Application of a Mathematical Method to Assess Associations Between Work Exposures and Severe Back and Hand Pain Using Job Exposure Matrices in the CONSTANCES Cohort : Journal of Occupational and Environmental Medicine

Secondary Logo

Journal Logo


Application of a Mathematical Method to Assess Associations Between Work Exposures and Severe Back and Hand Pain Using Job Exposure Matrices in the CONSTANCES Cohort

Deltreil, Guillaume MSc; Tardivel, Patrick PhD; Graczyk, Piotr PhD; Escobar-Bach, Mikael PhD; Fadel, Marc MD; Descatha, Alexis MD

Author Information
Journal of Occupational and Environmental Medicine 65(6):p e441-e442, June 2023. | DOI: 10.1097/JOM.0000000000002841

To the Editor:

Musculoskeletal disorders (MSDs) are a very important concern with major impact on individuals and companies and represent one of the main source disabilities at work.1 Ergonomic factors impact the risk of MSDs all along work history.2,3 However, having an unbiased assessment of past work exposure is difficult. Thus, to consider life course exposure, job exposure matrices (JEMs) focused on ergonomic factors have been suggested, although it is still not perfectly clear how JEMs' exposure rating and the duration of exposure interact with each other.4,5 To tackle this issue, a statistical modeling G has been developed to compare the precision of duration of work and intensity/frequency associations in application to knee disorders, as a simple example of degenerative disorders, with interesting results.6 This model is based on the interaction between the total duration of the career range and the time-weighted average of the exposure level.

However, we don't know if such modeling G could be applied to other types of MSDs, such as back pain, which has similar risk factors and cumulative physiopathology, or hand pain, which has different risk factors and short-term physiopathology. To pursue on this issue, we aimed to apply this model G to low back and hand pain.

Data from the CONSTANCES cohort, a French general population-based cohort, were used.7 At baseline, we retrieved relevant variables: age sex, body mass index, depression, inflammatory arthritis, leisure activities, and severity of pain (defined as intensity of pain >5/10, or daily pain or pain lasting at least 1 month) both on low back pain and hand pain. JEM CONSTANCES was combined with work history to assess work exposure: carrying heavy loads (10 to 25 kg) for back pain and repetitive task for hand pain.8 Analyses included model comparison and adjusted odds ratios similar to previous modeling. The G model is a logistic regression model that focuses on duration and intensity of exposure. Each subject belongs to either the control group (the category with the lowest exposure level and lowest duration of exposure) or to the exposed groups, which are characterized by a duration of exposure and an intensity of exposure, with each group having longer duration of exposure and/or higher intensity of exposure. The categories of the exposure time intervals are determined by the first and second quartiles of this variable, those of the JEM CONSTANCES exposure (intensity of exposure) by their clinical significance, and the number of intervals by a statistical criterion (Akaike information criterion). R codes for using this model are provided in supplemental materials,

The average age of the 63,938 subjects was 48.5 years (18 to 69 years, 21 years of employment); they were mostly women (n = 34,212 [53.51%]). Among them, 11,951 subjects had back pain, and 6623 subjects had hand pain.

The new modeling G that was first experimented on knee pain gave similar results, with our model G being slightly better than the regular modeling for back and hand pain and both sexes (score lower for our modeling vs models for duration multiplied intensity/frequency for hand and knee pain). In Table 1, application of the previous modeling work gave similar results for back pain and men, and lower for hand pain and women.

TABLE 1 - Association Using Our Modeling for Using Job Exposure Matrices
Reference is low—[0,1] Back Pain and Carrying Loads Hand Pain and Repetitive Tasks
Men Women Men Women
Duration Intensity/Frequency aOR (95% CI) aOR (95% CI) aOR (95% CI) aOR (95% CI)
Low [1,2] 1.75([1.5–2.03])*** 1.41([1.27–1.57])*** Reference Reference
Low [2,3] 2.18([1.86–2.56])*** 1.72([1.45–2.04])*** 1.67([1.26–2.22])*** 1.57([1.29–1.91])***
Low [3,4] 2.41([1.86–3.12])*** 1.8([ 1.3–2.49])*** 2.12([1.65–2.72])*** 1.90([1.60–2.27])***
Low {4} 3.14([2.48–3.99])*** 1.49([0.85–2.59]) 2.47([1.93–3.16])*** 2.20([1.82–2.67])***
Medium [0,1] 1.04([0.9–1.2]) 0.89([0.81–0.97])* a a
Medium [1,2] 1.59([1.35–1.86])*** 1.35([1.2–1.51])*** 1.06([0.85–1.34]) 1.10([0.94–1.30])
Medium [2,3] 2.61([2.21–3.07])*** 1.54([1.27–1.88])*** 1.93([1.47–2.53])*** 1.70([1.42–2.03])***
Medium [3,4] 2.88([2.18–3.8])*** 1.61([1.04–2.5 ])* 2.75([2.16–3.51])*** 1.80([1.51–2.15])***
Medium {4} 3.7([2.68–5.12])*** 1.8([0.83–3.88]) 2.63([1.98, 3.48])*** 2.56([2.07–3.17])***
High [0,1] 1.15([0.99–1.33]) 0.85([0.76–0.95])* a a
High [1,2] 1.75([1.51–2.04])*** 1.15([1.02–1.3])* 1.58([1.28–1.96])*** 1.22([1.04–1.44])
High [2,3] 2.42([2.07–2.81])*** 1.14([0.92–1.41]) 2.27([1.80–2.86])*** 1.59([1.33–1.89])***
High [3,4] 2.62([2.08–3.29])*** 1.34([0.99–1.81]) 2.24([1.78–2.82])*** 1.62([1.36–1.93])***
High {4} 2.77([2.12–3.63])*** 1.27([0.52–3.09]) 2.45([1.91–3.15])*** 2.19([1.77–2.72])***
CI, confidence interval; aOR, adjusted odds ratio.
aORs between outcome and category of exposure, adjustment on age, body mass index, leisure activity, depression or inflammatory osteoarthritis at baseline. low = 0–13 years, medium = 13–25 years, high = 25–51 years.
*p < 0.05.
***p < 10−4.
aFor hand pain, the scale goes from 1 to 4.

