The decision of the Cleveland Clinic to start a Center for Functional Medicine was big news. That the $9-billion system gave the initiative significant visibility suggested arrival for functional medicine. A few hurdles still existed. Cleveland Clinic’s new partners needed to clarify and create a clinical model that could be measured. That was the caveat. This Center was a bet – a pilot based on a largely untested belief that functional medicine could outperform regular medicine, and at lower cost. Most in the field assumed this would prove a slam dunk. Care from a team of functional medicine physician/nutritionist/health coach and then behavioral specialist became the unit for which outcomes would be measured. Now in a publication in JAMA Network, the first results are in. The headlines were positive – but what do the data really say?
I asked a select group of researchers and clinician researchers – mainly Associate Editors at JACM – to provide their analysis and perspectives. That respondents leaned heavily on naturopathic researchers speaks both to who chose to respond and the extent to which their research, like that at the Center for Functional Medicine, focuses on patient-reported and other outcomes of multiple component interventions. Like functional medicine practitioners, the naturopathic doctors need to show value not just of individual modalities, but of the care model. What have we learned? What would these veteran researchers suggest for next steps?
While each honored Cleveland Clinic and the functional medicine team for engaging the research, their conclusion do not have the same feeling of lift-off that came with the announcement of the founding of the Center for Functional Medicine. There’s plenty of work ahead. Here are the perspectives.
Wayne Jonas, MD
Wayne Jonas, MD, the first influential director of the NIH Office of Alternative Medicine and then the director of the Samueli Institute, had already written on the JAMA study when I reached him with my query about it. He had sent this mention in his “What We Are Reading” weekly newsletter:
Functional medicine (a form of integrative medicine with a major focus on diet) showed improved quality of life compared to standard primary care at 6 months but was about the same at 12 months. The practice of functional medicine looks promising but needs a randomized controlled study that lasts at least 12 to 24 months.
Short, and too the point. But how would Jonas propose to control for the functional medicine intervention. Sham coaching plus sham nutritional counseling and sham behavioral health – not to mention the some sham to parallel the influence of the functional medicine context? I asked Jonas, a methodologist, what he would propose. He responded:
Here are my suggestions. The comparison groups should be usual care without functional medicine using the NCCIH PRIME guidelines for pragmatic trials. Costs are an essential secondary outcome measure with clinical outcomes a primary and compliance/satisfaction as an additional outcome. The study should last a year.
Carlo Calabrese, ND, MPH
Carlo Calabrese, ND, MPH is a former member of the NIH National Advisory Committee on Complementary and Integrative Health whose professional career has included co-director of the Bastyr University Research Institute. He is a pioneer in the naturopathic and integrative space in promoting and designing whole systems research and effectiveness trials. He is presently based in Portland, Oregon, where, semi-retired, he maintains a part-time practice focused on “naturopathic psychological medicine.”
First, it is amazing that Cleveland Clinic has a Center for Functional Medicine, a sign of progressive thinking in a leading medical institution and without which this study would have been impossible. Given its existence, this analysis of outcomes is a responsible handling of potentially important clinical differences resulting under medical cultures or, as the authors put it, ‘operating systems’, which largely define how and which medical science is put into practice—in this case, comparing a conventional system of usual care in which a capitalistically-driven disease-drug model plays a role versus an alternative such as functional medicine’s systematic manipulation of physiology by generally unpatentable substances and lifestyle practices. Observational studies are almost a requirement to evaluate individualized multi-component whole alternative clinical practices for several reasons. Though they don’t have the cachet of RCTs, observational studies yield high external validity, ie, they point towards meaning in the real world. PROMIS outcomes measures are a rational choice for the assessment of whole health by an authoritative national standard. The health gains for functional medicine seem to be small, though over only a year, and we can imagine the improvements ideally concatenating into a potentially healthier lifespan. The authors might have said more about the principles of functional medicine toward treatment fidelity and replicability. While it may be considered something of an ethical obligation to evaluate non-standard clinical practice variants, it is also essential to shape the methods to the practice in order not to destroy it by attempting to force it into a randomized trial. The investigators do a good job of the compromises needed in this study, employing best current methods for observational studies.
Melinda Ring, MD
Melinda Ring, MD is a clinician and researcher who serves as director of the Northwestern Osher Center for Integrative Medicine where her clinic is part of the Bravenet network of integrative centers developing outcomes information. Ring has also served as core faculty at the Duke Leadership Program in Integrative Medicine and is co-director of a conference associated with the Cleveland Clinic’s separate integrative and lifestyle medicine center.
