Patients with Workers' Compensation Claims Have Worse Outcomes After Rotator Cuff Repair

Transcription

1 This is an enhance PDF from The Journal of Bone an Joint Surgery The PDF of the article you requeste follows this cover page. Patients with Workers' Compensation Claims Have Worse utcomes After Rotator Cuff Repair R. Frank Henn, III, Lana Kang, Robert Z. Tashjian an Anrew Green J Bone Joint Surg Am. 2008;90: oi: /jbjs.f This information is current as of ctober 7, 2008 Supplementary material Reprints an Permissions Publisher Information Commentary an Perspective, ata tables, aitional images, vieo clips an/or translate abstracts are available for this article. This information can be accesse at Click here to orer reprints or request permission to use material from this article, or locate the article citation on jbjs.org an click on the [Reprints an Permissions] link. The Journal of Bone an Joint Surgery 20 Pickering Street, Neeham, MA

2 2105 CPYRIGHT Ó 2008 BY THE JURNAL F BNE AND JINT SURGERY, INCRPRATED Patients with Workers Compensation Claims Have Worse utcomes After Rotator Cuff Repair By R. Frank Henn III, MD, Lana Kang, MD, Robert Z. Tashjian, MD, an Anrew Green, MD Investigation performe at the Department of rthopaeic Surgery, The Warren Alpert Meical School of Brown University, Provience, Rhoe Islan Backgroun: Previous stuies have emonstrate varying correlations between Workers Compensation status an the outcome of rotator cuff repair. However, none of those stuies have formally accounte for potential confouning factors with multivariable analysis. We hypothesize that patients with Workers Compensation claims who unergo rotator cuff repair have worse outcomes, even after controlling for confouning factors. Methos: ne hunre an twenty-five patients (incluing thirty-nine with Workers Compensation claims) who unerwent unilateral primary repair of a chronic rotator cuff tear by a single surgeon were stuie prospectively an were evaluate one year postoperatively, prior to the settlement of any claims. utcomes were assesse with the Simple Shouler Test (SST); the Disabilities of the Arm, Shouler an Han (DASH) inex; three visual analog scales (shouler pain, shouler function, an quality of life); an the Short Form-36 (). Results: Patients in the Workers Compensation group were significantly younger, ha greater work emans, an ha lower marital rates, eucation levels, an preoperative expectations for the outcome of treatment as compare with those in the non-workers Compensation group (p = to 0.016). Preoperatively, patients in the Workers Compensation group ha significantly lower scores on the SST, the Physical Function scale, an the Social Function scale (p = 0.01 to 0.038). ne year postoperatively, those patients reporte worse performance on the SST, the DASH, all three visual analog scales, an the (p = to 0.05) an ha worse improvement on the DASH, the visual analog scales for shouler pain an function, an the Boily Pain an Role Emotional scales (p = to 0.038). Multivariable analysis controlling for age, sex, comorbiities, smoking, marital status, eucation, uration of symptoms, work emans, expectations, an tear size confirme that Workers Compensation status was an inepenent preictor of worse DASH scores. Conclusions: Patients with Workers Compensation claims report worse outcomes, even after controlling for confouning factors. The present stuy provies further evience that the existence of a Workers Compensation claim portens a less robust outcome following rotator cuff repair. Level of Evience: Prognostic Level I. See Instructions to Authors for a complete escription of levels of evience. Rotator cuff tears are a common cause of isability relate to the shouler. Surgical treatment an repair of chronic tears is inicate when nonoperative treatment fails. verall, rotator cuff repair results in gooto-excellent outcomes in most patients, with significant improvement in the mean scores on self-assessment questionnaires Multiple factors have been shown to be associate with a less favorableoutcome following rotator cuff repair, incluing age 2,4,5,10, sex 2,5,10, smoking 3, larger tear size 5,7,9, poor tenon quality 9, fatty egeneration of cuff muscle 13, poor repair integrity 14,15, an Workers Compensation status 1,4,11,12. Disclosure: The authors i not receive any outsie funing or grants in support of their research for or preparation of this work. Neither they nor a member of their immeiate families receive payments or other benefits or a commitment or agreement to provie such benefits from a commercial entity. No commercial entity pai or irecte, or agree to pay or irect, any benefits to any research fun, founation, ivision, center, clinical practice, or other charitable or nonprofit organization with which the authors, or a member of their immeiate families, are affiliate or associate. A commentary is available with the electronic versions of this article, on our web site ( an on our quarterly CD-RM/DVD (call our subscription epartment, at , to orer the CD-RM or DVD). J Bone Joint Surg Am. 2008;90: oi: /jbjs.f.00260

