Introduction

Upper limb musculoskeletal disorders (UL-MSD) represent unquestionably a challenge to occupational health. They are the cause of veritable epidemics in some companies, with significant health care, social and economic consequences (Carrivick et al. 2002; Morse et al. 1998). In spite of the societal impact of these occupational diseases, knowledge about the natural course of UL-MSD in the working population remains far from complete. In France, the occupational health organization imposes compulsory medical monitoring of all employees based on an annual medical consultation held by the occupational physician in charge of the company. This organization enables systematic surveillance of occupational pathologies. With regard to UL-MSD, implementation of a standardized diagnostic approach guarantees the validity of data collected, in particular, its exhaustiveness and reproducibility (Meyer et al. 2002; Sluiter et al. 2001). From 2000 to 2002, the occupational physician for a household appliance assembly company implemented this approach within the framework of a UL-MSD prevention action (Aptel and Gerling 2000). Beyond its significance in relation to prioritizing prevention actions for implementation, this monitoring allowed the individual clinical history of employees to be detailed with respect to UL-MSD.

The aim of this study is to describe the dynamic individual course of UL-MSD over a 3-year period amongst the employees of a household appliance assembly company. The hypothesis is that the individual course of UL-MSD is not a stable status but fluctuates with time with improvements and worsening and may influence prevention strategies.

Materials and methods

Job description

The study took place in a manufacturing company producing electrical ovens, toasters and meat-grills. The company, situated in a provincial town in the East of France, has 627 employees, working in five departments of which one is administrative and the others concern the production area: in (1) “ovens” and (2) “meat-grill / toasters” departments, the employees work on assembly lines, connecting electric cables, screwing hoods, resistances, doors, hinges on household appliances. The work is physically demanding, often performed in awkward postures with repetitive movements and the use of powered screwdrivers. They handle materials from a few grams to 16 kg (ovens to be turned). Most of them rotate between the work tasks, but some of them don’t because of insufficient skills, health problems or unwillingness. In (3) “power press” and (4) “electrical resistance” departments, metal sections and resistances are produced for the assembly work. Pieces are manually lifted and carried onto racks before and after processing. Working heights vary from floor level to about 1.80 m. In these departments (3) and (4), most of the activity is manual handling of heavy loads. The production of the company is running 16 h a day, from Monday to Friday. The working hours are divided into two work-shifts per day.

Design and sampling schedule

The medical surveillance of UL-MSD was based on an annual standardized examination that was repeated three successive years and concerned only personnel permanently employed at the company. All employees were monitored by the same occupational physician during the 3-year follow-up. The clinical examination consisted of two steps: the employees were first asked about pain or symptoms of UL-MSD in the year before measurement, then in case of pain or symptoms, a physical examination was done, depending on the anatomical location that was concerned. The occupational physician used a standardized clinical examination guide, drawn up by the Occupational Health Department of Besançon Hospital, to characterize UL-MSD. For each of the specific disorders, proposed criteria are based on symptoms and clinical signs included because of their diagnostic value, their general acceptance and wide use in clinical practice. For example, carpal tunnel syndrome is attested by intermittent paresthesias in at least two of digits I, II, or III at the time or during the week before the examination, associated with a Tinel’s sign or a positive Phalen test (Table 1).

Table 1 Diagnosis criteria for UL-MSD and grouping by anatomical region
Full size table

UL-MSD taken into account were cervicobrachialgies, shoulder tendonitis, lateral and medial epicondylitis, cubital tunnel syndrome, hand and finger flexor/extensor peritendinitis/tenosynovitis, Guyon canal syndrome and carpal tunnel syndrome. For statistical analysis, diagnosed pathologies were grouped into four categories, based on anatomical location, because of the high number of potential UL-MSD diagnoses (Table 1). These categories are: (1) the neck, (2) the shoulder, (3) the elbow, and (4) the wrist and hand region. Each location disorder refers to a subject having either one or several of the defined disorders for the anatomical location. The medical data collection period extends from January 1, 2000 to October 30, 2002. The first two data collection years are therefore complete, whilst the third year is 2 months shorter due to retirement of the occupational physician. Data collection and processing are performed according to the recommended professional ethics and good epidemiological practice (Déontologie et bonnes pratiques en épidémiologie 1998).

Statistical analysis

Statistical analysis is based on the sample of employees systematically examined in occupational medicine consultation each year between January 1, 2000 and October 30, 2002. Employees seen at medical examination are compared with all company employees according to main available sociodemographic criteria. In the same way, comparison of employees affected by UL-MSD, during at least 1 in 3 years monitored, with employees unaffected by UL-MSD, during medical monitoring, allows the population exposed to the risk of UL-MSD within the company to be characterized. Then, the time trends of the individual clinical status of employees regarding to UL-MSD is described and related to work characteristics.

