• Why the Medical Research Grant System Could Be Costing Us Great Ideas

    The following originally appeared on The Upshot (copyright 2018, The New York Times Company).

    The medical research grant system in the United States, run through the National Institutes of Health, is intended to fund work that spurs innovation and fosters research careers. In many ways, it may be failing.

    It has been getting harder for researchers to obtain grant support. A study published in 2015 in JAMA showed that from 2004 to 2012, research funding in the United States increased only 0.8 percent year to year. It hasn’t kept up with the rate of inflation; officials say the N.I.H. has lost about 23 percent of its purchasing power in a recent 12-year span.

    Because the money available for research doesn’t go as far as it used to, it now takes longer for scientists to get funding. The average researcher with an M.D. is 45 years old (for a Ph.D. it’s 42 years old) before she or he obtains that first R01 (think “big” grant).

    Given that R01-level funding is necessary to obtain promotion and tenure (not to mention its role in the science itself), this means that more promising researchers are washing out than ever before. Only about 20percent of postdoctoral candidates who aim to earn a tenured position in a university achieve that goal.

    This new reality can be justified only if those who are weeded out really aren’t as good as those who remain. Are we sure that those who make it are better than those who don’t?

    recent study suggests the grant-making system may be unreliable in distinguishing between grants that are funded versus those that get nothing — its very purpose.

    When a health researcher (like me) believes he has a good idea for a research study, he most often submits a proposal to the N.I.H. It’s not easy to do so. Grants are hard to write, take a lot of time, and require a lot of experience to obtain.

    After they are submitted, applications are sorted by topic areas and then sent to a group of experts called a study section. If any experts have a conflict of interest, they recuse themselves. Applications are usually first reviewed by three members of the study section and then scored on a number of domains from 1 (best) to 9 (worst).

    The scores are averaged. Although the bottom half of applications will receive written comments and scores from reviewers, the applications are not discussed in the study section meetings. The top half are presented in the meeting by the reviewers, then the entire study section votes using the same nine-point scale. The grants are then ranked by scores, and the best are funded based on how much money is available. Grants have to have a percentile better than the “payline,” which is, today, usually between 10 and 15 percent.

    Given that there are far more applications than can be funded, and that only the best ones are even discussed, we hope that the study sections can agree on the grades they receive, especially at the top end of the spectrum.

    In this study of the system, researchers obtained 25 funded proposals from the National Cancer Institute. Sixteen of them were considered “excellent,” as they were funded the first time they were submitted. The other nine were funded on resubmission — grant applications can be submitted twice — and so can still be considered “very good.”

    They then set up mock study sections. They recruited researchers to serve on them just as they do on actual study sections. They assigned those researchers to grant applications, which were reviewed as they would be for the N.I.H. They brought those researchers together in groups of eight to 10 and had them discuss and then score the proposals as they would were this for actual funding.

    The intraclass correlation — a statistic that refers to how much groups agree — was 0 for the scores assigned. This meant that there was no agreement at all on the quality of any application. Because they were concerned about the reliability of this result, the researchers also computed a Krippendorff’s alpha, another statistic of agreement. A score above 0.7 (range 0 to 1) is considered “acceptable.” None were; the values were all very close to zero. A final statistic measured overall similarity scores and found that scores for the same application were no more similar than scores for different applications.

    There wasn’t even any difference between the scores for those funded immediately and those requiring resubmission.

    It would be easy to mistake this study as a death knell for the peer review process. It’s not. A careful reader must note that all of the grants in this study were exceptional. They succeeded, after all. Since the N.C.I. funds only about 10 percent of grants, we’re looking only at proposals in the best decile, and it’s likely that there might be less variability in scores among those than among grants occupying the full spectrum of quality.

    This should still concern us greatly. This system was devised back when more than half of submitted grants were funded. That’s very different than what we see today.

    The current system favors low-risk research. If you’re going to fund only a small percentage of proposals, you tend to favor the ones most likely to show positive results. You don’t want to have to defend null findings as a “waste of money.”

    The current system favors experienced researchers over new ones. They have thicker curriculum vitae, more preliminary data and name recognition. Moreover, they know how to work the system. At this point in my career, I know how to write multiple grants efficiently. I’m better at it than I used to be.

