Mapping physical functions of the shoulder to American Shoulder and Elbow Surgeons and Patient-Reported Outcomes Measurement Information System scores

J Shoulder Elbow Surg (2019) -, 1–12

www.elsevier.com/locate/ymse

Mapping physical functions of the shoulder to American Shoulder and Elbow Surgeons and Patient-Reported Outcomes Measurement Information System scores Aaron M. Roberts, MD, Ilya Voloshin, MD* Department of Orthopaedic Surgery and Rehabilitation, University of Rochester, Rochester, NY, USA Background: We sought to correlate physical functions of the shoulder to American Shoulder and Elbow Surgeons (ASES) and Patient-Reported Outcomes Measurement Information System (PROMIS) scores. Methods: We reviewed 3300 patient encounters with completed ASES scores, representing 2447 patients. Patients were seen for shoulder-related complaints. The most common diagnoses were rotator cuff disease (56%) and arthritis (9%); 54% and 46% of encounters were in operatively and nonoperatively treated patients, respectively. A total of 2632 PROMIS Physical Function (PF), 2574 PROMIS Pain Interference (PI), and 959 PROMIS Upper Extremity (UE) scores were simultaneously collected with the ASES form. The ASES form specifically asks about the ability to perform 8 physical functions. Receiver operating characteristic curves were calculated to determine 90% positive predictive value (PPV) and 90% negative predictive value (NPV) cutoffs for the ability to perform at high function for the ASES and PROMIS-PF, -UE, and -PI scores for the entire shoulder cohort and for rotator cuff disease and arthritis subgroups. Results: ASES scores had consistently excellent ability, PROMIS-UE scores had reasonable to excellent ability, and PROMIS-PF and PROMIS-PI scores had overall reasonable ability to determine high- and lowfunction states. For reaching a high shelf in the rotator cuff disease subgroup, the 90% NPV and PPV cutoff scores were 41 and 66, respectively, for the ASES instrument. For reaching a high shelf in the arthritis subgroup, the cutoff scores were 50 and 78, respectively, for the ASES instrument. The 90% NPV and PPV cutoffs for each score, physical function, and diagnosis group were depicted by visual representations (‘‘maps’’) for easier interpretation. Conclusion: Shoulder physical functions were mapped to outcome scores. Physical function mapping adds clinical meaning to the orthopedic literature, facilitating improved, more-informed decision making between physicians and patients. Level of evidence: Basic Science Study; Validation of Outcome Instruments Ó 2019 Journal of Shoulder and Elbow Surgery Board of Trustees All rights reserved. Keywords: Patient-reported outcomes; mapping; physical function maps

This study was presented as a Rapid Fire presentation at the 2018 Society of Military Orthopaedic Surgeons Annual Meeting, Dec 10-14, Keystone, CO, USA and as a poster presentation at the 2019 American Academy of Orthopaedic Surgeons Annual Meeting, March 12-16, Las Vegas, NV, USA.

Institutional Review Board approval for this study was received from the University of Rochester. *Reprint requests: Ilya Voloshin, MD, University of Rochester Medical Center, 601 Elmwood Ave, Rochester, NY 14642, USA. E-mail address: [email protected] (I. Voloshin).

1058-2746/$ - see front matter Ó 2019 Journal of Shoulder and Elbow Surgery Board of Trustees. All rights reserved. https://doi.org/10.1016/j.jse.2019.08.017

2

A.M. Roberts, I. Voloshin Table I

Shoulder cohort demographic characteristics Data

Encounters, n Patients, n Age, mean  SD, yr Sex, % Male Female Management, n (%) Nonoperative Operative Preoperative encounter Postoperative encounter Diagnosis, % Rotator cuff disease Arthritis Instability Labrum or biceps pathology Adhesive capsulitis AC arthritis or separation Fracture Calcific tendinitis Scapulothoracic disorder AVN Pectoralis tear Sternoclavicular injury Neuritis Infection Tumor Undetermined

3300 2447 54  27 62 38 1530 (46) 1770 (54) 836 (47) 934 (53) 56 9 7 6 3 2 1 1 <1 <1 <1 <1 <1 <1 <1 14

SD, standard deviation; AC, acromioclavicular.

