MINIMAL CLINICALLY IMPORTANT DIFFERENCE
The minimum clinically important difference (MCID) may provide a simple and efficient metric to empower physicians to maximize the utility of PROs in the clinical setting.
In Orthopedics and other intervention-based specialties, patient-reported outcome measures (PROMs) are collected on a routine basis and the trend continues to rise. The question is “how the heck do use the data now that you have it?”
Understanding and interpreting the PRO score can be complicated and it does not help that every tool has a unique scoring system. Most clinicians and patients lack the ability to place the PROs in context and as a result, do not use PROs in clinical decision making, which is a shame because that is why these tools exist in the first place! It isn’t negligence on the provider's part, it’s just complicated and nuanced. Surgeons receive hundreds if not thousands of hours in training, refining the ability to complete physical exams and evaluate a patient’s recovery clinically, but really limited exposure to using PROs in practice. The MCID may provide a simple and efficient metric to empower physicians to maximize the utility of PROs in the clinical setting.
The MCID is based on the theory that a change in PRO score is correlated with clinically meaningful improvement. Take for example a patient after receiving a knee replacement. The patient completes the KOOS (Knee Osteoarthritis Outcome Score) assessment before surgery and at 3 months post-operatively. The patient’s pre-op KOOS score was 40 and post-op score was 81. What does that mean? Does the patient actually feel better? Was the surgery worth it? That’s what the MCID is trying to define. Rather than the surgeon being presented with data displaying an arbitrary score, they are presented with the plus/minus of the MCID. Using the KOOS example above, say that the MCID for a KOOS score is 20- is the patient above the MCID? Yes. By how much? 1 point. Think of it like a basketball game. It’s like saying ‘your team scored 90 points last night’ vs. ‘your team won by 10 pts’. Knowing if they won or lost is the most relevant piece of data, followed by the margin. Did they squeak by with a couple of points or blow the other team up. Breaking the data down like this in a ‘did you win: yes or no’ and ‘by how much,’ is much more manageable, and potentially meaningful, than trying to memorize a bunch of PRO tools scores.
There are two (2) basic ways to calculate the MCID:
The distribution method: This method consists of halving the standard deviation (it is similar to an average variation in a group of patients) for the change in PRO scores. This is math plain and simple (or not that simple for most of us).
The anchoring method: This method uses averaging the change in PRO scores for all patients that reported a minimum of a one (1) point increase in another quality of the life-based score. One of the most common anchors is a question from PROMIS10 relating to a patient’s interpretation of their overall physical health or quality of life.