art%3A10.1023%2FB%3AJOOR.0000015008.25177.8b

P1: JLSJournal of Occupational Rehabilitation [jor] pp1091-joor-479243 December 19, 2003 15:43 Style file version Nov 28th, 2002

Journal of Occupational Rehabilitation, Vol. 14, No. 1, March 2004 ( C 2004)

Identifying Work Organization Targets for aWork-Related Musculoskeletal SymptomPrevention ProgramGrant D. Huang1;2;3 and Michael Feuerstein2;3;4;5

While research linking work organization factors to work-related musculoskeletal disordershas been increasing, there is still a need to delineate specific dimensions to be targeted byintervention programs. The present cross-sectional investigation identified work organiza-tion risk factors for low back (LB) and upper extremity (UE) symptoms and determinedthe magnitudes of such associations. Questionnaires containing items on ergonomic, indi-vidual psychosocial, and occupational psychosocial factors were administered to a sampleof workers (n D 248 U.S. Marines) in previously identified high-risk job categories formusculoskeletal disorders. Study participants were categorized into groups of either hav-ing LB symptoms only, UE symptoms only, concurrent LB and UE symptoms, or beingasymptomatic on the basis of self-report. Additionally, measures of pain intensity, phys-ical function, and mental health were obtained. Linear regression analyses adjusting fordemographics, ergonomic factors, and individual psychosocial factors indicated that deci-sion authority and experienced responsibility for work were significant correlates for painintensity during the week. Logistic regression analyses indicated that ergonomic stressorswere a risk factor for all symptomatic groups (OR D 1.02 per point increase; 95% CI:1.01.1). Time pressure (OR D 1.2 per point increase; 95% CI: 1.01.4) was also a signif-icant risk factor for all symptomatic groups, while cognitive processing placed workers athigher risks for concurrent LB and UE symptoms (OR D 1.2; 95% CI: 1.01.4). Interper-sonal demands placed individuals at a lower risk for LB symptoms (OR D 0.8; 95% CI:0.51.0). Findings highlight the importance of intervention approaches that address timepressure, cognitive processing factors, and interpersonal demands at work. In light of pastbiobehavioral studies, these results also suggest that job redesign and interventions that

1Office of Research & Development, U.S. Department of Veterans Affairs, Washington, District of Columbia.2Department of Preventive Medicine & Biometrics, Uniformed Services University of the Health Sciences,Bethesda, Maryland.

3Department of Medical & Clinical Psychology, Uniformed Services University of the Health Sciences, Bethesda,Maryland.

4Division of Behavioral Medicine, Department of Psychiatry, Georgetown University School of Medicine,Washington, District of Columbia.

5Correspondence should be directed to Michael Feuerstein, PhD, MPH, Department of Medical & ClinicalPsychology, Uniformed Services University of the Health Sciences, 4301 Jones Bridge Road, Bethesda, Maryland20814; e-mail: [email protected].

13

1053-0487/04/0300-0013/0 C 2004 Plenum Publishing Corporation

P1: JLSJournal of Occupational Rehabilitation [jor] pp1091-joor-479243 February 4, 2004 9:55 Style file version Nov 28th, 2002

14 Huang and Feuerstein

address a workers workstyle when faced with increased work demands may help reducethe likelihood of musculoskeletal symptoms and/or their intensity.KEY WORDS: job stress; work organization; ergonomics; low back pain; upper extremity; symptoms;intervention.

INTRODUCTION

Work-related musculoskeletal disorders (WMSDs) can have a major impact on workerhealth, function, performance, and productivity (13). In turn, the associated burden andcosts have given rise to a number of research and clinically based efforts designed tohelp develop innovative prevention strategies. However, there still remains a need to morespecifically identify those workplace physical and psychosocial factors to be targeted insecondary prevention approaches aimed at reducing fatigue, discomfort, pain, and/or func-tional limitations in the lower back, hand, wrist, arm, elbow, shoulder, and/or neck regions(25).