Recently, increased emphasis has been placed on using patient-reported outcome (PRO) measures to assess the effectiveness of treatments.3 The Patient-Reported Outcomes Measurement Information System (PROMIS) was developed by the National Institutes of Health to allow for efficient and effective assessment of PROs across a variety of disease states. PROMIS seeks to overcome the limitations of traditional PROs, including the narrow focus for a given PRO, the use of multiple PROs for a given diagnosis, and the burden of administering these PROs. Thus far, PROMIS has shown utility in evaluating outcomes in orthopedic and shoulder-specific conditions.5,15 Specifically, PROMIS instruments have been analyzed in patients with rotator cuff pathology, shoulder instability, and shoulder arthroplasty, showing validity against traditional shoulderspecific PROs with a decreased question burden.1,2,4,9 Traditional shoulder-specific PROs include the Simple Shoulder Test, Oxford Shoulder Score, and American Shoulder and Elbow Surgeons (ASES) score, which are commonly used outcome scores validated for a variety of shoulder conditions. Given the increasing use of PROs, the ability to correlate PRO scores to specific shoulder physical functions would be valuable. Thus far, multiple methods of interpreting the

Table II Subgroup demographic characteristics of rotator cuff disease subgroup and arthritis subgroup Data Rotator cuff disease subgroup Encounters, n Patients, n Age, mean  SD, yr Sex, % Male Female Management, n (%) Nonoperative Operative Preoperative encounter Postoperative encounter Arthritis subgroup Encounters, n Patients, n Age, mean  SD, yr Sex, % Male Female Management, n (%) Nonoperative Operative Preoperative encounter Postoperative encounter

1861 1391 57  12 61 39 752 1109 518 591

(40) (60) (47) (53)

291 240 66  11 53 47 128 163 74 89

(44) (56) (45) (55)

SD, standard deviation.

clinical significance of PROs have been proposed, including the minimal clinically important difference, substantial clinical benefit, and patient acceptable symptomatic state (PASS).6,8,13,18,19,22,23,24,26 However, these measures all analyze clinical significance in numeric terms. Physical function mapping, the ability to relate specific physical functions to numeric scores, would assist in the clinical interpretation of scores and could help counsel patients in clinical practice. Thus, we sought to correlate specific physical functions of the shoulder to ASES and PROMIS scores.

Materials and methods A retrospective review of patients with prospectively collected PRO scores was performed. We analyzed patient encounters from January 2016 to October 2017 for patients presenting with shoulder-specific complaints to 2 academic-practice orthopedic surgeons: 1 sports medicine fellowship trained and 1 shoulder fellowship trained. PROMIS and ASES scores were routinely collected at each encounter as part of practice. A total of 3300 patient encounters with completed ASES forms, representing 2447 patients, were available for review. Patient demographic characteristics including age and sex were recorded. A retrospective review using International Classification of Diseases and Current Procedural Terminology codes was performed to determine the diagnosis associated with the encounter and whether patients were operatively or nonoperatively treated. In total, 2632 PROMIS Physical Function (PF), 2574 PROMIS Pain Interference (PI), and

Mapping shoulder physical functions to outcome scores Table III

3

Area under curve in shoulder cohort, rotator cuff disease subgroup, and arthritis subgroup

Shoulder cohort Managing toileting Combing hair Putting on a coat Reaching a high shelf Washing one’s back/doing up bra Sleeping on affected side Lifting 10 lb (4.54 kg) Throwing a ball Rotator cuff disease subgroup Managing toileting Combing hair Putting on a coat Reaching a high shelf Washing one’s back/doing up bra Sleeping on affected side Lifting 10 lb (4.54 kg) Throwing a ball Arthritis subgroup Managing toileting Combing hair Putting on a coat Reaching a high shelf Washing one’s back/doing up bra Sleeping on affected side Lifting 10 lb (4.54 kg) Throwing a ball