Despite some limitations on representativeness (low number of agriculture and craft workers), the use of a JEM for work exposure assessment, and the outcome based on subjective pain, we could conclude that our modeling is reliable for other MSDs than knee disorders, which have different pathophysiology and risk factors.

Guillaume Deltreil, MSc
Université d'Angers
CHU Angers
Univ Rennes
Inserm, EHESP
Irset (Institut de recherche en santé
environnement et travail)—UMRS 1085
Angers, France
Université d'Angers
49000 Angers, France
Patrick Tardivel, PhD
Université de Bourgogne Franche-Comté
Dijon, France
Piotr Graczyk, PhD
Mikael Escobar-Bach, PhD
Université d'Angers
49000 Angers, France
Marc Fadel, MD
UNIV Angers
CHU Angers
Univ Rennes
Irset (Institut de recherche en santé
environnement et travail)—UMR_S 1085
F-49000 Angers, France
Alexis Descatha, MD
UNIV Angers
CHU Angers
Univ Rennes
Irset (Institut de recherche en santé
environnement et travail)—UMR_S 1085
F-49000, Angers, France
Epidemiology and Prevention
Donald and Barbara Zucker
School of Medicine
Hofstra Northwell
Uniondale, NY


1. Forouzanfar MH, Alexander L, Anderson HR, et alGBD 2013 Risk Factors Collaborators. Global, regional, and national comparative risk assessment of 79 behavioural, environmental and occupational, and metabolic risks or clusters of risks in 188 countries, 1990–2013: a systematic analysis for the Global Burden of Disease Study 2013. Lancet 2015;386:2287–2323.
2. Descatha A, Evanoff B, Leclerc A, Roquelaure Y. Occupational determinants of musculoskeletal disorders [Internet]. In: Bültmann U, Siegrist J, éditeurs. Handbook of Disability, Work and Health. Switzerland AG: Springer International Publishing; 2020. Available at: Accessed January 20, 2020.
3. Hulshof CTJ, Pega F, Neupane S, et al. The prevalence of occupational exposure to ergonomic risk factors: a systematic review and meta-analysis from the WHO/ILO Joint Estimates of the Work-Related Burden of Disease and Injury. Environ Int 2021;146:106157.
4. Descatha A, Fadel M, Sembajwe G, Peters S, Evanoff BA. Job-exposure matrix: a useful tool for incorporating workplace exposure data into population health research and practice. Front Epidemiol [Internet] 2022. Available at: Accessed May 21, 2022.
5. Descatha A, Evanoff BA, Andersen JH, et al. JEMINI (Job Exposure Matrix InterNatIonal) initiative: a Utopian possibility for helping occupational exposure assessment all around the world? J Occup Environ Med 2019;61:e320–e321.
6. Deltreil G, Tardivel P, Graczyk P, Escobar-Bach M, Descatha A. How to use biomechanical job exposure matrices (JEM) on job history for musculoskeletal disorders? Mathematical modeling on the example of severe knee pain in CONSTANCES. 2022;19:16217.
7. Goldberg M, Carton M, Descatha A, et al; CONSTANCES Team. CONSTANCES: a general prospective population-based cohort for occupational and environmental epidemiology: cohort profile. Occup Environ Med 2017;74:66–71.
8. Evanoff BA, Yung M, Buckner-Petty S, et al. The CONSTANCES job exposure matrix based on self-reported exposure to physical risk factors: development and evaluation. Occup Environ Med 2019;76:398–406.

Supplemental Digital Content

Copyright © 2023 American College of Occupational and Environmental Medicine