• It’s possible that price cuts lead to more care, not less

    This is a guest post from Adrianna McIntyre, a graduate student studying health policy at the University of Michigan. Adrianna currently blogs at Project Millennial, tweets at @onceuponA, and will be joining TIE soon as the team’s first intern. 

    We can thank former Vermont governor Howard Dean for reviving the “IPAB-as-a-rationing-body” meme again (you should read response pieces from Peter Orszag and Jonathan Cohn). The reality is that the advisory board is expressly prohibited from rationing care; all it’s really empowered to do is fiddle with reimbursements, and only if Medicare rates are climbing above a certain speed (they haven’t yet). But what if provision of care went up, not down, when reimbursements are cut? That’s what a new NBER paper (ungated) suggests, based on Medicare reforms and lung cancer treatment.

    In 2005, Medicare cut reimbursements for physician-administered (Part B) drugs. The resulting changes to chemotherapy drug payment had an especially acute impact on physicians’ profit margins.

    The 2005 reform, which replaced a reimbursement system based on list prices with one based on national average transaction prices, sought to remedy well known Medicare overpayments for Part B drugs. For chemotherapy, reimbursements averaged 29% above transaction prices in 1997 (Office of Inspector General (1997a)) and was 22% above by 2004, even after an initial fee reduction took effect (Government Accountability Office (2004)). Beginning in 2005, margins were administratively set at 6% above the national average transaction price for each drug. Thus, margins fell by nearly a factor of five, although this reduction was not uniform across drugs. Since chemotherapy accounted for over three-quarters of oncologist practice revenues in 2005 (Akscin and Barr (2007)), the drug mark-up accounted for more than half of their reported net income.

    While Aaron’s post yesterday covered a problem of overdiagnosis and overtreatment in cancer, broadly speaking, for certain cancers in certain populations—say, lung cancer in the over-65 cohort—literature suggests undertreatment. And following the 2005 payment changes, the authors found an increased likelihood of chemotherapy treatment within 30 and 90 days of diagnosis. Moreover, they found that these changes in treatment improved health outcomes—the likelihood of death at 3, 6, and 9 months after diagnosis declined by 2-3%, relative to patients who were diagnosed before the reform. While these are modest improvements, they are statistically significant. (Obligatory lead-time bias caveat: to any extent that diagnoses occurred earlier following the reform, survival changes might not reflect actual longevity gains.)

    Putting that together:

    These results indicate that, in the presence of significant physician agency effects, an over-generous payment scheme can lead to physician rationing — i.e. the under-provision of care — and that better aligning payments to costs can lead to better outcomes at lower total cost. While we discuss under-provision in terms of rationing, the results are also consistent with doctors mistakenly believing they were providing the optimal amount of care prior to the reform and thus intentionally providing excess care post-reform. This alternative scenario is consistent with our model, but, if true, suggests an additional policy lever. Specifically efforts to promote evidence-based medicine might have improved outcomes even in the absence of payment reform, although without the cost-savings generated by the reform.

    Austin has commented before that these changes likely reflect the latter, and the policy implications—that lower prices might improve otherwise underprovided care—strike me as particularly interesting. In the FFS system that we have, price levers are part of a rather limited tool set for encouraging evidence-based care and cost containment. Of course, it’s tougher in practice than in theory: with the 2005 Medicare reform, reimbursement fell for all chemotherapy drugs, and the researchers found that physicians shifted to drugs that caused the least shock to their bottom lines. Lowering the cost of certain treatments in isolation might just encourage a shift toward their more profitable counterparts. This also leaves us in a lurch on how to effectively to handle problems of overtreatment.

    Dean was right about one thing: fee-for-service sets up incentives that complicate efforts to encourage cost-effective care or change our system from reactionary to one that actually promotes health. Few would quarrel with his assertion that we need to change those incentives—but eliminating the IPAB is hardly the place to start.

    • Nice post on an interesting topic. It is interesting because it seems counterintuitive to classically trained economists. But it makes perfect sense if you consider a provider trying to maintain a stable income stream.

      Orszag wrote about this “volume effect” and “physician’s behavioral response” to rate changes in 2007. He called this “behavioral” to link to behavioral economics, which differs from classic. Here’s the CBO report – http://www.cbo.gov/sites/default/files/cbofiles/ftpdocs/81xx/doc8193/06-06-medicarespending.pdf

    • A key point is that even if there are volume offsets, spending will almost certainly be lower if prices are lower. The volume offsets just blunt the savings somewhat.