ASES score

PROMIS-PF score

PROMIS-UE score

PROMIS-PI score

0.890 0.888 0.898 0.903 0.858 0.856 0.872 0.850

0.783 0.800 0.808 0.800 0.764 0.738 0.798 0.765

0.834 0.860 0.873 0.858 0.816 0.748 0.862 0.803

0.816 0.813 0.830 0.820 0.781 0.794 0.784 0.773

0.891 0.892 0.895 0.895 0.841 0.854 0.856 0.842

0.804 0.817 0.816 0.792 0.749 0.726 0.789 0.752

0.815 0.840 0.871 0.855 0.792 0.738 0.848 0.793

0.830 0.820 0.832 0.815 0.765 0.785 0.774 0.768

0.859 0.786 0.853 0.863 0.840 0.826 0.830 0.819

0.748 0.676 0.761 0.749 0.779 0.703 0.789 0.707

0.790 0.810 0.813 0.811 0.838 0.668 0.821 0.758

0.790 0.710 0.836 0.821 0.779 0.745 0.787 0.745

ASES, American Shoulder and Elbow Surgeons; PROMIS, Patient-Reported Outcomes Measurement Information System; PF, Physical Function; UE, Upper Extremity; PI, Pain Interference. P  .001 for all measurements.

959 PROMIS Upper Extremity (UE) scores, simultaneously collected with the ASES form, were also available. Patients completed the PROMIS assessment on an iPad (Apple, Cupertino, CA, USA) using the computer adaptive testing (CAT) format. Full item banks utilized for PROMIS-PF, PROMIS-PI and PROMISUE CATs can be found in Supplementary Appendixes S1-S3. Substantially fewer PROMIS-UE scores were collected as the PROMIS-UE instrument was not routinely administered until later in the study period. Demographic characteristics of the 3300 patient encounters are listed in Table I. Of the 3300 encounters, 1530 (46%) were for patients managed nonoperatively and 1770 (54%) were for patients managed operatively. Among the operatively managed patients, 836 (47%) and 934 (53%) of the encounters were preoperative and postoperative encounters, respectively. As the goal of the study was to correlate specific physical functions to scores, encounters were excluded if they occurred at less than 90 days postoperatively, when postoperative activity restrictions could have been a confounder. The most common specific shoulder diagnoses were rotator cuff disease (56%) and arthritis (9%). In 467 encounters (14%), a specific diagnosis was not determined. These encounters were associated with a generic International Classification of Diseases code such as shoulder pain or injury or with a nonspecific Current Procedural Terminology code including shoulder arthroscopy with d
ebridement or diagnostic shoulder arthroscopy, or a clear diagnosis could not be

determined from retrospective review. Subgroups of patients with rotator cuff disease and arthritis were identified. The rotator cuff disease group included patients with encounter diagnoses of subacromial bursitis, impingement, or rotator cuff tendinitis and/ or tear or who underwent shoulder arthroscopy with rotator cuff repair. The arthritis group included patients with encounter diagnoses of shoulder arthritis or rotator cuff arthropathy or who underwent anatomic total shoulder arthroplasty (TSA) or reverse total shoulder arthroplasty (RTSA), excluding patients who underwent RTSA for fracture. Demographic characteristics of these subgroups are listed in Table II.

Statistical analysis The ASES form specifically asks about physical functions: putting on a coat, sleeping on the affected side, washing one’s back/doing up a bra, managing toileting, combing hair, reaching a high shelf, lifting 10 lb (4.54 kg), and throwing a ball. These were the physical functions to which we then sought to map PRO scores. Responses were stratified as follows: unable to do, very difficult to do, somewhat difficult to do, and not difficult to do. For our investigation, responses were grouped into 2 categories: a lowfunction response (unable or very difficult to do) and a highfunction response (somewhat difficult or not difficult to do). Receiver operating characteristic (ROC) curves were then

4

A.M. Roberts, I. Voloshin Table IV Prevalence of high function and low function in shoulder cohort, rotator cuff disease subgroup, and arthritis subgroup

Shoulder cohort Managing toileting Combing hair Putting on a coat Reaching a high shelf Washing one’s back/doing up bra Sleeping on affected side Lifting 10 lb (4.54 kg) Throwing a ball Rotator cuff disease subgroup Managing toileting Combing hair Putting on a coat Reaching a high shelf Washing one’s back/doing up bra Sleeping on affected side Lifting 10 lb (4.54 kg) Throwing a ball Arthritis subgroup Managing toileting Combing hair Putting on a coat Reaching a high shelf Washing one’s back/doing up bra Sleeping on affected side Lifting 10 lb (4.54 kg) Throwing a ball

