Abstract
Upper-extremity peripheral nerve disorders (UE–PNDs) occur often; they account for 30% of all workplace injuries and illnesses with concomitant lost work time (U.S. Bureau of Labor Statistics, 2012). The costs for treatment of UE–PNDs are higher than for other conditions of the upper extremity (Centers for Disease Control and Prevention, 2005). UE–PNDs can result in the inability to move, generate force, and manipulate objects, as well as detect touch, temperature, and pressure (Campbell, 2008). Many people with a UE–PND have significant difficulty performing activities of daily living, including self-care, household responsibilities, and leisure and social activities (Chemnitz, Dahlin, & Carlsson, 2013; Thorsén, Rosberg, Steen Carlsson, & Dahlin, 2012). Work is often negatively affected, including loss of employment and decreased productivity, change in employment, modified job demands, and poor return-to-work outcomes (Bruyns et al., 2003; Jaquet et al., 2001). Evidence of socioemotional issues in this population is growing. Thirty-nine percent of patients with a UE–PND scored in the clinically depressed range on a depression screen (Bailey, Kaskutas, Fox, Baum, & Mackinnon, 2009), which is much higher than the 16% reported in the general population (Strine et al., 2008).
Rehabilitation for UE–PNDs has historically focused on remediation of sensorimotor deficits, specifically pain control, muscle strengthening, sensory reeducation, and orthotic fabrication (Ruijs, Jaquet, Kalmijn, Giele, & Hovius, 2005; Skirven, Osterman, Fedorczyk, & Amadio, 2011). Occupational therapy practitioners commonly administer outcome measures to track functional changes in this population, yet strategies to improve performance of daily activities are not routinely emphasized in therapy (Kaskutas & Powell, 2013). Emerging evidence demonstrates poor quality of life (QOL; Bailey et al., 2009; Novak, Anastakis, Beaton, & Katz, 2009). However, these studies often have small sample sizes. For example, in one study, disability was measured in 37 patients with brachial plexus injury (Liu, Lao, Gao, Gu, & Zhao, 2013); another examined 71 patients with median nerve injuries (Vordemvenne, Langer, Ochman, Raschke, & Schult, 2007). A different study examined QOL and activity participation in 49 patients with UE–PNDs (Bailey et al., 2009). Furthermore, few studies have predicted QOL in a multivariate model in this population, and none have explored how disability and QOL relate (Ahmed-Labib, Golan, & Jacques, 2007; Novak et al., 2009).
There is a need to better understand the prevalence of the issues just identified in a large sample of patients presenting for treatment of a wide range of UE–PNDs. Early identification of occupational disruptions, socioemotional problems, and work participation limitations can help clinicians design interventions to address these unmet needs, resulting in improved long-term outcomes in this population. Therefore, the purpose of this study was to identify baseline predictors of disability, work disability, and physical and mental QOL in people with UE–PNDs. We also explored differences among UE–PND diagnostic groups.
Method
Study Design and Participants
Patients with UE–PNDs presenting to one hand surgeon for evaluation between December 2010 and October 2013 were asked to participate in this research. Medical data from 627 consenting patients over age 17 yr with a confirmed UE–PND diagnosis were entered into a database on a secured institutional server. This retrospective cross-sectional research was approved by the institutional review board at the Washington University School of Medicine.
Demographics and Routine Health Information
Self-reported demographic and health history data included age, sex, marital status, body mass index, hand dominance, duration of symptoms, and smoking and drinking status. Responses on an intake pain questionnaire (Chen, Novak, Mackinnon, & Weisenborn, 1998) routinely administered in this surgeon’s practice were examined. This questionnaire includes an abbreviated version of several standardized assessments that measure self-reported levels of pain, depression, stress at home and work, and ability to cope with stress, with ratings given on a 100-point scale. Patients also reported whether they had a history of psychiatric disorders or abuse (or both), whether pain affected sleep quality or intimate relationships (or both), and whether they were employed and involved in workers’ compensation, litigation, or both. They also reported the number of prior nerve surgeries performed and number of current medications they were taking. The remaining variables were categorized as follows:
• Injury onset: slow progressive, slow with exacerbation, sudden with or without event
• Weather affects pain: usually, occasionally, or never
• Work status: every day/same job, every day/modified duties, occasional, or not working
• Ability to perform household chores: same level of chores without pain, same level of chores with pain, reduced amount of chores, or most chores now performed by others.