3 2106 T HE J URNAL F B NE &JINT S URGERY JBJS. RG VLUME 90-A UMBER 10 CTBER 2008 PATIENTS WITH WRKERS CMPENSATIN C LAIMS H AV E WRSE UTCMES A FTER RTATR C UFF R EPAIR TABLE I Baseline utcome Instrument Associations* ä DASH VAS Pain VAS Function VAS Quality of Life Physical Function SST 0.70 (<0.001) 0.51 (<0.001) 0.50 (<0.001) 0.48 (<0.001) 0.42 (<0.001) DASH 0.65 (<0.001) 0.64 (<0.001) 0.58 (<0.001) 0.55 (<0.001) VAS pain 0.87 (<0.001) 0.66 (<0.001) 0.46 (<0.001) VAS function 0.65 (<0.001) 0.48 (<0.001) VAS quality of life 0.34 (<0.001) Physical Function Role Physical Boily Pain General Health Vitality Social Function Role Emotional *The values are expresse as the Spearman correlation coefficient for each test, with the corresponing p value in parentheses. Associations that i not reach significance (p > 0.05) are shown in bol. SST = Simple Shouler Test; DASH = Disabilities of the Arm, Shouler an Han; VAS = visual analog scale; an = Short Form-36. While multiple stuies have emonstrate that Workers Compensation claims are associate with less favorable outcomes after rotator cuff repair, other stuies have foun no ifference in the outcomes of shouler surgery between patients with an without a Workers Compensation claim 3,16.Inaition,the reasons for the observe ifferences between patients with an without a Workers Compensation claim are not well unerstoo. To our knowlege, no stuy has formally accounte for potential confouning factors with rigorous multivariable analysis. Recent multicenter work has emonstrate marke variability in the patient populations treate by ifferent shouler surgeons 17. Consequently, the finings of stuies of ifferent populations might be expecte to vary. Multivariable analysis provies a statistical metho for accounting for ifferences in patient populations by controlling for potentially confouning variables. Regaring Workers Compensation, for example, some investigators have suggeste that psychosocial factors, seconary gain issues, work emans, comorbiities, an smoking may explain the ifferences 1,3,18,19. Controlling for some of these factors with multivariable analysis provies a means to etermine the relative effect of these factors as compare with the effect of a Workers Compensation claim. The purpose of the present stuy was to evaluate the effect of a Workers Compensation claim on the outcome of rotator cuff repair. It was our hypothesis that patients with Workers Compensation claims have worse outcomes, even after controlling for confouning factors. Materials an Methos We stuie 125 patients with a chronic rotator cuff tear who subsequently unerwent unilateral primary rotator cuff repair by one fellowship-traine shouler surgeon (A.G.) from January 1998 to September 2001 at a tertiary care teaching hospital. Thirty-nine patients ha an open Workers Compensation claim (Workers Compensation group), an eightysix patients i not (non-workers Compensation group). All patients ha ha symptoms for more than three months, which efine a chronic rotator cuff tear The inication for surgery was the failure of nonoperative treatment, which inclue a physical therapy program an, in some cases, a corticosteroi injection. The stuy was approve by the hospital institutional review boar. Institutional review boar policy an review i not require specific consent from patients 22. The criterion for inclusion in the stuy was a primary repair of a unilateral symptomatic chronic full-thickness rotator cuff tear that ha faile to respon to nonoperative treatment. Patients were exclue if they ha ha any previous shouler surgery, ha an incomplete (partial) repair of a massive tear, or ha glenohumeral arthritis. The prospective preoperative evaluation inclue a etaile meical history, a physical examination, an completion of a series of questionnaires as previously escribe Patients complete four limb-specific instruments: the Simple Shouler Test (SST); the Disabilities of the Arm, Shouler an Han (DASH); a visual analog scale for shouler pain, anchore with none on one en of a 10-cm line an with isabling on the other; an a visual analog scale for shouler function, anchore with comfortable an can t use it. General health was assesse with use of the Short Form-36 (SF- 36) an a visual analog scale for overall quality of life anchore with little or no problem an very ba. These three specific

4 2107 T HE J URNAL F B NE &JINT S URGERY JBJS. RG VLUME 90-A UMBER 10 CTBER 2008 PATIENTS WITH WRKERS CMPENSATIN C LAIMS H AV E WRSE UTCMES A FTER RTATR C UFF R EPAIR TABLE I (continue) Role Physical Boily Pain General Health Vitality Social Function Role Emotional Mental Health 0.46 (<0.001) 0.42 (<0.001) 0.09 (0.33) 0.32 (<0.001) 0.34 (<0.001) 0.14 (0.13) 0.13 (0.17) 0.62 (<0.001) 0.65 (<0.001) 0.14 (0.12) 0.39 (<0.001) 0.48 (<0.001) 0.34 (<0.001) 0.25 (0.006) 0.56 (<0.001) 0.62 (<0.001) 0.10 (0.25) 0.26 (0.004) 0.34 (<0.001) 0.39 (<0.001) 0.21 (0.024) 0.51 (<0.001) 0.60 (<0.001) 0.16 (0.09) 0.21 (0.02) 0.34 (<0.001) 0.35 (<0.001) 0.15 (0.08) 0.51 (<0.001) 0.46 (<0.001) 0.21 (<0.001) 0.35 (<0.001) 0.45 (<0.001) 0.32 (<0.001) 0.24 (0.010) 0.53 (<0.001) 0.57 (<0.001) 0.49 (<0.001) 0.43 (<0.001) 0.47 (<0.001) 0.31 (<0.001) 0.33 (<0.001) 0.66 (<0.001) 0.28 (0.002) 0.44 (<0.001) 0.47 (<0.001) 0.47 (<0.001) 0.27 (0.002) 0.30 (<0.001) 0.35 (<0.001) 0.43 (<0.001) 0.39 (<0.001) 0.23 (0.01) 0.50 (<0.001) 0.47 (<0.001) 0.26 (0.003) 0.34 (<0.001) 0.46 (<0.001) 0.28 (0.001) 0.50 (<0.001) 0.48 (<0.001) 0.39 (<0.001) 0.37 (<0.001) visual analog scales have been use in many shouler outcome stuies but have not been formally valiate in the literature For ease of graphical comparison, the SST, DASH, an visual analog scale instrument scores were converte to a percentage of a perfect score, with 0% corresponing with the worst state of health an 100% corresponing with the ieal state of health 22. The raw scores for the eight subscales of the (Physical Function, Role Physical, Boily Pain, General Health, Vitality, Social Function, Role Emotional, an Mental