Tool development to assess the work related neck and upper limb musculoskeletal disorders among female garment workers in Sri-Lanka

Abstract

BACKGROUND: Musculoskeletal disorders (MSDs) have been identified as a predisposing factor for lesser productivity, but no validated tool has been developed to assess them in the Sri- Lankan context.

OBJECTIVE: To develop a validated tool to assess the neck and upper limb MSDs.

METHODS: It comprises three components: item selections, item reduction using principal component analysis, and validation. A tentative self-administrated questionnaire was developed, translated, and pre-tested. Four important domains – neck, shoulder, elbow and wrist – were identified through principal component analysis.

RESULTS: Prevalence of any MSDs was 38.1% and prevalence of neck, shoulder, elbow and wrist MSDs are 12.85%, 13.71%, 12%, 13.71% respectively. Content and criterion validity of the tool was assessed. Separate ROC curves were produced and sensitivity and specificity of neck (83.1%, 71.7%), shoulder (97.6%, 91.9%), elbow (98.2%, 87.2%), and wrist (97.6%, 94.9%) was determined. Cronbach’s Alpha and correlation coefficient was above 0.7.

CONCLUSION: The tool has high sensitivity, specificity, internal consistency, and test re-test reliability.

Keywords

Ergonomic assessment occupational shoulder pain

1 Background

1.1 Work related musculoskeletal disorders

Musculoskeletal disorders (MSDs) are an important public health and occupational health problem in both developed and developing countries, with substantial impact on quality of life and productivity due to lost working hours which imparts a substantial economic burden. In the United States, Canada, Finland, Sweden and England, disorders have caused more work absenteeism or disability than any other group of diseases [1]. Furthermore Van Eerd [1] reported that aches, pain and restriction of motions due to musculoskeletal disorders among female garment workers have hindered normal performance at work.

World Health Organization has characterized “work-related musculoskeletal disorders” as a multi-factorial disease because its etiology associates with personal, physical, work-organizational, psychosocial and socio-cultural, etc. factors [2].

Occupational health research has identified a strong relationship between work and work load and leisure time activities [1]. It is also important to consider workload and the match between the work tools and man in order to prevent musculoskeletal disorders among the workers [3 –5].

According to Punnett and Wegman [6], textile, apparel, and shoe manufacturing industries are considered high-risk occupational settings for MSDs.

MSDs are the single largest category of work-related illness, representing a third or more of all registered occupational diseases in the United States, the Nordic countries, and Japan [7 –9]. Numerous surveys of working populations have reported prevalence of upper extremity symptoms of 20 to 30% or even higher.

1.2 Garment industry and MSDs among female garment workers in Sri Lanka

The garment industry has become the driving force of the Sri Lankan economy and more than 2,918 garment factories are currently in operation. Of these, 26% are small scale industries with less than 100 employees and 51% of them belong to the medium scale. Only 23% belong to the large and very large categories with 500 or more employees, accounting for 62 per cent of the total employment in the industry [10]. By 2010 Sri Lanka’s textile and garment sector accounted for 8 per cent of GDP, 42 per cent of industrial production, 35 per cent of manufacturing employment, 56 per cent of total exports and 78 per cent of industrial exports.

The total Sri Lankan labour force, excluding Northern and Eastern provinces approximates to 9.9 million and more than 5 per cent of them are employed in the garment industry. According to the job title, job descriptions and task performance, the Sri Lankan garment industry worker can be divided in to two broad categories; a) factual garment workers and b) auxiliary garment workers. Where their educational standards are concerned, around 74% of the female workers have been identified to be secondary school dropouts and only 20% had completed secondary education [10].

1.3 Justification

Only two studies have been carried out to investigate work related musculoskeletal disorders among garment workers in Sri Lanka. The first study was on “Occupational overuse disorders in upper limbs of female garment workers and rapid ergonomic assessment of individual work areas” by Perera [11]. The second was on “The effects of posture and musculoskeletal disorders among automated sewing machine operators and productivity among female garment workers” by Haniffa [12]. These studies had many limitations and recommendations, focusing on the need for future research on MSDs and specifying the need of a validated tool to assess the work related MSDs for Sri Lanka.

