by Elizabeth Hofheinz, M.P.H., M.Ed., August 30, 2019
You might say the Minimum Clinically Important Difference (MCID) is a way of tracking any movement of the needle on patient satisfaction. But the complexity of using it merits caution. A recent study on this topic, “A Comparison of Treatment Effects for Nonsurgical Therapies and the Minimum Clinically Important Difference in Knee Osteoarthritis,” appears in the August 9, 2019 edition of JBJS Reviews.
Co-author Andrew Concoff, M.D. told OSN, “The MCID is
a patient reported outcome (PRO) measure that attempts to identify the smallest
change that is meaningful to patients. This can be considered the level
of improvement that the patients would first consider noticing if they felt
slightly ‘better.’ The use of the MCID represents part of a wider movement in
clinical research to include, even to focus upon, the patient experience as a
driver of treatment evaluations.” 
“We performed a systematic review of the available guidelines and meta-analyses that evaluate the impact of non-operative interventions on the pain experienced by patients with knee osteoarthritis (OA). We included the latest guidelines from the American Academy of Orthopaedic Surgeons, the European League Against Rheumatism, the National Institute for Healthcare Excellence as well as numerous meta-analyses and network meta-analyses including recent Cochrane reviews.”
There are benefits and there are BENEFITS.
“The benefit derived from treating patients suffering from knee OA with intra-articular hyaluronic acid injections has been controversial. It has long been recognized that the improvement of a group of patients treated with these injections is statistically significant, but the magnitude of the average benefit has been considered to be of questionable significance. That is, patients were numerically better than those, for instance, receiving placebo injections, but the difference was small enough to raise questions as to whether such a small change mattered.”
Dr. Concoff: “Further controversy and confusion have been inserted into this already enigmatic circumstance by comparing the MCID of those receiving certain treatments versus those receiving others. Several aspects of the proper application of the MCID as an evaluation of the efficacy of treatments evaluated in guidelines and meta-analyses have given rise to this uncertainty. First, while certain researchers have imagined the MCID to be a fixed value for a given parameter, herein, change in OA pain, the MCID can actually be established through a variety of methods, as described briefly in our paper. These include expert-based, distribution, and anchor-based approaches.”
Context is everything…
“Even among the most often applied, ‘anchor-based’ approach, techniques used to determine the MCID may differ which, in turn, may yield a variety of thresholds to define what degree of improvement represents the least, meaningful amount of response. Further, when using an anchor-based approach to determine an MCID, context is everything. A patient’s circumstances and expectations may dramatically affect the amount of improvement that they consider minimally but perceivably ‘better.’ Consider, for example, if we were trying to determine how much water a patient would have to drink to feel minimally ‘better.’ If one group of patients had been relaxing in the shade, that amount would be far different than those who had just returned from having been lost in the desert without access to water. Yet certain guidelines have ‘borrowed’ the MCID from prior research studies out of context or from groups of patients with a variety of different disease states and applied it to a novel circumstance without regard for the differences in the populations or context under study.”
Use this measure in a measured fashion…
“Additionally, the MCID is a very personal measure and, ‘My MCID may not be the same as your MCID.’ Thus, the MCID performs better when comparisons are made among the same individuals over time, that is for ‘within-groups’ comparisons, because the ‘me’-ness of the response is controlled for by repeating the question to me. The MCID has not been as well-validated for use across different groups of patients, so-called ‘between groups’ comparisons. Thus, the MCID is best used as a measure of how a given sample of patients change in their own responses to an intervention over time but not necessarily how one group of patients’ responses compare to a different group that has undergone a different treatment.”
“At times, however, these distinctions have been disregarded and the MCID has been applied to circumstances beyond those for which it has been validated. The results of such applications of the MCID are subject to considerable uncertainty. Several of these concerns were noted when the AAOS re-evaluated its recommendation regarding intraarticular hyaluronic acid (IAHA) injections for knee OA. Through use of an historical MCID and comparing between groups, the AAOS re-analyzed the same studies that had yielded an indeterminate prior recommendation and changed to strongly recommending against IAHA injections. This change, and the science upon which it was based, has resulted in considerable controversy, particularly as several payers have subsequently determined that IAHA are not ‘medically necessary,’ citing the AAOS guideline as evidence.”
Asked how to use this information to counsel patients, Dr. Concoff commented, “This is a challenging question. Most patients are not particularly interested in issues of outcome measurement, study design and methods, or even guideline recommendations. Patients simply want to feel better. Yet when we, as healthcare providers, advocate for one or another treatment for the specific patient before us in clinic, we have several tasks. First, we must understand the results of the clinical trials that have investigated the impact of the treatment in question. Second, we must recognize that patients included in clinical trials often are not reflective of the typical patient in our clinics, as the latter would often have been excluded from clinical trials on the basis of co-morbidities, disease severity, prior treatments, or other factors.”
Study results ≠ real world results…
“Finally, at least in part as a consequence of these differences, it is often very difficult to translate the results of clinical trials, much less guidelines, to the likelihood of benefit, or harm, from the treatment of an individual patient in the clinic. Perhaps this explains why compliance with the AAOS Guideline from a recent survey of orthopedists from the American Association of Hip and Knee Surgeons was remarkably low as, despite the recommendation against IAHA, the most frequently recommended treatment for moderate (Kellgren and Lawrence Grade 2 or 3) knee OA was IAHA. (Carlson VR, Ong AC, et al. Compliance with the AAOS guidelines for treatment of osteoarthritis of the knee: A survey of the American Association of hip and knee surgeons. J Am Acad Orthop Surg, 2018; 26: 103-17). Regardless, for patients who had previously experienced dramatic benefit from IAHA injections but who bear insurance that no longer cover this treatment, we are left to attempt to explain the vagaries of the science and coverage decisions even as we seek to support such patients efforts to obtain the medications that had previously been transformative for them.”
A tool is at the mercy of its user…
“The era of PROs is upon us. We must consider in granular detail the subjective experience of the patient in response to the treatments that we provide. Yet measures such as the MCID are merely tools. These tools may be applied in the context for which they were intended, in which case they serve to clarify the value of medications that we provide. However, such tools may also be used bluntly, in circumstances alternative to that for which they were designed. In the worst such misapplication of purpose, the classic adage applies, ‘If you have a hammer everything looks like a nail.’”
“In the end, the entire controversy that orbits about the non-operative treatment of OA of the knee indicates that our present armamentarium is woefully inadequate. Only through further research leading to the development of medications or treatments that not only address patients’ suffering but also slow the damage to the joint, or even reverse such pathology, will we be free from the vexing questions of how much benefit matters in a given clinical circumstance. With such new treatments in hand, we can move on from treatments that generate the MCID to its companion PRO measure, the Patient Acceptable Symptoms State (PASS), which relates that degree of improvement that a patient determines to be ‘good enough,’ … although, upon reflection, that too may give rise to debate, controversy and confusion.”