Health) were ajuste to a percentage of reporte normative values for age-matche an sex-matche controls 26. Preoperative expectations of treatment were also assesse with use of six questions from the Musculoskeletal utcomes Data Evaluation an Management System (MDEMS) The questionnaire aske, What results o you expect from your treatment? with regar to six items: (1) relief from symptoms, (2) to o more everyay househol or yar activities, (3) to sleep more comfortably, (4) to go back to my usual job, (5) to exercise an o more recreational activities, an (6) to prevent future isability. The patients circle one response for each item: 1 = not at all likely, 2 = slightly likely, 3 = somewhat likely, 4 = very likely, 5 = extremely likely, or not applicable. Therefore, a score of 1 correspone with the lowest level of expectations, whereas a score of 5 correspone with the highest level of expectations. The average response to the six questions was calculate for each patient to generate a mean expectations score. Questions marke not applicable were not inclue in the calculation of mean expectations. A variety of emographic information was collecte for each patient. Patients were consiere smokers if they reporte having smoke cigarettes uring the previous six months. The patients work emans were categorize into one of three groups: retire/seentary, light emans, or strenuous emans. The number of meical comorbiities for each patient was assesse as previously escribe 20. The follow-up evaluation, performe at a mean (an stanar eviation) of 54.1 ± 7.6 weeks (range, 32.7 to 88.7 weeks), consiste of a physical examination an completion of the same series of outcome instruments. An improvement from baseline score for each outcome instrument was calculate for each patient by subtracting the baseline score from the follow-up score 21,22. Surgical Technique Three ifferent repair techniques were use. f the 125 patients, twenty-six (20.8%) ha an open repair, sixty-two (49.6%) ha a mini-open repair, an thirty-seven (29.6%) ha an arthroscopic repair. f the thirty-nine repairs in the Workers Compensation group, seven (17.9%) were open repairs, nineteen (48.7%) were mini-open repairs, an thirteen (33.3%) were arthroscopic repairs. f the eighty-six repairs in the non-workers Compensation group, nineteen (22.1%) were open repairs, forty-three (50.0%) were mini-open repairs, an twenty-four (27.9%) were arthroscopic repairs. With the numbers available, no significant ifferences between the Workers Compensation an non- Workers Compensation groups coul be emonstrate with respect to repair technique (p > 0.53). All of the patients ha a subacromial ecompression with acromioplasty at the time of rotator cuff repair. With regar to ajunctive proceures, thirty-four (27.2%) of the 125 patients (incluing twenty [23.3%] of the eighty-six patients in the non- Workers Compensation group an fourteen [35.9%] of the thirty-nine in the Workers Compensation group) ha istal clavicular resection for the treatment of acromioclavicular joint arthritis. Five (4.0%) of the 125 patients (incluing three [3.5%] of the eighty-six patients in the non-workers Compensation group an two [5.1%] of the thirty-nine patients in the Workers Compensation group) ha biceps tenoesis. Five (4.0%) of the 125 patients (incluing three [3.5%] of the eighty-six patients in the non-workers Compensation group

5 2108 T HE J URNAL F B NE &JINT S URGERY JBJS. RG VLUME 90-A UMBER 10 CTBER 2008 PATIENTS WITH WRKERS CMPENSATIN C LAIMS H AV E WRSE UTCMES A FTER RTATR C UFF R EPAIR TABLE II Characteristics of Patients in Workers Compensation an Non-Workers Compensation Groups Workers Compensation Group (N = 39) Non-Workers Compensation Group (N = 86) P Value Age* (yr) 52.5 ± 1.6 (32 to 79) 57.8 ± 1.3 (35 to 84) Male 24 (61.5%) 48 (55.8%) 0.54 Right-han ominant 36 (92.3%) 81 (94.2%) 0.69 Right shouler rotator cuff tear 24 (61.5%) 58 (67.4%) 0.52 Dominant shouler rotator cuff tear 23 (59.0%) 59 (68.6%) 0.29 Smoker 12 (30.8%) 14 (16.3%) 0.06 Marrie 23 (59.0%) 71 (82.6%) College grauate 4 (10.3%) 32 (37.2%) Physically strenuous job 15 (38.5%) 9 (10.5%) Number of comorbiities* 1.8 ± 1.5 (0 to 5) 2.0 ± 1.5 (0 to 6) 0.48 Number of previous surgeries* 1.9 ± 1.7 (0 to 7) 2.2 ± 1.7 (0 to 8) 0.48 Duration of symptoms* (mo) 13.0 ± 13.9 (3 to 63) 17.5 ± 29.9 (3 to 210) 0.38 Boy mass inex* 29.1 ± 5.1 (19.4 to 45.6) 28.3 ± 5.0 (17.9 to 39.2) 0.42 Average expectation score* 4.20 ± 0.89 (2.0 to 5.0) 4.55 ± 0.54 (2.7 to 5.0) *The values are given as the mean an the stanar eviation, with the range in parentheses. The ifference between the groups was significant. The values are given as the number of patients with the ientifie trait, with the percentage in parentheses. an two [5.1%] of the thirty-nine patients in the Workers Compensation group) ha biceps relocation. Two (1.6%) of the 125 patients (incluing one [1.2%] of the eighty-six patients in the non-workers Compensation group an one [2.6%] of the thirty-nine patients in the Workers Compensation group) ha biceps tenotomy. With the numbers available, no significant ifferences between the Workers Compensation an non- Workers Compensation groups coul be emonstrate with respect to these ajunct proceures (p > 0.15). The size of the rotator cuff tear was measure intraoperatively prior to repair. Postoperative Management an Rehabilitation The upper extremity was place into a sling after surgery. The majority of patients were ischarge to home without requiring overnight amission. No analgesic catheters were use. Patients with