2 Objective

To develop and validate an instrument to assess the prevalence of work related neck and upper limb musculoskeletal disorders among the female garment workers in Sri Lanka

3 Methods

The study was designed to be carried out in two stages including the following two components:

Component 1 – Develop and validate an operational definition of work related neck and upper limb musculoskeletal disorders was carried out.

Component 2 – Development and validation of a tool to identify work related neck and upper limb MSDs and assess their prevalence.

Accordingly, an operational definition was developed to define ‘Work related neck and upper limb musculoskeletal disorders’ based on “symptoms” and “occuupational relatedness” respectively. Accordingly pain was identified as one of the necessary symptoms as per ICD 9 WHO [13], Canadian Centre for Occupational Health & Safety [14], Viikari-Juntura et al. [15], Stetson et al. [16], Baron et al. [17], Grandjean [18], Punnett [19]. Concurrently 22.09.10 “onset of the discomfort relates with the employment in the current job” was identified as an absolute criterion from occupational relatedness definingcriteria [20 –23].

3.1 Operational definition of “Work related neck and upper limb MSD”

A person to be diagnosed of having “work related neck and upper limb MSD” should complain of “pain” in the neck and or upper limb and “onset of the discomfort relates with the employment” as an absolute criterion, together with one or more of following symptoms; ache, numbness, burning sensation, edema, difficulty in moving, swelling, redness of the location and one or more of the following occupational relatedness defining criteria.

Discomfort does not relate with the past history of prior accident or sudden injury in the given location, discomfort episodes start with work or after few days of working, severity of the MSDs will be reduced during the off days. Accordingly Fig. 1 shows the absolute and probable symptoms and occupational relatedness criteria used in the operational definition of thisstudy.

The operational definition was modified by a panel of experts. The final modified operational definition was created including “pain” and “onset of the discomfort relates with the employment” as absolute criteria and ache, numbness, burning sensation, restriction of movements and swelling as probable symptoms. Discomfort does not relate with the past history of prior accident or sudden injury in the given location, discomfort episodes start with work or after few days of working, severity of the MSDs will be reduced during the off days as probable occupational relatedness criteria was used in the definition of the study.

The principal investigator has decided to develop a new study instrument based on the available NMQ, UEQ and identified symptoms and occupational related criteria titled as the “Neck and Upper limb Musculoskeletal Disorders Questionnaire” (NULMSDQ).

3.2 Consensus on translation of the Neck and Upper Limb MSDs Questionnaire

There were no major discrepancies between the English version and the Sinhala translation of the draft NULMSDQ. The validity was assessedemploying the following methods.

a) Assessment of content and consensual validity

Content and consensual validity was examined through a modified Delphi process [24].

b) Assessment of construct validity

Principal component analysis was carried out to 1) explore the relationship among the variables, 2) to see if the pattern of results can be explained by the underlying constructs 3) test some hypotheses about the data, and 4) reduce the number of variables to a more manageable size [25].

c) Assessment of criterion validity

The clinical criteria for the diagnosis of work related MSDs according to the ICD 9 classification [13], has been used in the clinical diagnosis of MSDs as well as in research as a “gold standard’ for the detection of MSDs [14, 20]. The participants of the validation study underwent a clinical examination conducted by a consultant rheumatologist where each one was diagnosed as either having “work related MSDs” or “not having work related MSDs” based on the above criteria.

3.3 Study design for the validation of NULMSDQ

A descriptive, cross sectional validation study was conducted.

3.4 Study population for the validation study

Females who fulfilled the eligibility criteria and employed for a period of six months or more in the selected garment factory, were defined as the study population.

3.5 Sample size for the validation study

The required minimum sample size for the validation study was calculated based on the expected sensitivity or the specificity of the instrument (p), required level of precision (w) and the confidence level (α) [26].

