Abstract
Background: Substantial differences between disability and impairment are commonplace and puzzling. Subjective (psychosocial) factors may be paramount given that pain is a more important determinant of perceived overall arm-specific disability than is objective elbow impairment. To further evaluate the relationship between impairment and disability, we tested the hypothesis that objective loss of elbow motion predicts perceived elbow-related task-specific disability better than does pain after elbow trauma.
Methods: One hundred and fifty-eight patients were evaluated at a median of twenty-six months after a traumatic elbow injury and completed the Disabilities of the Arm, Shoulder and Hand (DASH) questionnaire. Predictors of the total DASH score and of the scores for individual DASH items that were expected to be related to elbow function were evaluated with univariate and multivariate analyses.
Results: Motion accounted for 35% of the variability in the total DASH score, for 11% to 12% of the variability in the responses to questions specific to hand-based activities, and for 24% to 33% of the variability in the scores for tasks depending on elbow motion. Pain accounted for 41% of the variability in the total DASH score and was a better predictor than motion of disability associated with three tasks: opening a tight jar (with pain and motion accounting for 24% and 11% of the variability, respectively), pushing open a door (25% and 12%, respectively), and placing an object overhead (28% and 25%, respectively). None of the multivariate models explained more than 53% of the variability in the DASH scores.
Conclusions: Objective physical elbow impairment correlated with self-reported disability with respect to specific tasks, but a large proportion of disability remains unexplained. Further research is needed to better understand the differences between objective impairment and perceived disability.
According to the International Classification of Functioning, Disability and Health (ICF) published by the World Health Organization1, impairments are the manifestations of an underlying pathological condition and represent objective physical deviation or loss due to problems in bodily function or structure. Disability refers more generally to activity limitations, which can be the result of objective phys-ical impairments or psychosocial factors.
Prior studies have documented a wide variation in disability among patients with similar levels of hand or arm impairment2-4. For instance, the ranges and standard deviations for disability perceived by patients with trigger finger2,3 or carpal tunnel syndrome2-4—diagnoses with relatively limited variation in objective measurable impairment—have been large.
When a condition is associated with a broader range of objective physical impairment, that impairment explains a relatively small percentage of perceived disability5-10. For instance, the objective physical impairment of patients recovering from a fracture of the distal part of the radius explained only 25% of the reported disability8.
Doornberg et al.10 found that only 17% of the variation in disability after elbow trauma was explained by impairment. It is unclear if this relationship holds true when the relationship between impairment and disability is evaluated at the level of specific functional tasks. The current investigation addressed the hypothesis that elbow-related task-specific disability is determined more by impairment of elbow motion than by pain. A better understanding of the relationship between impairment, pain, and disability may help to achieve better treatment outcomes in the future.
Patients
During a four-year period, physical examinations were performed and health-status data were collected during evaluations of patients at various stages of recovery after a complex elbow injury. These evaluations were done as part of nine prospective and retrospective studies, all approved by the human research committees at our institutions in the United States and in The Netherlands. Of 324 eligible patients in those studies, twenty-three declined to participate in the current investigation, thirty-seven had died, and thirty-eight could not be located, leaving 226 patients who were evaluated for inclusion in the present study. The inclusion criteria for this study were (1) an operatively treated elbow fracture, (2) an age of at least eighteen years at the time of injury, (3) performance of the evaluation more than four months following the most recent injury or surgery, and (4) a completed Disabilities of the Arm, Shoulder and Hand (DASH) questionnaire available in our research records. Patients with rheumatoid arthritis and/or elbow instability were excluded. One hundred and fifty-eight patients (ninety-six men and sixty-two women) who met these criteria form the study cohort for the current investigation.
