Three weeks after a shoulder replacement, many patients ask, “Am I recovering the way I should be?” A study published recently in Clinical Orthopaedics and Related Research (CORR) offers a way to answer the question with a chart [1].
Borrowing the idea of the pediatric growth chart, researchers at the University of Colorado and Intermountain Health used records of 1,392 patients who had anatomic or reverse total shoulder arthroplasty to match each new patient with the 25 most similar earlier patients, and drew a personalized curve of expected DASH disability scores over the first 120 days after surgery. A free online tool puts the method in any clinician’s hands [2].
The idea is sound and the predictions were unbiased. For an individual patient, though, they are imprecise: a typical prediction missed by 11 points on the 100-point scale — about the size of the minimal clinically important difference — and the band meant to capture 80 percent of patients was 34 points wide. The curve leans heavily on the patient’s first postoperative score; when we tried the tool, changing a single day-14 score moved the predicted day-120 score by about 9 points.
More important, the chart stops at 120 days. Patients considering surgery want to know how their shoulder is likely to be doing early on, as well as in ten or twenty years. As detailed in a recent post [3], getting that information is critical but challenging. That post pointed out that in a 2026 Utah study in which the surgeon reached out to 1,028 patients by telephone, text and email, followed by automated messages, only 49 percent completed two-year follow-up; younger patients and those having instability surgery were the most often lost [4]. Without follow-up, we cannot know how the missing patients are doing.
In our study of rotator cuff tears, patients who stopped returning mailed questionnaires had a mean Simple Shoulder Test score of 3.3 of 12 on their last one, compared with 7.9 for those who kept responding — yet when 48 of them were reached by telephone about three years later, their mean was 8.9 [5]. A patient’s last report may not describe the patient now. And some failures occur late: survival free of radiographic glenoid loosening with one anatomic design fell from 99 percent at five years to 34 percent at fifteen [6].
So what are the simplest questions we could ask each patient to track their recovery?
Here is a short set of questions, asked the same way every time, that a patient can answer by text in about a minute:
· Are you satisfied with your recovery after your shoulder surgery, Yes or No?
· Compared with before surgery, is your shoulder better, the same or worse?
· Have you had, or been advised to have, another operation on this shoulder, Yes or No?
· The 12 yes-or-no questions of the Simple Shoulder Test, which show which functions the patient has and which they lack [7].
The plan is equally simple. Before surgery, ask permission to stay in touch and record several ways to reach the patient. At one year, send the questions from the surgeon’s own office. Repeat them at 2, 5, 10, 15 and 20 years by scheduled message. Follow anyone who does not answer with a reminder, a phone call, a mailed card and a family contact, and record results along with the percentage of eligible patients responding. One staff member, a texting app and an hour or two a month may be sufficient.
Even if fewer patients respond than in the Utah study, the answers provide information important to prospective surgical patients and to surgeons, who can learn from their suboptimal results [8].
In detail
In the CORR study, the authors used Intermountain Health’s rehabilitation outcomes database, using data collected during therapy sessions in 28 outpatient physical therapy clinics in Utah from 2017 to 2022 [1]. Of 2,322 eligible patients who had anatomic or reverse total shoulder arthroplasty by one of 26 surgeons, 1,392 (60%) had at least three postoperative DASH scores, one of them within 45 days of surgery. They were split into a training group (928) and a testing group (464).
Each test patient was matched with the 25 training patients most similar in age, gender, BMI, insurance, a Utah-specific area deprivation index, type of arthroplasty, and the first postoperative DASH score and its timing (median, day 19). The matches’ actual DASH scores were used to model patients’ expected curve, with 50%, 80% and 90% bands. How the model performed in the testing group:
Measure | Result | What it means for one patient |
Mean error (bias) | −2 points | Predictions were not systematically high or low. |
Mean absolute error | 11 points | A typical prediction missed by about one MCID. |
50% band | ±9 points (49% coverage) | Half of patients landed within 9 points of their curve. |
80% band | ±17 points (76% coverage) | Slightly fewer patients fell inside than intended. |
90% band | ±22 points (87% coverage) | The band spans 44 points of the 100-point scale. |
The authors’ case examples show why the comparison group matters. One patient’s DASH scores were consistently better than the typical (median) score of all the earlier patients, which would suggest he was recovering unusually well. But compared with the 25 patients most like him, his scores were just average. Another patient’s scores were close to the typical score of all the earlier patients, which would suggest he was on track. But compared with the 25 patients most like him, he had some of the worst scores between 60 and 90 days after surgery — a signal that something might be wrong.
Trying the tool
The online tool [2] asks for age, BMI, gender, type of arthroplasty, payer, a deprivation quintile and one or more postoperative DASH scores; there is no field for a preoperative score. We entered the tool’s default patient (68-year-old woman, anatomic arthroplasty, BMI 30, commercial insurance, average deprivation) with a DASH of 32 at day 28, and changed only her day-14 score. Values are read from the plotted curves and are approximate.
Day-14 DASH | Predicted DASH at day 120 | How the same day-28 score of 32 looks |
46 | about 14 | On the predicted curve (about the 50th percentile) |
70 | about 23 | Better than about 90% of similar patients |
Strengths and limits
Strengths. The question is the right one. Minimal clinically important differences are anchored at distant time points and vary widely with the method used to derive them [9]. A recovery percentile is easy to share among surgeon, therapist and patient, the model was tested on patients it had not seen, and the tool is openly available.
