Expanding the market region

The Petersen-Kaiser Family Foundation Health System tracker has a really interesting graphic on the average price paid for full knee replacements in metropolitan statistical areas in 2016 by large group insurers.  There is wild variation:


One of the first things I think about here is the possibility of regional trade. St. Louis is, on average, a little ($1,200) less expensive than Kansas City, Missouri for the same procedure. Cincinnati is, on average, a little ($1,800) cheaper than Columbus Ohio. Dallas is about $6,000 cheaper on average than Fort Worth. New York City is about $16,000 more expensive than the Philadelphia suburbs centered on Montgomery County, Pennsylvania.

It might not be worth sending someone from St. Louis to Kansas City for a knee replacement once travel, hotel and upkeep costs are factored in. It might be worth spending $2,000 to send someone 35 miles from Fort Worth to Dallas to save $6,000. It might be worth spending $10,000 to send someone two hours south in order to cut their knee open in Montgomery County instead of in Manhattan.

There are a couple of major challenges to expanding the geographic market region for health insurance. The first is that most plans are required to have network adequacy of so many surgeons within so many miles. Urban areas will often use a ten mile radius. In large metro areas, there are dozens of surgeons in a plan. For instance, within 5 miles of my former employer’s headquarters in the narrowest Exchange network (UPMC Partners) there are 70 orthopedic surgeons. Many can do a knee replacement if needed.

Benefit design is another challenge. The least expensive metro area in the Peterson/KFF tracker is Louisville at $22,606 as the average amount paid. Most benefit designs will see the patient already maxed out. High Deductible Health Plans (HDHP) will have hit their limit $15,000 ago, and deductible plus co-insurance plans will have very low to no cost sharing on the last dollar of spending. In New York City, patients will have reached their out of pocket maximum under almost any plausible plan design. There is no economic reason for people to decide that going to a hospital two hours away makes sense as they don’t see any gains. If they are already capping out their out of pocket expenses, the incentive then is to go “easy” which usually would mean close by, familiar and expensive.

Some insurers and large companies (Walmart) use a center of excellence approach where they routinely fly people across the country for very high cost procedures with all patient costs covered (including travel, living expenses and deductibles). This is effectively a preferred tier network design and that might be the way to encourage expanding the competition regions for some procedures with large price variation. The deal could be a person pays $4,000 for surgery in New York City or $0 plus nothing for a hotel room for a partner in suburban Philadelphia.

15 replies
  1. 1
    Consuela Bananahammock says:

    Outside my area of expertise but bear with me.

    Is there any way to determine how strong the correlation between procedure price variation and regional cost of living (or labor costs in general) is from a dataset like this? From all the information I’ve seen, I expect the correlation would not be very strong (especially after the Cooper/HCCI paper that shows price variation for the same procedures at the same facilities!). I’m just thinking that cost of living variation ought to have some impact on prices (i.e. it costs more to pay a competitive wage in NYC than in Sheboygan) such that either: (a) the range of price variation attributable to profit-maximization is narrower than reflected by these values or (b) the correlation is weak/non-existent indicating that pricing is generating margins so large (or inconsistent) that the impact of cost of living is undetectable. Or maybe this is all proof that providers are kinda firing blind when it comes to setting their prices…

    This is me bs’ing at economics but maybe you or somebody else at Margolis have ideas or further reading materials on these threads I’m thinking about. Wouldn’t be surprised if there had already been literature generated on it too, though probably in the limited context of a single facility or system.

  2. 2
    MomSense says:

    The other benefit, other than initial cost, is that surgeons who specialize in certain procedures have a much lower failure rate. I’ve known several people who have had to have new hip and knee replacements to repair failed replacements.

  3. 3

    @MomSense: That is definately true, specialization in particular procedures does produce a significant skill gain which leads to fewer clusterfucks and complications. However, knee replacements are a common enough procedure that the depth of skill can be locally developed. It is not like a one-off procedure that is done 22 times a year nationally of which 16 happen on a single floor in a single hospital by a single team.

  4. 4
    Victor Matheson says:

    @Consuela Bananahammock: Ok, this is my wheelhouse. Just eyeballing the data, there is probably some correlation between cost of living and the prices here. However, regional cost of living differences can only explain a very small part of the difference.