Physical Examination Results
We explored data from the physical examination, including grip strength, pinch strength, and diagnosis. We categorized participants into seven diagnostic groups: (1) median nerve disorder, (2) ulnar nerve disorder, (3) radial nerve disorder, (4) proximal nerve injury (axillary, long thoracic, suprascapular, or musculocutaneous), (5) dual-nerve compression (two or more different nerves), (6) thoracic outlet syndrome, and (7) brachial plexus injury.
Standardized Measures of Disability and Quality of Life.
The Disabilities of the Arm, Shoulder and Hand (DASH; Beaton, Davis, Hudak, & McConnell, 2001) is a 30-item measure of upper-extremity disability that assesses difficulty performing daily tasks; symptoms over the past week; and the impact of the injury on social activities, sleep, work, and self-esteem. The Work DASH is a four-item module that measures difficulty performing work-related tasks. For both DASH measures, 0 = no disability and 100 = maximal disability. We also explored responses on individual DASH items. Detailed scoring and psychometrics are available at http://www.rehabmeasures.org.
The Short-Form 8 (SF–8; Ware, Kosinski, Dewey, & Gandek, 2001) is a shortened version of the SF–36 Health Survey (Ware, Kosinski, Dewey, & Gandek, 2000) that measures health-related QOL. General health, physical functioning, bodily pain, vitality, social functioning, mental health, and role limitations due to physical and emotional problems are self-reported on a 5- or 6-point Likert scale. The resulting physical component scores and mental component scores were converted to a 0–100 scale for ease of interpretation, with higher scores indicating poorer QOL. Test–retest reliability is as follows: intraclass correlation coefficient = .73–.74, alternative-form reliability r = .85–.90, and convergent validity with the SF–36 r = .93.
Statistical Analysis
All statistical analyses were performed using IBM SPSS Statistics (Version 22; IBM Corp., Armonk, NY). We ran descriptive statistics, computing measures of central tendencies and evaluating normality. We corrected variables that were not normally distributed using a log10 transformation (Portney & Watkins, 2000). DASH and SF–8 scores were calculated as directed in the administrator’s manual (Beaton et al., 2001; Ware et al., 2001) and compared with U.S. population norms using unpaired-samples t tests. We evaluated differences in DASH and SF–8 scores among diagnostic groups using Kruskal–Wallis tests with Bonferroni correction or Mann–Whitney U tests, as appropriate for levels of data.
We used stepwise multivariate linear regression to build four predictive models of (1) disability, (2) work disability, (3) physical QOL, and (4) mental QOL. Variables that correlated with outcome measures on univariate analysis with Spearman’s correlation coefficient ≥.25 (p ≤ .01) were included in respective regression models. Age and sex were forced into all models. We excluded predictors that were collinear with dependent variables (correlation coefficient ≥ .70) or closely tied to the dependent variable. For example, employment status and workers’ compensation were not used to predict work disability. An a priori power analysis indicated that 261 participants were needed to achieve an effect size of F 2 = 0.15 with power of .90 and two-sided α = .01.
Results
Demographics and Participant Characteristics
A total of 627 participants (293 women and 334 men) with a mean age of 49.8 yr (standard deviation [SD] = 15.6 yr) were included in the dataset. Most of the participants were diagnosed with ulnar (23%) and median (21%) nerve disorders. Fifty-five percent reported a sudden onset of symptoms. Two-thirds (67%) were classified as either overweight or obese (see Table 1). The median duration of symptoms was 10.88 mo, and the mean number of current medications was 5.18. Mean ratings on a 100-point scale included pain 52 (SD = 28); depression, 33 (SD = 33); stress at home, 42 (SD = 31); and stress at work, 42 (SD = 32). Of the 436 employed participants (70% of the sample), 59% worked every day at the same job, 11% worked daily at the same job with modified job demands, 10% worked occasionally, and 20% were not working. Little difference in work status was noted across diagnostic groups, with the exception of participants with brachial plexus injuries, of whom only 25% were working every day at the same job and 48% were not working. Most of the 61 workers’ compensation participants who were not working (53%, compared with 14% of non–workers’ compensation participants) scored, on average, 14 points higher on depression (p = .003) and, on average, 9 points higher on stress at home (p = .058) than non–worker’s compensation participants.