Epidemiological studies have linked a number of demographic, medical, ergonomic,and psychosocial variables to work-related musculoskeletal outcomes (2,3,5,6). More re-cently, there has been a growing awareness of the link between occupational psychosociali.e., job stress) factors and WMSDs (25). Occupational psychosocial factors that havegenerated particular interest by researchers have been in relation to characteristics of workand how it is organized, or work organization (7). In the National Occupational ResearchAgenda (NORA), the National Institute for Occupational Safety and Health (NIOSH) de-scribes work organization as being comprised of six major components (8). These com-ponents are scheduling, job design, interpersonal, career concerns, management style, andorganizational characteristics. These categories closely resemble those proposed in an ear-lier model by Cooper and Marshall (9) on the dynamics of work stress. In this model, thesources of stress or exposures are classified as those intrinsic to the job, role in the organiza-tion, relationships at work, career development, organizational structure and climate, and thehomework interface. Descriptions of work organization have also included other psychoso-cial aspects of work such as interpersonal factors, supervision, practices, and organizationalcharacteristics (7,8,10). However, it is unclear whether certain categories play a more salientrole in WMSDs than others and whether prevention efforts should be concentrated on one,some, or all dimensions of work organization. Research that seeks to differentiate thosework organization factors that can impact work and health outcomes related to WMSDscan, in turn, assist with elucidating potential mechanisms and the development of strategicintervention approaches.

Given the recognition of the multidimensional nature of work-related musculoske-tal outcomes, there still remains a lack of intervention efforts that effectively incorpo-rate work organization factors. Exposures to adverse biomechanical conditions such ashigh repetition, sustained awkward postures, excessive force placed on tendons, joints, andmuscles, and inadequate recovery of musculoskeletal structures have been widely estab-lished in the literature as risk factors for both back and upper extremity symptoms anddisorders (2,3,11,12). Compared to ergonomic factors though, significantly less researchin the field has been conducted on occupational psychosocial factors. Evidence exists tosuggest that workstation redesign, exercise, employee involvement, and management sup-port have positive impacts on back and upper extremity outcomes (3). However, reviews


Musculoskeletal Symptom Prevention Program Targets 15

of individual- or organizational-level stress management programs have produced mixedresults (1315).

The present investigation was intended to help provide an empirical basis for guidingfuture efforts to reduce/eliminate the impact of specific workplace stressors on workermusculoskeletal health and associated costs. It builds upon a body of work designed to assistin the development of a secondary prevention program for musculoskeletal disorders withinan organization (i.e., U.S. Marine Corps) for which these problems present a significantchallenge (16,17). In addition to the role of ergonomic and individual psychosocial factors,it was hypothesized that different components of work organization could be identified asrisk factors. In turn, findings could highlight those work organization dimensions deservingparticular attention in secondary prevention efforts for low back (LB) and upper extremity(UE) symptoms.

METHODS

Study Participants

Study participants were recruited at selected locations in Washington, DC, Virginia,and California through news bulletins, electronic mail, and section heads that announced aresearch study on musculoskeletal disorders. Specifically, those invited to participate wereactive duty, enlisted U.S. Marines involved in jobs identified as having relatively higherrates for musculoskeletal-related clinic visits (e.g., Auditing & Accounting, Disbursing,Image Interpretation, Information & Education, and Transportation) (16). These jobs rep-resented ones included in the U.S. Department of Defense occupational categories of Func-tional Support & Administration and Communications & Intelligence (18). Respondentsto recruitment announcements attended an information session that further provided studydetails. Of the 400 individuals attending these sessions, 307 provided informed consent toparticipate.

Eligibility Criteria

In the present investigation, Marines who reported having a U.S. Department of De-fense occupational category of Infantry were excluded given the studys focus on in-dividuals in high-risk jobs that were primarily office-based in nature (16,18). Individualsnot providing their occupational category or in an occupational category with less thanthree individuals (among all participants) were also not included in the study. If a personindicated the experience of musculoskeletal symptoms, he/she could not also have self-reported a non-work-related accident/trauma (e.g., sports, traffic accident) to the anatomiclocation for which they had symptoms. From these criteria, 59 people (n D 32 Infantry;n D 9 no occupational category or in a category with less than three individuals; n D 18non-work-related accident/trauma) were excluded, resulting in a final study sample of 248.