According to Perera [11] the prevalence of upper limb MSDs was 12%. Based on that the sample size required for the validation study was 292workers.

3.6 Sampling technique

The calculated sample size was selected by systematic random sampling technique.

3.7 Data collection for the validation study

The developed, self-administrated draft Neck and upper limb MSDs questionnaire was administrated among 292 female garment workers and they underwent the clinical examination conducted by a Consultant Rheumatologist. Based on this clinical examination they were specially designed diagnosed as ‘having MSDs,” or ‘not having MSDs “and the diagnosis was marked in the diagnosis card according to the ICD 9 [13] criteria for work related MSDs, without looking at the information provided in the questionnaire. Concurrently work related MSDs, thoracic outlet syndrome, tension neck syndrome, rotator cuff syndrome, epicondylitis, DeQuervain’s disease and carpal tunnel syndrome were diagnosed by the Consultant Rheumatologist.

Criterion validity was assessed by comparing the total score of each anatomical region with the clinical diagnosis (presence or absence of MSDs) as the gold standard.

4 Results

a) Content validity

Content and consensual validity was examined through a Delphi process [24]. According to the panel of experts, nearly all items assessing the construct of interest in the draft Sinhala version of the NULMSDQ were judged to be adequate and appropriate to the target population. At this stage of validation process the draft NULMSDQ consisted of 52 questions including 47 questions related to symptom criterion and five questions related to occupation related criteria.

b) Construct validity

These results show evidence of factor loading in to 5 domains such as neck, shoulder, elbow, and wrist symptoms and occupation related criteria.

4.1 Results of criterion validity

The clinical diagnosis of work related neck MSDs was diagnosed by a Consultant Rheumatologist using a specifically developed diagnosis card to maintain the uniformity of the diagnosis. There were 74, 69, 62 and 28 workers who were clinically diagnosed as having work related neck, shoulder, elbow and wrist MSDs respectively and 218, 223, 230 and 264 as clinically “not” having work related neck shoulder, elbow or wrist MSDs respectively. ROC curve for work related MSDs was drawn using this diagnosis. Finally there were 103 workers who had any (one or more) neck and upper limb MSDs. Tables 1 and 2 show the distribution of the clinical diagnosis of the work related neck MSDs and the distribution of the clinical diagnosis of work related MSDs syndromes of 292 female garment workers respectively.

Table 2 shows the clinical syndromes that were diagnosed, according to ICD-9 criteria by a Consultant Rheumatologist. It was used as the “gold standard” criterion in drawing the ROC curve. The total score of each domain was used as a test variable in the ROC curve which was calculated by adding the scores of each symptom according to the severity scale. Separate ROC curves were created using the separate scores of work related neck, shoulder, elbow, and wrist as test variables and the clinical diagnosis as the criterion variable which determined the ability of each symptom score to predict work-related neck, shoulder, elbow and wrist MSDs compared to the clinicaldiagnosis. Accordingly, the diseases positives’ against test positives of work related neck, shoulder, elbow, and wrist MSDs among the 292 workers were assessed, and there the sensitivity and specificity corresponding to different cutoff values of the total score of work related musculoskeletal disorders were identified. Optimum cut-off value for each body part was selected based on the shortest distance from the curve. Accordingly, the best tradeoff between the sensitivity and specificity was corresponded with the cut off value of 23.0 and, at this cut off value the total score had a sensitivity of 91.9% and a specificity of 89.9% for work related neck MSDs. The disease positives against test positives of work related shoulder MSDs among the 292 workers show the sensitivity, and specificity which was corresponded with the cut off value of 20. At this cut off value, the total score had a sensitivity of 94.2% and a specificity of 99.1%. Furthermore, the diseases positives’ against test positives of work related elbow MSDs among the 292 workers show the sensitivity, and specificity as 79.0% and 95.2% at the cut off values of the 30. The sensitivity and specificity for work related wrist musculoskeletal disorders were 78.6% and had a specificity of 58.3%.