avance age or major meical comorbiities were amitte for overnight observation. No patient require reamission for pain control. All patients were referre to physical therapy. Passive range-of-motion exercises were initiate on the ay after surgery. During the first five weeks postoperatively, active use of the upper extremity was iscourage an patients were instructe to wear a sling. For patients who ha ha mini-open or open repair, passive range of motion commence with penulum circumuction, supine forwar elevation, supine external rotation, supine horizontal auction, an staning internal rotation. For patients who ha ha arthroscopic repair, passive motion was limite to penulum circumuction an supine passive external rotation for the first four weeks; after the fifth week, passive stretching was performe in all irections. Slings were iscontinue after five weeks for all treatment groups. Statistical Methos Thelevelofsignificancewassetatp<0.05.TheWorkers Compensation an non-workers Compensation groups were compare with use of two-taile t tests for continuous variables an chi-square tests for ichotomous emographic variables. Preoperative an postoperative scores for each variable were compare over time within the Workers Compensation an non-workers Compensation groups with use of paire two-taile t tests. Associations between the outcome parameters utilize in this stuy were assesse with use of Spearman correlations (see Table I for baseline associations an tables in the Appenix for one-year an improvement associations). The Spearman correlation coefficient can range from 21 to 1 an inicates both the irection an strength of the linear relationship. The DASH was selecte as the primary outcome variable because this instrument has been well characterize an some ata suggest that a ifference of 10 to 15 points is clinically important 27,28. As shown in Table I, the DASH, SST, an visual analog scale scores were all significantly correlate. The subscales were also all significantly correlate. However, the an other outcome instruments were variably correlate. Consequently, the subscale for general health was utilize as a seconary outcome parameter for multivariable analysis. Multivariable analyses were performe for the DASH an the general health subscale with use of a linear multivariable regression analysis to control for confouning variables Inepenent variables initially entere into each moel

6 2109 T HE J URNAL F B NE &JINT S URGERY JBJS. RG VLUME 90-A UMBER 10 CTBER 2008 PATIENTS WITH WRKERS CMPENSATIN C LAIMS H AV E WRSE UTCMES A FTER RTATR C UFF R EPAIR TABLE III Mean Preoperative Scores Parameter* Workers Compensation Group Non-Workers Compensation Group P Value SST 34.4 ± ± DASH 56.0 ± ± VAS Shouler pain 37.7 ± ± Shouler function 39.1 ± ± Quality of life 39.1 ± ± # Physical Function 78.8 ± ± Role Physical 31.2 ± ± Boily Pain 54.2 ± ± General Health ± ± Vitality 91.3 ± ± Social Function 82.1 ± ± Role Emotional 86.8 ± ± Mental Health 97.6 ± ± *SST = Simple Shouler Test; DASH = Disabilities of the Arm, Shouler an Han; VAS = visual analog scale; an = Short Form-36. The values are given (in points) as the mean an the stanar eviation. The p values were erive from the t test comparing the Workers Compensation group with the non-workers Compensation group for each parameter. The ifference between the two groups was significant (p < 0.05). #The scores are given as a percentage of normative values for age an sex-matche controls. inclue Workers Compensation status, age, sex, smoking status, mean expectations, number of comorbiities, college eucation, marital status, work emans, the size of the rotator cuff tear, an repair technique. A backwars-elimination stepwise technique was employe to remove noncontributory inepenent variables. The significance of each inepenent variable in the moel was evaluate after calculation, an, at each run, the single least significant variable was remove an the moel was recalculate. This process of removing the least significant variable was repeate in a stepwise fashion until a moel incluing only variables with significance (p < 0.05) was achieve. This was the final moel if Workers Compensation status was foun to be significant. If Workers Compensation was eliminate, the moel was recalculate to inclue Workers Compensation along with the remaining significant inepenent variables. This secon step was performe in orer to generate a Workers Compensation multivariable regression coefficient an p value for that epenent variable. There were no instances in which Workers Compensation status was foun to be significant after this secon step. Results For the stuy population of 125 patients, the average age at the time of surgery was 56.2 ± 11.4 years (range, thirty-two to eighty-four years). Seventy-two patients (57.6%) were male. ne hunre an seventeen patients (93.6%) were right-han ominant, an eighty-two (65.6%) unerwent surgery on the ominant shouler. The mean uration of symptoms was 16.0 ± 25.9 months (range, three to 210 months). There were thirty-seven small tears (<1 cm), forty-two meium tears ( 1 cm but <3 cm), thirty-two large tears ( 3 cmbut<5cm),an fourteen massive tears ( 5 cm). There were no perioperative complications. Two (1.6%) of the 125 patients ha complications postoperatively. ne (2.6%) of the thirty-nine patients in the Workers Compensation group ha evelopment of postoperative stiffness following mini-open repair an ultimately require arthroscopic capsular release. ne (1.2%) of the eighty-six patients in the non-workers Compensation group