It was a big win to get the paper in JAMA. I found the discussion very good about what the biases and confounding factors were in the study. One key issue that they mention has to do with how the attitudes of those who received functional medicine were already engaged. They could have used a patient activation measure to look at that. They saw excitement in the first six months, then a drop. These PROMIS measures are so broad. It would be interesting if they could go in and look at sub-populations and how they fared. How do the more seriously ill fare relative to those less so. We had a study with the Bravenet network that found serious increase in patient activation after a visit. It makes sense from a functional and integrative medicine approach. That would be a good study to pair with this one. The strength in the Cleveland Clinic study is the direct comparison with the primary care patient there. Of course, it would be interesting to see the outcomes if every patient who saw the primary care doctor also saw a nutritionist and a health coach. Is it functional medicine that has the positive outcomes found here, or is it the lifestyle medicine? Finally, they don’t discuss the financial aspects of it. What does it cost the system when these functional medicine doctors spend more time with patients. They don’t generate as much revenues as regular primary care doctors in short visits.
I noted to Ring that while this is true in the current volume based industry, there is no clinical data that shows that medical doctors in short visits are actually having a more valuable impact on patient care, and long-term costs saving, than functional and integrative doctors who spend more time with patients. But that is another story.
Ryan Bradley, ND, MPH
The present director of the Helfgott Research Institute at National University of Natural Medicine, Ryan Bradley, ND, MPH developed the naturopathic medical profession’s first research – working with Dan Cherkin, PhD, Karen Sherman, PhD and others – on their whole person and whole system approach to the treatment of individuals with diabetes. Their population of “naturopathically naive” patients was also from a large medical delivery organization – in their case what is now the Kaiser Health Washington State system. They showed multiple, positive patient-centered outcomes. Bradley notes, among other things, that it is hard to generalize anything from this “functional medicine” to community practitioners of functional medicine who, for instance, are unlikely to be surrounded by the Cleveland Clinic’s team.
Kudos to the Functional Medicine community for critically evaluating their demonstration project at the Cleveland Clinic! The recent publication by Beidelschies [Michelle, PhD] and colleagues causes me to reflect on our efforts to publish similar research related to clinical outcomes from exposure to naturopathy, and the common criticisms we received, i.e., issues of “black box” types of analyses, scalability and fidelity. We ultimately responded to many of these critiques by carefully describing the components of the care delivered as “naturopathic” in both retrospective and prospective publications, as well as publishing algorithms on the clinical thinking behind the use of interventions, creating replicable individualization of treatments. Additional measures to generalize our results included engaging a community network of providers. Unfortunately, in this case the reader is left unable to interpret the exposure, “Functional Medicine”, as generalizable, or having fidelity to a foundational conceptual or philosophical model. Regardless of what was delivered, the manuscript by Beidelschies, et al., also demonstrates major questions on the impact of attrition/selection bias (acknowledged by the authors) due to limited follow-up data (~13% of the original sample at 12-months), and the overall impact of the intervention due to a lack of significant difference in the proportion of patients who had at least a 5-point change at 6-and 12 months (despite significant small mean changes in physical health scores). Combined these results call for intensifying efforts to collect follow-up data to ensure results can be extended more generally to “Functional Medicine”. Finally, the cost of the intervention is undisclosed, and would be a major consideration for future prospective research. The CIH community benefits from all critical appraisals, and this publication will aid in the refinement of future research and of functional medicine practices, to the benefit of many who will be served in the future.
Jeff Dusek, PhD
Researcher Jeff Dusek, PhD left a position leading research at the Mind-Body Medical Institute at Harvard Medical School in 2007 when he leapt into real world exigencies and ambiguities as research director for the Penny George Institute, part of the massive Allina medical organization. The Penny George effort was, through its outcomes, seeking to be a model for the nation based on researched outcomes. Dusek presently continues to toil in these same effectiveness and implementation domains as research director for University Hospitals Connor Integrative Health Network in Cleveland while continuing also to lead to the BraveNet practice-based research network (PBRN) that is generating outcomes in collaboration of academic health center integrative clinics.