Evaluation
An investigator who was not involved in patient care evaluated each patient at a median of thirty-eight months (average, 101 months; range, four to 364 months) after the injury and a median of twenty-six months (average, sixty-four months; range, four to 363 months) after the most recent surgery. Patients with incomplete recovery from the injury were included in order to (1) capture impairment and disability experienced by patients at various stages of recovery, (2) increase the variation in impairment, and (3) limit ceiling effects (i.e., too many patients with little impairment). The evaluation consisted of an interview, a physical examination (measurement of ulnohumeral and forearm motion), and administration of the DASH questionnaire. The DASH questionnaire is a validated outcomes instrument specific to the upper extremity and is applicable to a wide variety of problems11. The questionnaire contains thirty items evaluated on 5-point Likert scales: twenty-one items evaluate difficulty with performance of certain activities (some ask about specific tasks such as "turning a key," whereas others address difficulty with activities in general, such as "managing transportation needs"), five evaluate symptoms (e.g., pain and stiffness), and one each evaluates limitations of social function, work function, sleep, and confidence secondary to the extremity under evaluation. The total DASH score is scaled from 0 to 100 points, with higher scores indicating greater disability. We selected eight questions that measure a patient's ability to perform specific tasks that are likely to be affected by loss of elbow or forearm motion: (1) opening a tight or new jar, (2) turning a key, (3) pushing open a heavy door, (4) placing an object on a shelf above the head, (5) changing a light bulb overhead, (6) washing or blow-drying hair, (7) washing the back, and (8) putting on a pullover sweater. Patients completing the DASH questionnaire are asked to rate the degree of their ability to perform the tasks on a Likert scale ranging from 1 to 5, with higher scores indicating more difficulty (1 = no difficulty, 2 = mild difficulty, 3 = moderate difficulty, 4 = severe difficulty, and 5 = unable to perform the activity). In addition, perception of pain and stiffness is rated on the DASH questionnaire (1 = none, 2 = mild, 3 = moderate, 4 = severe, and 5 = extreme). Each of the eight tasks requires substantial elbow or forearm motion in one or more directions, which provided us with an opportunity to evaluate discrepancies between objective impairment (measured motion) and perceived disability and perceived stiffness as rated on the DASH. We thought that it would be more reliable to use the 5-point Likert scales of the DASH to measure pain and stiffness than to use a second, repeat question separate from the DASH.
Statistical Analysis
Univariate Analysis
The dependent (or response) variables were the total DASH score, the scores for the eight individual DASH questions that were thought likely to be affected by loss of elbow and forearm mobility, and the score for the individual DASH question regarding perceived stiffness. The independent (or explanatory) variables that we investigated included flexion, extension, pronation, supination, age at the time of follow-up, number of operations, pain, arthrosis, time since the last surgery, secondary gain, associated armn injuries, sex, limb dominance, distal humeral fracture, coun-try of residence, ulnar neuropathy, and occupation.
Associations between continuous explanatory variables and the response variables were evaluated with use of Spearman correlations. Associations between dichotomous explanatory variables and the response variables were evaluated with use of the Mann-Whitney U test. Associations with a p value of <0.05 were considered significant.
Multivariate Analysis
Multiple linear regression analysis was used to assess the ability of the explanatory variables to account for variation in the response variables, while accounting for any confounding between the explanatory variables. A multiple linear regression model produces a statistic called the adjusted R2, which reflects the percentage of the overall variability in the response variable that can be explained or accounted for by the explanatory variables included in the multiple linear regression model.
We ran several models, including two backward stepwise multiple linear regression models (a model that includes all of the entered variables initially and then iteratively removes variables from the model until the best-fit model is achieved according to set criteria), one including pain and one excluding pain; a model with pain as the only explanatory variable; and a model with only those explanatory motion variables thought to be related to the specific task. Comparison of the variability accounted for by each model (the adjusted R2) provides a measure of the relative influence of each explanatory variable on the overall variation in the response variable.
The number of explanatory variables that can be included in a multivariate model is limited by the overall sample size of the study. Therefore, instead of entering all of our potential explanatory variables into the backward stepwise models, we chose to enter only those variables that were either significant (p < 0.05) or nearly significant (p < 0.10) in the univariate analysis, a common cutoff value for inclusion of variables in regression modeling.
For each multivariate model, a multivariate analysis of variance was performed to assess significance, which indicates a linear relationship between at least one of the explanatory variables and the dependent variable. The statistical methodology is summarized in Table I.