Measures

After providing informed consent, participants completed a questionnaire containingitems on individual, work-related, and health-related variables. Demographic information



collected included age, gender, marital status, education level, rank, length of service, mili-tary occupational specialty (MOS), and length of time in MOS. Additionally, biomechanicalexposures/ergonomic factors, characteristics of work organization, individual psychosocialstressors, musculoskeletal symptoms, and general health were assessed. These items aredescribed further in the following sections.

Ergonomic Factors

Frequency of movements, postures, loads, and environmental factors while performinga work task were determined by the 38-item Job FactorsDescription of Work subscale ofthe U.S. Air Force Job Requirements and Physical Demands Survey (JRPD) (19). Marcotteand colleagues found the JRPD to have adequate validity in relation to worksite assessmentsperformed by an ergonomist (19). In a study of office workers (n D 92), a subset of theseitems were found to distinguish individuals with UE pain, symptom severity, and func-tional limitation to a greater extent than an observational ergonomic risk factor checklistassessment that was completed by an ergonomist (20).

Work Organization

The frequency one experienced job stress was determined by an item from theU.S. Army Health Risk Appraisal (HRA). This item has response categories of never,seldom, sometimes, often and has been used in previous studies on musculoskeletal-related outcomes (21,22). Additional items and scales involved specific aspects of workorganization previously examined by factor analyses (17). These items were selected onthe basis of previous reported associations with work-related musculoskeletal symptoms,disorders, or disability and/or had been used in previous studies, having acceptable psycho-metric properties (e.g., internal consistency, discriminant validity), ease of administration(i.e., relatively few items), and suggestions of potential areas for modification through in-tervention efforts. These items were found to represent dimensions of work organizationsimilar to those categories described by the NIOSH but were not explicitly labeled in thequestionnaire (8,9).

Time Pressure Pressure to keep working, sense of urgency at work, and ability torelax were assessed by three questions adapted from the Work Environment Scale that havealso been used in a NIOSH study on WMSDs (23,24).

Job Design (Skill Discretion, Cognitive Demands, Cognitive Processing)Items fromthe Job Content Questionnaire were used to determine skill discretion (25). Cognitiveaspects of the job related to information processing, memory, and routine were assessed by10 questions obtained from a prior NIOSH study on job stress factors in musculoskeletaldisorders (24).

Interpersonal DemandsPersonal interaction required on the job was assessed bythree items from the Dealing with Others subscale of the Job Diagnostic Survey (JDS)(26).

Management Style Items on the level of involvement a worker has in making work-related decisions and feedback from supervisors and coworkers and comprised the man-agement style factor. Specifically, items were from Hales et al. and the Feedback fromAgents subscale of the JDS (24,26).



Job ResponsibilityPerceptions of individual and work group responsibility for per-forming a job task and how it is completed were determined by the Experienced Respon-sibility for the Work subscale of the JDS (26).

Career ConcernsFive NIOSH questions originally developed by Caplan were alsoused to determine ones certainty on future career picture, opportunities for promotion,value of job skills in the future, future responsibilities, and a workers ability to supportoneself if he/she lost his/her job (24,27).

Individual Psychosocial Factors

Questions related to individual psychosocial factors were included on the basis ofpast research indicating that life-related worries and family conflict are associated withmusculoskeletal-related pain and disability (21,22,28). Therefore, a question from the HRAon the frequency that worries interfered with ones life and the Conflict Subscale of theFamily Environment Scale were included in the questionnaire (29). To avoid potentialincrimination of military personnel, one item on physical abuse within the family wasexcluded.

Health Status

Musculoskeletal Symptoms

A modified NIOSH symptom survey was used to ask individuals about the type(s) ofmusculoskeletal symptoms one had, time of first onset, frequency, duration, intensity, workinterference, and associated limited duty within the past 12 months (30). These questionswere specifically directed to symptoms in the lower back, neck, shoulders, elbows/forearms,and wrists/hands regions. A visual analog scale (VAS) that measured pain intensity overthe past week was also used (31).

