|Home | About | Journals | Submit | Contact Us | Français|
M. Koehoorn1, F. Xu2, J. Village3, C. Trask3, K. Teschke1. 1Department of Health Care and Epidemiology, University of British Columbia; 2Centre for Health Services and Policy Research, University of British Columbia; 3School of Occupational and Environmental Hygiene, University of British Columbia
ObjectivesDespite considerable research efforts, reducing the incidence of and work disability associated with back disorders has remained challenging. One of the major reasons is that the natural course of these disorders is poorly understood and therefore identifying new episodes of morbidity remains difficult. The objective of this study was to propose a definition of new onset back disorder for use in occupational studies.
MethodsMedical services, hospitalisations and workers' compensation data were linked to create a person‐specific, longitudinal database (1992–2001) of health care contacts for a cohort of 116268 workers employed in heavy industry. Contacts with back‐related ICD‐9 diagnoses were analysed using a semi‐parametric mixture model (Nagin, 1999) to investigate trajectories of morbidity over time.
ResultsDuring the follow‐up period, 22.9% of the cohort had no back‐related health care contacts. Among workers with at least one contact, four back morbidity trajectories provided the best fit of the data. The first trajectory of individuals had a high probability (>60%) of back morbidity every year of follow‐up. The other three groups had episodic trajectories of increasing and decreasing probability of back morbidity over time (ranging from 0% to 55% per year). The three episodic groups are postulated to represent the same trajectory with different start dates. Based on health care contacts within these trajectories, a back disorder episode was defined by a minimum of two outpatient visits per year. Separate episodes were defined by a 3‐year gap with no health care contacts. Individuals with chronic back disorders were defined by a minimum of four health care contacts per year with no gaps in contacts from year to year.
ConclusionThese results provide useful information for defining new episodes of morbidity for use in aetiological studies of occupational back disorders. In this heavy industry cohort, those with no history of back disorders could be followed for future incidence in cohort studies, or could serve as a control group in case‐control designs. The episodic disorder group could serve as a source of incident cases in either cohort or case‐control designs, as long as they had been symptom free for at least 3 years.
Key wordsback disorders; epidemiologic methods; heavy industry
A. Leclerc, J. Gourmelen, J. F. Chastang, J. L. Lanoe. INSERM U687
ObjectivesIn occupational epidemiology distal factors such as level of education are often considered only as confounding factors. Our objective was to study the causal pathway leading to low back pain (LBP), taking into account both the proximal factors (physical exposure at work) and the level of education.
MethodsIn the national survey “Enquête décennale santé” conducted in 2002–2003 in France, a self‐administered questionnaire included questions on LBP, and past or present exposure to handling of heavy loads and awkward postures. Level of education was also recorded. Among the 13824 subjects in the age group 30–69 years, LBP for more than 30 days in the past year was studied in relation to various factors. Three logistic models were compared, separately for men and women, with the following variables: model 1 – age, other personal factors, physical work‐related risk factors; model 2 – age, other personal factors, education, physical work‐related risk factors; model 3 ‐ age, other personal factors, education.
ResultsPrevalence of LBP more than 30 days was 15.4% for men and 19.4% for women. Controlling for age, smoking and body mass index, physical factors were strongly associated with LBP (ORs for handling of heavy loads 1.76 for men and 1.71 for women; ORs for awkward postures 1.48 for men and 1.31 for women, in model 1). Adding education as a confounding factor (models 2) did not change these conclusions much. In model 3, the level of education was strongly associated with LBP. The OR was equal to 1.67 for men without a diploma. Additional analyses indicated that physical exposure was strongly associated with education.
ConclusionThe results suggest that the level of education has direct consequences on occupational exposure, especially for men. From a prevention point of view, focusing on occupational risk factors is relevant, since primary prevention is feasible and effective at this level; however, a better understanding of the causal pathways leading to occupational health disorders, especially concerning the role of social factors, is also needed.