    The current system can also be biased against women and minorities in ways that could keep them out of funding range. The system is not blinded, and many studies have shown that even after controlling for other factors, the ways in which grants are discussed, scored and funded can favor men over women, and whites over minorities.

    If researchers are getting into the top 10 percent more than others based on such factors, especially with less and less money available, many great proposals — and many great researchers — are being sidelined inappropriately.

    We may be missing out on a lot of excellent, and perhaps novel, work that can’t break into the top 10 percent because of structural problems. There are things we could do to fix that. One might be, of course, to increase funding across the board. John Ioannidis has proposed that we fund researchers, not research. A group of informaticists from Indiana University has suggested a percentage of funding be put to all scientists for a vote.

    Other solutions are more radical. One might involve a modified lottery. The current system seems to do reasonably well at discriminating between “bad” and “good” grants. Once those good ones are put aside, we might do better by assigning funding through chance. Ferric Fang and Arturo Casadevall, who are researchers and journal editors, have proposed that such a system could reduce bias and increase diversity among researchers, suggesting that seniority and other factors still play too large a part in funding decisions.

    They make the case that we already have a de facto lottery now, except it’s not random, and therefore unfair.

    The current granting system doesn’t just fund the researchers of today — it also steers the careers of tomorrow. Should it fail, the repercussions will be felt for decades.

    @aaronecarroll

    Comments closed
     
  • Healthcare Triage: Why Don’t We Invest More in Public Health?

    Many, many studies conclude that investing in public health is more effective than continually increasing spending on expensive treatments. So why doesn’t the US spend more on public health?

    This episode was adapted from a column Austin and I wrote for the Upshot. Links to sources can be found there.

    @aaronecarroll

    Comments closed
     
  • JAMA Pediatrics Podcast: Dietary Supplement Use in Kids

    As you know, I’m now the Web and Social Media Editor at JAMA Pediatrics. We’ve got a podcast where I discuss a paper from the journal. I do my best to pick good ones.

    Please consider giving this a listen, and subscribe! Doing so makes it more likely that I’ll be able to keep doing this.

    This week, I’m covering “Prevalence of Dietary Supplement Use in US Children and Adolescents, 2003-2014”:

    Comments closed
     
  • Child Separation and History: the Canadian Residental Schools

    The scenes of children being separated from their families are sadly familiar to Canadians. American readers should know about the history of child separation in Canada and its costs to both families and the Canadian nation.

    First, the US policy. Dara Lind at Vox reports that

    As a matter of policy, the US government is separating families who seek asylum in the US by crossing the border illegally. Dozens of parents are being split from their children each day — the children labeled “unaccompanied minors” and sent to government custody or foster care, the parents labeled criminals and sent to jail. …at least 2,700 children have been split from their parents. …at present, an average of 45 children are being taken from their parents each day.

    This policy is opposed by the American Academy of Pediatrics, the American Academy of Child and Adolescent Psychiatry, the American Psychological Association, the American Psychiatric Association, the American Medical Association, and the American Public Health Association. This list may not be complete.

    The American Bar Association criticized the putative legal justification for this policy.

    It is apparent from the public comments of several high-ranking Administration officials that a primary purpose of the “zero tolerance” policy is to serve as a deterrent for migrant parents who enter the United States without authorization accompanied by their children. These statements make clear that family separation is not a collateral consequence of regular law enforcement under this policy; it is an explicitly intentional goal. Although the Supreme Court has never addressed a case involving the exact facts presented by the current practice of family separation, existing law suggests the policy violates rights to family integrity and due process.

    Some have compared this to Nazi human rights abuses. I disagree. The Nazis intended to murder children and the connection to the death camps rests on a slippery slope argument. It is sufficient to point out that the policy is inhumane and that it violates human rights to harm children to deter misdemeanour offences by their parents.

    There is a closer historical parallel than Nazi Germany: the treatment of indigenous children in Canada. Here is what happened (you can find links here).

    Beginning in 1831, Canada began setting up boarding schools to educate indigenous children. The practice grew and became official federal policy in 1884. Education is, of course, a good idea. However, many of the schools were located far from First Nations communities. This was by design: the program was intended to ‘civilize’ the children, which was thought to require erasing their cultures: including their languages, art, clothing, customs, and religions. Erasing culture, in turn, required separating children from their families. Residential schools were established far from tribal lands to achieve this separation.