High function, %

Low function, %

91 86 83 58 54 51 42 40

9 14 17 42 56 49 58 60

92 87 84 57 54 51 40 39

8 13 16 43 46 49 60 61

89 80 70 41 34 42 28 26

11 20 30 59 66 58 72 74

calculated to determine 90% positive predictive value (PPV) and 90% negative predictive value (NPV) cutoffs for the presence of a high ability to perform a function for the ASES and PROMIS-PF, -UE, and -PI scores for the entire shoulder cohort and for the disease-specific subgroups. This analysis is an application of a common function of ROC curves, to assess diagnostic tests. The ROC curve is a plot of all the sensitivity and specificity pairs resulting from continuously varying the decision threshold over the entire range of results observed.28 As the ROC curve displays all possible cutoff points, optimal cutoff points for correctly identifying subjects with disease vs. subjects without disease can be determined.14 These analyses are commonly used in medicine, such as in radiology to evaluate the performance of radiologic tests.11,16,17,25 In our study, we did not assess the presence or absence of disease but rather the presence or absence of a physical function. Often, ROC curves are used to determine sensitivity, specificity, and the tradeoff between sensitivity and specificity for diagnostic tests. However, we used PPV and NPV cutoffs, as predictive values are influenced by the prevalence of disease in the given population and thus reflect the probability that a subject really does or does not have the disease (or physical function). This model was chosen to increase the clinical applicability of the results. For example, the ASES 90% PPV cutoff score for reaching a high

shelf would be the ASES score above which at least 90% of individuals gave responses indicating high function in reaching a high shelf and thus were true-positive findings (10% of individuals had a score above the cutoff but gave responses indicating low function in reaching a high shelf; they were falsepositive findings). The ASES 90% NPV cutoff score for reaching a high shelf would be the ASES score below which at least 90% of individuals gave responses indicating they did not have high function in reaching a high shelf and thus were true-negative findings (10% of individuals had a score below the cutoff but gave responses indicating high function in reaching a high shelf; they were false-negative findings). The area under the curve (AUC) is a measure of accuracy. An AUC of 0.5 signifies accuracy similar to chance, whereas an AUC of 1 signifies perfect accuracy with a sensitivity and specificity of 100%. An AUC of 0.7 was considered reasonably accurate, and an AUC of 0.8 was considered excellent.10,12,27 All statistical analysis was performed using XLSTAT software (Addinsoft, Long Island City, NY, USA).

Results AUC measurements for each ROC curve are shown in Table III. ASES scores had a consistently excellent ability to determine high- and low-function states, whereas PROMISUE scores were reasonable to excellent and PROMIS-PF and PROMIS-PI scores were overall reasonable. The prevalence of high or low function in performing a given activity for the entire shoulder cohort and for the diseasespecific subgroups is shown in Table IV. The 90% PPV and 90% NPV cutoffs for the presence of a high ability to perform a function for the ASES and PROMIS-PF, -UE, and -PI scores for the entire shoulder cohort and for the disease-specific subgroups are shown in Table V. It is important to note that higher ASES, PROMISPF, and PROMIS-UE scores signify increased function whereas a higher PROMIS-PI score indicates increased pain and thus decreased function. Toileting is not included because even at the lowest thresholds of all scores, more than 90% of individuals reported high function in toileting themselves. In some cases, a 90% PPV or 90% NPV cutoff could not be reached. Visual representations, or ‘‘maps,’’ of the PPV and NPV cutoff scores were then created for easier interpretation. An example map with the cutoff values transposed at the cutoff points is shown in Figure 1. The green area is the area in which at least 90% of individuals would be expected to have high function for a given activity (the activity is either somewhat or not difficult to do). The red area is the area in which at least 90% of individuals would be expected to not have high function for a given activity (the activity is very difficult to do or unable to be done). Physical function maps for the ASES and PROMIS-PF, -UE, and -PI scores for the entire shoulder cohort and for the disease-specific subgroups are shown in Figures 2-5, respectively.