Function

Functional limitation in work and daily activities was measured by the SF-12 (32).These items were derived from the longer SF-36 which have been used in past researchon WMSD populations (3335). The SF-12 is a 12-item questionnaire containing twosubscales: the Physical Component Summary (PCS) and Mental Component Summary(MCS). These relate to physical and mental health function, respectively.

Case Definitions

On the basis of responses to the symptom survey, four groups were identified forpurposes of analyses. These groups included an asymptomatic group (i.e., reporting nosymptoms in any anatomic location), cases with LB symptoms only, cases with UE (i.e.,neck, shoulders, elbows/forearms, and wrists/hands) symptoms only, and cases with bothLB and UE symptoms. Symptomatic cases must have reported experiencing symptoms



since entering into their current occupational specialty. Those persons who reported havinghad a non-work-related accident or trauma to a given anatomic region had already beenexcluded from the study on the basis of previously described eligibility criteria.

Analyses

Since less than 3% of the total sample had missing data for a given item, missing datawere replaced with the mode of that respective question. This procedure was performedso that all subjects could be included in subsequent analyses. All analyses were performedusing SPSS v.10.1 (SPSS, Chicago, IL).

Differences between each symptomatic group and the asymptomatic group on demo-graphic characteristics were determined by Pearson chi-square and Students t tests. Subse-quently, analyses of variance (ANOVAs), linear trend analyses, and Scheffe and Tamhanespost hoc analyses were conducted to identify any group differences on and patterns for themeans for pain intensity over the last week and function associated with physical and mentalhealth (SF-12). Linear regression analyses were also conducted to determine those signifi-cant work organization correlates of pain intensity and physical function. In addition to thework organization variables, the linear regression model adjusted for age, gender, educa-tion level, individual psychosocial factors (i.e., life-interfering worries), and biomechanicalexposures/ergonomic factors.

Associations among the work organization variables and the HRA question on fre-quency of job stress, family conflict, and life-related worries were determined by computingPearson correlation coefficients. These correlations were computed to determine the extentof any potential overlap between each of these variables.

To identify work organization risk factors associated with the occurrence of LB, UE,and concurrent LB and UE symptoms, separate logistic regression analyses were conductedthat involved the following: cases with only LB symptoms and asymptomatic individuals;cases with only UE symptoms and asymptomatic individuals; and, cases with simultaneousback and UE symptoms and asymptomatic individuals. For each set of groups, all workorganization variables were initially examined in univariable logistic regression analyses.Those variables with a significance level (p value) equal to or less than 0.25 in the univari-able logistic regression analyses were then further examined in final multivariable logisticregression models that adjusted for age, gender, and education level, individual psychoso-cial factors (i.e., life-interfering worries), and biomechanical exposures/ergonomic factors(36). From all logistic regression analyses, odds ratios, 95% confidence intervals (CIs),parameter estimates (fl coefficients), and significance levels were obtained.

RESULTS

Subject Characteristics

Distributions of the study participants according to group status were 30.2% (n D 75)asymptomatic; 20.2% (n D 50) LB symptoms only; 21.0% (n D 52) UE symptoms only;and, 28.6% (n D 71) with concurrent LB and UE symptoms. The average age of all studyparticipants was 27.9 years old (SD D 7.4) and the majority of participants were male



(87.1%). Specific information on the demographic composition of the sample is given inTable I.

Given the study goal to target high-risk occupations in the Marine Corps, examina-tion of the Department of Defense occupational category distribution indicated that 74.6%(n D 185) of participants were involved in Functional Support & Administration. For theother occupational categories, 30 (12.1%) were in Service and Supply, 16 (6.5%) in Com-munications & Intelligence, 13 (5.2%) in Electrical/Mechanical, and 4 (1.6%) in ElectronicEquipment Repair.

Chi-square tests on gender, race, education level, marital status, rank, and occupationalcategory indicated no significant differences between the asymptomatic group and any of thesymptomatic case groups. Those with concurrent LB and UE symptoms were significantlyolder than asymptomatic individuals (t D 3:50, df D 144, p < 0:01).