    The system lasted well into the 20th century. About 150,000 children were taken from families and sent to the residential schools. Many children suffered physical and sexual abuse. The Canadian Truth and Reconciliation Commission estimated that 6,000 children died, but the count may be much higher. Many families were never notified that their child had died. That the number of dead children is uncertain is itself an indicator of how the system neglected them.

    What is the lesson from Canada?

    Most importantly, mass separation of children from their families can be grievously harmful to those children. Unless a child is in imminent danger of abuse or neglect, don’t do it!

    However, all of Canada was harmed by the residential schools. The policy greatly added to the social inequity and distrust between the colonial and indigenous communities. The practice had the intended consequence of weakening the cultures, religions, and languages of Canadian indigenous peoples, to their and our enormous loss.

    Moreover, the cruel mistreatment of children damaged the legitimacy of Canadian institutions. Many of these schools were run by the Roman Catholic, Anglican, or other Christian denominations, others by universities or provincial governments. Some of this institutional harm may be reparable by the recent apologies from Prime Minister Stephen Harper, provincial premiers, church leaders (including the Pope), and a Truth and Reconciliation process. Time will tell.

    What should Americans do? The most important thing is to vote.

    Unfortunately, the next election is months away, a long time in a child’s life. If we cannot get these children returned to their families, we need to keep our eyes on them and record whether they are harmed. They will be treated better if the authorities know that the world is watching.

    But we should not assume that the world will continue to watch. The national attention span is measured in hours, or at most days. There will be another media event soon, and our memory that children are being taken from their parents will fade. Keep watching.

    @Bill_Gardner

    Comments closed
     
  • Healthcare Triage News: Medicaid Report Cards Are Out, and Life Isn’t About Grades

    The Trump administration has released Medicaid report cards, which allow citizens and health care journalists to look at the results of Medicaid, state by state. The good news: our beloved Indiana is doing OK! The bad news, there’s a lot of missing data in this reporting process, so it’s hard to really know what these report cards mean.

    @aaronecarroll

    Comments closed
     
  • The case that could end the Texas lawsuit.

    A slew of amicus briefs were filed yesterday in the Texas lawsuit, almost every one of them pushing back on the argument that the Affordable Care Act should be invalidated, in whole or in part. (I filed one myself, together with a bipartisan group of law professors.)

    Today, I wanted to highlight one of those briefs: this one by the American Medical Association and other medical societies. It’s the brief that the United States would have filed if it hadn’t abandoned its duty to defend the statute.

    The brief is excellent throughout, especially in arguing that the plaintiffs don’t have standing to bring the suit (an issue I flagged last week). But its most important contribution is identifying a Fifth Circuit case—one that binds the district court that’s hearing the lawsuit—that seems to dispose of the argument that a penalty-free mandate is unconstitutional.

    In United States v. Ardoin, 19 F.3d 177 (5th Cir. 1994), the Fifth Circuit held that Congress may constitutionally exercise its taxing power without actually raising “some revenue” for the government. Ardoin involved the intersection between the National Firearms Act of 1936 (NFA) and a subsequently-enacted statute, the Firearms Owners’ Protection Act of 1986 (FOPA). The NFA regulated machine guns under Congress’s taxing power; it barred the “making of machineguns without having filed a written application or paying the making tax.” Id. at 179. FOPA made possession of machineguns illegal. Id. After FOPA was enacted in 1986, the Bureau of Alcohol, Tobacco, and Firearms refused to “accept applications to register or to pay the tax” on machineguns. Id. at 179. When the defendant was convicted of violating the NFA, he contended that the “constitutional authority for provisions of the NFA dealing with the registration and taxing of post-1986 machineguns [was] gone.” Id.

    The Fifth Circuit rejected this argument. It explained that “the basis for ATF’s authority to regulate—the taxing power—still exists; it is merely not exercised.” Id. at 180. In so doing, the Fifth Circuit held that the NFA could “be upheld on the preserved, but unused, power to tax.” Id. (emphasis added). And in so holding, the Fifth Circuit flatly rejected the dissent’s claim—identical to the Plaintiff’s here—that “[t]o remain legitimate . . . a measure enacted under the tax power must raise some revenue.” Id. at 187 (Wiener, J., dissenting).