Mapping shoulder physical functions to outcome scores Table V

5

Ninety percent PPV and 90% NPV cutoff scores in shoulder cohort, rotator cuff disease subgroup, and arthritis subgroup ASES score

Shoulder cohort Combing hair Putting on a coat Reaching a high shelf Washing one’s back/doing up bra Sleeping on affected side Lifting 10 lb (4.54 kg) Throwing a ball Rotator cuff disease subgroup Combing hair Putting on a coat Reaching a high shelf Washing one’s back/doing up bra Sleeping on affected side Lifting 10 lb (4.54 kg) Throwing a ball Arthritis subgroup Combing hair Putting on a coat Reaching a high shelf Washing one’s back/doing up bra Sleeping on affected side Lifting 10 lb (4.54 kg) Throwing a ball

PROMIS-PF score

PROMIS-UE score

PROMIS-PI score

NPV

PPV

NPV

PPV

NPV

PPV

NPV

PPV

9 13 43 39 40 56 53

20 28 66 78 78 83 85

d d 32 d d 39 39

32 36 52 d d d d

d 23 26 26 d 33 29

25 26 41 44 d 44 d

0 0 68 67 67 63 63

68 65 51 47 47 0 0

13 18 41 36 39 55 51

18 25 66 80 77 86 86

d d 28 27 d 39 39

32 35 52 d d d d

d d 26 25 d 32 27

26 25 40 d d d d

d d 69 69 69 65 62

68 66 51 48 45 d d

d d 50 50 27 54 52

43 50 78 91 76 83 83

d d 34 36 d 39 34

43 42 53 d d 53 d

d 23 26 32 d 33 33

29 d 42 d d d d

d d 63 62 d 60 62

58 60 d d d d d

ASES, American Shoulder and Elbow Surgeons; PROMIS, Patient-Reported Outcomes Measurement Information System; PF, Physical Function; UE, Upper Extremity; PI, Pain Interference; PPV, positive predictive value; NPV, negative predictive value.

Discussion Specific physical functions of the shoulder were correlated to ASES and PROMIS scores using 90% PPV and 90% NPV cutoffs, and visual maps of these cutoffs were created for easier interpretation. Given the increasing use of PROs,

the ability to map PRO scores to specific shoulder physical functions has multiple benefits. These data can assist in evaluating the clinical meaning of outcome scores and other measures of patient satisfaction. As analytical models are developed to predict outcomes after treatment interventions, these data could also help counsel patients in

Figure 1 Example physical function map. The American Shoulder and Elbow Surgeons (ASES) score physical function map for the shoulder cohort is displayed with 90% positive and negative predictive value cutoff values transposed at the cutoff points. The green area is the area in which at least 90% of individuals would be expected to have high function for a given activity (the activity is either somewhat or not difficult to do). The red area is the area in which at least 90% of individuals would be expected to not have high function for a given activity (the activity is very difficult to do or unable to be done).

6

A.M. Roberts, I. Voloshin

Figure 2 American Shoulder and Elbow Surgeons (ASES) score physical function maps. (A) Shoulder cohort. (B) Rotator cuff disease subgroup. (C) Arthritis subgroup.

clinical practice. Ultimately, physical function mapping aids the clinician in interpreting the medical literature to then apply it to the patient sitting in front of him or her in the clinic. Physical function mapping can assist in evaluating the clinical meaning of outcome scores. Simovitch et al20 found that the average ASES score improved to 86.6 and 84.2 after TSA and RTSA, respectively. Applying these values to the ASES physical function map for the arthritis subgroup (Fig. 6) adds clinical meaning to this finding. On average, these patients can be expected to have high function in activities such as reaching a high shelf, lifting

10 lb (4.54 kg), and throwing a ball. As the use of PROs has grown, other measures of patient satisfaction, such as the PASS, have become prevalent. The PASS is the level of symptoms beyond which patients consider themselves well. The PASS for the ASES score has been found to be 76 in shoulder arthroplasty patients6 and 86.7 in rotator cuff repair patients.8 These values can also be interpreted through the ASES physical function maps (Fig. 7). Thus, the PASS for a patient undergoing shoulder arthroplasty is a state in which the patient can sleep with, at worst, some difficulty and will likely be able to reach a high shelf. On the other hand, a patient undergoing rotator cuff repair will

Mapping shoulder physical functions to outcome scores

7

Figure 3 Patient-Reported Outcomes Measurement Information System Physical Function (PROMIS PF) score physical function maps. (A) Shoulder cohort. (B) Rotator cuff disease subgroup. (C) Arthritis subgroup.