Limits. The bands are wide. They captured slightly fewer patients than intended. Only 20 of 1,392 patients had a preoperative DASH, so the curve is anchored to early recovery rather than to the starting point, and it cannot be used to inform patients before surgery. The cohort includes only patients who stayed in therapy long enough to record three scores; those who recover quickly may stop early, and those who struggle contribute more scores. Diagnosis, prior surgery, comorbidity, smoking, mental health, the surgeon and the rehabilitation protocol were not modeled. The window is 120 days, and the effect of the chart on treatment decisions or longerterm outcomes is untested [1].
Why long-term follow-up matters — and why it is hard
Our recent post [3] adds three points to those above. Long-term studies lose most of their patients: ten-year studies of reverse arthroplasty lost a weighted mean of 63% [10]. Implant survival is not the same as patient outcome: in one reverse arthroplasty series, survival free of revision was 93% at ten years, yet clinical scores had declined significantly from the medium-term evaluation [11]. And a simple, repeated self-assessment can reveal failure: in shoulders later revised for glenoid loosening, SST scores fell back to preoperative levels, often more than a year before the patient returned to the office for revision surgery[12]. For each surgeon, the goal is to identify and learn from every poor result, wherever and whenever it occurs [8]. In this way the surgeon can improve their method.
A practical plan
· Before surgery. Ask permission to check in “for as long as you’ll let me.” Record mobile number, email, mailing address and one family contact, and a baseline SST.
· The first months. Use whatever helps judge whether recovery is on track, including tools like the growth chart, with their limits in mind [1].
· At one year. One text from the surgeon’s office, one patient at a time, with a call for anyone not doing well. A practice doing 100 cases a year has about eight patients due each month.
· Later. The same questions at 2, 5, 10, 15 and 20 years, sent as scheduled messages, so that staff time goes to the patients who do not answer and those not doing well.
· When no one answers. One reminder, then a call from a real person, then a mailed card, then the family contact.
· Deaths and outside revisions. Check the medical record and shared-record systems. A death or a revision is an outcome, not a reason to drop a patient.
· Every quarter. Report the capture rate — patients who answered divided by patients due — with every result, and review each poor result by name.
Following the curve
Long-billed Curlew
Follow on twitter: https://twitter.com/shoulderarth
Follow on facebook: https://www.facebook.com/frederick.matsen
Follow on LinkedIn: https://www.linkedin.com/in/rick-matsen-88b1a8133/
The author of this post has no financial relationship with any orthopaedic company.
References
1. Graber J, Jin X, Juarez-Colunga E, Dibblee P, Woodfield D, Minick KI, Chamberlain AM, Bade M, Stevens-Lapsley J. Do personalized recovery trajectories enhance the utility of patient-reported outcome measures after orthopaedic surgery? Clin Orthop Relat Res. 2026;484(8):1568-81. doi:10.1097/CORR.0000000000003901
2. University of Colorado RESTORE group. Personalized DASH recovery trajectories after TSA [web tool]. https://cu-restore.shinyapps.io/TSA_recovery_v1/ (accessed November 14, 2026).
3. Matsen FA 3rd. How do our surgical patients do in the long term? A critical but difficult question. Shoulder Arthritis Blog, November 11, 2026. https://shoulderarthritis.blogspot.com/2026/11/long-term-followup-of-our-patients.html
4. Soule S, Smith A, Staten T, Kirkham M, Zhang C, Presson A, Joyce CD, Tashjian RZ, Chalmers PN. Predictors of loss to long-term follow-up after shoulder surgery. J Bone Joint Surg Am. 2026;108(19):1536-42. doi:10.2106/JBJS.25.00767
5. Norquist BM, Goldberg BA, Matsen FA 3rd. Challenges in evaluating patients lost to follow-up in clinical studies of rotator cuff tears. J Bone Joint Surg Am. 2000;82(6):838-42.
6. Young A, Walch G, Boileau P, Favard L, Gohlke F, Loew M, Molé D. A multicentre study of the long-term results of using a flat-back polyethylene glenoid component in shoulder replacement for primary osteoarthritis. J Bone Joint Surg Br. 2011;93(2):210-6. doi:10.1302/0301-620X.93B2.25086
7. Lippitt SB, Harryman DT 2nd, Matsen FA 3rd. A practical tool for evaluating function: the Simple Shoulder Test. In: Matsen FA 3rd, Fu FH, Hawkins RJ, eds. The Shoulder: A Balance of Mobility and Stability. Rosemont, IL: American Academy of Orthopaedic Surgeons; 1993:501-18.
8. Menendez ME, Matsen FA 3rd. Learning from surgical failures. J Bone Joint Surg Am. 2026;108(8):547-8. doi:10.2106/JBJS.25.01110
9. Kolin DA, Moverman MA, Pagani NR, et al. Substantial inconsistency and variability exists among minimum clinically important differences for shoulder arthroplasty outcomes: a systematic review. Clin Orthop Relat Res. 2022;480(7):1371-83.
10. Biner M, Klotz S, Andronic O, Karczewski DC, Zingg L, Karl W, Kriechling P. Long-term outcomes following reverse total shoulder arthroplasty: a systematic review with a minimum follow-up of 10 years. JBJS Open Access. 2025;10(2):e25.00025. doi:10.2106/JBJS.OA.25.00025
11. Bacle G, Nové-Josserand L, Garaud P, Walch G. Long-term outcomes of reverse total shoulder arthroplasty: a follow-up of a previous study. J Bone Joint Surg Am. 2017;99(6):454-61. doi:10.2106/JBJS.16.00223
12. Buckingham BP, Parsons IM 4th, Campbell B, Titelman RM, Smith KL, Matsen FA 3rd. Patient functional self-assessment in late glenoid component failure at three to eleven years after total shoulder arthroplasty. J Shoulder Elbow Surg. 2005;14(4):368-74.
