Patterns of multisite pain and associations with risk factors
aMedical Research Council Lifecourse Epidemiology Unit, University of Southampton, Southampton, UK
bSchool of Nursing, University of São Paulo, São Paulo, Brazil
cCorporación para el Desarrollo de la Producción y el Medio Ambiente Laboral–IFA (Institute for the Development of Production and the Work Environment), Quito, Ecuador
dDepartment of Industrial Engineering, School of Engineering, Pontificia Universidad Javeriana, Bogotá, Colombia
eSouthwest Center for Occupational and Environmental Health, The University of Texas Health Science Center at Houston School of Public Health, Houston, TX, USA
fCenter for Disease Control and Prevention/National Institute for Occupational Safety and Health, Atlanta, GA, USA
gMedical Research Council Social, Genetic and Developmental Psychiatry Centre, Institute of Psychiatry, Kings College, London, UK
hCenter for Research in Occupational Health (CiSAL), Universitat Pompeu Fabra, Barcelona, Spain
iEpidemiology and Preventive Medicine Research Center, University of Insubria, Varese, Italy
jDepartment of Social Medicine, Medical School, University of Crete, Heraklion, Greece
kDepartment of Public Health, University of Tartu, Estonia
lDepartment of Environmental Health, Faculty of Health Sciences, American University of Beirut, Beirut, Lebanon
mDepartment of Occupational Health, Faculty of Health, Shahroud University of Medical Sciences, Shahroud, Iran
nDepartment of Community Health Sciences, Aga Khan University, Karachi, Pakistan
oDepartment of Medical Education and Health Sciences, Faculty of Medical Sciences, University of Sri Jayewardenepura, Gangodawila, Nugegoda, Sri Lanka
pClinical Research Centre for Occupational Musculoskeletal Disorders, Kanto Rosai Hospital, Kawasaki, Japan
qNational Institute for Occupational Health, National Health Laboratory Service, Johannesburg, South Africa
rFaculty of Health Sciences, University of Witwatersrand, Johannesburg, South Africa
sDepartment of Epidemiology and Preventive Medicine, School of Public Health and Preventive Medicine, Monash University, Melbourne, VIC, Australia
tDepartment of Preventive and Social Medicine, University of Otago, Dunedin, New Zealand
uSchool of Nursing of Ribeirão Preto, University of São Paulo, São Paulo, Brazil
vFederal University of Paraná, Curitiba-PR, Brazil
wInstitute for Studies on Toxic Substances (IRET), National University of Costa Rica, Heredia, Costa Rica
xCIBER of Epidemiology and Public Health, Barcelona, Spain
yOccupational Health Service, Parc de Salut MAR, Barcelona, Spain
zDepartment of Clinical Sciences and Community Health, Università degli Studi di Milano, Milan, Italy
aaFondazione Ca’ Granda Ospedale Maggiore Policlinico, Milan, Italy
abDepartment of Psychiatry, Medical School, University of Crete, Heraklion, Greece
acCentre for Research in Environmental Epidemiology (CREAL), Barcelona, Spain
adIMIM (Hospital del Mar Research Institute), Barcelona, Spain
aeNational School of Public Health, Athens, Greece
afNorth Estonia Medical Centre, Tallinn, Estonia
agPõlva Hospital, Põlva, Estonia
ahKlinikum Leverkusen, Leverkusen, Germany
aiDepartment of Physiology, Faculty of Medical Sciences, University of Sri Jayewardenepura, Gangodawila, Nugegoda, Sri Lanka
ajSection of Epidemiology and Biostatistics, School of Population Health, Faculty of Medical and Health Sciences, University of Auckland, Auckland, New Zealand
akDepartment of Epidemiology, School of Public Health, University of Alabama at Birmingham, Birmingham, AL, USA
alFaculty of Medicine,University of Kalaniya, Kelaniya, Sri Lanka
amDepartment of Joint Disease Research, University of Tokyo, Tokyo, Japan
anCentre for Occupational and Environmental Health, Department of Social and Preventive Medicine, Faculty of Medicine, University of Malaya, Kuala Lumpur, Malaysia
aoInjury Prevention Research Unit, Department of Preventive and Social Medicine, University of Otago, Dunedin, New Zealand
Received March 25, 2013; Revised May 16, 2013; Accepted May 20, 2013.
This document was posted here by permission of the publisher.
At the time of the deposit, it included all changes made during peer review,
copy editing, and publishing. The U. S. National Library of Medicine is responsible
for all links within the document and for incorporating any publisher-supplied
amendments or retractions issued subsequently. The published journal article,
to be such by Elsevier, is available for free, on ScienceDirect, at: http://dx.crossref.org/10.1016/j.pain.2013.05.039
The CUPID study sample comprised workers aged 20–59 years from 47 occupational groups (office workers, nurses, and “other workers”) in 18 countries (). During 2006–2011, participants completed a standardised questionnaire about musculoskeletal pain, associated disability, and possible risk factors, either at interview (25 groups), by self-administration (18 groups), or a combination of interview and self-administration (4 groups). Response rates among those invited to take part were mostly higher than 80% (33 groups), but were lower than 50% in 5 groups. For logistic reasons, data collection was earlier in some countries than in others.
Occupational groups included in the CUPID study.