    Ardoin thus refutes Plaintiffs’ arguments, and its unequivocal reasoning requires this Court to uphold the constitutionality of § 5000A. Under this controlling precedent, a law need not raise “some revenue” to qualify as a tax under Congress’s taxing power. All that is required is that the authority to tax be preserved, even if no revenue is actually raised.

    Q.E.D. If there’s no constitutional flaw with the penalty-free mandate, there’s no need to address the (meritless) arguments about severability.

    To be sure, it’s possible to distinguish Ardoin: an unexercised tax is formally different from a tax that Congress has zeroed out. Functionally, however, the $0 tax in Ardoin is identical to the $0 mandate penalty—which is why the case lends powerful support to the argument that the penalty-free mandate is constitutional.

    How is it, then, that the Justice Department can say that it couldn’t come up with a single non-frivolous argument to defend the statute? Didn’t it have a duty to at least mention the case to the Texas judge who’s hearing the lawsuit?

    @nicholas_bagley

    Comments closed
     
  • Amicus brief in the Texas ACA case

    Together with a bipartisan group of law professors—Jonathan Adler, Abbe Gluck, Ilya Somin, and Kevin Walsh—I submitted an amicus brief today in the Texas litigation over the constitutionality of the Affordable Care Act. Joe Palmore at Morrison and Foerster graciously helped to pull the brief together; he did characteristically excellent work.

    The amicus brief doesn’t take a position on the constitutionality of the penalty-free mandate. Instead, it argues that it’d be inappropriate to sever any part of the ACA if the mandate is constitutionally defective:

    The cornerstone of severability doctrine is congressional intent. Under current Supreme Court doctrine, a court must offer its best guess on what Congress would have wanted for the rest of the statute if a single provision is rendered unenforceable. But this guessing-game inquiry does not come into play where, as here, Congress itself has essentially eliminated the provision in question and left the rest of a statute standing. In such cases, congressional intent is clear—it is embodied in the text and substance of the statutory amendment itself. Under these circumstances, a court’s substitution of its own judgment for that of Congress would be an unlawful usurpation of congressional power and violate basic black-letter principles of severability. Yet that is what the plaintiff States and the United States invite this Court to do.

    As a group, we don’t all see eye to eye on questions pertaining to constitutional interpretation, statutory construction, or the Affordable Care Act. Members of the group were on both sides of NFIB v. Sebelius and King v. Burwell; indeed, many of us have publicly sparred with each other over the right outcome in those cases.

    But we all agree on this: “The arguments of both the plaintiff States and the United States on the severability of the insurance mandate from the other provisions of the ACA are inconsistent with settled law.”

    At a time when it’s hard to find bipartisan agreement on anything, I’d like to think that the amicus brief testifies to the outlandishness of the severability arguments in the Texas litigation. But you can be the judge of that. In the meantime, I’m just glad this group was able to come together.

    @nicholas_bagley

    Comments closed
     
  • A Big Loss for Insurers in the Risk Corridor Litigation.

    Today was a bad day for insurers seeking more than $12 billion in risk corridor money from the United States. Over a dissent, the Federal Circuit ruled that the federal government doesn’t owe them a cent. (I’ve been following the litigation for a long time; you can catch up on my prior posts here.)

    The opinion starts on a bright note for insurers. The ACA, the Federal Circuit reasoned, created an enforceable obligation to pay risk corridor claims in full: the statute is “unambiguously mandatory.” The court brushed back the government’s arguments that the ACA created only a budget-neutral risk corridor scheme: “Nothing in section 1342 indicates that the payment methodology is somehow limited by payments in.”

    That’s a big deal. If nothing else other than the ACA were in play, insurers would be entitled to recover everything they say they’re owed.

    * * *

    Unfortunately for insurers, that’s not all that’s in play. Appropriations riders for fiscal years 2015 and 2016 limited funds available to pay risk corridor claims. Here’s the text of one of those riders (the other is identical):

    None of the funds made available by this Act from [two available funds], or transferred from other accounts funded by this Act to the “Centers for Medicare and Medicaid Services-Program Management” account, may be used for payments under section 1342(b)(1) of [the ACA] (relating to risk corridors).

