How much improvement in the SST, ASES, and VAS score is clinically significant?

Determining the minimal clinically important difference for the American Shoulder and Elbow Surgeons score, Simple Shoulder Test, and visual analog scale measuring pain after shoulder arthroplasty.

Primary anatomic total shoulder arthroplasty (TSA), primary reverse TSA, or hemiarthroplasty was performed in 326 patients. The SST score, ASES score, and VAS pain score were collected preoperatively and at a minimum of 2 years postoperatively (mean, 3.5 years).

The minimal clinically important differences (MCIDs) were calculated for the ASES score, SST score, and VAS pain score using a 4-item anchor question evaluating improvement after treatment. Patients were asked the following: “Since your shoulder replacement surgery, please rate your response to treatment: A, none—no good at all, ineffective treatment; B, poor—some effect but unsatisfactory; C, good—satisfactory effect with occasional episodes of pain or stiffness; D, excellent—ideal response, virtually pain free.” Patients were classified by the anchor question as having “no change” (A group [none] and B group [poor] combined) or “change” (C group [good]). The D group (excellent) was not included in the analysis because this was considered beyond minimal change.

The MCIDs for the ASES score, SST score, and VAS pain score were 20.9 (P < .001), 2.4 (P < .0001), and 1.4 (P = .0158), respectively. 

Duration of follow-up and type of arthroplasty (anatomic TSA vs reverse TSA) did not have a significant effect on the MCIDs (P > .1) except shorter follow-up correlated with a larger MCID for the ASES score (P = .0081). 

Younger age correlated with larger MCIDs for all scores (P < .024). Female sex correlated with larger MCIDs for the VAS pain score (P = .123) and ASES score (P = .05).

Patients treated with a shoulder arthroplasty require a 1.4-point improvement in the VAS pain score, a 2.4-point improvement in the SST score, and a 21-point improvement in the ASES score to achieve a minimal clinical importance difference from the procedure.

Comment: MCID is one way of looking at the amount of improvement, but it has a problem. Consider two patient having a shoulder arthroplasty, each with an improvement of 3 in the SST score (both exceeding the 2.4 MCID improvement).

Their outcomes are not the same. For that reason we use both the preoperative to postoperative change in the SST as well as the percent of maximal possible improvement to characterize the result:

Here we can see that Smith only improved by 27% of the maximal possible improvement, whereas Jones improved by 75% of the maximal possible improvement (even though the improvement in both cases exceeded the MCID).

We've found that the concept of %MPI is easier to explain to patients than MCID.

===

Consultation for those who live a distance away from Seattle.

Click here to see the new Shoulder Arthritis Book.

Click here to see the new Rotator Cuff Book

Information about shoulder exercises can be found at this link.

Use the "Search" box to the right to find other topics of interest to you.

You may be interested in some of our most visited web pages including:shoulder arthritis, total shoulder, ream and run, reverse total shoulder, CTA arthroplasty, and rotator cuff surgery as well as the 'ream and run essentials'

See from which cities our patients come.

See the countries from which our readers come on this post.

Rotator cuff surgery, measuring outcome: the MCID vs the I/MPI

Investigating minimal clinically important difference for Constant score in patients undergoing rotator cuff surgery.

The minimal clinical important difference (MCID) is the smallest change in an outcome score for the treatment of a clinical condition that is meaningful and important for the patient. In order to determine the minimal clinically important difference for the Constant Score and cuff repair, the authors studied a prospectively collected cohort of 802 consecutive shoulders with arthroscopically treated partial- or full-thickness rotator cuff tears. The Constant score was measured preoperatively and at 3 months and 1 year postoperatively. At follow-up visits, the patients were asked whether the shoulder was better or worse after the operation compared with the preoperative state? This 2-level question was used as an indicator of patient satisfaction to calculate the MCID for the Constant score.

The preoperative Constant score was 56.2 (SD 17.4) in male patients, and 48.2 (SD 15.6) in female patients. Postoperatively at 3 months, the scores were 65.1 (SD 16.1) in male patients, and 56.8 (SD 15.5) in female patients. At 1 year, the scores were 79.0 (SD 14.9) in male patients, and 71.0 (SD 14.3) in female patients. At 3 months postoperatively, 92.2% of male patients and 87.2% of female patients were satisfied with the outcome (P = .027); at 1 year, the satisfaction was 93.2% and 89.5%.

The 3-month mean change estimate of MCID was 10.4 points.