Key wordsmusculoskeletal disorders; causality; social determinants of health
WAhrens Socioeconomic status: more than a confounder? Int J Epidemiol200433 pp 806-7
H. J. Lipscomb, K. Kucera, C. A. Epling, J. M. Dement. Duke University Medical Center
ObjectivesWe evaluated health effects of employment in poultry processing on women in a rural area of the southern United States. The region is economically depressed with an African–American majority population; the poultry processing industry is the largest single employer of women. Because of a poor occupational safety history and concerns of women in the community, the study was designed to circumvent the need to involve the employer.
MethodsData were collected from a cohort of 291 women through interviews and physical exams conducted at 6‐month intervals over 3 years. Community‐based recruitment of participants was staggered resulting in variable tenure in the plant and follow‐up time. The primary outcomes of interest were musculoskeletal symptoms and disorders. An index of cumulative exposure, based on departmental rankings and work history, was used as the primary exposure variable. We also evaluated associations with work organisation factors, such as strain, social support, and job insecurity and satisfaction as well as other medical conditions, depressive symptoms, children in the home, and hand intensive home activities. Poisson regression with generalised estimating equations was used to evaluate factors associated with upper extremity symptoms and disorders separately.
ResultsMusculoskeletal problems increased with increasing age and the presence of depressive symptoms at baseline. However, even after adjusting for these things risk increased with cumulative exposure, a stronger dose response was observed for disorders (adjusted rate ratios low to high cumulative exposure 1, 0.7, 2.0, 2.2) rather than for symptoms (1, 1.2, 1.2, 1.4). Symptoms that made it difficult to maintain work speed or quality at baseline were associated with musculoskeletal problems as well. Although multiple work organisation factors were crudely associated with musculoskeletal problems, most effects were diminished when considering other risk factors in the multivariate analyses. The association of job insecurity at baseline with disorders at follow‐up may reflect that women who were insecure about their jobs continued working with symptoms – perhaps without treatment or accommodation – making them more likely to develop disorders.
ConclusionAmong this highly exposed population, these findings illustrate the highly contextual nature of complex relationships among musculoskeletal problems, physical work exposures and psychosocial factors.
Key wordscohort study; musculoskeletal disorders; women's health
K. T. Palmer1, M. Calnan2, I. Reading1, D. Coggon1. 1MRC Epidemiology Resource Centre; 2MRC Health Services Research Collaboration, Department of Social Medicine, University of Bristol
ObjectivesStatistics from the UK Labour Force Survey are widely quoted as evidence for the scale of occupational illness in Great Britain. Occupational attribution depends on whether participants believe their health problem is caused or aggravated by work, but personal beliefs may be unreliable. We assessed the potential for error for work‐associated arm pain.
MethodsWe mailed a questionnaire to 4998 working‐aged adults, randomly chosen from general practice registers. We asked about occupational activities, mental health, self‐rated health, arm pain and beliefs about its causation. Those in work were asked about four physical activities that we judged likely to give rise to arm pain and from these we derived a single variable for exposure to any “arm‐straining” occupational activity. Associations of arm pain with occupational activity were examined using a modified Cox model. We estimated an adjusted prevalence ratio (PR) for the effect of arm‐straining activity on arm pain risk and derived the population attributable fraction (PAF). We compared self‐reported and calculated attributable fractions, overall and for subsets defined according to demographic and other characteristics.
ResultsAmong 1769 respondents who gave full information and were in work, 817 reported arm pain in the past 12 months. Some 54% of these cases said their symptoms were caused/made worse by work. Arm pain was more common in those with arm‐straining work (PR 1.24), but the corresponding calculated PAF was substantially less than indicated by self‐report (13.9%). Over‐attribution was higher below age 50 (ratio 5.4) than at older ages (ratio 3.0), and also higher in the worst versus best bands by self‐rated health (ratios 7.4 and 2.4, respectively) and mental health score (ratios 6.3 and 2.2).