Demographics (N = 627)
Disability and Quality of Life
Our sample’s mean disability level was 43 (SD = 22), 33 points higher (p < .001) than the general population’s mean (Hunsaker, Cioffi, Amadio, Wright, & Caughlin, 2002) but similar to levels previously reported in participants with nerve injuries (Liu et al., 2013; Novak, Anastakis, Beaton, Mackinnon, & Katz, 2011). Regarding individual DASH items, 9% of participants were unable to participate in sexual activities and social activities, 10% were unable to use a knife to cut food, and 23% were unable to open a jar. Work DASH results indicated that 25% of participants were completely unable to perform their jobs as they like. Disability and work disability were significantly higher (p < .001) in participants without jobs, on workers’ compensation, with a sudden onset of symptoms, and with depression ratings above the sample’s mean (Table 2). No significant differences were found in disability and work disability between participants with bilateral versus unilateral injuries or dominant versus nondominant hand injuries. Work disability was 7 points higher in participants with unilateral injuries. In a comparison of disability by diagnoses, participants with brachial plexus injuries demonstrated the highest disability, scoring 20 points higher on the DASH and 25 points higher on the Work DASH than participants with median nerve disorders (p ≤ .01). Participants with median nerve disorders demonstrated the least amount of disability and work disability. Cronbach’s α was high for both measures of disability (DASH α = .97, Work DASH α = .98), suggesting that these measures were internally consistent.
Mean DASH and SF–8 Scores and Group Comparisons (Mann–Whitney U Tests)
Note. Higher scores indicate greater disability and poorer quality of life. Score range = 0–100. DASH = Disabilities of the Arm, Shoulder and Hand; MCS = mental component score; PCS = physical component score; SF–8 = Short-Form 8.
p < .05.
Mean physical (44, SD = 20) and mental (37, SD = 22) QOL ratings were significantly poorer (p < .001) than general population norms (Ware et al., 2001), yet they were consistent with previous samples of participants with UE–PNDs (Ahmed-Labib et al., 2007; Novak et al., 2009). Poorer physical and mental QOL were noted in participants who were female, were unemployed, were on workers’ compensation, had a sudden injury onset, and had depression ratings above the mean (p < .001; see Table 2). Mental QOL was poorest in participants with brachial plexus injuries compared with other diagnosis groups (p < .01). Cronbach’s α for the SF–8 was .90.
When we examined the relationship between disability and QOL we found strong, positive correlations between the DASH and the Work DASH (r = .76, p < .01). SF–8 physical component scores were negatively correlated with DASH (r = −.62, p < .01) and Work DASH scores (r = −.59, p < .01), demonstrating that, as disability increases, physical QOL decreases. A moderate correlation was found between the SF–8 mental component score and all other measures.
Disability and Quality of Life Final Regression Models
As noted in Table 3, 23 of the 39 candidate variables we examined in univariate analysis had Spearman’s correlation coefficients ≥.25 (p ≤ .01) with at least one of the four outcome measures. Variables found to predict disability in prior research—for example, workers’ compensation, diagnosis, and symptom duration (Novak et al., 2011)—did not meet criteria for inclusion in multivariate analysis because they poorly correlated with disability in univariate analysis. Final stepwise linear regression models are presented in Table 4.
Variables Correlating With Outcome Measures on Univariate Analysis (Spearman ρ Coefficients)
Note. — = variables that did not meet threshold of ρ ≥ .25 for inclusion in the multivariate model; DASH = Disabilities of the Arm, Shoulder and Hand; MCS = mental component score; NA = correlation was not computed because of a relationship with the outcome variable; PCS = physical component score; SF–8 = Short Form–8.
Final Stepwise Linear Regression Models of Disability and Quality of Life
Note. DASH = Disabilities of the Arm, Shoulder and Hand; QOL = quality of life; R 2 = coefficient of determination; β = unweighted β coefficient.
Work status reference category: work every day/same job.
