middle of the pack using my standard analysis but to anoint a state as “best” or “worst” I’d like to have all the data first.

    I think HCUP will be out with its 2010 outcomes data within a month or two,and will be adding some more states to their scrum, and I promise to post at that point as requested. If some states are still missing outcomes data at that point I simply won’t include them in the analysis.

  16. Ah, the link is broken…and I thought it was just me. Would love a copy. Bring it on!

    And I promise not to tell the other candidates. This could kill Romney. Republicans don’t mind when you take an Obama quote completely out of context to imply he said the exact opposite of what he said, but they hate it when a government program is successful.

  17. Well if we ever get to purchase health insurance across state lines, I now know where to shop!

  18. With responses like Mr. Drug’s (and the others as well), how can I not follow up on these suggestions? I like the concept of controlling the entire study (not just outcomes) for family income, Because right now we see that as in sports, the “big market teams” with greater resources generally outperform th “Small market teams” that are poorer.

    I’ll try to adjust for that and re-post.

  19. If you do as George says, I would guess (based on the current HISS list) that your top-ranked states would include Vermont, Iowa and also Alabama (the only low-income state absent from the low-HISS score states) .

    I would opine that any analysis that has those three states in the top ten of healthcare or anything else overcomes the usual conundrum where — no matter what you are ranking — the rich (usually Blue) states do better than the poor (usually Red) states.

    I would really look forward to a refinement of this analysis that takes wealth out of the equation.

  20. I would like to see which states do best with the resources they have. Even with the adjustment for income it seems like the rich states do better. Good start and please dont stop here.

  21. Oh, and did anyone else notice another correlation? These are blue states and the poor value states are mostly red states. Some track record of success.

  22. Mary, not surprisingly the top 10 states for health care value are well represented among the top 10 states for savings. http://en.wikipedia.org/wiki/List_of_U.S._states_by_savings_rate

    In fact, the overlap is so strong it reinforces that high health care costs are eroding our retirement savings. Maybe people in Texas “want to” pay more for health care in order to have less to spend in their retirement, but I kind of doubt it.

  23. Interesting way to look at cost, comparing it to income. People here in Mass spend more money on a lot of things because we have higher incomes and we want to. Clearly we want more healthcare too because we’ve elected to and we are happy with the result.

    A true test of affordability would compare savings rates to healthcare spending by state. That would determine whether people are spending more because they want to or because they have to. Barring that this is the next best thing. How come no one has noticed this until now?