    That’s all it says. By its terms, it doesn’t change the scope of the ACA’s underlying promise. It’s just a restriction on the use of certain funds to pay certain claims.

    Nonetheless, the Federal Circuit says that Congress “clearly indicated its intent” in the appropriations riders to amend the risk corridor program:

    [Congress] asked GAO what funding would be available to make risk corridors payments, and it cut off the sole source of funding identified beyond payments in. It did so in each of the three years of the program’s existence. And the explanatory statement regarding the amendment containing the first rider of House Appropriations Chairman Rogers confirms that the appropriations language was added with the understanding that HHS’s intent to operate the risk corridors program as a budget neutral program meant the government “will never pay out more than it collects from issuers over the three year period risk corridors are in effect.” 160 Cong. Rec. H9838 (daily ed. Dec. 11, 2014). Plainly, Congress used language similar to the appropriations riders in [prior cases] to temporarily cap the payments required by the statute at the amount of payments in for each of the applicable years—just as those decisions altered statutory payment methodologies.

    What else could Congress have intended? It clearly did not intend to consign risk corridors payments “to the fiscal limbo of an account due but not payable.”

    This is pretty weak tea. As the dissent points out, Congress tried and failed to amend the ACA to make the risk corridor program budget neutral. The appropriations rider was the best it could do. Chairman Rogers may have wanted to change the scope of the entitlement, but he wasn’t able to lock that down into statutory language. It’s a bad look for the court to privilege the bombastic statement of a bill’s proponent over the text of the law that Congress enacted.

    Besides which, it’s completely plausible to think that Congress meant to “consign risk corridors” to “unpayable obligations.” Even as congressional Republicans kicked up a kerfuffle over “bailouts,” HHS was reassuring insurers that the agency would “record risk corridors payments due as an obligation of the United States Government for which full payment is required.” Knowing full well that it had an obligation to pay, Congress shut off the funding stream in a deliberate effort to sabotage the ACA.

    And it worked! Loads of co-ops went under in response to the unexpected financial hit. Did Republicans in Congress care that they were reneging on a promise? Not a bit. They saw a chance to hurt Obamacare, and they took it.

    * * *

    The Federal Circuit’s next move is to reject the insurers’ claim that HHS had made a contractually binding commitments to pay. Insurers argued that HHS used those commitments to “induce insurers to offer plans in the exchanges without an additional premium accounting for the risk of the dearth of data about the expanded market, in reliance on the presence of a fairly comprehensive safety net.”

    Without quarreling with any of that, the court held that “the overall scheme of the risk corridors program lacks the trappings of a contractual arrangement that drove the result in [a prior case].” Why? The court thought it was relevant that HHS didn’t employ certain words in entering into this commitment—words like guarantee, offer, or contract. Instead, risk corridors were an “incentive program” for the benefit of third parties, not a “traditional quid pro quo.”

    I’m again at a bit of a loss. You don’t have to use magic words to enter into a contract. You just have to make a promise. What’s more, every contract operates as an “incentive program”: my promise to perform is an incentive for you to perform. And loads of contracts are entered into primarily for the benefit of third parties. That doesn’t make them unenforceable.

    * * *

    What’s next? The insurers will probably seek en banc review. This is a plausible candidate for full court rehearing: it’s a big money case, it carves a lot of crucial Federal Circuit cases, and (to my eyes) the reasoning is pretty dubious. But en banc review is rare, so insurers can’t pin their hopes on it.

    From there, it’s on to the Supreme Court. Who knows if it’ll agree to hear the case? It’s not an implausible candidate for review, and lord knows the Court can afford to take more cases. But the Court might be gun-shy about wading into another case about the ACA.

    So this isn’t the end of the road for insurers. But it’s a Michigan-sized pothole in that road. And the opinion reflects a disturbing unwillingness to hold the government accountable for its promises. That’s a point Judge Newman made in concluding her dissent:

    The government’s ability to benefit from participation of private enterprise depends on the government’s reputation as a fair partner. By holding that the government can avoid its obligations after they have been incurred, by declining to appropriate funds to pay the bill and by dismissing the availability of judicial recourse, this court undermines the reliability of dealings with the government.