    • Given far less attention is the differential in reimbursement rates between hospitals and physicians offices/free standing facilities, the former being much higher than the latter. CMS has adopted a counter-intuitive strategy: CMS has been reducing the physician offices/free standing facilities reimbursement rates rather than reducing the hospital reimbursement rates, actually increasing the differential and providing an even greater incentive for migrating treatment to the hospital. It’s counter-intuitive, but only if you ignore the other policy consideration: consolidation/integration. Many health care policy experts believe efficiency (and lower costs) will be promoted if physicians consolidate (into larger group practices) or, even better, integrate (into hospitals). Of course, achieving one goal (integration) is contrary to another (lower health care spending). The administration has chosen consolidation/integration as a higher priority; indeed, last year we passed the threshold of over 50% of physicians now being employed by hospitals. We’ve spent the past 25-30 years promoting outpatient services, because patients prefer them and they are cheaper; now we’ve turned 180 degrees and are encouraging migration of services back to the hospital. It puts an enormous faith in hospitals, which are by far the most inefficient of the providers.

    • -The paper states the following:

      Furthermore, we show the mix of drugs used to treat patients changed in wayspredicted by economic theory: drugs that lost the most margin were used less frequently among the chemotherapy-treated population and conversely, expensive drugs favored by the reform’s 6% margin on all drugs, were used more often than previously.”

      E.g. post reform, the patient population was being treated with a different set of drugs. It’s entirely possible that the drugs that weren’t subject to the price controls to the same degree (newer drugs and biologics) may have had greater efficacy than the cohort of drugs used prior to the reform.

      If that’s the case, then it’s not clear that the timing of the treatment relative to diagnosis, which may or may not have been influenced by the payment reform (can anyone out there provide a logically coherent and clinically realistic explanation for why doctors suddenly started treating cancer patients more promptly after their margin on drugs was cut?) influenced survival, or if the changes in the drugs they were treating the cancer with are responsible for the improvements in survival.

      It’s also not clear that the changes actually saved money. It’s quite likely that the cohort of drugs that oncologists administered post-reform (less subject to the margin cap) are actually more expensive than the mix of drugs that they would have administered had the reform never occurred.

      Are we paying oncologists less as a consequence of these reimbursement reforms? Yes. Are we spending less? Maybe not. Is the care we are delivering more efficient? Also not clear.

      -IIRC one unintended consequence of the payment reforms was a significant shift in the locus of treatment from community outpatient to hospital outpatient settings – most likely due to the emergence of a reimbursement differential favoring treatments delivered at hospitals. Did this improve care? Possibly. Did it save money? Most likely not.

      -Contacting practicing oncologists working in both hospital and outpatient settings and asking them to comment on the theories regarding price, physician behavior, and patient care in the wake of the 2005 reforms would provide some valuable real-world perspective and context for both the original article and your commentary on it.

      Related tidbits:


    • Adrianna
      A very interesting post–very counterintuitive. Most striking to me I excerpted below:

      “Specifically efforts to promote evidence-based medicine might have improved outcomes even in the absence of payment reform, although without the cost-savings generated by the reform.”

      I tend to think individuals back into their income. If one generates 400K in 2004, take home should be 415K in 2005. If oncologists must administer more chemo to get there, they will achieve their goal (probably benevolently and with best intentions despite $ drivers).

      The effects of payment change may be a positive externality unique to chemo pricing and the uniqueness of the specialty.

      Literally and figuratively, I find the EBM approach as a solution the investigators suggest above (“absent reform”) a bit of a reach and a tough pill to swallow.

      Money drove the change and the overly optimistic prognostication of “education”as policy lever way too rosy.



    • @ BradF:

      Is there any evidence that the reimbursement policies have actually generated cost savings?

      If there isn’t, I encourage you to read through the testimony by Brooks as you parse the evidence and arrive at your own conclusion.

      As I said above, it’s not even clear how much EBM was driving changes in care and clinical efficacy, as opposed to changes in reimbursements driving clinicians to administering a new mix of drugs in a new setting. If I had to bet, I’d bet on the latter but it’s just an intuition.

      I’d also encourage anyone reading TIE piece to read the testimony from Brooks regarding all of the operational expenses that have to be financed by the 6% differential. Storing, preparing, administering, and disposing of cancer meds are all non-trivial aspects of care that require a considerable overhead in terms of equipment and fairly highly trained staff. Neither are cheap. Read about this in more detail and the fact that community outpatient oncology clinics are imploding by the hundreds every year won’t come as much of a surprise. The fact that the reimbursement formula lags behind the market price of the drug by 6-8 months can and does substantially erode the actual magnitude of the 6% margin. Toss in the fact that physicians have to front the money for the drugs long before they’re reimbursed and financing costs further erode the margin.

      There are a number of clear consequences of reimbursement reform – saving money and improving patient care aren’t amongst them.

    • “But what if provision of care went up, not down, when reimbursements are cut?”

      But what if provision of care went up, not down, when reimbursements are cut?

      data from the oecd countries – where prices are much lower and utilization moderately higher than in the u.s. – would seem to support that conclusion.