As you may know, I’ve argued that that the quality and safety of healthcare have traditionally been underregulated. But regulators are like patients with Parkinson’s: it’s hard to get them unglued, but once they’re moving, it’s hard to stop them.
Welcome to Exhibit A.
Last month, I described Atul Gawande’s thrilling New Yorker article recounting the seminal work by Peter Pronovost and his Johns Hopkins colleagues in Michigan. By implementing checklists to prompt people to follow safety procedures (like using barrier precautions during catheter insertions), they were able to slash the infection rate in over 100 Michigan ICUs. The results were tremendous savings – of over 1500 lives and $200 million. The original Pronovost study was published a year ago in the New England Journal of Medicine – AHRQ Patient Safety Network, the federal government’s patient safety portal, deemed it an “instant classic,” a status granted only 1-2 articles each year.
So you’d think we’d be talking today about how to migrate this work to other states, or to operating, delivery, or emergency rooms. Instead, the news is far more depressing. The study has been shut down by a handful of bureaucrats, acting in the name of patient protection. Huh?
The shocking saga was recounted by Gawande in a Sunday NY Times op-ed piece, and analyzed nicely in a handful of blog postings by Maggie Mahar (here and here) and others (here). In a nutshell, a whistleblower – apparently concerned that patients and providers were not consented and IRB approval was not obtained at each of the 70 hospitals – alerted the federal Office of Human Research Protections, the office whose job it is to protect patients from the next Tuskegee, Nuremberg, or Vioxx. But from checklists?
As Maggie (who has done some behind-the-scenes digging) describes, OHRP began its crusade by sending threatening letters to Michigan and Hopkins. Hopkins investigators and campus leaders protested, arguing that they had received approval for the study from their own Institutional Review Board (IRB), which waived patient and provider consent requirements or the need for individual hospital IRB approvals. But OHRP strongly disagreed, insisting that the study needed to be approved by every participating hospital’s IRB, and that every participating patient and provider needed to provide informed consent.
So instead of saving more lives, Pronovost and his colleagues are presently jumping through bureaucratic hoops, getting approval from 70 Michigan IRBs. Some hospitals doubtless lack IRBs (since they have no research infrastructure) and will choose not to participate. Meanwhile, the study’s data collection is more frozen than Lake Huron.
Peter Pronovost is a can-do, eternally optimistic guy (you can’t be as effective as he is if you’re not), and he and Hopkins have the wherewithal to plow ahead. I’m more worried about the chilling effect of this ruling on quality improvement work elsewhere. Gawande shares this concern. In a series of email exchanges with me, he made the following points (all quotes are Atul’s, except for my insertions in italics):
“(1) The central difficulty for those of us trying to improve care is to change how people practice medicine day to day. We’re trying to ingrain in medical and surgical culture that we will more consistently adhere to at least the minimum standards of care, and to do this we will likely need to use checklists and other strategies, track our performance, learn from the results, and share the lessons as a regular part of everyday practice.
[Parenthetically, I’ll add that Atul made this point brilliantly in another New Yorker article.]
(2) This should not require IRB approval. It’s hard enough to get clinicians to do this as it is. For example, following Peter’s lead, I am working on introducing a surgical safety checklist and a key part is helping everyone with data collection to ensure success. My local Harvard School of Public Health IRB considered this non-research. The federal ruling disagrees with this logic. If the IRB had insisted on review, we’d have to get approval in every hospital signing up, which would be impossible. Similarly, Peter’s efforts to extend his program to the whole of the US, Spain, Netherlands will be simply impossible – just from requiring local IRB review. (Keep in mind, Peter did submit to his own IRB – it judged his work non-research and that judgment was the target of the federal ruling). That’s not what we want for patient care.
[A clarification: according to Pronovost, the Hopkins IRB actually did deem the Michigan work to be “research” – it’s just that they did not feel that it was feasible or necessary to get informed consent from the involved participants or individual hospital IRB approval.]