not report satisfaction unless he or she is able to lift and throw postoperatively. Although PRO-related investigations are helpful to the clinician when interpreting the literature, there is a limited ability to translate PROs when counseling patients. A physician may be able to relate to a patient that undergoing TSA will improve the average patient’s ASES score by 51 points to a final score of 86.6,20 but

what does this mean in the everyday life of that particular patient? Some studies have sought to determine whether certain patients will be more or less likely to benefit from surgery based on preoperative PROs.7,12,21,27 Solberg et al21 developed a predictive model that uses a patient’s preoperative VAS pain score, ASES function score, Veterans RAND 12-Item Health Survey mental health component summary score, age, sex, and type of

8

A.M. Roberts, I. Voloshin

Figure 4 Patient-Reported Outcomes Measurement Information System Upper Extremity (PROMIS UE) score physical function maps. (A) Shoulder cohort. (B) Rotator cuff disease subgroup. (C) Arthritis subgroup.

arthroplasty to predict outcomes after shoulder arthroplasty, including ASES scores. Their main goal was to generate a clinical tool to help set patient expectations. They gave an example of preoperative PROs and baseline characteristics for 2 sample patients and the corresponding 1-year postoperative results these patients should expect according to the model. Patient 1 achieves a theoretical ASES total score of 67, and patient 2

achieves a theoretical ASES total score of 90. Although Solberg et al discussed that these patients should be counseled differentlydwhere, if patient 1 expects a ‘‘normal’’ shoulder after surgery, he or she likely will be disappointeddphysical function mapping can take this 1 step further. On the basis of the ASES physical function map (Fig. 8), patient 1 can be counseled that his or her ability to perform functions postoperatively beyond

Mapping shoulder physical functions to outcome scores

9

Figure 5 Patient-Reported Outcomes Measurement Information System Pain Interference (PROMIS PI) score physical function maps. (A) Shoulder cohort. (B) Rotator cuff disease subgroup. (C) Arthritis subgroup.

combing hair and putting on a coat is uncertain. On the other hand, patient 2 can be counseled that he or she can reasonably expect to be able to perform all of these functions at 1 year postoperatively. When we analyzed the predictive accuracy of the models by the AUC, the ASES and PROMIS-UE ROC plots were consistently more accurate than the PROMIS-PF and -PI ROC plots. We found that 23 of 24 ASES ROC plots and 17 of 24 PROMIS-UE ROC plots had AUC values greater than 0.8 and were excellent. In contrast, the majority of the

PROMIS-PF and PROMIS-PI ROC plots had AUC values between 0.7 and 0.8 and were reasonable. This finding could result from the ASES and PROMIS-UE scores being more shoulder-specific PROs. A better predictive accuracy could also be expected from the ASES models as the physical functions were derived from the ASES forms. Compared with the ASES form, the PROMIS instruments demonstrated a limited ability to determine NPV cutoffs for lower-level functions and PPV cutoff values for higherlevel functions. Overall, the 90% PPV and 90% NPV

10

A.M. Roberts, I. Voloshin

Figure 6 Physical function map applied to reported outcomes. Average American Shoulder and Elbow Surgeons (ASES) scores after total shoulder arthroplasty (TSA) and reverse total shoulder arthroplasty (RTSA) are transposed onto the ASES score physical function map for the arthritis subgroup. If the black line falls in the green area, the patient would be expected to have high function. If the black line falls in the red area, the patient would be expected to have low function.

cutoffs were similar between the overall cohort and the rotator cuff subgroup. This finding could be expected as the rotator cuff subgroup made up the majority of the overall cohort and the prevalence of high function and low function was similar for the given physical functions when comparing groups. However, differences in cutoff values were found when comparing these groups with the arthritis

subgroup. The arthritis subgroup had a higher prevalence of low function for all the physical functions, which might be expected in an older and potentially more disabled population. Limitations of this study include the retrospective nature of the review to determine the diagnosis associated with each encounter. In addition, PPV and NPV are influenced by

Figure 7 Physical function maps applied to patient acceptable symptomatic state (PASS). (A) The PASS for total shoulder arthroplasty (TSA) is transposed onto the American Shoulder and Elbow Surgeons (ASES) score physical function map for the arthritis subgroup. (B) The PASS for rotator cuff (RC) repair is transposed onto the ASES score physical function map for the rotator cuff disease subgroup.

Mapping shoulder physical functions to outcome scores

11

Figure 8 Physical function map applied to predicted outcomes. Predicted American Shoulder and Elbow Surgeons (ASES) scores for patient 1 and patient 2 are transposed onto the ASES score physical function map for the arthritis subgroup.

the prevalence of the disease or condition in a population and thus are unique to the study population. These findings may not be applicable to other populations. Further studies that use mapping are warranted to corroborate our findings and to expand the use of physical function mapping in the orthopedic literature.