Further ROC curves were created using the total score as test variable and the clinical diagnosis of musculoskeletal disorder as the criterion variable. Separate ROC curves were plotted for the different cut off values of the total score of the work related tension neck syndrome, thoracic outlet syndrome, rotator cuff syndrome, elbow epicondylitis, DeQuarions and carpel tunnel syndromes, giving the sensitivity against different values for (1- specificity) using test variables and the clinical diagnosis as the criterion variable.

The summery of the criterion validity of the NULMSDQ at cut off scores for work related neck, shoulder, elbow and wrist musculoskeletal disorders and specific syndromes sensitivity, specificity, positive and negative predictive values for all MSDs are given in the Table 3.

Reliability of the study instruments were assessed by establishing test-retest reliability and internal consistency (Cronbach alpha). Internal consistency estimates having a magnitude of 0.70 or greater were considered satisfactory [27].

A random sample of 20 female garment workers was conveniently selected from a factory at Katunayaka EPZ. To assess the test-retest reliability, a 14-day interval was allowed.

As presented in the Table 4 the Cronbach Alpha estimates for the subscales ranged between 0.76 and 0.98.

4.2 Assessment of test-retest reliability

Test-retest reliability measures the ability to generate stable results which are reproducible [28, 29]. A correlation coefficient (Spearman’s r) of 0.70 or above was considered as a good level of agreement between the scores [29]. This was assessed by re-administering the NULMSDQ after two weeks to a sub sample of 20 female garment workers who were conveniently selected from a garment factory at Katunayaka EPZ and the Table 5 shows the test-retest reliability of the NULMSDQ.

The translated tentative NULMSDQ was then pre-tested and the social acceptability was assessed.

5 Conclusions

The Neck and Upper Limb Musculoskeletal Disorders Questionnaire was developed based on the work relatedness criteria and considering the respective symptoms related to the disease condition. Content validity and criterion validity of the final version of the tool was assessed. Physical examination by a rheumatologist was considered the “Gold standard”. Separate ROC curves for each body part were produced and sensitivity and specificity of each subsection with symptoms of neck (83.1%, 71.7%), shoulder (97.6%, 91.9%), elbow (98.2%, 87.2%), and wrist (97.6%, 94.9%). Cronbach’s Alpha and correlation coefficient of each subsection was above 0.7. Therefore, it was concluded that this tool has a very high sensitivity and specificity and good internal consistency and test re-test reliability. Finally, it was considered a validated questionnaire to assess the work related neck and upper limb MSDs inSri Lankan female garment workers.

Conflict of interest

The authors have no conflict of interest to report.

Footnotes

Acknowledgments

I sincerely thank Prof. Rohini De A. Senevirathne the senior professor, Department of Community Medicine, Faculty of Medicine, Colombo for thegenerous guidance, support and advisors and supervising my thesis to complete my MD community Medicine. The time spent on correcting the thesis cannot be evaluated and valued. I always acknowledge the guidance and encouragement given by Prof. Alwis in completing this research.

References

Van Eerd

, Beaton

, Cole

, Lucas

, Hogg-Johnson

, Bombardier

. Classification systems for upper-limb musculoskeletal disorders in workers. Journal of Clinical Epidemiology2003;56(10):925–36.

The main risk factors for work related musculo skeletal disorders. World Heath Organization [Internet]. 1985 [cited 8 January 2014];. Available from: http://www.ilo.org/safeworkbookshelf

Andersson

. The epidemiology of spinal disorders principles and practice. 2nd ed. New York: Raven Press Ltd; 1996.

Araujo

, Karasek

. Validity and reliability of the job content questionnaire in formal and informal jobs in Brazil. Scand J Work Environmental Health Supplement2008;6:52–9.

Bjelle

, Hagberg

, Michaelson

. Occupational and individual factors in acute shoulder-neck disorders among industrial workers. Occupational and Environmental Medicine1981;38(4):356–63.

Punnet

, Wegman

. work related musculoskeletal disorders: The epidemiological evidence and the debate. J ElectromyogrKinesiol2004;14:13–23.