    Craig Garthwaite and I sounded the same theme in the New York Times last year. That’s why this decisions disturbs me so much: the costs of being cavalier about our debts extend far beyond a fight over risk corridor money.

    @nicholas_bagley

    Comments closed
     
  • The Texas lawsuit could end some of the ACA’s protections for employer coverage.

    The Trump administration’s refusal to defend portions of the Affordable Care Act is shocking enough. Equally shocking is how little it seems to care what happens if it gets what it’s asking for.

    One question in particular: what about legal protections for the 160 million people who get insurance through their employers? Will their insurance still cover their preexisting conditions, even if they switch jobs? I honestly have no idea.

    In its brief, the Justice Department argues that the community rating and guaranteed issue provisions of the ACA must be invalidated. But it never mentions that those provisions apply not only to individual health plans, but also to employer plans.

    So should those rules give way across the board? Or only for individual insurance plans?

    Maybe it should be the latter. The mandate isn’t critical to securing the health of the employer market, so the ACA rules that protect employees aren’t inextricably linked to the mandate and shouldn’t be invalidated. But it could also plausibly be the former: if the rules governing community rating and guaranteed issue are inseverable, maybe the court shouldn’t do micro-surgery to save some subpart of those rules.

    But guess what? In its brief, the Justice Department doesn’t say which approach it endorses.

    Actually, it’s worse than that.  When the Justice Department identifies the rules governing community rating and guaranteed issue, it doesn’t cite the ACA itself (Public Law 111-948). Instead, it cites parts of the U.S. Code that codify portions of the ACA (e.g., 42 U.S.C. 300gg). The implication is that the Justice Department wants the court to enjoin those code provisions.

    But the code provisions were on the books long before the ACA was adopted. Prior to the ACA, they listed protections for employer-sponsored plans that had been adopted in the Health Insurance Portability and Accountability Act. Among other things, HIPAA limited the circumstances under which an employer could refuse to cover an employee’s preexisting conditions. The protections weren’t perfect, but they were something. The ACA patched HIPAA’s gaps by amending those code provisions.

    So if the U.S. Code provisions are enjoined altogether—which, again, is what the Justice Department appears to be asking for—some of the HIPAA-era protections would be wiped from the books too.* Is that really what the Justice Department wants? Because that’s the thrust of its brief.

    The confusion may reflect a basic legal mistake, one that Tobias Dorsey highlighted in Some Reflections on Not Reading the Statutes: the U.S. Code is a codification of existing laws, but it’s not itself the law. That’s why code provisions shouldn’t themselves be the target of any injunction. Any injunction should run against the ACA itself. If that’s what the Justice Department really wants, then it has to clarify what it’s really asking for. Failing to do so could wreak havoc in the employer-sponsored market.

    Even if the injunction only runs against portions the ACA, however, that still wouldn’t resolve whether the ACA’s protections would still apply to employer-sponsored plans. If they don’t, that’s a big deal: HIPAA’s protections are porous.

    So far, however, the Trump administration hasn’t said a word—leaving 160 million people in the lurch.

    @nicholas_bagley

    * UPDATE: Some HIPAA protections—including those for people who switch jobs—are found in other code provisions that wouldn’t be the subject of any injunction. That includes the protections for people who switch jobs and COBRA. They’re still on the books, even though they were superseded by the ACA. So the effect of enjoining some of those code provisions may be somewhat less dramatic than I initially thought.

    Comments closed
     
  • In health care, less is not always more

    The following originally appeared on The Upshot (copyright 2018, The New York Times Company). It was jointly authored by Dhruv Khullar and Austin Frakt.

    How much you spend on medical care depends on what you get, but also where you get it.

    Confoundingly to many, the cost of the same procedure on the same patient by the same physician can vary by thousands of dollars depending on whether it’s performed in a hospital, a hospital’s outpatient department, an ambulatory surgical center or a doctor’s office. It can also vary by who’s paying the bill — which insurer or public program.

