Pricing variations

There are three major types of pricing variations in healthcare.  The first is general product level pricing variations.  Medicaid tends to pay less than Medicare which tends to pay less than large group Commercial.


There is also regional variation.  New York City is more expensive than its suburbs.  North Dakota and other very rural places which have a hard time attracting and holding onto docs are more expensive than medium sized cities.

Finally there is in-region, in-product idiosyncratic pricing variation. A new paper in Health Affairs looks at the pricing variance within markets and between markets for commercial insurers. **

Our study included prices for up to 242 services in each of forty-one states and the District of Columbia. Prices for 162 of these services were reportable in all forty-one states and the District of Columbia. We found that the ratios of average state prices to the average national price for these 162 services varied from a low of 0.79 in Florida to a high of 2.64 in Alaska. Ratios at the twenty-fifth and seventy-fifth percentiles—Oklahoma (0.97) and New Mexico (1.25)—differed by 0.28….

Average prices were computed for 242 services, some of which are standardized collections of common groupings of diagnostic and procedure codes.9 Some services have a single code (for example, Current Procedural Terminology [CPT] code 76811 is for pregnancy ultrasound). Other services encompass an episode of care, such as knee replacement, which includes a specialist’s evaluation, surgery, physical therapy, and follow-up evaluation.

Examining price variation by service provides an understanding of the impact of the variation on patients and insurers. We selected three services—pregnancy ultrasound, knee replacement, and, again, cataract removal—for this examination because they exemplify the range of services and the extent of price variation that exist for common medical services….

Based on the interquartile range ratio, knee replacement prices appear to have the least variation: 1.32, compared to 1.54 for pregnancy ultrasound and 1.47 for cataract removal (Exhibit 3). However, the national average price for knee replacement is more than a hundred times higher than the national average price for pregnancy ultrasound and ten times higher than the price for cataract removal (see the Appendix).11 Thus, even though knee replacement has less variation in price than the other two services do, its variation can have a substantial impact on total expenditures and on patient cost sharing….

Price variation within states was examined though MSA-level prices….We also found considerable variation in the average price for pregnancy ultrasound (Exhibit 6). The average price in Cleveland ($522) was almost three times that in Canton ($183), even though these two Ohio MSAs are only 60 miles apart. Conversely, Virginia Beach ($275) and Richmond ($271), both in Virginia and 107 miles apart, had nearly identical average prices.

I would want to overlay the pricing variations with some type of medical provider market concentration factor.  I would bet that areas within a state that have higher levels of pricing than other areas in the state are also areas where the providers are relatively more concentrated than the payers.  Elective procedures that are deferrable (knee replacements)  and quasi-elective procedures that are fairly low skill and generic (pregnancy ultrasounds) should have variance in pricing due to local general price levels (New York City should be more expensive than Buffalo on this logic) but the wild swings should not be present if the medical services markets were vaguely efficient or functional.

** Newman, D., Parente, S. T., Barrette, E., & Kennedy, K. (2016). Prices For Common Medical Services Vary Substantially Among The Commercially Insured. Health Affairs, 35(5), 923-927. doi:10.1377/hlthaff.2015.1379

It takes a local party

Building on Betty’s post this morning on how we need to build a leftward pendulum swing at all levels of government and society, I want to endorse and fundraise for promising candidate for Oakland County Commissioner in Michigan:

Charles Gaba is the guy behind which is the go-to source for all enrollment information about the Exchanges and a very good clearinghouse for lots of other health wonkery.

He is running for County Commissioner in a suburban county in the Detroit metro area where the Presidential top-lines have the county as a lean Democratic area but the local government has a significant Republican presence.  He is doing something about that and I want to help a colleague and a fellow wonk.  So if you can spare a few bucks, help elect a wonky progressive to a county government board:


Here is Charles’ Act Blue link:

On-exchange profitability

It has been a busy day for me, so this is just a quick note on Exchange profitability:

Also, Centene, the holding company for Ambetter is reporting profitability on the Exchange markets as well:

What does this mean besides more hookers and blow for the C-level?

Profitability outside of California is increasingly plausible. And this makes sense. 2014 was a year where there were only guesses about both the Exchange population, the market structure, and federal policy structure (specifically the risk corridor revenue neutrality restrictions.  2015 had a bit more clarity on who was coming into the market, what was working and what was not working, and what federal policy on risk corridors would actually be.  2016 is the first year where the policies are priced on functionally decent real information and some of the amazingly dumb strategic decisions have been unwound through either course changes or through exiting the market.