Statistical analyses for representativeness employ conventional techniques. Variance analysis (ANOVA) is applied to test the hypothesis of equal means and frequency distributions are compared with the χ² test. Multivariate analysis uses logistic regression. Significance is defined as P<0.05. Analyses are performed using Statgraphics (version 5.1) software.

Results

Sociodemographic characteristics of study population

Six hundred and twenty-seven employees were invited to participate on a voluntary basis during the annual medical consultation. Four hundred and fifty-nine (73%) were systematically examined each year during the follow-up period. At baseline, the mean age is 43 years (SD 7.0), and duration of employment is 19 years (SD 7.6). These employees include significantly more women than men, compared with employees irregularly or never examined (P<0.05). They are also significantly older (P<0.0001) and their length of service is greater (P<0.0001), whilst there is no marked difference in age and length of service between men and women. There are significantly less employees from the administrative department compared with the other ones.

Irrespective of the year, reasons for absence from medical examinations, given by the latter employees, comprise mainly personnel movements and parental leave. A minority of absences, approximately 10%, are due to illnesses but not linked with UL-MSD. Finally, reasons for absence from medical examination remain unknown for 25 (4.0%) employees.

In the sample of employees systematically examined from 2000 to 2002, univariate analysis shows that employees affected by UL-MSD at least 1 in 3 years are significantly older (P<0.01) then unaffected employees and that women are more represented than men amongst employees affected by UL-MSD (P<0.0001). Most of them work in the production area and significantly in higher proportion in “meat-grill/toasters” and “Ovens” departments compared with the other ones (P<0.0001). Multivariate analysis, using a logistic model, was performed by considering sociodemographic characteristics that are significantly linked to suffering from UL-MSD. Analysis results show that the probability of having at least one UL-MSD during the three monitoring years is four times higher (P<0.00001) for women than for men and 4.5 times higher (P<0.0001) in production departments compared with the administrative department. Furthermore, the probability of suffering increases significantly with age.

Variability in employee clinical status with respect to UL-MSD

Detailed study of UL-MSD time trends over 3 years led to classifying the employees, seen at medical examination during three consecutive years, into eight clinical course typologies in relation to time (Table 2). These eight typologies represent all possible clinical status combinations during the 3 years follow-up. It turns out that 34.0% of employees seen at medical examination during three consecutive years present at least one UL-MSD during the monitoring period. Amongst these, 19.2% are affected by UL-MSD during three consecutive years (type 2), 25.6% are affected by UL-MSD during two consecutive years in three (types 3 and 4) and 55.1% are irregularly affected by UL-MSD with time (types 5–8).

Table 2 Clinical course typologies of employees monitored by occupational medicine from 2000 to 2002
Full size table

Amongst employees seen at medical examination during three consecutive years, the annual prevalence of UL-MSD cases is high and increases slowly during the medical monitoring period, from 17.6 to 19.2 and 21.6 in 2000, 2001 and 2002, respectively. Whatever the year, the annual prevalence remains high in the production departments (20.1, 21.9 and 24.0 in 2000, 2001 and 2002, respectively) compared with the administrative one (1.6, 3.3 and 6.6 in 2000, 2001 and 2002, respectively). Moreover, irrespective of the year considered, a high proportion of employees present simultaneously several UL-MSD, as illustrated in Fig. 1. Indeed, 31.0–42.0% employees affected by UL-MSD are effectively suffering from two UL-MSD or more, depending on the year. Finally, restrictions at work allocated by the Occupational physician concern about half of the UL-MSD cases every year (from 46.2% to 58.1% depending on the year) whereas they concern about 10% of employees without UL-MSD.

Fig. 1
figure 1

Annual distribution (%) of employees affected by UL-MSD with respect to the number of anatomical locations

Full size image

The proportion of incident cases of UL-MSD, defined as the number of employees newly affected in year n related to the number of employees remaining unaffected in year n−1, is 9.8% in 2001 and 13.5% in 2002. They do not differ from the other employees for age, gender and length of service, whatever the year. However, in 2002, there are significantly more incident cases in “ovens” department than in the others (P<0.05); they concern all anatomical locations.

The proportion of recoveries from UL-MSD, defined as the number of employees affected in year n and unaffected in year n+1, is also high in both 2001 and 2002, as might be expected from the distortion existing between a small increase in prevalence over 3 years and a high proportion of incident cases during the same period. Indeed, 37.0% of employees affected by UL-MSD in 2000 become unaffected by UL-MSD in 2001 and 44.3% of employees affected by UL-MSD in 2001 become unaffected in 2002. There are significantly more women than men concerned by recoveries in 2002 (P<0.05). No difference is observed concerning department affiliation, age and length of service whatever the year.