Interestingly, the pathology and surgical procedures were varied:  27% of the patients had "partial" cuff tears and were treated with "subacromial decompression", 13% had irreparable cuff tears and were treated with "partial" cuff repairs. 96% had involvement or the supraspinatus, 24% had involvement of the infraspinatus, and 38% had involvement of the subscapularis. 36% had a distal clavicle excision and 44% had biceps tendon surgery. So this is not a study of patients having cuff repair only. Even more interestingly, the mean change in Constant score was not affected by the severity of the cuff tear, working status of the patient, history of trauma.

A limitation of the MCID is that it it is 'fickle'. Even in in this study, different statistical approaches yielded 5 different values ranging from 2 to 16. Another limitation is that it needs to be estimated for each different diagnosis and each different outcome metric. A final limitation is that a defined value of MCID does not mean the same thing across the spectrum of values, for example an improvement of the Constant score from 0 to 11 is not the same thing as an improvement of the Constant score from 85 to 96; a patient with a preoperative score of 90 would not be capable of achieving the MCID of 10.4.

An alternative is the ratio of improvement to the maximal possible improvement (I/MPI). With this approach a change in Constant score from 0 to 11 would represent  11/100 or 11% of the possible improvement and a change from 85 to 96 would represent 11/15 or 73% of the maximal possible improvement. The I/MPI does not require complex statistics, can be applied to any pathology, and can be applied to any outcome measurement tool. Finally, it answers the question patients have, "How much better am I likely to be after this procedure?"

==
Consultation for those who live a distance away from Seattle.

**Check out the new (under construction) Shoulder Arthritis Book - click here.**

To see the topics covered in this Blog, click here

Use the "Search" box to the right to find other topics of interest to you.

You may be interested in some of our most visited web pages including:shoulder arthritis, total shoulder, ream and run, reverse total shoulder, CTA arthroplasty, androtator cuff surgery as well as the 'ream and run essentials'

See from which cities our patients come.

See the countries from which our readers come on this post.

minimal clinically important difference (MCID) vs I/MPI

Investigating minimal clinically important difference for Constant score in patients undergoing rotator cuff surgery

There has been a recent fascination with the minimal clinically important difference as a unit of treatment effectiveness. These authors prospectively analyzed collected cohort of 802 consecutive shoulders with arthroscopically treated partial- or full-thickness rotator cuff tears.. The Constant score was measured preoperatively and at 3 months and 1 year postoperatively. At follow-up visits, the patients were asked a simple 2-stage question: Is the shoulder better or worse after the operation compared with the preoperative state? This single 2-level question was used as an indicator of patient satisfaction and as an anchor to calculate the MCID for the Constant score.

The preoperative Constant score averaged 53.1 (SD 17.2) in all patients, 56.2 (SD 17.4) in male patients, and 48.2 (SD 15.6) in female patients. Postoperatively at 3 months, the scores were 61.7 (SD 16.4) in all patients, 65.1 (SD 16.1) in male patients, and 56.8 (SD 15.5) in female patients.  At 3 months postoperatively, 92.2% of male patients and 87.2% of female patients were satisfied with the outcome (P= .027). The 3-month mean change estimate of MCID was 10.4 points.

Comment: As surgeons, our goal is to restore as much of the patient's lost function and comfort as possible. The problem with the MCID is that an improvement of 10 points from a Constant score of 10 to 20 is not the same as an improvement of 10 points from a Constant score of 60 to 70, in that it does not reflect the amount of improvement in relation to the improvement possible.  As an alternative, we have used the I/MPI or percent of possible improvement as metric for measuring the result of treatment. The I/MPI is easily calculated as:

(score at the time of follow-up - score before surgery) 

divided by 

(perfect score - score before surgery) 

The beauty of the I/MPI  is that it can be used for any scoring system (Constant, SST, WOMAC, UCLA,  ASES, etc). Thus, taking the example above, if a perfect score is 100 and the patient records 60 before and 70 after surgery (improving by the MCID), the I/MPI  is

(70 - 60)/(100 - 60) 

or 

25% (the patient improved 25% of the possible improvement).

On the other hand, if the patient records 10 before and 20 after surgery (again improving by the MCID), the I/MPI  is

(20-10)/(100 - 10) 

or 

11% (the patient improved 11% of the possible improvement).

The I/MPI  (percentage of possible improvement) has the following advantages:

(1) It measures what is most important to the patient (how much better is the shoulder after treatment).

(2) It enables comparison of studies done using different outcome metrics (i.e. 'normal' is converted to 100%)

(3) It avoids the problem of having different MCIDs for each condition (arthritis, instability, cuff disease), for different genders, and for different instruments (ASES, SST, Constant). 

