Search tips
Search criteria 


Logo of oenvmedOccupational and Environmental MedicineVisit this articleSubmit a manuscriptReceive email alertsContact usBMJ
Occup Environ Med. 2007 December; 64(12): e28.
PMCID: PMC2095378

Psychosocial 3

143 Predictors of delayed return to work for Washington State union carpenters: a case‐control analysis

K. L. Kucera1, H. J. Lipscomb1, W. Cameron2. 1Division of Occupational and Environmental Medicine, Duke University Medical Center; 2Center to Protect Workers' Rights, AFL‐CIO, Washington, DC

ObjectivesConstruction workers have high rates of work‐related back pain and injury and limited opportunity for light duty work. Using a case‐control approach we identified factors associated with delayed return to work after a back injury.

MethodsUnion administrative records were used to identify a dynamic cohort of 20 642 union carpenters who worked in Washington State from 1989 to 2003. The Department of Labor and Industries provided records of workers' compensation claims and associated medical care. Data were linked without personal identifiers. Work‐related back claims (n = 4241) were identified by American National Standards Institute (ANSI) codes (back, trunk, or neck/back) or by International Classification of Disease (ICD‐9) codes relevant to medical care for claims consistent with a back injury. Cases (n = 738) were defined as back injury claims with more than 90 days of paid lost time; controls (n = 699) resulted in return to work within 30 days. Logistic regression models estimated odds ratios and 95% confidence intervals (OR, 95% CI) of delayed return to work.

ResultsThirty per cent of case claims and 8% of control claims were identified by an ICD‐9 code (ANSI body code was missing or multiple injuries were sustained). Delayed return to work after back injury was associated with being female (2.3, 95% CI 1.1 to 4.7), age 30–44 (1.5, 95% CI 1.1 to 1.9) and age over 45 (2.0, 95% CI 1.5 to 2.8), 4 or more years union experience (1.6, 95% CI 1.3 to 2.0) and evidence of more acute trauma (struck against (1.8, 95% CI 0.7 to 4.4) or by an object or person (2.7, 95% CI 1.6 to 4.5), fall from elevation (2.3, 95% CI 1.6 to 3.3) or same level (1.8, 95% CI 1.2 to 2.8), and motor vehicle accident (MVA) (12.0, 95% CI 1.5 to 94.0)).

ConclusionUse of ICD‐9 codes identified claims with multiple injuries that would otherwise not be captured by ANSI codes alone. Though carpenters of younger age and inexperience were at increased risk for a paid lost time back injury claim, older carpenters and more experienced workers, once injured, were more likely to have delayed return to work as were those who experienced acute events such as falls, being struck by objects or persons, and MVAs.

Key wordsworkers' compensation; back injury claim; paid lost work time

144 Benchmarking work disability management: preliminary results of utility and outcomes

I. Rivilis1, B. C. AmickIII1, D. C. Cole1, A. Clarke2, L. Scott3. 1Institute for Work & Health; 2Clarke Brown Associates, 3Organizational Solutions

ObjectivesInconsistencies in measures across work disability management (WDM) data systems pose challenges for researchers in estimating the burden of work disability and improving WDM practices. We describe preliminary results of a pilot work disability management benchmarking (WDMB) collaborative involving privately insured workplaces, WDM consultants, and researchers, intended to develop meaningful metrics to evaluate Canadian WDM practices.

MethodsIn phase one of the pilot, organisations reviewed 265 metrics covering outcomes, processes and satisfaction domains. Outcomes were largely based on the US Employer Measures of Productivity, Absence, and Quality standards, while the process metrics were developed based on current practice and available literature. Key informants rated each metric on perceived utility and ease of provision. Twelve companies representing financial, education, healthcare, and pharmaceutical sectors participated in phase 1. Metrics assigned high benchmarking utility and ease of provision by members of a peer group proceeded to phase 2. Organisations self‐assessed their WDM outcomes based on metrics developed in phase 1. We conducted a descriptive analysis of the utility data from all companies and the outcomes data from the largest peer group in the pilot (financial services group, n = 5 companies), as well as ANOVAs seeking sources of variation in the outcome metrics.