Table I. Demographic Characteristics

Back only Upper extremity Back and upper TotalControls cases only cases extremity cases sample(n D 75) (n D 50) (n D 52) (n D 71) (n D 248)

Age in (years)Mean (SD) 26.3 (6.5) 28.0 (7.9) 26.5 (6.8) 30.4 (7.8)b 27.9 (7.4)

n (%)a n (%)a n (%)a n (%)a n (%)aGender

Males 66 (88.0) 42 (84.0) 48 (92.3) 60 (84.5) 216 (87.1)Females 9 (12.0) 8 (16.0) 4 (7.7) 11 (15.5) 32 (12.9)

RaceWhite/Caucasian 35 (46.7) 26 (52.0) 30 (57.7) 30 (42.3) 121 (48.8)Black/African American 19 (25.3) 15 (30.0) 11 (21.2) 21 (29.6) 66 (26.6)Hispanic/Latino 19 (25.3) 7 (14.0) 8 (15.4) 17 (23.9) 51 (20.6)Asian 2 (2.7) 2 (4.0) 2 (3.8) 1 (1.4) 7 (2.8)American Indian 0 (0) 0 (0) 0 (0) 1 (1.4) 1 (0.4)

EducationH.S. Grad/GED 29 (38.7) 25 (50.0) 21 (40.4) 28 (39.4) 103 (41.5)Some college/ 35 (46.7) 20 (40.0) 24 (46.2) 35 (49.3) 114 (46.0)

other post H.S.2-year degree 6 (8.0) 1 (2.0) 1 (1.9) 4 (5.6) 12 (4.8)4-year degree/college 2 (2.7) 4 (8.0) 1 (1.9) 1 (1.4) 8 (3.2)Some graduate work 1 (1.3) 0 (0) 2 (3.8) 2 (2.8) 5 (2.0)

Marital statusSingle 28 (37.3) 20 (40.0) 24 (46.2) 24 (33.8) 96 (38.7)Married 27 (36.0) 21 (42.0) 14 (26.9) 33 (46.5) 95 (38.3)Separated 3 (4.0) 0 (0) 1 (1.9) 6 (8.5) 10 (4.0)Divorced 3 (4.0) 5 (10.0) 6 (11.5) 1 (1.4) 15 (6.0)

RankE2 Private 5 (6.7) 2 (4.0) 3 (5.8) 2 (2.8) 12 (4.8)

First ClassE3 Lance Corporal 18 (24.0) 13 (26.0) 13 (25.0) 13 (18.3) 57 (23.0)E4 Corporal 15 (20.0) 8 (16.0) 16 (30.8) 7 (9.9) 46 (18.5)E5 Sergeant 18 (24.0) 13 (26.0) 7 (13.5) 19 (26.8) 57 (23.0)E6 Staff Sergeant 9 (12.0) 4 (8.0) 6 (11.5) 9 (12.7) 28 (11.3)E7 Gunnery Sergeant 6 (8.0) 4 (8.0) 4 (7.7) 9 (12.7) 23 (9.3)E8 Master Sergeant 2 (2.7) 3 (6.0) 3 (5.8) 8 (11.3) 16 (6.5)

or First SergeantE9 Master Gunnery 2 (2.7) 3 (6.0) 0 (0) 4 (5.6) 9 (3.6)

Sergeant or Sergeant MajoraPercentage of group; total n for category may not equal total n for group because of missing data.bt Test comparing group to controls: t D 3:50, df D 144, p < 0:01.



