Shoulder pain among male industrial workers: Validation of a conceptual model in two independent French working populations.

This study aims to validate a conceptual model for shoulder pain risk factors in two independent samples of male industrial workers: the Cosali cohort (n = 334) and one pharmaceutical company (n = 487). Direct and indirect relationships between work organization factors (automatic speed of a machine or movement of a product and work pace dependent on customers' demand), psychosocial factors (Job strain model), biomechanical factors (working with abducted arms, working with arms at or above shoulder level, and perceived physical exertion), perceived stress, and shoulder pain were explored using structural equation models. Shoulder pain was positively associated with biomechanical exposure in both samples, and with perceived stress only in the pharmaceutical preparation manufacturer, while factors related to work organization and psychosocial factors had indirect impacts on the risk of chronic shoulder pain in both samples. The results provide a deeper understanding of the complex relationships between workplace risk factors and shoulder pain.

Reliability and validity of the Functional Assessment of Cancer Therapy General (FACT-G) in French cancer patients.

This report describes the reliability and validity of a French version of the Functional Assessment of Cancer Therapy - General (FACT-G) with a French sample of 493 cancer patients. The FACT-G consists of 27 items and four subscales: Physical (PWB), Functional (FWB), Social/Family (SFWB) and Emotional well-being (EWB). The study sample includes 64% with localized disease, 26% with metastases, 11% in remission, and 71% receiving radiation/chemotherapy. Internal consistency Cronbach alphas of the global FACT-G scale (0.90) and subscales (>0.75) are satisfactory (n = 126). Test-retest reproducibility is satisfactory for all subscales and the global scale (n = 87 to 93, r = 0.74 to 0.90). ANOVA models show that PWB differentiated between the three disease stages; the global FACT-G and FWB discriminated between patients with metastases and others with localized disease or in remission; EWB only discriminated between metastases and localized disease; while SFWB did not discriminate between groups at different stages of cancer. Only the PWB subscale discriminated between patients with no history from those receiving chemotherapy (p < or = 0.05). None of the scales discriminated between groups based on radiotherapy. These results may be useful in the design and interpretation of clinical trials involving French patients when the FACT-G is the outcome measure.

Development of a preference-weighted health status classification system in France: the Health Utilities Index 3.

The Health Utilities Index is a generic multiattribute preference-based system for assessing health-related quality of life, devised by Torrance et al. It is being used in cost-effectiveness evaluations in North America and in international multicentre studies but was not available in France. Following adaptation of the HUI3 classification in France, the purpose of the reported investigation was to derive French preference weights. This article provides a reminder of the theoretical foundations used to model the multiattribute utility function. Within this framework, a multiattribute multiplicative aggregate utility function was constructed in accordance with the explicitly decomposed approach. The study took place in June 1999 over a sample of 365 persons from the French general population, aged between 20 and 65, and not suffering from any chronic or incapacitating illness. The recruitment procedure was based upon a random selection of individuals, using the phone book. Interviews took place in the homes of the interviewees. The methods of revelation (VAS and SG) were applied by setting the value of the best possible state of the HUI3 a priori at 1, and leaving a choice between two states (worst possible state, death) for 0. The aggregated individuals (person-mean and median) were calculated and the multiplicative utility functions constructed. A comparison of the calculated utilities with the observed ones provides a primary indicator of the validity of the person-mean or median functions constructed. The slight absolute differences obtained between observed and calculated utilities and the low RMSE scores lead us towards a favourable conclusion.