(3) The agency’s language in finding against Johns Hopkins is forceful in saying all involved institutions IRBs must review and that the agency thinks the failure to obtain consent from patients and providers is itself a violation. The question of patient consent is NOT the central issue. I’m sure they’d back down on that if local IRBs came down differently. The fundamental issue, though, is whether being systematic about clinical care requires doctors and hospitals to get special permissions from all the local IRBs – and it shouldn’t. This would be crazy, a hindrance to good care, and costly.
(4) This does not mean that quality improvement is ethics free. The responsibility for clinical care and quality have always been with clinicians – including responsibility to protect privacy. Just because we make a shift from doing things by the seat of our pants to following clinical standards more scientifically does not absolve us of those same ethical clinical responsibilities. But it should not require people to have to seek additional permissions. This is the way care should be expected to be done – and that should not be made more difficult to do.
(5) HOW then can we act? Part of the question is whether there is room under current legislation for the agency to issue a different clarification of the rules than the one it is currently preparing. The Hastings Center thought there is room. So I actually do think pressure from Congress and higher ups at HHS would have real impact. Specifically, we’d want to ask congressional representatives to request that the Office for Human Research Protections let quality improvement research proceed unencumbered and reverse the ruling on Peter’s program [I have some addresses for you at the end of this posting]. Work to monitor, study and improve performance should become the norm in medicine, and doctors and hospitals should be expected to pursue this work – without federal research regulations blocking the way.”
I’ve already heard from several colleagues doing QI projects who now wonder whether they need to stop collecting data – or even discontinue their project altogether – if they haven’t consented the patients or providers or obtained IRB approval at every practice site. Atul describes a Nigerian physician learning statistics in a Harvard Public Health School class. “As a class project, he came to me for mentorship in looking at the patterns of his hospital’s surgical death and complication rates… Given the ruling against Johns Hopkins, however, we were advised (when I stupidly inquired) that he could NOT do this research work without submission to an IRB – likely at both Harvard and at each institution. It’s hard to see how this ruling could mean otherwise. But this is a 4-week class. The result: the project was dropped. That can’t be good for patients.”
Just think about it. We used to believe that hospital-acquired infections were an inevitable consequence of hospital care – bad things happen to good people. Then research, mostly the collection of day-to-day clinical data by committed people looking for improvement opportunities, demonstrated that certain commonsensical interventions, like more religious hand washing, could largely prevent these infections. Furthermore, we learned that using antibacterial alcohol gel rubs works well and is easier to perform than traditional hand washing.
Let’s say I now want to put an alcohol hand gel dispenser in every patient’s room, and encourage people to use it – with checklists, teamwork training… hell, I’ll drink the stuff if that’s what it takes. Do I need to obtain IRB approval before I install the dispensers? Or before I measure my infection rates after installation? Do I need the consent of the docs and nurses to “participate” in this “intervention”? How about the patients?
I think you can see where this takes us.
Having written about the unanticipated consequences of quality and safety practices and critiqued (along with Pronovost) some aspects of the IHI’s 100,000 Lives Campaign for vigorously promoting practices with weak evidence, I recognize that these are not easy questions. Even commonsensical practices can backfire, and Pronovost believes, as do I, that serious quality improvement interventions – when accompanied by a hypothesis and data collection – are research. The question here is one of balance and feasibility. If the OHRP ruling stands, then what we’ve ostensibly done is place a huge bureaucratic tax on quality improvement work – a tax extracted in money, time, hassles, and frustration. The results will be predictable: less good QI work, fewer participating providers and institutions, fewer smart young people drawn to careers in quality and safety, and ultimately more harm and deaths.
In the end, the question is this: will more patients benefit from the protections afforded by extra bureaucratic obstacles to QI work, or by wider dissemination of rigorous quality improvement projects and studies? If you want to see a close race, look to Huckabee vs. McCain, or Obama vs. Clinton. But don’t look here: this contest is a landslide.
If you agree, there’s something you can do. Please write to one or all of the following:
- Your US representative (click here)
- Your US senator (click here)
- And the Office for Human Research Protections:
Kristina Borrer, PhD
Director, Division of Compliance Oversight, OHRP
Ivor Pritchard, PhD
Acting Director, OHRP