First, I was impressed with the use of the PROMIS Global Health affectionately known as the PROMIS-10 by Beidelschies and colleagues. Second, conducting a comparative analysis with a propensity score matched cohort of patients from a family medicine setting is a major strength of this publication. Finally, that the functional medicine patients reported significantly larger improvements in the Physical Score at 6 months and Mental Score at 6 and 12 months than family medicine patients is important and should not be completely overshadowed by my following concerns.
From the article, it is clear that the functional medicine studied required patients to spend time with a clinician (60-75 minutes), dietician (unknown duration) and health coach (unknown duration). As written, it is difficult to assess the differences between the care provided by functional medicine and integrative medicine providers. Accordingly, it was troubling that the authors ignored my team’s research with implementing the PROMIS-10 as part of clinical care in an Integrative Medicine clinic.
Briefly, my team found that patients receiving integrative consultative visits reported improvements akin to those seen in the current article. We collected PROMIS-10 data on 59% of all eligible patients, whereas the denominator of patients from the current study is unknown. If there were only data on 1,595 out of 8,000 or 9,000 possible patients, that would diminish my view of their findings. Finally, my team used regression analyses to explore baseline demographic characteristic associated with lower PROMIS-10 scores at baseline and longitudinally. These analyses provide valuable insights into factors such as age, gender, race, marital status and insurance coverage may have on PROMIS-10 longitudinal changes. The lack of regression analyses in the current publication is a weakness. In all, the concept was great, but there are some important missing components that limit my enthusiasm.
Readers following sequentially will recognize that Ring, the Northwestern Osher Center-based commentator above, is part of the Bravenet network that generated the outcomes Dusek references here. Ring commented similarly on what they believe was a prior study from which they could have learned, or at least referenced in the discussion. I wonder if a tendency of the functional medicine world to hold “integrative medicine” at arm’s length in asserting functional medicine’s own, separate profile, may account for this oversight.
Patricia Herman, ND, PhD
The leading cost-effectiveness researcher in integrative health in the United States is Patricia Herman, ND, PhD, a senior behavioral and social scientist at the RAND Corporation. Herman’s work has informed key explorations at the National Academy of Medicine, the US Veteran’s Administration, and elsewhere. She just completed a 3-year term as a member of the NIH National Advisory Committee on Complementary and Integrative Health. On what she admitted was yet a cursory look, Herman had questions.
This outcomes study is as close as anybody has come to measuring the impact of integrative medicine for awhile, and so that’s good. Nevertheless, it’s not a big study though, and has some problems. They reference a large number of patients – 1595 – but only compare characteristics on 395. That’s less than 25% – and at 12 months there were only 200 remaining, and just 90 who completed both the 6 and 12 month surveys. I think the matching process [with the functional medicine and primary care patients] could have been better, e.g. “double robust matching” method in a recent study that might have been good. The drop off in results at 12 months could suggest that there was some “placebo enthusiasm” for this care when the functional medicine group began treatment. These are hard studies to do and I appreciate the level of effort here but it’s hard to generalize from these results. It’s also hard to know what happened. How long did the patients have treatment and what made up the treatment? Good for them to get it into JAMA. Good that they tried to look at outcomes, but it is unfortunate that costs were not considered. We need even and better studies on these integrative systems of care.
I was reminded, on reading the study, and again on reading these comments, of a moment at a conference a half decade ago when I was listening to a report on an NIH-funded study that examined a whole system of naturopathic care. As I was leaving, I asked another naturopathic researcher what she thought of the outcomes. The outcomes included, like this, positive patient-reported measures – and outcomes on a biomedical indicator that were trending positive but not statistically significant. The response: “We (naturopaths) can do better than this.”
I liked that blunt passion to take the learning and figure out what to shift to enhance the patient impact outcomes. Commentator Calabrese – my first mentor on the value of health services , outcomes and effectiveness research – is fond of saying in his promotion of investigations researching real-world practice such as this at the Cleveland Clinic that they provide service in practice improvement. Such studies ]mark paths toward improvement of both patient care and research modeling.
The results can be sobering as the field moves from claims that our medicine has fewer adverse effects and costs less to actually examining how the ducks look as they line up after a study. Credit the Cleveland Clinic for engaging the research, and getting it published. Now – how will the next trial be carried out? Clearly the financial component is key. And speaking of which – indirectly – how about those group services functional medicine outcomes from Cleveland Clinic that early indications say outperformed the functional medicine team?
Note: For a follow-up piece, I will seek some response from the research team and functional medicine community.