• All-payer rate setting in JAMA: a “Maryland miracle?”

    Another article is praising Maryland’s all-payer hospital price control system, this time in JAMA. Here’s the key claim:

    In 1976, when the system began full operation, the adjusted costs per admission in Maryland hospitals were approximately 26% higher than the national average.​ Between 1977 and 2009, Maryland hospitals experienced the lowest cumulative increase in cost per adjusted admission of any state in the nation. For fiscal year 2009, the average cost per admission at Maryland hospitals increased by only 2% compared with an estimated 4.5% increase for the rest of the nation.

    We’ve talked about this extensively here at TIE (all-payer tag), including a recent podcast. I’m keeping an open mind on this topic as the evidence develops, but this JAMA article repeats a weakness in some prior attempts to promote the Maryland story:  success is being measured in unit prices (cost per case) rather than overall prices:

    That’s a big “Per Case” caveat. How does Maryland compare on Medicare hospital discharge volume? Glad you asked:

    Or on hospital reimbursements per Medicare enrollee?

    (Medicare only; from Dartmouth Atlas online query)

    In Robert Murray’s 2009 Health Affairs article, he clearly describes the volume problem and how Maryland has tried to respond:

    However, although costs per admission were well controlled, the same cannot be said for hospital admissions and overall hospital volume. Rate regulation was not structured to have oversight over individual physicians’ decision making, and there is no legislation currently that allows the HSCRC to establish regional hospital spending limits, which would be needed to curtail case volume increases.  There was a limited break on admission growth over the period 1978–2001, when changes in the volume of hospital admissions triggered the application of fixed/variable cost adjustments to payment rates. This adjustment was eliminated in 2000 as part of a rate negotiation with the hospital industry (the expectation that managed care would control volume growth prompted the HSCRC to remove volume adjustment in exchange for a lower update formula for 2001–03). Immediately, admission rates began to increase, quickly outpacing national rates. During the period 2001–07, admissions grew at an annual average rate of 2.7 percent in Maryland versus an average annual rate of 1 percent nationally (Exhibit 2).

    I’d like to hear more about tools to control overall costs and hospital volume before we all jump on the all-payer bandwagon.

    • Kevin
      To extend your point further, from KFF:

      Hospital costs are one piece of the puzzle. Maryland searches for the cost control golden chalice like all the rest. They may be “winning” in one domain, but are lagging in others.


    • Many years ago, I was the CEO of a hospital that participated in the all payer, globally budgeted revenue cap for nine hospitals known as the Hospital Experimental Payment program (HEP) in Rochester, New York. The design of HEP included incentives to reduce unnecessary admission and days of inpatient care. During the first phase of HEP 1980-1984, the hospitals reduced admissions/1000 by a little more than 8%. Inappropriate admissions were substantially below the national average. According to the GAO, Rochester’s community approach resulted in premiums that were 1/3 less than the national average.

      Sarah Liebschutz’s June 2011 book “Communities and Health Care: The Rochester, New York Experiment” notes that the Rochester business community was a major driver of the HEP experiment which was something like market pressure that CEOs of publicly traded companies experience. I wonder if the Maryland system exposes hospitals to that kind of pressure.

    • “I’d like to hear more about tools to control overall costs and hospital volume before we all jump on the all-payer bandwagon.”

      Ditto. I’ll even set the bar lower. Show me an all-payer price control regime anywhere at any time that has ever worked in any sector of the economy.

      I’ll give a consolation prize to any respondent who can correctly identify the difference between the price and the cost of a given good or service, and identify which of the two can be controlled by bureaucratic fiat and which cannot.

      • The law of one price is fairly common, isn’t it? Do you pay a different price than I do at your nearby grocery store?

        Maryland’s system is only one variant of all-payer. There are others that harness market signals and are not based on administrative pricing. I’ve discussed them on this blog. Search “all payer” and you should find the posts.

        • That comparison isn’t entirely fair, the law of one price doesn’t preclude bulk discounts. A provider could argue that they do charge everyone the same price, but they give insurers varying discounts based on the volume of patients they deliver.

        • 1)There’s considerably heterogeneity in pricing, even in consumer staples. You can actually pay quite a bit more per unit of weight for the same brand of cheese at one grocery store than you do at another. Expand the suite of comparisons from wholesale clubs to convenience stores and the differences in the prices get even larger. The clear implication here is that retailers compete on a variety of margins – only one of which is price.

          2)There are extremely important differences between the process that generate prices that are forced into a narrow band by competitive pressures, and prices that are forced into an equally narrow band by bureaucratic fiat.

        • Actually, Maryland’s system is far more along the lines of a quasi-negotiated German all-payer system than you give it credit for. Every year, the HSCRC convenes major payers, labor and hospitals and negotiates out the annual update to rates. The Commission also acts as a forum for cooperation and negotiation around virtually every major payment policy change in the system.

          The Maryland system also makes heavy use of more functional market based dynamics – such as making information on relative cost and quality performance of hospitals available to purchasers of care. One contemplated use of the Commission’s extensive data is the development of tiered rankings of hospitals (perhaps for use by major insurers, the health insurance exchange, the State employee benefit program and medicaid in benefit design. The Commission also has been working on ranking hospitals by product line using risk adjusted and quality adjusted cost metrics (rank cardiology, orthopedics, maternity services on this tiered basis). In addition, the Commission was working toward the development of “reference prices” for clean procedures, CABGs, Stents, Knees, Hips, etc. showing hospital/physician virtual bundles of care to help guide both primary care physicians under more risk-based payment structures (such as primary care advanced medical home models), insurers and patients.

          Finally, having established rational unit prices and per case/per visit prices and having gotten a handle on the overall level of pricing, the Commission has now moved on to address the “volume” problem by moving hospitals into broader episode-based payment structures. In 2010, the Commission negotiated with 10 rural hospitals (accounting for about 10% of system revenue) to place them under 100% fixed payment structures (enforceable global budgets regardless of volume). These systems can work because these hospitals have isolated catchment areas, they are conditioned to provide services to their associated (and captive) communities and they have a good understanding of their costs (because the Commission’s system forces all hospitals to account for the cost of all care in a very detailed fashion). The caps established for these hospitals are thus reasonable and reflect the underlying risk profile and cost profile of the patients served. These are arguably the most powerful payment structures to address the excess admissions, re-admissions and outpatient hospital care that was an unintended consequence of controlling unit prices (per case, per visit and unit prices for ancillaries).

          The Commission is also working on a version of these global budgets that can be applied to suburban hospitals and ultimately urban facilities as well.

          In 2011 (as of late September) an additional 31 hospitals in the state (including Johns Hopkins and University of Maryland and their affiliate hospitals) have entered into agreements with the Commission to go at risk for admissions (per case constraint) AND all all-cause 30 re-admissions. The program is structured much like the Geisinger warranty program – except it covers all DRGs and all re-admissions (not just related re-admissions).

          These three-four models of expanded payment episodes along with other Commission policy changes are having some impact on overall expenditures it appears (although it is difficult to sort out the declines in volumes experienced in the past 2 years from other macro effects of a stagnant economy). But year to year expenditures on hospital care was in excess of 10% in 2007. As of the end of calendar year 2010 they were approximately 2.2%.

          Capitation of 10 facilities and expanded episode based payment for an additional 31 large hospitals will certainly start to address Maryland’s volume issue.

          The CMMI bundled payment initiative along with Blue Cross of Maryland’s advance Medical Home model for PCPs now give Maryland the opportunity to coordinate the favorable HSCRC cost and quality initiatives in the hospital sector with similar developments in the physician sector.

          The regulatory structure and governance in the State is flexible enough to respond to the development of problems and unintended consequences. As long as there is sufficient social/economic imperative to overcome political interests.

          I believe Maryland now has the benefits of establishing rational and equitable prices (that reflect cost – what a concept in health care!) and also payment structures (beyond the 4 walls of the hospital) geared toward improving the overall value received for dollars expended on health services and treating populations of patients.

          Would be happy to provide more detail should you wish to contact me.

          Robert Murray

    • I got to this article via a link on Mark Thomas’s “Economist’s View”. This blog is a credit to Robert Woods Johnson and I hope more laymen like me find it – I’ll do what I can.

    • I agree the JAMA article was not as descriptive as it could have been. It seemed to be a re-write of previous pieces on the Maryland system.

      Per your comment: I’d like to hear more about tools to control overall costs and hospital volume before we all jump on the all-payer bandwagon.

      See my response to Austin Frakt above.

      Maryland recognized the volume behavioral response and is now moving most all hospitals into episode-based payment structures. The system is also coordinating with other programs CMMI bundled payment project and Maryland Blue Cross PCHM program to help integrate and align these beneficial hospital payment structures with the physician community in the State.

      Would enjoy sharing more detail about these reforms.

      Robert Murray


    • Robert.,

      Thanks for the lengthy and generous response. As you have conveyed so much, perhaps I can persuade you to continue with a bit more detail. Its very helpful:

      This paragraph caught my eye:

      “In 2011 (as of late September) an additional 31 hospitals in the state (including Johns Hopkins and University of Maryland and their affiliate hospitals) have entered into agreements with the Commission to go at risk for admissions (per case constraint) AND all all-cause 30 re-admissions. The program is structured much like the Geisinger warranty program – except it covers all DRGs and all re-admissions (not just related re-admissions).”

      Can you elaborate:
      –planned readmits for same DRG, or unplanned for new DRG <30 days post d/c?

      –SES factors ("unmeasured") and MSA variations—how to account?

      –defining outliers, ability to make hospitals whole for bonafide hits on their bottom line?

      Additionally, quality adjusted rankings are crude, to say the least, given our limitations with current data sets and their inherent narrowness. The measures either don’t exist, or for those that do, lack robustness and have limited utility, ie, what exists for CHF, etc.

      In addition, with reference pricing, you cited ortho, cards, ob service lines—predictable utilization and costing structures, and the challenge is grappling with medical diagnoses. The science of accounting for those DRGs, and the chronic illnesses and the cost driving challenges embedded within them are not eminently solvable in the 3-5 year time horizon.

      Maryland is ahead of the game, but would you agree that the science of adjustment and what is in the armamentarium to "measure" and rank/pay is still rudimentary at best?

      I see this as a valid obstacle in moving forward. Hospitals may gripe, but in this instance, some of the grievances might be legitimate. After all, they want an apple to apples comparison when it is time to settle their tab. The system is not their yet to meet that expectoration.


    • yipes.

      last word. ? expectoration.

      expectation 🙂

    • McClatchy has this article on provider CEO pay. Does the Maryland system depend on competition to reduce administrative costs? Does Germany’s?

      My brother feels the Massachusetts providers are gaming their system.