Onset type reference category: slow progressive.
Household chores reference category: same level of chores without pain.
p < .001.
Ten variables predicted 64% of the variance in disability: (1) depression, (2) pain level, (3) not working at present, (4) grip strength (right), (5) difficulty sleeping, (6) intimate relationships affected, (7) modified job demands, (8) pinch strength (left), (9) sudden onset with acute event, and (10) stress at work. Unstandardized betas for the 2 work status variables were high: The mean DASH scores of participants not working at present and participants working with modified job demands were 16.70 and 9.70 points higher, respectively, than the sample as a whole. Five variables predicted 46% of the variance in work disability: (1) intimate relationships affected, (2) others perform household chores, (3) performs a reduced amount of household chores, (4) difficulty sleeping, and (5) performs same level of household chores with pain. Three of these 5 variables were categories of the ability to perform household chores: Relying on others to perform household chores contributed to DASH scores 49.49 points higher than the sample as a whole (β = 49.49). Of the 17 variables significant in univariate analysis, 4 variables predicted 52% of physical QOL scores in the multivariate model (p < .001): (1) DASH score, (2) pain level, (3) number of medications, and (4) Work DASH score. Four variables predicted 46% of the variance for mental QOL: (1) ability to cope with stress at home, (2) DASH score, (3) stress at home, and (4) difficulty sleeping. Difficulty sleeping yielded a mean mental QOL score 7.21 points higher than the sample as a whole, which is suggestive of poorer QOL.
Discussion
This research supports findings from several smaller studies that found high levels of disability (Ahmed-Labib et al., 2007; Kitajima, Doi, Hattori, Takka, & Estrella, 2006) and poorer QOL (Bailey et al., 2009; Novak et al., 2009) in participants with UE–PNDs. Our cohort of 627 participants with UE–PNDs demonstrated higher disability than reported for participants with upper-extremity amputations (Davidson, 2004), poorer physical QOL than reported for participants with multiple sclerosis (Buchanan, Zhu, Schiffer, Radin, & James, 2008), and poorer mental QOL than a sample of participants with schizophrenia (Kolotkin et al., 2008). Disability was a strong predictor in multivariate models of both physical and mental QOL. QOL and disability were strongly associated in participants with UE–PNDs, with poorer QOL found in participants with higher levels of disability. Disability and work disability correlated closely, as did physical and mental QOL.
Difficulty sleeping was the only variable that made it into three of the four multivariate models: (1) disability, (2) work disability, and (3) mental QOL. Previous studies that have examined sleep have focused primarily on people with carpal tunnel syndrome (Patel et al., 2014); in our sample with a variety of UE–PNDs, 73% reported sleep disturbances. Poor sleep quality has been found to negatively affect QOL and daily functioning, including work performance (Swanson et al., 2011; Zeitlhofer et al., 2000). Sleep is an area that needs to be further explored in research and should be addressed in clinical practice. Occupational therapy practitioners can help patients develop restful routines, such as performing relaxing activities and preparing the physical environment for sleep (American Occupational Therapy Association [AOTA], 2014).
Problems with intimate relationships were also predictive of disability and work disability in multivariate models. Our exploration of individual DASH items showed that many participants needed help with intimate self-care tasks, and a large percentage no longer participate in, or have severe difficulty performing, sexual and social activities. Meaningful intimate relationships and social engagement are critical for emotional well-being and overall QOL (Reis, Sheldon, Gable, Roscoe, & Ryan, 2000; Robinson & Molzahn, 2007). Clinicians should discuss these areas with patients and brainstorm strategies to help them maintain meaningful relationships while dealing with temporary or permanent consequences of UE–PNDs.
Depression and its effect on disability and QOL are just beginning to be understood in this population. Depression accounted for 28% of the variance for disability in the first step of the multivariate model and was the top predictor in the final step. This finding is consistent with those of other studies highlighting that depression is closely associated with disability and patient satisfaction (Adogwa et al., 2012). Bailey et al. (2009) found that depression accounted for a large proportion of the variance in QOL and should be routinely screened in patients with UE–PNDs.