ha persistent shouler stiffness after arthroscopic rotator cuff repair an was manage with arthroscopic capsular release an biceps tenotomy for the treatment of a 50% partial-thickness biceps tenon tear. In both cases, the rotator cuff was well heale an the arthroscopic capsular release was successful. Both patients were inclue in the analysis. The specific characteristics of the patients in the Workers Compensation an non-workers Compensation groups are shown in Table II. The patients in the Workers Compensation group were younger (p = 0.016), were less likely to be marrie (p = 0.003), were less likely to have grauate from college (p = 0.002), an ha greater work emans (p = 0.001) as compare with those in the non-workers Compensation group. While both groups ha high preoperative expectations for surgery, the patients in the non-workers Compensation group ha significantly greater mean expectations than those in the Workers Compensation group (4.6 compare with 4.2; p = 0.008). The percentage of patients who were smokers was greater in the Workers Compensation group than in the non-workers Compensation group (30.8% compare with 16.3%; p = 0.06). With the numbers available,

7 2110 T HE J URNAL F B NE &JINT S URGERY JBJS. RG VLUME 90-A UMBER 10 CTBER 2008 PATIENTS WITH WRKERS CMPENSATIN C LAIMS H AV E WRSE UTCMES A FTER RTATR C UFF R EPAIR TABLE IV Mean ne-year Follow-up Scores Parameter* Workers Compensation Group Non-Workers Compensation Group P Value SST 66.4 ± ± DASH 72.9 ± ± VAS Shouler pain 69.2 ± ± Shouler function 69.2 ± ± Quality of life 70.1 ± ± # Physical Function 89.1 ± ± Role Physical 66.1 ± ± Boily Pain 80.4 ± ± General Health 99.1 ± ± Vitality 96.3 ± ± Social Function 90.1 ± ± Role Emotional 78.2 ± ± Mental Health 96.2 ± ± *SST = Simple Shouler Test; DASH = Disabilities of the Arm, Shouler an Han; VAS = visual analog scale; an = Short Form-36. The values are given (in points) as the mean an the stanar eviation. The p values were erive from the t test comparing the Workers Compensation group with the non-workers Compensation group for each parameter. The ifference between the two groups was significant (p < 0.05). #The scores are given as a percentage of normative values for age an sex-matche controls. no significant ifference coul be emonstrate between the groups with respect to sex (p = 0.54), the number of meical comorbiities (p = 0.48), the uration of symptoms (p = 0.38), or the size of the rotator cuff tear (p = 0.67). Preoperatively, the two groups emonstrate a similar level of impairment on the majority of instruments (Table III). With the numbers available, no significant ifferences coul be emonstrate with respect to scores on the DASH an all visual analog scales. However, the patients in the Workers Compensation group were able to perform fewer SST functions preoperatively as compare with those in the non- Workers Compensation group (p = 0.010). As shown in Table III, there was a tren for the patients in the Workers Compensation group to have lower scores on the eight subscales, but the ifferences were only significant for Physical Function (p = 0.032) an Social Function (p = 0.038). In contrast to the preoperative finings, follow-up outcome evaluation emonstrate consistent ifferences between the two groups. The patients in the Workers Compensation group ha significantly worse scores on the SST (p = ), the DASH (p = ), an the visual analog scales for shouler pain (p = ), shouler function (p = 0.013), an quality of life (p = 0.019) (Table IV). Furthermore, eighteen (46%) of the thirty-nine patients in the Workers Compensation group ha a postoperative DASH score that was >10 points worse than the mean postoperative non-workers Compensation group score of The scores for the patients in the Workers Compensation group were lower than those for the patients in the non-workers Compensation group on all subscales; the ifference was significant on all scales except Physical Function (Table IV). f note, in the non-workers Compensation group, the average scores on each subscale approache or exceee the mean age an sex-matche values. Both groups ha significant improvement from the preoperative evaluation on the SST, DASH, an all visual analog scale scores (p < ). The Workers Compensation group showe less improvement on each instrument, but this ifference was significant only for the DASH (p = 0.020) an the visual analog scale scores for shouler pain (p = 0.006) an shouler function (p = 0.006) (Table V). The non-workers Compensation group showe improvement from the preoperative evaluation on all eight subscales (Table V). The improvement was significant for the Role Physical (p < ), Boily Pain (p < ), Vitality (p = 0.002), an Role Emotional (p = 0.001) subscales. In contrast, the Workers Compensation group ha worse postoperative scores compare with baseline on the subscales for General Health (p = 0.45), Role Emotional (p = 0.49), an Mental Health (p = 0.81). The Workers Compensation group emonstrate significant improvement from baseline on the subscales for Role Physical (p = 0.009) an Boily Pain (p = ). The Workers Compensation group ha less improvement as compare with the non-workers Compensation group on each subscale except Physical Function (p = 0.79), an the ifferences were significant for the Boily Pain (p = 0.038) an Role Emotional (p = ) subscales (Table V). The multivariable analysis confirme that a Workers Compensation claim was an inepenent preictor for worse postoperative DASH scores (p = 0.015). The multivariable correlation coefficient inicate that a Workers Compensation