Comparisons Between Groups on Symptoms, Physical Function, and Mental Health

The ANOVAs comparing all groups (i.e., asymptomatic, LB only, UE only, concurrentLB and UE symptoms) indicated significant group differences on pain intensity over thelast week (F D 46:9, df D 3, p < 0:001). ANOVAs on physical function also indicatedsignificant differences between groups (F D 15:3, dfD 3, p < 0:01). No significant groupdifferences were found for mental health. While the assumption of homogeneity of vari-ances was not met in the ANOVAs for physical function (as indicated by Levenes testfor homogeneity of variances), findings can still be accepted on the basis of the level ofsignificance and the nature (i.e., robustness) of the analyses (37). Post hoc analyses usingTamhanes test found that all three musculoskeletal symptom groups significantly differedfrom the asymptomatic group on pain intensity and physical function. The group with con-current LB and UE symptoms was also significantly different from the UE symptoms onlygroup on pain intensity according to Tamhanes test.

Significant linear trends were also found for pain intensity over the last week andphysical function. Cases with both LB and UE symptoms had the highest level of painintensity (mean D 4.7), followed by LB symptom only cases (mean D 3.6) and cases withUE symptoms only (mean D 3.1), respectively (F D 10:0, df D 1, p < 0:01). Physicalfunction levels in descending order were asymptomatic (mean D 57.5); UE symptom onlycases (meanD 53.2); LB symptom only cases (meanD 51.1); and, cases with both LB andUE symptoms (meanD 49.3) (F D 42:6, dfD 1, p < 0:01). Figures 13 graphically depictthe group means for pain intensity, physical function, and mental health, respectively.

Linear regression analyses for pain intensity over the past week indicated age, er-gonomic factors, life-interfering worries, decision authority, and experienced responsibilityfor work as significant correlates (p < 0:01). This model accounted for 14% of the variance.Regression analyses for physical function (SF-12) only found ergonomic factors to be asignificant correlate in the model (adjusted R2 D 0:08; p < 0:01).

Fig. 1. Pain intensity over the last week by musculoskeletal group.



Fig. 2. Physical function by musculoskeletal group.

Correlation Analyses

Table II lists the correlation coefficients computed for the work organization factors,HRA item relating to the frequency of job stress experienced, and the individual psychosocialvariables on family conflict and life-related worries for all study participants. This tableshows that aside from the correlation between skill discretion and cognitive demands,statistically significant correlations between work organization items were low to moderate(0.170.36). While in this study, skill discretion and cognitive demand had a relatively higherlevel of correlation (r D 0:63), it should be noted that previous factor analyses indicated thatthese items loaded highly on a single factor (17). However, they were examined separately

Fig. 3. Mental health by musculoskeletal group.



Tabl

eII

.Co

rrela

tions

Am

ong

Indi

vid

uala

nd

Occ

upat

iona

lPsy

chos

ocia

lVar

iabl

es

Skill

Cogn

itive

Man

agem

ent

Tim

eJo

bCa

reer

Cogn

itive

Inte

rper

sona

lFr

eque

ncy

of

Fam

ilydi

scre

tion

dem

ands

styl

epr

essu

rere

spon

sibili

tyco

nce

rns

proc

essin

gde

man

dsjob

stre

ssco

nfli

ct

Cogn

itive

dem

ands

0.6

3

Man

agem

entS

tyle

0.2

6

0.30

Tim

epr

essu

re0

.03

0.20

0.0

6Jo

bre

spon

sibili

ty0

.36

0.35

0.33

0.1

5Ca

reer

con

cern

s0

.25

0.24

0.38

0.2

0

0.33

Cogn

itive

proc

essin

g0

.07

0.0

70

.23

0.06

0.1

40

.21

Inte

rper

sona

ldem

ands

0.1

9

0.24

0.17

0.09

0.21

0.18

0.1

0Fr

eque

ncy

ofjo

bstr

ess

0.1

00.

21

0.1

9

0.43

0.1

20

.18

0.18

0.18

Fam

ilyco

nfli

ct0.

110

.13

0.2

5

0.13

0.1

50

.21

0.07

0.0

70.

11Li

fe-re

late

dw

orr

ies

0.0

50.

060

.18

0.15

0.1

8

0.1

70.

13

0.14

0.47

0.16

Note

.nD

248.

p

art%3A10.1023%2FB%3AJOOR.0000015008.25177.8b

Documents

Transcript of art%3A10.1023%2FB%3AJOOR.0000015008.25177.8b