The Minimal Important Difference in Dermatology Clinical Trials
In this interview, Reinhart Speeckaert, MD, PhD, met with The Dermatologist to discuss the importance and reliability of the minimal important difference (MID) in dermatology clinical trials, the inappropriate generalization of previously reported MIDs for patient populations with different disease characteristics, and steps that can be used to avoid this pitfall.
Dr Speeckaert’s clinical work focuses on inflammatory skin disorders, particularly vitiligo, but also on a wide range of other immune-mediated diseases. He holds 2 PhDs: the first in dermatology, which focused on the immunosurveillance of melanocytes in conditions ranging from melanoma to vitiligo, and a second in clinical chemistry, which dealt with the clinical implications of the genetic polymorphism in the haptoglobin gene.
The Dermatologist: Can you start by discussing patient-reported outcome measures, or PROMs, and their importance in clinical trials?
Dr Speeckaert: To know whether a treatment is working, we need instruments to measure the disease severity at a certain moment, and different aspects of a skin disease can be assessed. For example, how extensive a disease is and how the lesions look. Are they very red or are they very thick? Also, how bothersome the symptoms are for the patient. In most cases, symptoms such as itch or loss of sleep can best be measured by the patient. Then we speak about PROMs or patient-reported outcome measures. For more objective evaluations measured by clinicians, for example, to assess the extent of a skin disorder, we then talk about ClinROMs or clinician-reported outcome measures. These PROMs and ClinROMs are used to evaluate the evolution of skin disorders in clinical trials, so it is very important that they are accurate to know if you can trust the results of a clinical trial.
The Dermatologist: You state that the minimal important difference, or MID, represents the point at which a difference in an outcome measure, such as the Dermatology Life Quality Index, is important enough to warrant a change in treatment. Can you expand on this concept and discuss the use of MIDs in clinical trials?
Dr Speeckaert: I will give an example. If we have a patient with extensive skin disease, perhaps 80% of their whole body is covered, even when the treatment makes only a small difference, it can be statistically significant. For example, from 80% to 78%. For the patient, a 2% improvement will likely not make a difference in the burden of the disorder, but you have a difference that is statistically significant. To know whether the difference impacts the patient’s quality of life, the concept of the MID was developed. It is the value at which most patients would be willing to have the treatment. That is why MIDs are so popular now in clinical trials, because they say it is not only statistically significant, but patients will also benefit from the treatment.
The Dermatologist: Part of your investigation was to determine the credibility of MIDs in dermatology literature and whether it is reasonable to extrapolate previously reported MIDs to other patient groups. Can you tell us about your results in this area?
Dr Speeckaert: We looked at articles with MIDs, and the values were often unreliable. In some cases, it was just that there was a lack of detailed information in the manuscript, or the number of patients was not high enough. We just need more additional research to solve this. However, there was another problem. In some articles, only the opinion of the physician was asked. Then we cannot know if the treatment was important to the patient. Other times, patients were only asked whether they could notice a difference and not to measure how the difference impacted their lives. In other cases, patients were not asked whether this was the smallest difference they would consider important. Then the MID value could be overestimated for some treatments. There were also different interpretations of the definition of the MID, so readers could not know for sure what the value meant. Is this really the value at which point the treatment becomes beneficial for the patient? Or have the researchers used another definition?
   Concerning the extrapolation of results, many researchers took the MID value of a certain publication and then used it for their study. But it was often a different diagnosis from the earlier study and the patients had a different disease severity. This can have important implications because, for example, if you have a patient with severe itch, 10 out of 10 itch severity on a scale of 0 to 10, maybe they will not benefit from a small reduction. If the patient goes from 10 to 8, they will still have loss of sleep. And if the other patient population is specific to mouth itch, for example, in a patient with 4 out of 10 or 2 out of 10 itch severity, a reduction of the same 2 points will have a major implication. You cannot just take a value and apply it to another patient population without comparing the same disease or severity.
The Dermatologist: You note that inappropriate generalization of previously reported MIDs to patient populations with different disease characteristics is a major concern. Can you share steps that can be used to avoid this pitfall?
Dr Speeckaert: The first thing everyone must acknowledge is that there is not a single MID value, it depends on the patient population. I think researchers must look at this, and if there is not a real MID value for their population available, then they can say, “We have this value, but it will not be exactly the same, so it can be under- or overestimated, but it is the best we have.” I think that is the best approach, so the reader is well-informed.
The Dermatologist: Is there anything else you would like to share with your colleagues related to this topic?
Dr Speeckaert: When there are several definitions of the same concept, it can be complicated and then it becomes messy. The MID has important implications because it is determining whether a treatment is useful. If we have practical tables, we can look at which MID value is indeed correct and to which patient population it applies. This will let us know that the MID value can be trusted. We must be careful when interpreting results, and we hope to be helpful in giving some guidance on that with our publication.
Reference
- Speeckaert R, Belpaire A, Herbelet S, Speeckaert MM, van Geel N. The meaning and reliability of minimal important diff erences (MIDs) for clinician-reported outcome measures (ClinROMs) in dermatology—a scoping review. J Pers Med. 2022; 12(7):1167. doi:10.3390/jpm12071167