8 2111 T HE J URNAL F B NE &JINT S URGERY JBJS. RG VLUME 90-A UMBER 10 CTBER 2008 PATIENTS WITH WRKERS CMPENSATIN C LAIMS H AV E WRSE UTCMES A FTER RTATR C UFF R EPAIR TABLE V Mean Improvement from Baseline Scores Parameter* Workers Compensation Group Non-Workers Compensation Group P Value SST 32.5 ± ± DASH 17.0 ± ± VAS Shouler pain 30.9 ± ± Shouler function 30.2 ± ± Quality of life 30.1 ± ± Physical Function 11.3 ± ± Role Physical 33.0 ± ± Boily Pain 26.1 ± ± General Health 25.7 ± ± Vitality 6.0 ± ± Social Function 8.0 ± ± Role Emotional 29.1 ± ± Mental Health 21.4 ± ± *SST = Simple Shouler Test; DASH = Disabilities of the Arm, Shouler an Han; VAS = visual analog scale; an = Short Form-36. The values are given (in points) as the mean an the stanar eviation. The p values were erive from the t test comparing the Workers Compensation group with the non-workers Compensation group for each parameter. The ifference between the two groups was significant (p < 0.05). claim accounte for an 8.7-point reuction in the postoperative DASH score. In contrast, a Workers Compensation claim was not a significant inepenent preictor of the postoperative General Health score (p = 0.073). Discussion The results of the present stuy support our hypothesis that patients with Workers Compensation claims report worse outcomes, even after controlling for confouning factors. In our patient population, the Workers Compensation group was significantly younger an ha significantly lower marital rates, eucation levels, an preoperative expectations as compare with the non-workers Compensation group. The multivariable analyses controlle for these ifferences as well as for work emans, smoking, comorbiities, the uration of symptoms, the size of the tear, an repair technique. Consequently, we can conclue that these factors o not account for the ifference in outcome between the two groups. This suggests that ifferences observe after multivariable analyses are relate to an effect of a Workers Compensation claim. Several previous stuies have investigate the relationship between Workers Compensation status an preoperative functional status in patients with rotator cuff tears. Viola et al. reporte that patients with Workers Compensation insurance who ha a full-thickness rotator cuff tear ha significantly worse scores on five of eight subscales as compare with those without such insurance 18. Likewise, McKee an Yoo reporte that patients who ha file a Workers Compensation claim ha significantly lower mean preoperative physical an mental subscores as compare with those who ha not file such a claim 1. In the present stuy, we use a percentage of normalize age an sex-matche values an foun that a Workers Compensation claim is associate with lower scores on the preoperative subscales for Physical Function an Social Function. With regar to preoperative functional status, our results agree with those of Smith et al., who also foun that patients with a Workers Compensation claim who ha chronic rotator cuff tears ha worse preoperative SST scores than those without such a claim 29. In contrast, we foun that the patients in the Workers Compensation group i not have significantly worse preoperative pain or function as assesse with the DASH an the visual analog scale than i those in the non-workers Compensation group. This iscrepancy is at least partially explaine by the fact that one of the twelve SST questions asks specifically about the ability to work full time, whereas the DASH an visual analog scales o not. In the present stuy, only 30% of the patients in the Workers Compensation group were able to work full time, compare with 65% of the patients in the non-workers Compensation group. Multiple stuies have emonstrate that a Workers Compensation claim is associate with poorer outcomes after rotator cuff repair 1,4,11,12. ur results corroborate these finings. Like Watson an Sonnaben 4, we foun that patients in the Workers Compensation group reporte greater levels of postoperative pain on a visual analog scale as compare with the patients in the non-workers Compensation group. Like McKee an Yoo 1, we foun that patients in the Workers Compensation group ha lower postoperative scores as compare with those in the non-workers Compensation

9 2112 T HE J URNAL F B NE &JINT S URGERY JBJS. RG VLUME 90-A UMBER 10 CTBER 2008 PATIENTS WITH WRKERS CMPENSATIN C LAIMS H AV E WRSE UTCMES A FTER RTATR C UFF R EPAIR group. However, this relationship i not remain significant after multivariable analysis of the General Health subscore. We also foun that the functional status was significantly lower for the Workers Compensation group as assesse with the SST, DASH, an visual analog scales. Multivariable analysis confirme that a Workers Compensation claim is an inepenent preictor of worse postoperative DASH scores. An association between Workers Compensation status an the outcome of rotator cuff repair as assesse with the DASH has not been reporte previously, to our knowlege. Previous work has suggeste that a ifference of >10 points on the DASH is clinically important, an we i fin that eighteen (46%) of the thirty-nine patients in the Workers Compensation group ha a postoperative DASH score that was >10 points worse than the mean postoperative score for the non-workers Compensation group 27,28. To our knowlege, no ata have been reporte regaring the preoperative expectations of patients with a Workers Compensation claim who are scheule to unergo surgery. Preoperative expectations of treatment may have a