Try it, you will like it! 

--

--

To see the topics covered in this Blog, click here

Use the "Search" box to the right to find other topics of interest to you.

You may be interested in some of our most visited web pages including:shoulder arthritis, total shoulder, ream and run, reverse total shoulder, CTA arthroplasty, and rotator cuff surgery as well as the 'ream and run essentials'


See from which cities our patients come.


See the countries from which our readers come on this post.

The results of shoulder arthroplasty - how to measure. Impact of Total Shoulder Arthroplasty on Generic and Shoulder-Specific Health-Related Quality-of-Life MeasuresJBJS

The JBJS recently published: Impact of Total Shoulder Arthroplasty on Generic and Shoulder-Specific Health-Related Quality-of-Life Measures

This is an important article. It emblemizes the shift from evaluating results of shoulder joint replacement by physician-measured metrics (e.g. range of motion) to patient-measured metrics (e.g. comfort and function) as well as overall well-being. In this systematic review, the SF 36 self-assessment of overall physical function was highly significantly improved - meaning that the improvement in the shoulder resulted in a significant benefit to overall functioning. They also found highly improved scores on three shoulder-specific metrics: the Constant, American Shoulder and Elbow Surgeons, and the  Simple Shoulder Test .*

The value of patient self-assessment is that it focuses attention on what is important to the individuals we treat. As the authors point out, it is essential to use the same metric before and after treatment so that the difference can be measured.

We use the SF 36 not only as a benchmark for the patient's status before treatment, but also as an indicator of health issues that may be correlated with the outcome of surgery, such as emotional and social functioning. We use the Simple Shoulder Test because it is the only one of the three shoulder-specfic metrics in this study that does not require the patient to return to the office for an examination - by increasing the convenience of followup for the patient, it has the advantage of reducing the number of patients lost to followup (the downfall of many clinical studies).

The value of this study extends far beyond showing that the reports confirm the value of shoulder arthroplasty. In our opinion, its greatest value is in establishing patient self-assessment metrics (the SF 36 and the SST) by which we can determine which patients are NOT benefitting from shoulder arthroplasty so we can ask, as Codman said "Why Not?".  It is often stated that '85% of patients receive a good or excellent result'. We should use the tools described there to learn what is it about that other 15% that led to suboptimal results and what can we do to reduce that number. Only by studying failures, as we have tried to do here and here, can we learn out to get better. Imagine a world in which shoulder surgeons shared the data on their patients who did not improve their SST and SF 36 scores after shoulder joint replacement with a goal of identifying the features common to those shoulders and individuals!

*The authors state that the minimal clinically important difference has not been defined for the Simple Shoulder Test. Actually, this definition is included in our recent publication: The Prognosis for Improvement in Comfort and Function After the Ream-and-Run Arthroplasty for Glenohumeral Arthritis: An Analysis of 176 Consecutive Cases. To quote from that manuscript: "We utilized an alternative approach in evaluating the benefit of shoulder arthroplasty by determining the percentage of the total possible improvement realized by the patient. In this calculation, the improvement realized by the patient is divided by the total possible improvement for the patient (i.e., the difference between the maximal possible score on the SST [12 points] and the preoperative score). We refer to this as the I/MPI (improvement / maximal possible improvement). Thus, the percentage of the total possible improvement is calculated as:

(SST total score at the time of follow-up - SST total score before surgery X 100%) 

divided by 

(12 points - SST total score before surgery) 

For purposes of this analysis, we defined the MCID as an improvement of 30% of the total improvement possible. Thirty percent is the highest percent improvement required for an MCID in the shoulder literature. It is of interest that since the average preoperative SST score was 4 points, the average maximal possible improvement was 8 points. Thirty percent of this value would be 2.4 points, a value that lies between the values separately proposed by Roy et al. and Tashjian et al. The value of the method selected for our analysis lies in the fact that it avoids the ceiling effect, and the MCID is normalized by the maximal possible improvement for each patient, rather than consisting of a fixed value applied to all patients irrespective of their maximal
possible improvement."

----

If you have suggestions for topics you'd like us to address in this blog, please send an email to

shoulderarthritis@uw.edu.

Use the "Search the Blog" box to the right to find other topics of interest to you.

You may be interested in some of our most visited web pages including:shoulder arthritis, total shoulder, ream and run, reverse total shoulder, CTA arthroplasty,  and rotator cuff surgery.

See the countries from which our readers come on this post.