    And even for the same insurer, cataract surgery might cost twice as much in a clinic affiliated with a hospital compared with an independent surgery center. The cost of cancer care is significantly higher in hospital outpatient departments compared with community practices, partly because insurers often pay hospitals double for chemotherapy drugs. Delivering a baby in a teaching hospital costs about $2,000 more than in a community hospital.

    Some of this is a result of different prices. Some reflects differences in how much care is delivered: its intensity.

    Either way, such cost variation across care settings has led policymakers to consider paying more evenly for medical services regardless of where they’re delivered, and to shift care from expensive, high-intensity settings to cheaper, low-intensity ones. Doing so, the thinking goes, could result in more efficient use of resources by health systems.

    But new research also shows the downside of this approach: A study of Medicare hospitalizations found that almost all patients are more likely to survive at teaching hospitals, which tend to be more expensive. Amid our enthusiasm for more efficient care settings, we should be cleareyed about the limitations: Sometimes less is more, but sometimes more is more.

    To some extent the shift toward cheaper settings is already happening. Medicare has started to close the gap in payment rates between hospital-owned clinics and private doctor’s offices through site-neutral payments. The Massachusetts Health Policy Commission has recommended that more patients be diverted to low-cost community hospitals from high-priced academic medical centers. And insurers are encouraging health systems to shift hospital care to less expensive outpatient clinics, rehab facilities and even patients’ homes.

    It has long been common for lower-priced community hospitals to transfer patients to higher-priced and more intensive teaching hospitals. But some hospitals like Massachusetts General Hospital, a relatively higher-priced, academic medical center in Boston, now have programs in which they send stable patients from their emergency departments to affiliated, lower-priced community hospitals.

    Other academic centers are sending patients not to another hospital, but directly home. In 2014, the Mount Sinai Health System began a hospital-at-home program for patients sick enough to need a hospital but stable enough to be cared for at home. Patients get visits from doctors, nurses, physical therapists and social workers, and they can have intravenous antibiotics, lab draws and breathing treatments in the comfort of their home. Research suggests that care for patients treated at home costs less and results in fewer complicationshigher satisfaction and lower mortality.

    These efforts are part of a broader trend away from high-intensity-care settings. The use of inpatient care is declining across the United States, including a 6 percent drop in inpatient admissions for Medicare patients from 2004 to 2010. The Department of Veterans Affairs has experienced a 10 percent decline in inpatient use over the past decade, while outpatient visits have increased by 40 percent. New technologies are making outpatient care safer, and new payment models are encouraging a shift to lower-cost settings.

    But not all such efforts may serve patients well. There’s concern, for example, that outpatient surgical centers don’t always have the resources and staff needed to handle potential complications of the increasingly complex operations they’re taking on. Other research suggests that hospitals that spend more and do more may have better patient outcomes and lower mortality rates.

    Teaching hospitals are generally the costliest medical environments. Some have argued that only the sickest patients — for whom complex services and technologies are most likely to help — should be treated there, while relatively healthy patients should preferentially be cared for in less costly community hospitals. But do sicker patients really do better at teaching hospitals? And do healthier patients fare just as well in either setting?

    The new study on Medicare hospitalizations sheds some light. (Both of us were part of the group that conducted this study.) Led by Laura Burke and Ashish Jha at Harvard, the study analyzed more than 11 million Medicare hospitalizations and found that almost all patients — whether very sick or relatively healthy — had lower mortality rates at teaching hospitals. But there are also some differences.

    Among patients admitted for operations like hip replacements, the patients with the most health problems over all were the ones likeliest to benefit from a teaching hospital. On the other hand, among people admitted with conditions like pneumonia or heart failure, though all groups did better at the teaching hospitals, the difference was greatest for the relatively healthy patients.

    The more advanced technologies available at teaching hospitals explained some, but not all, of the difference. Other factors like subspecialty expertise, more clinicians involved in care, and greater availability of ancillary services may also be playing a role.

    Given the high — and sometimes unjustifiable — cost of some health care settings, it seems reasonable to pursue payment parity for comparable care delivered in different settings. And all other things equal, the shift toward lower-intensity, lower-cost settings is a worthy goal. But in some cases, outcomes may not be equal, and it seems we should make sure we’re not cutting quality when we’re cutting costs.

    Comments closed