In all, 57.5% of employees affected by UL-MSD in 2002 were unaffected in 2000. Conversely, 48.1% of employees affected in 2000 became unaffected in 2002. Thus, the small increase in UL-MSD case prevalence over the three monitoring years conceals high variability of employee individual status with respect to UL-MSD, as Fig. 2 illustrates. Renewal in the company population affected by UL-MSD with time is in fact very extensive during 3 years of monitoring in all departments.

Fig. 2
figure 2

Variability of UL-MSD individual clinical status with time

Full size image

Renewal of the population of employees affected by UL-MSD is linked to the fluctuation of UL-MSD in all anatomical locations considered in the study. For example therefore, we observe that the prevalence of shoulder MSD, defined by the number of shoulder MSD in relation to the number of employees examined by the occupational physician, is 13.5% in both 2001 and 2002. Yet, in-depth examination of the data shows that 41 new shoulders MSD appeared between 2001 and 2002, whilst 41 shoulders MSD disappeared during the same period. Thus, stability of shoulder MSD prevalence in 2002, compared with 2001, in fact conceals the clinical fluctuation in this anatomical location with time. In 2002, shoulder MSD recoveries are effectively, strictly compensated by the appearance of new shoulder MSD (Fig. 3).

Fig. 3
figure 3

Prevalent cases, incident cases and recoveries by anatomical location

Full size image

This observation is backed up by close examination of data concerning employees affected by UL-MSD three consecutive years. Indeed, amongst these 30 employees, 12 are in fact affected by UL-MSD from different anatomical locations from one year to the next whilst 18 employees are affected by at least one UL-MSD from the same anatomical location during three consecutive years as seen in Table 3. In two-third cases, these persistent UL-MSD are associated to UL-MSD from other anatomical locations. All of them work in production departments.

Table 3 Anatomical localization of UL-MSD in employees affected 3 consecutive years
Full size table

If the data examination is extended to the employees affected by UL-MSD in two consecutive years, one-third present UL-MSD, whose anatomical locations differ from one year to the next.

Beyond the variability of UL-MSD related clinical status observed for most employees, the fact remains that 10% of employees seen at medical examination present UL-MSD lasting two or three successive years. In most cases UL-MSD are shoulder MSD.

Discussion

Sociodemographic and work characteristics

The participation rate was very high in the company with a total of 459 employees out of 627 enrolled in the study. This is partly explained by the stability of the working population. Parental leave and personnel movements were the main causes of absenteeism and could explain the difference on age and length of service between participants and nonparticipants. Parental leave indeed concerned only young women whereas personnel movements concerned a high majority of young men.

Upper limb musculoskeletal disorders have a high prevalence in all production departments. Employees affected by UL-MSD at least 1 in 3 years are significantly older than employees who have never been affected. This phenomenon has been frequently reported and could be partly explained by the chronic overload of muscles and tendons induced by the combination of a lowered fatigue threshold for the ageing worker and a reduced recovery after muscle activity with age, as suggested by de Zwaart et al. (1995). Furthermore, amongst the number of employees seen systematically at medical examination, UL-MSD are significantly more prevalent in women than in men; that could be explained in several ways. In one way, medical diagnosis was first based on pain reports and then completed by a clinical examination. The over-reporting of pain by women compared to men was described in many studies (Dahlberg et al. 2004). It could have introduced a reporting bias in the present study. In another way, women were not employed at the same tasks as men in this company; women were indeed over-represented in production departments where the exposure to risk factors for UL-MSD is very high. Moreover, in production departments, most women are employed in very hand intensive and repetitive tasks on production lines, whereas men are more often employed at lifting and carrying heavy loads to supply workplaces and to package merchandise at the end of lines, with a longer cycle time. So the difference observed between men and women in relation to UL-MSD prevalence could be explained by a gender difference in occupational exposure. Thus, sociodemographic characteristics of the employees affected by UL-MSD reflect the results of many studies published in this field (de Zwaart et al. 2001).