powerful effect on outcomes 22. Cole et al. emonstrate that greater recovery expectations among patients with a Workers Compensation claim who ha a soft-tissue injury were associate with significantly ecrease pain levels an better general health status at the time of follow-up 30. The reasons for lower expectations in the Workers Compensation group in the current stuy are not clear but may be relate to eucation, previous meical experiences, an potential seconary gain. A prospective stuy, with consistent preoperative an postoperative evaluation, allows for the objective measurement of improvement from the time of surgery. We foun that the Workers Compensation group improve significantly from baseline on all shouler-specific instruments, the visual analog scale for overall quality of life, an the subscales for Role Physical an Boily Pain. However, we report for the first time that postoperative improvement from baseline after rotator cuff repair is less robust for patients with a Workers Compensation claim. WefounthatpatientsintheWorkers Compensationgroupha significantly less pain relief an more moest improvements in functional status as assesse with the DASH an the visual analog scales. Similarly, we foun significantly less improvement on the for the Boily Pain an Role Emotional subscales. It is clear that, in general, patients with Workers Compensation claims who unergo rotator cuff repair have a worse overall outcome. By stuying confouning variables or factors that are potentially relate to a Workers Compensation claim, we attempte to etermine if it is the process of a Workers Compensation claim that affects the outcome or something about the iniviuals who file claims. While the confouning variables play a role in etermining the outcome of the population that we stuie, there is evience in our ata analysis that the Workers Compensation claim itself has an important role. This suggests that the process of a Workers Compensation claim either is the cause of worse outcomes or may be serving as a marker of hitherto unientifie or unaccountefor factors that are actually causing worse outcomes. ne potential factor is seconary gain, either real or perceive, which may be variable in the Workers Compensation population. Seconary gain is very ifficult to separate from the Workers Compensation claim process an was not irectly assesse in the present stuy. Despite the fact that the present stuy provies evience that a Workers Compensation claim can have a negative effect on the outcome of rotator cuff repair, the stuy ha several limitations. First, the multivariable analyses that we utilize may not have inclue all of the confouning variables that coul influence the results. For example, we i not analyze the effect of anatomic factors such as repair integrity, fatty egeneration of muscle, or tenon quality on the outcome of rotator cuff repair. In aition, preoperative questionnaires were aministere to the patients only after the ecision to procee with surgery ha been mae. The knowlege that the shouler problem was severe enough to require surgery may have ha a negative effect on the baseline assessment that patients reporte. Consequently, our preoperative ata may not be irectly comparable to ata in stuies regaring the presentation of rotator cuff tears. Relative to the Workers Compensation claim, the patients in the present stuy were evaluate prior to the settlement of the claim. The effect of Workers Compensation status on outcome may actually change over time, an this may limit how our results can be generalize to stuies of patients after the settlement of claims. For the purpose of the present stuy, we chose to perform the follow-up outcome evaluation one year after surgery. We thought that this woul ensure that the open Workers Compensation claim still ha an effect on the self-reporte outcomes, while patients who unergo rotator cuff repair have usually reache an en result by twelve months after surgery 1. A final limitation of the present stuy is that the results may be specific to the population stuie. Not only o patients in ifferent geographic areas iffer, but Workers Compensation laws vary from state to state. These ifferences might also affect outcome. Future stuies are require to aress these limitations. Although we o not yet know how a Workers Compensation claim affects the long-term outcome of rotator cuff repair, these finings have important societal implications. Shouler isorers are very common an injuries are boun to occur at work. We strongly believe that the present stuy shoul not iscourage clinicians from recommening inicate rotator cuff repair to Workers Compensation patients. Rather, it shoul provie a framework for outcome evaluation for both patients an surgeons. In the future, our goal is to etermine how to optimize the outcome of rotator cuff repair for all patients, incluing those with Workers Compensation claims. Appenix Tables showing the associations between the outcome parameters utilize in this stuy are available with the electronic versions of this article, on our web site at jbjs.org (go to the article citation an click on Supplementary Material )