ResultsProcess metrics were all rated as potentially useful with mean utility ratings ranging across various metrics from 3.00 to 4.92 on a scale of 0–5 (5 = very useful). The utility ratings of the outcome metrics were more variable, with means ranging from 1.00 to 4.44. The lower ratings occurred for the Group Benefits and Emergency Leave domains. The first phase reduced the total number of metrics to 162 (61%). Within the financial services group, the highest rated outcome metrics from STD‐active claims, LTD‐active claims, Casual Absence‐total lost workdays, and Group Benefits‐program cost revealed significant variability across companies for all metrics and over time for a few.

ConclusionCompanies are generally interested in process and outcome metrics. Since outcomes varied across companies and over time, an opportunity exists to benchmark WDM performance if appropriate standards are achieved. The pilot also illustrates a model for effective collaboration between various stakeholders.

Key wordsdisability management; work disability; benchmarking

145 Psychometric properties of generic and specific items used to measure safety climate

H. S. Shannon1, G. R. Norman1, S. J. Maracle2. 1McMaster University; 2Electrical & Utilities Safety Association

ObjectivesSafety climate (SC) refers to workers' shared perceptions of the safety of their work environment. Items ask for perceptions of the immediate supervisor and of management generally. SC is used as a “leading indicator” of safety, to provide information about safety before injuries occur and measures of SC typically use generic items. Our aim was to determine the psychometric properties of SC when items specific to the economic sector were added.

MethodsFocus groups were held with key personnel in several electrical utility companies to identify safety topics of particular relevance in this sector. Items were written to reflect these topics and were added to the generic SC measure to produce a sector‐specific measure. The two measures were distributed to 250 employees among five participating companies, and returned to the research team. The reliabilities of the generic questions and the sector specific questions were estimated. Generalisability coefficients (also called reliability) were calculated using generalisability theory; this uses estimates of the components of variability in the measure, and ranges between 0 and 1. High values show that the questionnaire can distinguish between units of measurement; the coefficient increases with more items and if mean scores of interest are based on larger sample sizes.

ResultsThirteen new items were added to the existing 32. The generalisability coefficient for the overall supervisor score (mean of 10 workers) for the generic items only was 0.39, and for specific items only was 0.80. For the overall score, it was 0.84. The coefficients for the overall management score (mean of 50 employees under five supervisors) for the generic, specific and total items were 0.35, 0.85 and 0.93, respectively.

ConclusionEven though there were fewer items, the generalisability coefficient was much higher for the specific items than for the generic ones. We recommend that further studies check if this improvement in psychometric properties of the measures is repeated in other economic sectors.

Key wordsmeasurement; safety climate; generalisability theory

146 Inclusive workplace employment, disability and sickness absence

L. Foss1, K. Skyberg1, P. Kristensen1, B. Claussen2, H. M. Gravseth1. 1National Institute of Occupational Health; 2University of Oslo, Institute of General Practice and Community Medicine

ObjectivesIn 2001 the social partners and the government of Norway signed an agreement on Inclusive Working Life (IWL). The agreement was initiated to combat the trend of increasing sickness absence and disability pensioning. At present, 52% of Norwegian employees work in an enterprise with an IWL agreement. We wanted to examine individual and work related factors, and their relationship to sickness absence and disability pensioning. Furthermore, we examined differences in work environment, sickness absence and disability between the IWL and non‐IWL groups.

MethodsData were collected from a linkage between The Oslo Health Study and several national registers based on the personal 11‐digit national identification number. This provided longitudinal individual data updated through 2004. Persons born between 1940 and 1970 were selected and 13 317 persons were followed for 11 years (1994–2004). The cohort was divided into an IWL group (5502 persons ever employed in an enterprise with an IWL agreement) and a non‐IWL group (n = 7815). We computed sickness absence rates and duration of sickness absence and cumulative risks of disability pension 1994–2004.

ResultsIWL persons were younger and more highly educated, and with a female majority compared to non‐IWL persons. IWL persons reported feeling safer as to income and position 2 years on and received more recognition for their efforts from their superiors. The sickness absence rate increased for both IWL and non‐IWL persons during the study period. We found a larger increase for IWL persons (from 0.10 to 0.33 episodes per person per year between 1994 and 2004). The increase for the non‐IWL group was from 0.12 to 0.23 episodes per person per year. However, the average sickness absence duration was shorter for IWL persons (66 vs 94 days per episode). The cumulative incidence of disability pension during the study period was considerably higher for non‐IWL persons (12.4% vs 3.4%).