Variability of clinical status

The clinical examination is based on a standardized approach and was repeated by the same physician throughout the study; this is an argument for the internal validity of the measurements and the medical data can therefore be considered reliable. The annual prevalence of UL-MSD diagnosed by the occupational physician was high in 2000, 2001 and 2002. It seems not to be explained by too “relaxed” diagnosis criteria. Actually, they have systematically associated subjective signs and at least one positive clinical test, whatever the pathology concerned. Furthermore, during the 3-year monitoring periods, the annual prevalence of UL-MSD varies little whilst a high intra-individual variability of UL-MSD is observed in the company. In fact, out of 156 employees affected by UL-MSD at least once during the three years of monitoring, 80.8% change effectively their clinical status with time. These results reflect a very high renewal rate for the population of employees affected by UL-MSD, as illustrated by a high annual incidence of UL-MSD cases and a high proportion of recoveries. Roquelaure et al. (2002), Takala et al. (1992), Cole et al. (2002), Jonsson et al. (1988) and Kaergaard et al. (2000) also refer to the fluctuation of UL-MSD following prospective studies conducted in various sectors of activity such as the shoe industry, the banking sector, the press or the clothing industry. Depending on the study, state of health assessment methods vary. They include collection of symptoms by questionnaire and/or a standardized clinical examination, possibly complemented by functional explorations. Similarly, surveillance frequency varies between 3 months and 1 year. In spite of these differences, UL-MSD variability is observed by the authors, whatever the criteria considered. In the present study, more frequent clinical examinations during the follow-up could have shown even more variability. Medical assessment was actually done once a year and no information is provided in the interval. So, transient cases that might have developed and disappeared between the scheduled examinations were not detected. In fact, nothing is known about shorter period clinical status variations. Moreover, the health evaluation was done in a binary way (affected / unaffected). Consequently, no information was provided about the variability of the clinical severity degree within subjects.

Incident cases do not differ from the others regarding main socio-demographic characteristics but in 2002, they are significantly higher in “ovens” department compared with the other ones and concern essentially shoulders and wrist/hand disorders. It could be due to the several events. Indeed, in 2001, changes occurred in “ovens” department to supply the production lines by the front instead of right and left side of the operators, so that employees had to work most of the time with the hands at or above shoulder height to reach assembly pieces. In the same time, cycle time was reduced on the lines. No change was observed at a “macro-level” in the other departments. Moreover, “Ovens” department was the only one in 2002 to be concerned by the implementation of training sessions about symptoms of MSD and their main risk factors before the implementation of an ergonomic intervention in 2003. So, it could have influenced employees’ pain reporting.

In this study, clinical status variability is observed whatever the anatomical location concerned. The specific place of shoulder MSD should be considered. The shoulder is the anatomical location most frequently concerned by in-company UL-MSD and it is therefore not surprising that there are also more shoulder MSD fluctuations in absolute numbers. Paradoxically, in the present study, the shoulder is also the anatomical location most affected by MSD considered as persistent, i.e. observed by the occupational physician two or three consecutive years. Shoulder MSD therefore appears very fluctuating and the most persistent compared with other anatomical locations. These two modes of shoulder MSD clinical course involve employees who do not differ significantly from each other in terms of main sociodemographic criteria including the age. In a follow-up study of MSD in industrial and service workers, Bonde et al. (2003) indicate the transient nature of shoulder disorders; they observe that some 50% had recovered within 10 months. Nevertheless, shoulder pain and clinical signs continued for more than 18 months in some 25%. In our study, the clinical course of shoulder MSD is probably influenced by occupational factors such as changes observed in the workplace design in “ovens” department. Restrictions allocated by the occupational physician could also have changed the clinical course of shoulder MSD. Some workers probably used new working strategies to diminish muscular strain on their own initiative or because of the recommendations of the occupational physician to adapt working conditions to the observed disabilities. Employee therapeutic healthcare and extra-occupational factors such as unpaid work or recreation activities were not monitored in this study. So, their influence on the clinical course could not be assessed.

Nevertheless, the hypothesis according to which transition to chronicity of shoulder MSD is preceded by major variations in clinical signs should not be excluded. The underlying assumption is that both profiles of shoulder MSD observed in the company could correspond to different developing clinical stages in an ageing population.

The high prevalence of UL-MSD observed in this company has led to the implementation of an intervention program to prevent UL-MSD. The demonstration of the fluctuating nature of MSD was a strong argument to promote a global approach, taking into account people affected or not by UL-MSD, in all types of work situations. Intervention has begun first in the “ovens” department where the highest prevalence was observed and then has been developed gradually in other departments. The intervention programme is presently being continued with changes in work organization and technology.

Conclusion

In conclusion, UL-MSD were characterized in the present study by temporal changes in clinical status, whatever the anatomical location concerned. These results argue for a high variability of MSD clinical status with time. They reinforce the needs for global and durable interventions to prevent UL-MSD. Within a wider research perspective, the knowledge about the natural course of UL-MSD in the active population is far from complete and this area of investigation has in fact been little explored to date. Clarification of the transition from one clinical status to another appears necessary. Yet, closer understanding of the mechanisms influencing the clinical course of UL-MSD including incidence and recovery would allow new lines of prevention to be opened up.