Other person-level variables predicting disability included stress at work, sudden onset of symptoms, decreased strength, and increased pain. Previous studies have found that pain has a strong predictive effect on disability (Novak et al., 2009; Novak, Anastakis, Beaton, Mackinnon, & Katz, 2010). We found that pain was predictive of mental QOL, but not physical QOL. Pain should continue to be routinely emphasized in clinical practice, yet it should be addressed in the context of meaningful activity and its impact on life domains. Our finding that diagnosis, symptom duration, and dominant hand or bilateral involvement did not predict disability suggests that prevention of disability should be a focus with all patients.
Our sample’s mean work disability score of 52 was 21% higher than the mean disability score of 43, highlighting the greater problems performing work tasks than basic and instrumental activities of daily living. The importance of household task performance to work disability was striking in this research. All three categories signifying a reduced ability to perform household chores, along with difficulty sleeping and problems with intimate relationships, predicted 46% of work disability. Helping patients identify ways to maintain home and community roles may facilitate their resumption of worker roles. It is important to note that areas emphasized in work rehabilitation, such as strength and pain, did not predict work disability, nor did diagnosis. A patient’s ability to work should not be assumed on the basis of his or her diagnosis or impairment.
The DASH is one of the most widely used standardized outcome measures of overall disability and work disability; however, it also has a major role in identifying daily living tasks to target in a client-centered treatment plan. Left on their own to deal with functional deficits, patients may exceed restrictions or perform tasks in an unsafe manner (Kaskutas & Powell, 2013). Occupational therapy practitioners can help patients identify adaptive techniques, innovative strategies, and environmental modifications to compensate for temporary or long-term functional deficits, such as using one-handed dressing techniques or negotiating modified work duties with their employer. Expansion of routine therapy to address participation in meaningful tasks, in addition to interventions to improve sensorimotor deficits and manage pain, will influence occupational performance, disability, and QOL (Guzelkucuk, Duman, Taskaynatan, & Dincer, 2007).
This research has several limitations. All participants in our sample were from a single surgeon practicing in an urban setting. However, scores on a standardized outcome measure were similar to those reported by other researchers, suggesting more similarities than differences. Duration of symptoms varied widely in our sample, which may be of concern, yet symptom duration was not found to be predictive of disability when examined contiguously with other variables. This research identified several differences in patients receiving workers’ compensation; however, workers’ compensation was not a predictor in multivariate analyses, possibly because of the small portion of patients receiving workers’ compensation. We were unable to recruit every patient seen during recruitment because of scheduling and staffing. The retrospective cross-sectional design limited certain data to a questionnaire used in clinical practice (Chen et al., 1998) that has not been previously validated. Analysis of longitudinal data with this sample, in which we will explore other items, such as minimal clinically important differences, is underway.
Implications for Occupational Therapy Practice
Use of the Occupational Therapy Practice Framework: Domain and Process (3rd ed.; AOTA, 2014) to guide treatment of patients with UE–PNDs ensures that all domains that affect health, well-being, and participation in life are addressed. To deliver client-centered care, we recommend that occupational therapy practitioners working with clients with UE–PNDs do the following:
• Broaden treatment approaches past remediation to adapt environments and tasks, teach compensatory strategies, prevent disability, and promote health and QOL
• Administer disability and QOL measures to track outcomes, and use the results to identify performance problems that may be temporary or permanent
• Screen for client factors identified in this research, such as depression, stress, and coping
• Evaluate and develop occupation-based interventions to enhance patients’ performance of basic and instrumental activities of daily living, work, sleep, and social participation.
Conclusion
This sample of patients presenting to a hand surgeon for evaluation and treatment of UE–PNDs displayed a high degree of disability and work disability as well as poor physical and mental QOL. Many domains of occupational therapy predicted disability and QOL, including depression and stress, problems with sleep and intimate relationships, deficits in work and household performance, and pain. Early referral to occupational therapy services is recommended to help patients identify methods to perform meaningful daily occupations, develop adaptive coping strategies, and prevent disability. Future research to gain a deeper understanding of issues that patients with UE–PNDs experience can improve client-centered care, surgical and rehabilitation outcomes, and QOL in this population.
Footnotes
Acknowledgment
We thank the research participants and staff at the Department of Plastic and Reconstructive Surgery at Washington University School of Medicine for their assistance with this research.