ConclusionThe results indicate that employees in enterprises with an IWL agreement had shorter sick leaves and a lower risk of disability pensioning than persons never employed in an IWL company. However, the trend of increasing number of sick leaves was strongest in the IWL group. The two groups were different and comparisons should be made with caution.

Key wordssickness absence; disability; workplace exposure

147 The influence of health on productivity: relative perceptions of call centre and other white collar employees

T. Haines, B. Stringer, J. So. McMaster University

ObjectivesThe purpose of this analysis is to evaluate the perceptions of call centre workers of a large financial institution, relative to those of other white collar workers from the same institution and from a specialised government unit, regarding the influence of their health on productivity, in the context of perceived work organisation factors and stressors, willingness to raise concerns with management, return‐to‐work difficulties and stigma associated with mental health problems, and self‐rated health.

MethodsCall centre workers of a large financial institution (CC), and other white collar workers from the same institution (F) and from a specialised government unit (G), completed a questionnaire developed from: the Copenhagen Psychosocial Questionnaire (CPQ), the SF‐36, the Work Productivity and Activity Impairment Questionnaire, the Return to Work Perception Survey (RWPS), the Likelihood to Voice Scale (LVS), and the Stigma Scale for Receiving Psychological Help. Modified Dillman strategies were used to enhance response.

ResultsQuestionnaires were returned from 50 CC, 48 F and 44 G, with response rates of 46%, 84% and 76%, respectively. The response rate from CC was probably reduced due to questionnaire administration during Registered Retirement Savings Plan (RRSP) season. CC had more male (46%), non‐permanent (18%) and part‐time employees (52%). CC were younger (mean age 34 vs 46 years), with less work experience (3 vs 15 years). By linear regression, controlling for age, work experience and gender simultaneously, groups did not differ by CPQ, RWPS, LVS, or health status. Yet CC perceived a greater impact of health problems on productivity (p<0.001). A higher proportion of G (73%, p = 0.04) perceived that co‐workers had been off work for mental health reasons; they also perceived more stigma associated with seeking psychological help (p = 0.04). Older workers reported better health status (p = 0.007) and were more likely to voice concerns (p = 0.03); however, those with less work experience were less likely to do so (p<0.001).

ConclusionDespite comparable self‐rated health, call centre workers reported greater impact of health problems on productivity. Our group will evaluate the impact of a “Complex Issues. Clear Solutions.” workshop of the Canadian Mental Health Association on the outcomes reported here.

Key wordsmental health; stress; occupational

148 The effect of evaluations of psychosocial work conditions on sickness absence

T. Munch‐Hansen1, J. Wieclaw1, E. Agerbo2, N. Westergaard‐Nielsen3, M. Rosenkilde1, J. P. Bonde1. 1Department of Occupational Medicine, Aarhus University Hospital; 2National Center for Register Based Research, University of Aarhus; 3Aarhus School of Business

ObjectivesEvaluations of psychosocial work conditions are widely used as a tool for maintaining healthy organisations and satisfied employees. Sickness absence is a central outcome parameter for the performance of organisations. The objective is to examine if interventions at the workplace in the form of employee evaluations of the psychosocial work conditions has an impact on sickness absence.

MethodsThe participants were 3193 employees from 222 different public service workplaces in Aarhus County, Denmark. The sickness absence of the participants was calculated for the periods 6 months before and 6 months after the performance of evaluations of psychosocial work conditions. Controls were matched to each participant according to institution type, gender, age and occupation. Controls were employees who had not yet been subject to evaluations of psychosocial work conditions. The difference‐in‐difference method was used to calculate a single measure for the different amounts of sickness absence before and after the time of evaluation of psychosocial work conditions among cases and controls. The covariates included in the model were level of satisfaction with psychosocial work conditions, gender, age, occupation, size of workplace, family status and season (month).

ResultsThe development in the amount of sickness absence was generally more favourable at workplaces where an intervention in the form of an evaluation of psychosocial work conditions had taken place (1.44 days of sickness absence less per employee for a period of 6 months (95% CI 0.82 to 2.05)).The largest reduction was found at workplaces with medium levels of satisfaction with psychosocial conditions (1.57 days per 6 months (95% CI 0.92 to 2.22)) while the reduction was smaller among the most satisfied employees (1.13 days (95% CI 0.20 to 2.06) and the most dissatisfied employees (1.07 days (95% CI −0.06 to 2.21)).