A power analysis indicated that a minimum sample size of 100 patients would provide 90% statistical power (ß = 0.1, a = 0.05) to detect a moderate correlation (rho = 0.30), with excellent precision, between flexion and the total DASH score, and between pain and the total DASH score.
The mean age at the time of injury was forty-four years (range, eighteen to eighty years). Eighty-four left arms and seventy-four right arms were involved. Eighty-eight patients (56%) had involvement of the dominant arm. Of the 120 patients employed outside of the home at the time of the injury, ninety-two performed desk-based work and twentyeight were laborers. Five patients were students, and the remaining thirty-three patients did not have a job (sixteen were retired, fourteen were unemployed, and three were disabled). Seven patients had a secondary gain issue: four patients had filed a disability claim, one had filed a Workers' Compensation claim, and two patients had a narcotic addiction.
The injuries included a fracture-dislocation of the elbow in sixty-nine patients, an intra-articular fracture of the distal part of the humerus in forty-nine, an isolated radial head fracture in thirty-two, and an olecranon fracture in eight. Seventeen patients had associated injuries of the same upper extremity: nine had a distal radial fracture, three had a diaphyseal forearm fracture, and one patient each had a scaphoid fracture, scaphoid and distal radial fractures, a distal ulnar fracture, a metacarpal fracture of the ring and long fingers, and a scapular fracture. Eighty-six patients underwent an average of two subsequent operations on the affected elbow (range, one to sixteen operations).
At the index evaluation, the mean flexion arc (and standard deviation) was 101° ± 35° (range, 0° to 150°), with an average of 124° ± 18° (range, 55° to 150°) of flexion and an average flexion contracture of 23° ± 21° (range, 10° to 100°). The mean forearm rotation arc was 149° ± 38° (range, 0° to 180°), with 76° ± 20° (range, 0° to 90°) of pronation and 72° ± 25° (range, 0° to 90°) of supination. Thirty patients had symptoms or signs of ulnar neuropathy at the time of follow-up. Sixty-two patients showed radiographic signs of arthrosis; it was mild in thirty-six of them, moderate in sixteen, and severe in ten, according to criteria of Broberg and Morrey12.
The total DASH scores averaged 20 ± 20 points (range, 0 to 93 points). The mean score was 2.0 ± 1.2 points for item 1, 1.5 ± 0.9 points for item 2, 1.9 ± 1.1 points for item 3, 2.1 ± 1.3 points for item 4, 1.8 ± 1.2 points for item 5, 1.7 ± 1.0 points for item 6, 2.2 ± 1.3 points for item 7, and 1.7 ± 0.9 points for item 8. The scores for perceived stiffness and pain averaged 2.4 ± 1.2 and 2.2 ± 0.9 points, respectively.
Statistical Analysis
The details and results of the statistical analysis are presented in the Appendix and in Tables I, II, and III. By performing each of the multivariate analyses with four distinct models, we were able to determine the relative ability of (1) all of the explanatory variables, (2) all of the explanatory variables except pain, (3) pain alone, and (4) motion alone to account for variance in the DASH scores. These analyses demonstrated that pain accounts for more of the variance in the overall DASH score than do motion and other objective variables, but motion is more important than pain in predicting the level of some of the specific tasks—primarily those that could be most directly related to the need for elbow motion (Tables I, II, and III).
Our hypothesis was confirmed in large part: disability correlated with objective physical impairment of the ability to perform specific tasks, and impairment explained more of the variability in the DASH scores for most of these tasks than pain did. For instance, impairment in extension corresponded with difficulty in changing a light bulb overhead and placing an object on a shelf above the head, and impairment in forearm rotation corresponded with difficulty in turning a key or opening a tight jar. Flexion was among the strongest predictors in each model, except for those with activities that clearly require extension. Furthermore, objective physical impairments of the elbow explained a larger proportion of the variability in the disability when the task depended more on elbow motion (e.g., changing a light bulb overhead and blow-drying the hair) than when it was a hand-based activity (e.g., opening a jar and turning a key). As expected, there was a strong relationship between objective physical motion impairments and perceived stiffness; however, pain accounted for more of the variability in perceived stiffness scores than did objective impairments in motion. Finally, for most specific tasks, physical impairments explained more of the variability in the DASH scores than did pain.