10 2113 T HE J URNAL F B NE &JINT S URGERY JBJS. RG VLUME 90-A UMBER 10 CTBER 2008 PATIENTS WITH WRKERS CMPENSATIN C LAIMS H AV E WRSE UTCMES A FTER RTATR C UFF R EPAIR an on our quarterly CD/DVD (call our subscription epartment, at , to orer the CD or DVD). n R. Frank Henn III, MD Lana Kang, MD Hospital for Special Surgery, 535 East 70th Street, New York, NY aress for R.F. Henn: Frank_Henn@yahoo.com Robert Z. Tashjian, MD University of Utah rthopaeic Center, 590 Wakara Way, Salt Lake City, UT Anrew Green, MD Department of rthopaeic Surgery, Brown Meical School, Rhoe Islan Hospital, 2 Duley Street, Suite 200, Provience, RI References 1. McKee MD,Yoo DJ. The effect of surgery for rotator cuff isease on general health status. Results of a prospective trial. J Bone Joint Surg Am. 2000;82: Romeo AA, Hang DW, Bach BR Jr, Shott S. Repair of full thickness rotator cuff tears. Gener, age, an other factors affecting outcome. Clin rthop Relat Res. 1999;367: Mallon WJ, Misamore G, Snea DS, Denton P. The impact of preoperative smoking habits on the results of rotator cuff repair. J Shouler Elbow Surg. 2004;13: Watson EM, Sonnaben DH. utcome of rotator cuff repair. J Shouler Elbow Surg. 2002;11: Cofiel RH, Parvizi J, Hoffmeyer PJ, Lanzer WL, Ilstrup DM, Rowlan CM. Surgical repair of chronic rotator cuff tears. A prospective long-term stuy. J Bone Joint Surg Am. 2001;83: Golberg BA, Lippitt SB, Matsen FA 3r. Improvement in comfort an function after cuff repair without acromioplasty. Clin rthop Relat Res. 2001;390: Hollinshea RM, Mohtai NG, Vane Guchte RA, Waey VM. Two 6-year followup stuies of large an massive rotator cuff tears: comparison of outcome measures. J Shouler Elbow Surg. 2000;9: Galatz LM, Griggs S, Cameron BD, Iannotti JP. Prospective longituinal analysis of postoperative shouler function: a ten-year follow-up stuy of full-thickness rotator cuff tears. J Bone Joint Surg Am. 2001;83: Pai VS, Lawson DA. Rotator cuff repair in a istrict hospital setting: outcomes an analysis of prognostic factors. J Shouler Elbow Surg. 2001;10: Gartsman GM, Brinker MR, Khan M. Early effectiveness of arthroscopic repair for full-thickness tears of the rotator cuff: an outcome analysis. J Bone Joint Surg Am. 1998;80: Misamore GW, Ziegler DW, Rushton JL 2n. Repair of the rotator cuff. A comparison of results in two populations of patients. J Bone Joint Surg Am. 1995;77: Shinners TJ, Noorsij PG, rwin JF. Arthroscopically assiste mini-open rotator cuff repair. Arthroscopy. 2002;18: Goutallier D, Postel JM, Gleyze P, Leguilloux P, Van Driessche S. Influence of cuff muscle fatty egeneration on anatomic an functional outcomes after simple suture of full-thickness tears. J Shouler Elbow Surg. 2003;12: Harryman DT 2n, Mack LA, Wang KY, Jackins SE, Richarson ML, Matsen FA 3r. Repairs of the rotator cuff. Correlation of functional results with integrity of the cuff. J Bone Joint Surg Am. 1991;73: Klepps S, Bishop J, Lin J, Cahlon, Strauss A, Hayes P, Flatow EL. Prospective evaluation of the effect of rotator cuff integrity on the outcome of open rotator cuff repairs. Am J Sports Me. 2004;32: Nicholson GP. Arthroscopic acromioplasty: a comparison between workers compensation an non-workers compensation populations. J Bone Joint Surg Am. 2003;85: Harryman DT 2n, Hettrich CM, Smith KL, Campbell B, Siles JA, Matsen FA 3r. A prospective multipractice investigation of patients with full-thickness rotator cuff tears: the importance of comorbiities, practice, an other covariables on selfassesse shouler function an health status. J Bone Joint Surg Am. 2003;85: Viola RW, Boatright KC, Smith KL, Siles JA, Matsen FA 3r. Do shouler patients insure by workers compensation present with worse self-assesse function an health status? J Shouler Elbow Surg. 2000;9: Sallay PI, Hunker PJ, Brown L. Measurement of baseline shouler function in subjects receiving workers compensation versus noncompensate subjects. J Shouler Elbow Surg. 2005;14: Tashjian RZ, Henn RF, Kang L, Green A. The effect of comorbiity on selfassesse function in patients with a chronic rotator cuff tear. J Bone Joint Surg Am. 2004;86: Tashjian RZ, Henn RF, Kang L, Green A. Effect of meical comorbiity on selfassesse pain, function, an general health status after rotator cuff repair. J Bone Joint Surg Am. 2006;88: Henn RF 3r, Kang L, Tashjian RZ, Green A. Patients preoperative expectations preict the outcome of rotator cuff repair. J Bone Joint Surg Am. 2007;89: Kang L, Henn RF, Tashjian RZ, Green A. Early outcome of arthroscopic rotator cuff repair: a matche comparison with mini-open rotator cuff repair. Arthroscopy. 2007;23: Iannotti JP, Norris TR. Influence of preoperative factors on outcome of shouler arthroplasty for glenohumeral osteoarthritis. J Bone Joint Surg Am. 2003;85: Norris TR, Iannotti JP. Functional outcome after shouler arthroplasty for primary osteoarthritis: a multicenter stuy. J Shouler Elbow Surg. 2002;11: Ware JE Jr, Snow KK, Kosinski M, Ganek B. health survey manual interpretation guie. Boston: The Health Institute, New Englan Meical Center; Gummesson C, Atroshi I, Ekahl C. The isabilities of the arm, shouler, an han (DASH) outcome questionnaire: longituinal construct valiity an measuring self-rate health change after surgery. BMC Musculoskelet Disor. 2003;4: Beaton DE, Katz JN, Fossel AH, Wright JG, Tarasuk V, Bombarier C. Measuring the whole or the parts? Valiity, reliability, an responsiveness of the Disabilities of the Arm, Shouler an Han outcome measure in ifferent regions of the upper extremity. J Han Ther. 2001;14: Smith KL, Harryman DT 2n, Antoniou J, Campbell B, Siles JA, Matsen FA 3r. A prospective, multipractice stuy of shouler function an health status in patients with ocumente rotator cuff tears. J Shouler Elbow Surg. 2000;9: Cole DC, Monloch MV, Hogg-Johnson S; Early Claimant Cohort Prognostic Moelling Group. Listening to injure workers: how recovery expectations preict outcomes a prospective stuy. CMAJ. 2002;166: