• The opposite of cost shifting

    From “Hospital volume responses to Medicare’s Outpatient Prospective Payment System: Evidence from Florida,” by Daifeng He and Jennifer Mellor (JHE, 2012):

    On average, hospitals that experienced rate cuts under OPPS [Medicare’s Outpatient Prospective Payment System] either decreased or saw no change in Medicare volume but increased private FFS volume. Moreover, when we account for heterogeneous responses based on a hospital’s exposure to OPPS (measured by Medicare share at baseline), we find that highly exposed hospitals responded to payment cuts with larger increases in private FFS volume than did less exposed hospitals. In terms of Medicare volume responses to rate cuts, the highly exposed hospitals responded with smaller decreases, and in some cases increases, compared to less exposed hospitals. These results are consistent with highly exposed hospitals experiencing larger income effects in light of payment cuts, and we view these results as consistent with the theory of provider induced demand.

    The volume shift from Medicare to privately insured patients is also the opposite of what one would expect under cost shifting. Or, to put it another way, it is what one would expect if hospitals are profit maximizing. The way to attract more privately insured patients (i.e., to increase volume of them) is to offer lower prices, not higher ones as the cost shifting theory predicts.

    The profit-maximizing volume shift theory has been covered on TIE before. More on cost shifting of various types in the FAQ. If you’re in a hurry, you can’t go too far wrong in assuming cost shifting doesn’t happen. That might not always and everywhere be quite right, but it is a good, first-order approximation and, in the long-run, has to be true. (If you disagree, read the FAQ first.)

    An ungated, working paper version of the paper quoted above is here.


    • Austin
      Your prior review on cost shifting, as well as other papers concern mostly shifts in inpt volume/payment. Due to bed limitations (occupancy) and substitutions, as opposed to additions–like in outpt setting, one will see a difft dynamic. Hospitals also have more at risk on inpt side–they generate more revenue off of wards.

      Can you draw same conclusions re: how institutions will profit maximize and engage in cost shifting when you consider above?

      • Sounds far more like cost cutting than cost shifting. How is it different than the dynamics of reduced hospital stay lengths and increased outpatient utilization after Medicare inpatient PPS?

    • You can expand outpatient setting–add more FFS for lower rates, make it up on volume…

      With beds, payer mix less flexible and requires more planning. Capacity limits strategy.

      Firms profit maximize, but given realities of hospitals, if one facility draws 90% vs 70% from inpatient as opposed to outpt revenue, could behavior patterns differ? The prices they seek, and the costs they shift might vary.

    • Austin,

      I read this article differently. In the conclusion they state:

      ” Our results suggest that Medicare payment reforms that retain a FFS approach are unlikely to be effective in containing Medicare costs, because such reforms could create the incentive for certain hospitals to induce demand among Medicare patients. Even more worrisome, reductions in fee-for-service payments may even create incentives for hospitals to induce demand among non-Medicare patients. By thus affecting private payers as well, Medicare FFS-type payment reforms are further limited in their ability to reduce societal costs.”

      That looks to me like an explanation of a mechanism for cost shifting through provider induced demand, not a refutation of the concept. I’m interested in hearing your response.

      • The definition of “cost shifting” I’ve always used is an increase in private price caused by a decrease in public price. Volume shifting (by induced demand or otherwise) is NOT cost shifting. It is, in fact, the opposite according to standard economic theory. For, how do you raise volume? You cut price! My links will take you to the theory.

        • I have always understood the core of the cost shifting argument in the public policy debates to be that when public prices decrease (or uncompensated care increases) “costs” are shifted to private payers. The argument does not depend on whether or not the channel for this shift comes through “prices” or some other mechanism.

          If hospitals are in fact inducing demand by ordering unnecessary procedures rather than cutting price, than the result of lower Medicare reimbursement rates would be higher costs for non-Medicare patients.The paper does not provide evidence on how that increase in demand is induced. That is worth further research.

          • I think you have a mistaken impression of the literature, or the economics literature at the very least. Despite the term, it’s about prices. (The term is so abused outside economics that I can see how one might get the wrong impression.) I refer you to my review article and posts on this blog (follow the links). But if my authority isn’t good enough, I’ll quote Michael Morrisey’s 1994 book on the topic:

            Cost shifting occurs when providers raise prices to one group of payers because another group of payers is now paying less.

            For all that, I think we can agree that spending (not prices, but prices x volumes, or “costs” in your terms) can shift even if prices do not. So, we’re not disagreeing at all on the substance, just the terminology.

            • That’s helpful. Thanks.

              I find your argument about prices to be persuasive.

              If spending does in fact shift, even if prices do not, it still has important policy implications, though the solutions may be different (e.g., payment reform).

    • -(1)Increasing operational efficiencies and/or reducing compensation levels to the degree necessary to lower prices and maintain current margins at the hospital level is *very* difficult.

      -(2)The capacity to raise prices is constrained by fee schedules negotiated with insurers, and only overcome via increasing market power achieved by local provider consolidation.