ConclusionWe found a decrease in the number of days of sickness absence after the performance of evaluations of psychosocial work compared to workplaces where no such interventions had taken place. In subsequent analysis a measure of actions deployed to improve psychosocial work conditions as a consequence of the evaluations will be included.

Key wordssickness absence; intervention studies; psychosocial work conditions

149 Is overtime work a real predictor of stress response?

Y. Sato. Department of Occupational Health, McGill University

ObjectivesThe aim of this study was to clarify the influence of overtime work on psychological stress and how other stress related factors interact on the relationship between overtime work and stress response.

MethodsThe study was conducted among 24685 employees of a company in Japan using the Brief Job Stress Questionnaire (BJSQ). The BJSQ assessed stress response, job stressors and social supports. The subjects were divided into five groups of overtime work (0–19, 20–39, 40–59, 60+ and exempts).To study the association between overtime work, stress response and other stress factors, we used multiple logistic regression analysis.

ResultsThe non‐adjusted odds ratio for stress response for 60+ overtime hours per month in reference to 0–19 overtime hours was 1.58 (CI 1.42 to 1.76). After adjustment for age, job stressors and social supports, the association between overtime work and stress response disappeared. In the full multivariate logistic regression analysis, quantitative workload, qualitative workload and interpersonal conflict have a significantly increased odds ratio of 1.37 (CI 1.30 to 1.45), 1.57 (CI 1.48 to 1.66) and 1.50 (CI 1.43 to 1.57) for stress response. Job control, skill utilisation, job suitability, work satisfaction, social supports (supervisor, coworker and family/friend supports) have a significantly decreased odds ratio for stress response. In bivariate (overtime more than 60 h versus each variable independently) logistic regression analysis, quantitative workload and qualitative workload decreased the odds ratio of 60 h overtime‐stress response from 1.62 (CI 1.50 to 1.76) (non‐adjusted) to 0.91 (CI 0.84 to 1.00) and 1.27 (CI 1.17 to 1.38), respectively and job suitability, work satisfaction and supervisor support increased the odds ratio to 1.98 (CI 1.73 to 2.04), 2.02 (CI 1.86 to 2.19) and 1.90 (CI 1.75 to 2.07), respectively. The adjustment for interpersonal conflict did not change the odds ratio for stress response.

ConclusionThese results suggest that overtime work causes stress response through other stress factors, such as quantitative workload. Interpersonal conflict is an independent risk factor regardless of the association between overtime and stress response.

Key wordsovertime work; stress response; occupation and stress

150 Interaction between hazard exposures and organisational factors in women's occupational illnesses in manufacturing industries

J. Lu. National Institutes of Health, University of the Philippines, Manila

ObjectivesThe study aimed at examining the problems caused by organisational factors such as job autonomy, job content, nature of the task, hazard exposure and management styles to the mental health of women workers.

MethodsThis study was conducted in an export zone involving 31 industries and 613 women workers were interviewed. Industrial hygiene measurements were also carried out for work hazard analysis.

ResultsThe results showed that there were several interactions between variables which include the following: the need for new quality for products is associated with the need to upgrade knowledge and skills, fast‐paced work is associated with heavy physical load, poor health and safety policies at work with exposure to vapours, and participation in benchmarking with better policies on promotion and training. Those with autonomy in accomplishing jobs had lesser boring and repetitive tasks but had sufficient salary and compensation. Those with autonomy in taking rest such as slowing down of the pace and amount of work had less work pressure. Hazardous work was also correlated with boring and repetitive jobs, work pressure and those without autonomy in accomplishing work. Mental tasks were less repetitious and less boring but had greater work pressure. Results also showed that the overall good physical health of workers is affected by overtime and mental work. Workers who have autonomy to accomplish their work were 1.8 times more likely to have a better level of physical health (p = 0.05). Those under close monitoring were 76% more likely to have poor physical health. Those exposed to poor quality of work suffered 44% more poor physical heath than those who had good quality of work (p = 0.05). Another result was that workers who were under work pressure and work overtime were more prone to having skin allergies. Other illnesses such as hypertension, low back pain and respiratory problems had psychosomatic origins.

ConclusionThe study proposed a massive review of existing work regulations and their relationship to mental and psychological health. This study has policy implication and can serve as a basis for advocacy work and program implementation for the promotion of mental health counselling.

Key wordsorganisational factors; health hazard exposures; manufacturing women workers

Articles from Occupational and Environmental Medicine are provided here courtesy of BMJ Publishing Group