Although the amount of variability explained by objective physical impairment in the current study (35%) was substantially higher than that in the study by Doornberg et al.10 (17%), the current findings are consistent with those of Doornberg et al. in that pain was the strongest predictor of the total DASH scores (36% of the variability was explained by pain in the study by Doornberg et al. compared with 41% in our study). The important difference between the current study and the study by Doornberg et al. is that we found objective physical impairment to have a greater influence than pain on disability when the relationship between the task and the impairment is more direct and specific. In other words, subjective factors such as pain have a greater influence when disability is measured with respect to the entire arm rather than with respect to the specific anatomical site involved. This is to some degree intuitive and leads to the more general hypothesis that less specific measures of disability are more easily influenced by factors other than objective physical impairment.
Although we found the relationship between impairment and disability to be stronger when we measured disability at the level of specific tasks, the majority of the variability in the DASH scores could not be accounted for by our best multivariate models—even those including measures of pain. In other words, there is a substantial discrepancy between objective physical impairments and perceived disability related to the injured elbow that remains unexplained. This observation is consistent with those in prior investigations of the elbow13,14 and of the wrist8 and leg15. For instance, MacDermid et al.8 found that physical impairment accounted for only 25% of perceived disability (and pain) after wrist fractures. Mock et al.15 found that 23% of perceived disability after lower-extremity fractures was explained by the range of motion and strength and 29% was explained by pain. Thus, current scientific analyses suggest that health status measures such as the DASH seem to be measuring something beyond and independent of a patient's objective physical impairment or even their subjective experience such as pain.
Discrepancies between impairment and disability have also been documented in other fields of medicine. Prior research regarding chronic pain has established that perception of pain, other symptoms, and perceived disability are extremely variable and strongly psychosocially mediated16,17. Psychosocial factors influence the degree to which a variety of symptoms are perceived and expressed as disabling or painful18-20. For instance, among several clinical and sociodemographic variables, depression was the best predictor of disability among more than a thousand patients with osteoarthritis21. For some patients, the distress and illness behavior that develop secondary to an underlying physical problem can be just as disabling as the original physical problem9. In addition, secondary gain issues such as lawsuits, insurance claims, or claims for Workers' Compensation may influence reported pain and disability8,10.
In a recent meta-analysis of thirty-one articles that reported associations of impairment with patient-rated disability and health status22, only 36% of the variability in the disability scores and 13% of the variability in the health status scores were explained by impairment. The fact that 64% of the variation in disability and 87% of the variation in health status remain unexplained by objective disease factors and impairment implies that there are substantial, yet incompletely understood, opportunities for improvement in quality of life independent of physical impairment.
Recent data from the World Health Organization's World Health Surveys indicate that psychological distress (specifically depression) not only produced more decrement in health than did other chronic illnesses, but also was responsible for incrementally worse health in association with other chronic diseases23. Prior work has demonstrated that depression has a direct, moderate correlation with perceived disability related to arm conditions, and the slope of the relationship is the same for several common diagnoses such as distal radial fracture, carpal tunnel syndrome, and lateral elbow pain2. This line of evidence supports the concept that discrepancies between impairment and perceived disability are explained, to a large degree, by psychosocial factors.
We were able to assess culture-related differences among psychosocial factors since our study included large numbers of patients from two different cultures. The country of residence was strongly correlated with each of the questions in the univariate analysis, with Dutch patients reporting less disability than American patients; however, multivariate analyses showed that the country of residence was not an important predictor of perceived disability after controlling for diagnosis, time since the last surgery, objective physical impairments, and pain. This may be a result of the fact that most of the patients from The Netherlands were evaluated more than ten years after the injury, whereas the patients in Boston were evaluated after no more than a few years. In any case, we think that variations in perceived disability across cultures merit additional investigation. It has previously been reported, for instance, that complaints of pain may differ between cultures24-26, and differences in perception of health status between populations have been described by others27,28.