2.

3.

Conclusion

4.

We mapped shoulder physical functions to outcome scores by determining the ASES and PROMIS cutoff scores for high and low levels of function in rotator cuff disease and shoulder arthritis populations. Ultimately, these physical function maps can be used to add clinical meaning to the orthopedic literature and better counsel patients about treatment outcomes, and they can be a practical tool to support shared understanding of medical knowledge between the physician and patient.

5.

6.

7.

Disclaimer The authors, their immediate families, and any research foundations with which they are affiliated have not received any financial payments or other benefits from any commercial entity related to the subject of this article.

8.

9.

Supplementary Data Supplementary data to this article can be found online at https://doi.org/10.1016/j.jse.2019.08.017.

10. 11.

12.

References 1. Anthony CA, Glass N, Hancock K, Bollier M, Hettrich CM, Wolf BR. Preoperative performance of the Patient-Reported Outcomes

13.

Measurement Information System in patient with rotator cuff pathology. Arthroscopy 2017;33:1770-4. http://doi.org/10.1016/j.arthro. 2017.04.018 Anthony CA, Glass NA, Hancock K, Bollier M, Wolf BR, Hettrich CM. Performance of PROMIS instruments in patients with shoulder instability. Am J Sports Med 2017;45:449-53. http://doi.org/ 10.1177/0363546516668304 Ayers DC, Bozic KJ. The importance of outcome measurement in orthopaedics. Clin Orthop Relat Res 2013;471:3409-11. http://doi.org/ 10.1007/s11999-013-3224-z Beckmann JT, Hung M, Bounsanga J, Wylie JD, Granger EK, Tashjian RZ. Psychometric evaluation of the PROMIS Physical Function Computerized Adaptive Test in comparison to the American Shoulder and Elbow Surgeons score and Simple Shoulder Test in patients with rotator cuff disease. J Shoulder Elbow Surg 2015;24:1961-7. http://doi.org/10.1016/j.jse.2015.06. 025 Brodke DJ, Saltzman CL, Brodke DS. PROMIS for orthopaedic outcomes measurement. J Am Acad Orthop Surg 2016;24:744-9. http:// doi.org/10.5435/JAAOS-D-15-00404 Chamberlain AM, Hung M, Chen W, Keener JD, McAllister J, Ebersole G, et al. Determining the patient acceptable symptomatic state for the ASES, SST, and VAS pain after shoulder arthroplasty. J Shoulder Elbow Arthroplasty 2017;1. http://doi.org/10.1177/ 2471549217720042 Chen RE, Papuga MO, Nicandri GT, Miller RJ, Voloshin I. Preoperative Patient-Reported Outcomes Measurement Information System (PROMIS) scores predict postoperative outcome in total shoulder arthroplasty patients. J Shoulder Elbow Surg 2019;28:547-54. http:// doi.org/10.1016/j.jse.2018.08.040 Cvetanovich GL, Gowd AK, Liu JN, Nwachukwu BU, Cabarcas BC, Cole BJ, et al. Establishing clinically significant outcome after arthroscopic rotator cuff repair. J Shoulder Elbow Surg 2019;28:93948. http://doi.org/10.1016/j.jse.2018.10.013 Dowdle SB, Glass N, Anthony CA, Hettrich CM. Use of PROMIS for patient undergoing primary total shoulder arthroplasty. Orthop J Sports Med 2017;5. http://doi.org/10.1177/2325967117726044 Gortmaker S, Hosmer D, Lemeshow S. Applied logistic regression. Contemp Sociol 1994;23:159. Hajian-Tilaki K. Receiver operating characteristic (ROC) curve analysis for medical diagnostic test evaluation. Caspian J Intern Med 2013; 4:627-35. Ho B, Houck JR, Flemister AS, Ketz J, Oh I, Baumhauer JF. Preoperative PROMIS scores predict postoperative success in foot and ankle patients. Foot Ankle Int 2016;37:911-8. http://doi.org/10.1177/ 1071100716665113 € arimaa V. InvestiKukkonen J, Kauko T, Vahlberg T, Joukainen A, A€ gating minimal clinically important different for Constant score in

12

14. 15.