ConclusionOur findings indicate that counting cases of arm pain perceived to be work‐related can substantially overestimate the burden of illness attributable to arm‐straining occupational activities. Moreover, the degree of overestimation may not be constant among subgroups. This calls into question the validity of the LFS approach towards estimating the burden of this and perhaps other work‐related illness. In principle, physician‐led reporting schemes that rely on experts' beliefs about work causation could suffer from similar limitations.
Key wordsburden; RSI; attribution
A. Venema, E. M. M. De Vroome. TNO Quality of Life, Division of Work and Employment
ObjectivesThe Dutch workforce over 55 years of age has increased 60% over the last 10 years. The Dutch government is promoting this development, because of a growing need for a larger workforce. Mental and physical differences do exist between younger and older workers. This however, does not necessarily lead to a lesser level of employability. The main purpose of the study is to investigate possible effects of the aging workforce on occupational safety in The Netherlands, in order to come to conclusions on age‐specific safety and employment policy.
MethodsThe National Survey on Working Conditions is a large questionnaire on occupational safety and health issues. Eighty thousand employees were questioned by means of a written questionnaire. The response rate was 32.5%. A total of 20986 employees were included in the analysis of which 6.6% reported to have had an occupational injury in the last 12 months. Data on work accidents, worker and work characteristics are analysed using Cox proportional hazards regression model.
ResultsThe results show that the percentage of accidents and accidents with absence from work decline with age, but that the percentage of accidents with prolonged absence from work increases with age. The decline with age is absent in employees who report high levels of work autonomy and in employees who report high levels of aggression in their work. Also, there are clear differences in accident risk between sectors. The hotels and restaurants sector and the transport sector even show a small increase in accident rate by age.
ConclusionYoung workers are most at risk for occupational accidents. The aging workforce will therefore have a positive effect on the accident rate in The Netherlands. The number of occupational accidents resulting in prolonged absence from work, however, will increase. These are the most expensive for employers. In those sectors and those worker groups that are identified as showing a positive effect between accident risk and age, there is need to invest (age specific) preventive measures and effective and timely reintegration after an accident.
Key wordsoccupational safety; occupational accidents; aging workforce
T. Videman1, E. Levalahti2, M. C. Battié1. 1University of Alberta; 2University of Helsinki
ObjectivesAlthough recent studies have shown that heredity is a dominant factor in disc degeneration, the common notion that occupational physical loading is the major risk factor persists. However, substantial variations in disc degeneration, particularly at the lowest lumbar levels, remain unexplained by heredity or occupational physical demands. The exact goal was to examine the influences of lifetime physical activities at work and during leisure time and to compare these to the effects of body anthropometrics, axial disc area, and lifting strength on disc degeneration and age.
MethodsA cross‐sectional study design was used. Univariate methods and stepwise multiple regression modelling were used to estimate associations of body height, weight, fat content, axial disc area, isokinetic lifting performance and lifetime routine physical activities at work and leisure with disc height narrowing and disc signal (in T2 images) based on lumbar MRIs. These data were available from a population sample of 600 men, 35–70 years of age.
ResultsLower disc signal, representing disc desiccation, was associated with higher age, lower body mass and lifting strength, and larger axial disc area. Of the variance in disc signal, age explained 8.0% (p<0.001) and body weight/axial disc area, isokinetic lifting strength and occupational lifting history added 3.9%, 2.3% and 1.3%, respectively. Greater disc narrowing was associated with higher age, larger axial disc area, and higher occupational physical loading. Of the variance in disc narrowing, age accounted for 3.8% (p<0.001); axial disc area and occupational loading added 1.9% (p<0.004) and 1.3% (p<0.007), respectively.
ConclusionBody weight, lifting strength, and axial disc area were more highly associated with disc degeneration than occupational and leisure physical activity histories, although all had modest influences. Furthermore, higher body mass, greater lifting strength, and heavier work were all associated with more disc height narrowing but less disc desiccation contrary to current views. Smaller discs appeared to have beneficial effects.
Key wordsphysical loading; anthropometric factors; disc degeneration