These data should be interpreted in the light of their limitations. First, we used a convenience sample of data from long-term prospective and retrospective studies and did not follow a prospective protocol. Second, some may question the use of Likert scales for pain and stiffness that are part of the DASH questionnaire, but we thought that asking the same question outside of the DASH questionnaire might be confusing and would be unlikely to result in important differences. Third, since we made many comparisons, there is a substantial likelihood that some are spurious. Therefore, the results of the univariate analysis in particular should be interpreted with caution. We did not correct the significance level for multiple comparisons because the purpose of the univariate analysis was simply to limit the number of variables entered into the multivariate statistical analyses.
While the current study suggests that impairment is a better predictor of disability with respect to specific functional tasks than of overall arm-specific disability, there is still a substantial discrepancy between impairment and disability that is not accounted for either by pain or by objective physical impairments. Because some research suggests that the influence of psychosocial factors (depression in particular) may best explain the discrepancy between impairment and disability and since many psychosocial factors are amenable to treatment, additional research along these lines is merited.
A detailed description of the statistical analysis is available with the electronic versions of this article, on our web site at (go to the article citation and click on "Supplementary Material") and on our quarterly CD/DVD (call our subscription department, at 781-449-9780, to order the CD or DVD). 
Note: The authors are grateful to the AO Documentation Center in Davos, Switzerland, for managing the fracture database for the Departments of Orthopaedic Surgery and General Surgery at the Academic Medical Center in Amsterdam, The Netherlands, over the last decades. All Dutch cases included in the study were identified through this database. We thank the Departments of Surgery and Traumatology for their permission to use the data on their patients.
Ring D, Kadzielski J, Fabian L, Zurakowski D, Malhotra LR, Jupiter JB. Self-reported upper extremity health status correlates with depression. J Bone Joint Surg Am.2006;88:1983-8.881983
2006
[PubMed][CrossRef]
Ring D, Guss D, Malhotra L, Jupiter JB. Idiopathic arm pain. J Bone Joint Surg Am.2004;86:1387-91.861387
2004
Gummesson C, Atroshi I, Ekdahl C. The disabilities of the arm, shoulder and hand (DASH) outcome questionnaire: longitudinal construct validity and measuring self-rated health change after surgery. BMC Musculoskelet Disord.2003;4:11.411
2003
[CrossRef]
Acosta-Rojas ER, Comas M, Sala M, Castells X. Association between visual impairment and patient-reported visual disability at different stages of cataract surgery. Ophthalmic Epidemiol.2006;13:299-307.13299
2006
[CrossRef]
Rietman JS, Dijkstra PU, Debreczeni R, Geertzen JH, Robinson DP, De Vries J. Impairments, disabilities and health related quality of life after treatment for breast cancer: a follow-up study 2.7 years after surgery. Disabil Rehabil.2004;26:78-84.2678
2004
[CrossRef]
O'Connor D, Kortman B, Smith A, Ahern M, Smith M, Krishnan J. Correlation between objective and subjective measures of hand function in patients with rheumatoid arthritis. J Hand Ther.1999;12:323-9.12323
1999
MacDermid JC, Donner A, Richards RS, Roth JH. Patient versus injury factors as predictors of pain and disability six months after a distal radius fracture. J Clin Epidemiol.2002;55:849-54.55849
2002
[CrossRef]
Waddell G, Main CJ, Morris EW, Di Paola M, Gray IC. Chronic low-back pain, psychologic distress, and illness behavior. Spine.1984;9:209-13.9209
1984
[CrossRef]