16.

17.

18.

19.

20.

21.

A.M. Roberts, I. Voloshin patients undergoing rotator cuff surgery. J Shoulder Elbow Surg 2013; 22:1650-5. http://doi.org/10.1016/j.jse.2013.05.002 Kumar R, Indrayan A. Receiver operating characteristic (ROC) curve for medical researchers. Indian Pediatr 2011;48:277-87. Makhni EC, Meadows M, Hamamoto JT, Higgins JD, Romea AA, Verma NN. Patient Reported Outcomes Measurement Information System (PROMIS) in the upper extremity: the future of outcomes reporting? J Shoulder Elbow Surg 2017;26:352-7. http://doi.org/10. 1016/j.jse.2016.09.054 Mandrekar JN. Receiver operating characteristic curve in diagnostic test assessment. J Thorac Oncol 2019;5:1315-6. http://doi.org/10. 1097/JTO.0b013e3181ec173d Park SH, Goo JM, Jo CH. Receiver operating characteristic (ROC) curve: practical review for radiologists. Korean J Radiol 2004;5:11-8. http://doi.org/10.3348/kjr.2004.5.1.11 Simovitch R, Flurin PH, Wright T, Zuckerman JD, Roche CP. Quantifying success after total shoulder arthroplasty: the minimal clinically important difference. J Shoulder Elbow Surg 2018;27:298-305. http:// doi.org/10.1016/j.jse.2017.09.013 Simovitch R, Flurin PH, Wright T, Zuckerman JD, Roche CP. Quantifying success after total shoulder arthroplasty: the substantial clinical benefit. J Shoulder Elbow Surg 2018;27:903-11. http://doi.org/10. 1016/j.jse.2017.12.014 Simovitch RW, Friedman RJ, Cheung EV, Flurin PH, Wright T, Zuckerman JD, et al. Rate of improvement in clinical outcomes with anatomic and reverse total shoulder arthroplasty. J Bone Joint Surg Am 2017;99:1801-11. http://doi.org/10.2106/JBJS.16.01387 Solberg MJ, Alqueza AB, Hunt TJ, Higgins LD. Predicting 1-year postoperative visual analog scale pain scores and American Shoulder and Elbow Surgeons function scores in total and reverse total

22.

23.

24.

25. 26.

27.

28.

shoulder arthroplasty. Am J Orthop (Belle Mead NY) 2017;46:E35865. Tashjian RZ, Deloach J, Green A, Porucznik CA, Powell AP. Minimal clinically important differences in ASES and simple shoulder test scores after nonoperative treatment of rotator cuff disease. J Bone Joint Surg Am 2010;92:296-303. http://doi.org/10.2106/JBJS.H.01296 Tashjian RZ, Deloach J, Porucznik CA, Powell AP. Minimal clinically important differences (MCID) and patient acceptable symptomatic state (PASS) for visual analog scales (VAS) measuring pain in patients treated for rotator cuff disease. J Shoulder Elbow Surg 2009;18:92732. http://doi.org/10.1016/j.jse.2009.03.021 Tashjian RZ, Hung M, Keener JD, Bowen RC, McAllister J, Chen W, et al. Determining the minimal clinically important difference for the American Shoulder and Elbow Surgeons score, Simple Shoulder Test, and visual analog scale (VAS) measuring pain after shoulder arthroplasty. J Shoulder Elbow Surg 2017;26:144-8. http://doi.org/10.1016/j. jse.2016.06.007 Turner DA. An intuitive approach to receiver operating characteristic curve analysis. J Nucl Med 1978;19:213-20. Werner BC, Chang B, Nguyen JT, Dines DM, Gulotta LV. What change in American Shoulder and Elbow Surgeons score represents a clinically important change after shoulder arthroplasty? Clin Orthop Relat Res 2016;474:2672-81. http://doi.org/10.1007/s11999-016-4968-z Wong SE, Zhang AL, Berliner JL, Ma CB, Feeley BT. Preoperative patient-reported scores can predict postoperative outcomes after shoulder arthroplasty. J Shoulder Elbow Surg 2016;25:913-9. http:// doi.org/10.1016/j.jse.2016.01.029 Zweig MH, Campbell G. Receiver-operating characteristic (ROC) plots: a fundamental evaluation tool in clinical medicine. Clin Chem 1993;39:561-77.