Arthur

, Hawkins

, Pollard

, Catalano

, Baglioni

. Measuring risk and protective factors for substance use, delinquency and other adolescent problem behaviors: The communities care youth surveys. Evaluation Review2002;26(6):575–601.

David

. Ergonomic methods for assessing exposure to risk factors for work-related musculoskeletal disorders. Occupational Medicine2005;55(3):190–9.

Frank

, Pulcins

, Kerr

, Shannon

, Stansfeld

. Occupational back pain – an unhelpful polemic. Scandinavian Journal of Work, Environment & Health1995;21(1):3–14.

10.

Central bank of Sri Lanka, Annual Report: Future challenges of the apparel industry’s report on consumer finance and socio-economic survey Colombo. Central Bank of Sri Lanka; 2003.

11.

Perera

. Occupational overuse disorders in upper limbs of female garment workers and rapid ergonomic assessment of individual work areas. Post Graduate Institute of Medicine, University of Colombo; 1993.

12.

Haneefa

. The effects of posture and musculoskeletal disorders among automated sewing machine operators and also on productivity among female garment factory workers. Post graduate institute of Medicine, University of Colombo; 2003.

13.

International classification of diseases and related health problems [Internet] 10th ed. World Health Organization; 2010 [cited 22 March 2010]. Available from: http://www.who.int.

14.

Canadian center for occupational health and safety. Work related musculoskeletal disorders [Internet]. 2005 [cited 16 December 2009]. Available from: http://www.cicohs.ca/oshanswers/diseases/rmirsi.html.

15.

Viikari-Juntura

, Vuori

, Silverstein

, Kalimo

, Kuosma

, Videman

. A Life-long prospective study on the role of psychosocial factors in neck-shoulder and low-back pain. Spine1991;16(9):1056–61.

16.

Stetson

, Silverstein

, Keyserling

, Wolfe

, Albers

. Median sensory distal amplitude and latency: Comparisons between non exposed managerial/professional employees and industrial workers. Am J Ind Med1993;24(2):175–89.

17.

Baron

, Hales

, Hurrell

. Evaluation of symptom surveys for occupational musculoskeletal disorders. Am J Ind Med1996;29(6):609–17.

18.

Grandjean

. Ergonomics in computerized offices. London: Taylor & Francis; 1987.

19.

Punnett

, Fine

, Keyserling

, Herrin

, Chaffin

. Back disorders and no neutral trunk postures of automobile assembly workers. Scandinavian Journal of Work, Environment & Health1991;17(5):337–46.

20.

Encyclopedia of occupational safety and health. 4th ed. International Labour organization; 2004.

21.

Snook

, Campanelli

, Hart

. A study of three preventive approaches to low back injury. Journal of Occupational and Environmental Medicine1978;20(7):478–81.

22.

Hadler

. Repetitive upper-extremity motions in the workplace are not hazardous. The Journal of Hand Surgery1997;22(1):19–29.

23.

Hansen

, Jensen

. musculoskeletal disorders in the work place: Low back and upper extremities. Work Place Ergonomics. Washington DC: National Academy press; 1993.

24.

Jones

, Hunter

. Qualitative research: Consensus methods for medical and health services research. BMJ1995;311(7001):376–80.

25.

Katz

, Punnett

, Simmons

, Fossel

, Mooney

, Keller

. Workers’ compensation recipients with carpal tunnel syndrome: The validity of self-reported health measures. Am J Public Health1996;86(1):52–6.

26.

Hully

, Cummings

, Browner

. Designing clinical research. 5th ed. Lippincott Williams and Wilkins; 2001.

27.

Nunnally

, Bernstein . Psychometric theory. USA: McGraw Hill Inc.; 1994.

28.

Abramson

, Abramson

Survey methods in community medicine epidemiological research programme evaluation. Clinical Trials. 5th ed. New York: Churchill Livingstone Harcourt Brace and Company Limited; 1999.

29.

Litwin

. The survey kit: How to measure survey reliability and validity. California: University of California; 1995.