Doornberg JN, Ring D, Fabian LM, Malhotra L, Zurakowski D, Jupiter JB. Pain dominates measurements of elbow function and health status. J Bone Joint Surg Am.2005;87:1725-31.871725
2005
[CrossRef]
Hudak PL, Amadio PC, Bombardier C. Development of an upper extremity outcome measure: the DASH (disabilities of the arm, shoulder and hand) [corrected]. The Upper Extremity Collaborative Group (UECG). Am J Ind Med.1996;29:602-8. Erratum in: Am J Ind Med. 1996;30:372.29602
1996
[CrossRef]
Broberg MA, Morrey BF. Results of treatment of fracture-dislocations of the elbow. Clin Orthop Relat Res.1987;216:109-19.216109
1987
Angst F, John M, Pap G, Mannion AF, Herren DB, Flury M, Aeschlimann A, Schwyzer HK, Simmen BR. Comprehensive assessment of clinical outcome and quality of life after total elbow arthroplasty. Arthritis Rheum.2005;53:73-82.5373
2005
[CrossRef]
King GJ, Richards RR, Zuckerman JD, Blasier R, Dillman C, Friedman RJ, Gartsman GM, Iannotti JP, Murnahan JP, Mow VC, Woo SL. A standardized method for assessment of elbow function. Research Committee, American Shoulder and Elbow Surgeons. J Shoulder Elbow Surg.1999;8:351-4.8351
1999
[CrossRef]
Mock C, MacKenzie E, Jurkovich G, Burgess A, Cushing B, deLateur B, McAndrew M, Morris J, Swiontkowski M. Determinants of disability after lower extremity fracture. J Trauma.2000;49:1002-11.491002
2000
[CrossRef]
Elliott TE, Renier CM, Palcher JA. Chronic pain, depression, and quality of life: correlations and predictive value of the SF-36. Pain Med.2003;4:331-9.4331
2003
[CrossRef]
Gonzales VA, Martelli MF, Baker JM. Psychological assessment of persons with chronic pain. NeuroRehabilitation.2000;14:69-83.1469
2000
Meredith P, Strong J, Feeney JA. Adult attachment, anxiety, and pain self-efficacy as predictors of pain intensity and disability. Pain.2006;123:146-54.123146
2006
[CrossRef]
Meredith PJ, Strong J, Feeney JA. The relationship of adult attachment to emotion, catastrophizing, control, threshold and tolerance, in experimentally-induced pain. Pain.2006;120:44-52.12044
2006
[CrossRef]
Severeijns R, Vlaeyen JW, van den Hout MA, Weber WE. Pain catastrophizing predicts pain intensity, disability, and psychological distress independent of the level of physical impairment. Clin J Pain.2001;17:165-72.17165
2001
[CrossRef]
Rosemann T, Laux G, Kuehlein T. Osteoarthritis and functional disability: results of a cross sectional study among primary care patients in Germany. BMC Musculoskelet Disord.2007;8:79.879
2007
[CrossRef]
Weisscher N, de Haan RJ, Vermeulen M. The impact of disease-related impairments on disability and health-related quality of life: a systematic review. BMC Med Res Methodol.2007;7:24.724
2007
[CrossRef]
Moussavi S, Chatterji S, Verdes E, Tandon A, Patel V, Ustun B. Depression, chronic diseases, and decrements in health: results from the World Health Surveys. Lancet.2007;370:851-8.370851
2007
[CrossRef]
Watson PJ, Latif RK, Rowbotham DJ. Ethnic differences in thermal pain responses: a comparison of South Asian and White British healthy males. Pain.2005;118:194-200.118194
2005
[CrossRef]
Zborowski M. Cultural components in responses to pain. J Soc Issues.1952;8:16-30.816
1952
[CrossRef]
Vlaar AP, ten Klooster PM, Taal E, Gheith RE, El-Garf AK, Rasker JJ, van de Laar MA. A cross-cultural study of pain intensity in Egyptian and Dutch women with rheumatoid arthritis. J Pain.2007;8:730-6.8730
2007
[CrossRef]
Heslin JM, Soveri PJ, Winoy JB, Lyons RA, Buttanshaw AC, Kovacic L, Daley JA, Gonzalo E. Health status and service utilisation of older people in different European countries. Scand J Prim Health Care.2001;19:218-22.19218
2001
[CrossRef]
Hildingh C, Luepker RV, Baigi A, Lidell E. Stress, health complaints and self-confidence: a comparison between young adult women in Sweden and USA. Scand J Caring Sci.2006;20:202-8. Erratum in: Scand J Caring Sci. 2006;20:472.20202
2006
[CrossRef]