    For example, the Bureau of Economic Analysis, the government folks who calculate the GDP, releases what are called Regional Price Parity indices for every state and metropolitan area. OK City and Louisville, the cheapest cities in this data set, both have indices close to 100, which means they are right at the national average for cost of living. That may seem off to you but while Louisville is cheap compared to NYC or Boston, it is expensive compared to most of rural American or smaller cities like Morristown, TN. NYC has an RPP of 132, so it is 32% more expensive than average. Again, that may seem low, but not everywhere in the greater NY/Newark/White Plains metro area is Manhattan.

    Dallas/Fort Worth has an RPP of 108, so 8% more expensive than the country as a whole. But we are getting knee-replacement cost differences of 50% to 150% higher. NYC is 122% more expensive surgery-wise than Louisville but the cost of living is only 32% higher. Fort Worth is 100% more expensive surgery-wise than OK City but the cost of living is only 8% more.

  5. 5
    dr. bloor says:

    @Victor Matheson: Seems like operating costs for medical facilities would be a better standard than consumer prices. I’ll guess that real estate, prevailing wages to employees, utilities, etc.are more than 32% higher than national norms.

  6. 6
    Dan Myers says:

    Does this take into account the extensive rehab in order to have a successful replacement? If you don’t have help at home you need time in a care facillity. After that you need physical therapy. If it’s the right knee you will not be able to drive for an extended period. My wife has had both knees replaced and the outcome was good. A friend who did not have the kind of help I was able to provide did nat have as good an outcome.

  7. 7
    Yutsano says:

    FYI: Dallas and Fort Worth aren’t really 35 miles away from each other. The city cores might be but it doesn’t take but 20 minutes to get from one to the other. A lot of that is because of sprawl but really if there’s that much difference it’s not a hard choice to make.

  8. 8
    Mnemosyne says:

    No West Coast states at all? That seems weird.

  9. 9
    Yarrow says:

    @Yutsano: Dallas and Forth Worth are in different counties. If your insurance plan limits you to doctors in your county (yes, this is a real thing) then you cannot go from one to the other and stay in network.

  10. 10
    Victor Matheson says:

    @dr. bloor: Three comments here. First of all, you go to war with the data you have not the data you want. So, I have metro-area CPI data but not cost data specific to health care providers.

    Second, for most things, consumer prices are a pretty good proxy for what most things will cost. Basic labor? You need to pay them consistent with local consumer prices as reflected in the CPI. For example, RNs in NYC make 41% more than in Louisville and LPNs make 25% more, both broadly consistent with overall CPI being about 32% higher. Land prices? Reflected in CPI through housing costs among other things. Utilities? Reflected in CPI.

    Third, specialized labor may get paid more in certain areas than others, but it is not at all clear whether that is a cost or just that the monopoly-pricing profits in a particular area get shared in part with the docs. You can’t say that $50K for a knee is not overpriced because doctors in that area get paid $800K/year if the reason doctors in the area get $800K is because they can charge $50K for a knee.

  11. 11
    JAFD says:

    Dumb question, but how come Chester, Montgomery and Bucks counties get seperated out from Philadelphia and Delaware ? We’ve got 70 miles of suburbs engrossed, some of which are closer to the Lehigh Valley, some closer to Lancaster, some closer to Wilmington, and most closer to the four medical school/hospital complexes in Philly. The world wonders.

  12. 12
    J R in WV says:


    Yes, my wife and I had our replacements (knees for her and shoulders for me) done by an ortho who typically only does knees, hips and shoulders, twice a week in the OR, the other three days providing aftercare. Practice really makes a difference.

    Wife was in the hospital for two nights, as she needed to walk up a flight of stone steps to get to the house. We have both had great outcomes.

    PT really counts, also. Patients who don’t do their PT hard have many more failures than those who go at it as if it matters. I’ve watched patients on their first day of rehab, and known that they weren’t going to be successful nor happy because they didn’t want to stretch or exercise their new joint because that hurts! Not the medical people’s fault there…

  13. 13
    lahke says:

    Boston suburbs coming in lower than average? Really? And given the cities of Cambridge-Newton-Framingham, there’s at least one Partners hospital in there, which means I’d expect higher-than-average costs. I clicked through to the methodology explanation, but got no more illuminated.

  14. 14
    TriassicSands says:

    Afterthought: one could take a nice vacation to Thailand and come home with a “new” knee and lots of money left over.

  15. 15
    boatboy_srq says:

    Notice who’s third most expensive in that chart? Think about THAT next time someone mentions retiring to FL because “it’s cheaper.”


Leave a Comment

Leave a Reply

Your email address will not be published. Required fields are marked *