      -Since increasing operational/cost efficiency and raising prices are hard, what you tend to see in practice in cases of diminishing compensation from public payers is a mix of provider consolidation coupled with an attempt to maximize privately insured patients and minimize the number of publicly insured patients that make it through the door, rather than increasing prices on the privately insured.

      Holding publicly insured patients constant or reducing them on a percentage basis while increasing the percentage of privately insured patients allows you to increase your net margins without increasing your prices. You do so by competing on different margins – service mix, location, amenities, convenience, prestige, etc. Basically locate in affluent areas with lots of private-covereds and start adding capacity in the things that they value.

      -For a case study in payor-mix optimization gone wrong, take a look at the Swedish hospital group in Western Washington. They built out a high-end hospital in an affluent suburb East of Seattle and took a massive financial hit when the payor mix that actually materialized was way too heavy on Medicare patients to finance the cost of building and operating the hospital.

      “We have received many questions regarding our new Issaquah hospital, especially how it fits into our future and how we paid for it…In the ever changing healthcare marketplace, we need to ensure that we have “diversification.” This means we need geographic diversification, patient and payor diversification and service/specialty diversification. Issaquah and the ACCs in Redmond and Mill Creek are helping us diversify into suburbs where there are typically younger patients who have commercial insurance, which pays 40% to 60% more than government insurance (Medicare and Medicaid). At First Hill, Cherry Hill and Ballard, the payor mix is much less favorable than in the suburbs because patients tend to be older and rely on government insurance at our in-city locations.”

      After the payor-mix gambit failed, Swedish’s financial difficulties forced them into the arms of Providence.

      -Are you basing your claims about prices and cost shifting on standard neoclassical assumptions about perfect competition, etc, or another model where non-price factors influence where patients seek care, price sensitivity, etc?

      • Follow the links and you’ll find much of the same as you wrote, as well as all that underlies my conclusions.

        • -All of the models that you cover in your review of the cost shifting literature assume a greater capacity for providers to vary prices, and a greater price sensitivity amongst consumers than seems to exist amongst real providers and patients/payors.

          The statement from your earlier treatment of the issue “The private market will buy more only if the price is reduced” doesn’t seem to adequately address the role that price constraints, provider market power, and the vast litany of non-price factors that determine where patients seek care. Geographic proximity, convenience, prestige, peer-referrals, service mix, etc, etc, etc, etc.

          -In the review you linked you seem to at least provisionally accept the approximate figure of 20 cents in costs shifted to private payors (Wu). Yet above you write: “If you’re in a hurry, you can’t go too far wrong in assuming cost shifting doesn’t happen. That might not always and everywhere be quite right, but it is a good, first-order approximation and, in the long-run, has to be true.” 20 cents on the dollar is a long way from zero (as is 10 or 5 cents), and as the relative percentage of publicly insured patients increases, the total magnitude of this shift becomes more significant and problematic.

          -Finally – the reality is that for most hospitals and providers, profits from treating privately insured patients are necessary to offset losses from treating publicly insured patients. Whatever value distinctions between “cost-shifting ” and “volume shifting” have elsewhere, terminological distinctions of this kind do nothing to obscure or attenuate this reality on the front lines.

          • I don’t recognize most of what you wrote. I covered the literature thoroughly. Not a single paper was missed.

            20 is far closer to 0 than to 100, which is what some in the industry assume. Look, my writing on this subject speaks for itself. No need for you to find it adequate, but I’m not going to write more on demand from you.

            • Very well. The operational realities that real hospitals are contending with will present innumerable opportunities to revisit these issue, and soon.

            • I encourage those with other views than are presented in the literature to write them out in long form and seek publication. And, while awaiting it, send the document to me. I’m interested in all aspects of cost shifting.

              I also failed to mention that I view the 20% shift rate as a maximum. I’ve seen both published and more recent work (not yet published) that suggests it is much lower. Really, in a pinch, zero is not such a bad approximation.

    • “Swedish Medical Center is losing $250,000 a day and must take immediate steps to cut costs — including measures that may require layoffs, the hospital system’s Chief Executive Kevin Brown says in a letter to the Swedish staff released Monday…

      A shift in the “payor mix” is adding to the medical systems budget woes, Brown said. “We are seeing a reduction in commercially insured patients and an increase in uninsured or those covered by either Medicare or Medicaid, which do not adequately cover the cost of providing care. “

    • *Any* non-zero value for cost shifting has practical significance for providers, since the total magnitude of the costs that they must recoup elsewhere or reduce depends on the total magnitude of these costs. This magnitude is determined by the total volume of publicly insured patients they see, not the percentage of the cost-shift for each patient.

      The formal distinction between the cost-shifting by price adjustments and loss-mitigation by payor-mix adjustment doesn’t seem to have any practical significance for providers, most of whom are stubbornly clinging to the idea that the costs resulting from treating Medicare/Medicaid patients must be recouped elsewhere, despite claims to the contrary in the literature.

      • Surely the magnitude matters, and it is generally quite small, as many studies document. The studies with the least rigorous methods (laughable even) are sponsored by the industry. They imply or claim levels of cost shifting that simply do not exist: 50 to 100 cents on the dollar.