The questionnaire was originally drafted in English, and then translated into local languages where necessary. The accuracy of translation was checked by independent back-translation, and amendments were made if needed. Among other things, the questionnaire asked whether during the past month, pain had been present for a day or longer in each of 6 anatomical regions (low back, neck, shoulder, elbow, wrist/hand, and knee) depicted in diagrams, and for the limb regions, whether the pain had been on the right, left, or both sides. It also asked about sex, age, age at which full-time education was completed, smoking habits, somatising tendency, mental health, physical activities at work, psychosocial aspects of work, and fear-avoidance beliefs about musculoskeletal pain.
Somatising tendency was assessed using questions from the Brief Symptom Inventory 
, and graded according to the number of common somatic symptoms from a total of 5 (faintness or dizziness, pains in the heart or chest, nausea or upset stomach, trouble getting breath, hot or cold spells) that had been at least moderately distressing in the past week. Questions about mental health came from the relevant domain of the Short Form-36 questionnaire 
, and scores were classified to approximate thirds of the distribution in the full study sample (denoted good, intermediate, and poor).
Exposure to physical loading at work was scored according to how many of 5 activities (lifting weights of 25 kg or more by hand; working for longer than 1 hour in total with the hands above shoulder height; repeated bending and straightening of the elbow for longer than 1 hour in total; use of a computer keyboard or other repeated movements of the wrist or fingers for longer than 4 hours in total; and kneeling or squatting for longer than 1 hour in total) were reported in an average working day. Time pressure at work was considered to be present if a participant reported either a target number of articles or tasks to be finished in the working day, or working under pressure to complete tasks by a fixed time. Lack of support at work was deemed to occur if help with difficulties was seldom or never provided by colleagues or a supervisor/manager. Job dissatisfaction was classed as present if overall, the participant felt dissatisfied or very dissatisfied with their employment. Lack of control was considered to occur if there was seldom or never choice in all of: a) how work was done, b) what was done at work, and c) work timetable and breaks. Job insecurity was taken as present if the participant felt that the tenure of their employment would be “rather unsafe” or “very unsafe” if they were off work for 3 months with significant illness.
Questions concerning fear-avoidance beliefs were adapted from the Fear Avoidance Beliefs Questionnaire 
. Participants were deemed to have adverse beliefs about the work-relatedness of musculoskeletal pain if they completely agreed that either low-back pain or arm pain is commonly caused by people’s work; about physical activity if either for someone with low-back pain or for someone with arm pain, they completely agreed both that physical activity should be avoided as it might cause harm, and that rest was needed to get better; and about prognosis if either for low-back pain or arm pain, they completely agreed that neglecting such problems can cause permanent health problems, and completely disagreed that such problems usually get better within 3 months.
Further details of the methods of the CUPID study sample and methods of data collection have been reported elsewhere 
Statistical analysis was carried out with Stata 12.1 software (StataCorp LP, College Station, TX, USA). We first calculated the prevalence of pain in the past month at each of 10 anatomical sites (low back, neck, right shoulder, left shoulder, right elbow, left elbow, right wrist/hand, left wrist/hand, right knee, and left knee) and summarised the associations between pain at pairs of sites by odds ratios (ORs).
Next we classified subjects according to the number of anatomical sites (from 0 to 10) that they reported as having been painful in the past month, and compared the observed frequencies with the numbers that would have been expected given the overall prevalence of pain at each site by sex and age, and assuming that the occurrence of pain at any 2 sites was independent. For example, if within a specified sex and age group, the prevalence of pain in the 10 sites was P1
, … P10
, then in that group, the expected prevalence of no pain at any of the 10 sites would be
and that of pain at all 10 sites
. Ratios of observed to expected counts (O/E) were calculated for the full study sample, and broken down according to whether or not the distribution of pain was widespread (ie, it was reported in each of the trunk, upper limb and lower limb, and also on both sides of the body). This analysis was used to define “limited pain” involving a small number of sites with O/E < 1, and “extensive pain” involving a large number of sites with O/E clearly > 1.
We then explored personal risk factors for musculoskeletal pain affecting different numbers of anatomical sites. We used Generalised Linear Latent and Mixed Models (GLLAMM) to fit 2-level, random intercept Poisson regression models with robust standard errors, in which individuals were clustered by occupational group. Associations were summarised by prevalence rate ratios (PRRs) with associated 95% confidence intervals (95% CIs). To check the robustness of the findings, we repeated the analyses, using 2-level, random intercept logistic regression models.
To check whether a pattern of pain that was widespread (ie, in the trunk, on both sides of the body, and in both an upper and lower limb) showed additional association with risk factors after the number of sites with pain had been taken into account, we carried out a Poisson regression analysis with widespread pain as the outcome, adjusting for the number of sites with pain (treated as dummy variables).
Next, we constructed single-level Poisson regression models with limited and extensive pain as the outcome variables, and incorporating occupational group as an independent variable while adjusting for all of the personal risk factors examined previously. For this purpose, office workers in the UK were taken as the reference group for risk estimates, and the PRRs for limited and extensive pain (relative to no pain) were compared across the 47 occupational groups to see whether they correlated.
Finally, we carried out sensitivity analyses in which we repeated the Poisson regression analyses: a) excluding the 5 occupational groups with response rate < 50%; and b) adjusting also for the method by which the questionnaire was answered (interview or self-administration).