As a simple reminder, competetive markets should see some companies make money and some companies that offer more expensive and less attractive products lose money.  I would be extremely worried if everyone was making money after three years, just like I would be extremely worried that everyone was losing money after three years of increasingly better data.

Churn and Medicaid

Arielle Levin Becker reports about churn in Connecticut:

What this means is in the past two months, people who had qualified for Medicaid either through Expansion or through the legacy process are moving to Exchange and seeing their deductibles be reset to zero.

Medicaid is a combination of income determined and condition determined.  People with fairly low incomes that qualify for Medicaid will often experience significant income volalitity.  Someone who is making 130% of Federal Poverty line and thus is qualified for Medicaid Expansion could pick up an extra shift or get a $.50 an hour raise that bumps their income to 140% FPL.  That moves them from Medicaid qualified to Exchange qualified.  Moving to Exchange has three major implications.  The first is the individual is paying out higher premiums and higher cost sharing if they live in a simple Expansion state.  It is a fairly significant marginal income tax rate.  Secondly, the Medicaid and Exchange insurers have almost no incentive to focus on preventive care as it is highly probable that this class of individual will have a fairly short stay on whatever policy they currently have.

Finally, the biggest problem in American health financing is that we pay too much for each unit of service.  Medicaid is usually the lowest level of payments to hospitals and doctors.  Exchange, if the pricing is based on Medicare plus a little bit, will often pay 20% or more for the same service, the same test, the same brand name prescription.  If the Exchange plan is paying based on commercial rates, the Exchange plan could be paying twice as much per unit of service.

What is the policy solution?

The ideal solution would be to allow people to lock in for a year in whatever system that the qualify for during an open enrollment period.  If someone is Medicaid qualified in January, they would be Medicaid qualified for the entire year even if they tripled their income during the course of the year.  There could be a work-around that people would have to pay premiums based on a sliding scale after they leave normal qualification range.  The same would apply to Exchange plans where subsidies would be wrapped around and cost sharing reductions would be applied as people saw their incomes dropped.

Pricing plans on no data

Charles Gaba has a good post on the problems of pricing plans for the 2014 and 2015 Exchange years.

On the surface, it looks as though both the carriers as well as the regulators did a lousy job of predicting the costs for both 2014 and 2015, right? Well, perhaps…except for some important caveats.

For 2014, no one had the slightest clue what the market would look like. The ACA exchanges turned the entire individual market upside down;…

For 2015, however, the carriers should have had a full year of data under their belt so that they could adequately project expected costs over the 2nd year, right?…

The insurance carriers can’t wait until 5 minutes before the next open enrollment period begins to figure out their pricing. They’re legally required to submit their rate requests to regulators as much as 6 months earlier (June or July at the latest), depending on the state….

so instead of 12 months of data, they only had 3. That sucks, but at least it’s something, right?…

only around 2 million people were even enrolled in exchange policies in January 2014. Another 400K or so were enrolled in February, and around 800K more for March. The rest didn’t even start coverage until either April or May, by which point the actuaries for the various carriers were already furiously scrambling to cobble together the data from January, February and March….

As was pointed out to me this morning, not all of the claims are sent in right away. The claims for heart surgery performed in, say, mid-March might not actually be sent to the insurance carrier until May or June, and so on.

Policies that went live on 1/1/16 were the first policies where the insurers actually had decent data to price them correctly.  Policies that go live on 1/1/17 will be the first policies where the actuaries will be projecting off of reliable evidence instead of guessing.

This is one of the reasons why the decision by United Healthcare to stay off Exchange in 2014 and then get on Exchange in 2015 and 2016.  2014 was always a beta test year for the insurers that got on the market in 2014.  The 2015 plans had maybe 10% real data added to their pricing as we had to remember that we knew that people who signed up for 1/1/14 coverage starts were different and much sicker than people who signed up for 4/1/14 coverage starts.  Initial rate proposals were due in the late spring with a revised/updated proposal response to the state regulators by mid-summer when there was a little more claims data.  2016 plans were being priced out on 70% of the preferred data set that had incorporated most of the catch-up care.

UHC did not have that data advantage.  Their 2015 plans were priced on the combination of observation and conjecture and their 2016 plans were priced on 10% to 15% real data and 85% projections.