Questionnaire for low back pain in the garment industry workers.

Low back pain affects up to 90% of the world's population at some point in their lives. Until date no questionnaire has been designed for back pain in the garment industry workers. Therefore, the objective of this study is to design a questionnaire to determine the prevalence, risk factors, impact, health care service utilization and back pain features in the garment industry workers and gain preliminary experience of its use. The content validity and reliability of the questionnaire was established. Items showing acceptable internal consistency and moderate to high test re-test reliability were retained in the questionnaire. Items showing unacceptable internal consistency, low test re-test reliability or poor differentiation were reworded, redrafted and re-tested on the workers. It took 20 min to complete one interview schedule. Environmental factors such as the absence of the garment industry owner/supervisor or co-workers at the time of the interview and interview during leisure hours need to be standardized. Thus, final questionnaire is ready for use after necessary amendments and will be used on the larger sample size in the main study.

INTRODUCTION

Low back pain (LBP) is ranked first as a cause of disability and inability to work and is expected to affect up to 90% of the world's population at some point in their lives.[1] LBP is a poorly defined term and has been defined as “any back pain between the ribs and top of leg, from any cause;”[2] “pain, muscle tension or stiffness localized below the costal margin and above the inferior gluteal folds, with or without sciatica and is defined as chronic when it persists for 12 weeks or more.”[3] Thus defining LBP becomes essential for a particular study of interest. For this study LBP is defined as “pain in the region between the lower margins of the 12th rib and the gluteal folds.”[4]

The outcome measures for LBP vary from being general health status measures (SF 36) to region specific questionnaires (Oswestry disability index [ODI], Roland Morris disability scale). The most commonly used back pain specific measures such as Roland Morris disability questionnaire assesses the domain of physical disability from LBP, but is inadequate to assess psychological or social problems related to LBP. ODI permits investigation into patient's social problems as well and is recommended in patients of persistent severe disability. The Quebec back pain disability scale assesses only functional disability and sleep, but doesn’t evaluate pain. The Waddell disability index is a short nine item score consisting of questions about daily activities but needs to be associated with other functional scoring systems to obtain a complete assessment of disability.[5] Thus, each of these questionnaires needs to be supplemented with other questionnaires or physical examination features to cover the maximum constructs and risk factors for LBP.

As such no questionnaire specific to LBP in the garment industry workers has been found in the literature, therefore the objective of this study is to design a schedule of the questionnaire to study the prevalence, physical and psychosocial risk factors, impact, health care service utilization and back pain features in the garment industry workers and gain preliminary experience of its use.

METHODOLOGY

Generation of items for the draft questionnaire

The process commenced after review of literature and other questionnaires relevant to the area of interest. The domains of enquiry were: Demographic features; identification and screening of LBP; determining point, 12 months and lifetime prevalence of LBP; categorization of LBP; determining the causative factors in terms of personal physical (such as back and abdominal muscle strength and endurance, spinal range of motion); work related physical (such as frequent bending, twisting, lifting weight); personal psychosocial (such as mental stress at job, fear of loss of job) and life-style (smoking, alcohol consumption and physical activity) factors associated with LBP; determining the impact of LBP in terms of activity limitation due to current LBP episode [Table 2] and back pain in the past 12 months; cost of the work days lost; health related expenses; sickness absenteeism and sleep disturbance and the remedial measures sought for LBP [Appendix 1]. The personal physical factors were determined using a separate physical examination form.

Table 2

Activity limitation due to current LBP episode

APPENDIX 1

Questionnaire for back pain in garment industry

Amendments in the questions

The amendments for content validity of the questionnaire were established by a physiotherapist, occupational therapist and an expert from community medicine. The expert advice from academic and clinical experts in the field included a university presentation feedback session.

Classifications and coding legends for questionnaire

The demographic features were analyzed according to the category of responses given along with. Numerical scoring was given to each of the responses. Questions to be answered in terms of yes were given a score of 1 and in terms of no were given 2. Responses on likert scale were scored from 1 to 5. The socio-economic status was calculated and classified as per Kuppuswamy socioeconomic status scale (1976).[6]

Reliability of the questionnaire

The questionnaire was tested for internal consistency and test re-test reliability.

Internal consistency

The internal consistency for the construct of activity limitation due to current episode of LBP, activity limitation due to LBP in past 12 months, physical and psychosocial factors at work were tested using Cronbach alpha on a sample of 29 respondents (of which 21 male and 8 female working on different sections of garment industry.

Internal consistency was found to be excellent for the construct on activity limitation due to current LBP episode (Cronbach's alpha = 0.99) and LBP in the past 12 months (Cronbach's alpha = 0.99) and unacceptable for physical (Cronbach's alpha = 0.48) and psychosocial (Cronbach's alpha = −0.01) factors at work. The items found to have unacceptable internal consistency were reworded, redrafted and deleted. Physical factors at work such as sitting cross legged, kneeling or squatting and whole body vibration were deleted as these were not the part of working environment in the garments industry workers. Question on psychosocial factors such as “Do you suffer mental stress at job?” was reworded as “I am asked to do too much work,” “I have to work very hard" and “I enjoy my work.” The construct for physical and psychosocial factors was re-tested for internal consistency on a sample of 10 subjects. The internal consistency was found to be acceptable (Cronbach's alpha = 0.77) for physical factors but unacceptable (Cronbach's alpha = 0.42) for psychosocial factors. The probable reason for unacceptable internal consistency for psychosocial factors could be the environmental factors such as presence of a supervisor or owner at the time of interview, thus the environmental conditions were standardized and 7 subjects were interviewed in the absence of the supervisor/owner, interview was taken during the leisure hours/rest intervals and psychosocial factors tested for internal consistency. Upon standardization of environmental conditions, the internal consistency was found to be acceptable (Cronbach's alpha = 0.70) and the items were retained in the questionnaire.

Test re-test reliability

Test re-test reliability was computed using Pearson correlation coefficient. A total of 26 subjects responded for a second interview after an interval of ten days for test re-test reliability, giving a response rate of 89.65% and were used for the study. The items showing little or low correlation (r ≤ 0.49) were deleted or redrafted.

Demographic features such as age and level of education showed moderate (r = 0.50-0.69) whereas socio-economic status (r = 0.29) showed low test re-test correlation [Table 6]. The test retest correlation was high (r = 0.70-0.89) for pain features except for perceived cause of pain which showed moderate correlation. Test retest correlation was found to be moderate for point, lifetime prevalence and low for 12 month prevalence. Frequency of attacks of LBP in past 12 months showed high, presence of pain on most days in past 12 months showed moderate, duration of longest attack, more than one attack and 12 month interference with work showed low test re-test correlation [Table 6].

Table 6

Test re-test reliability for various variables under study

Test re-test correlation was found to be moderate for absenteeism from work due to LBP and high for the duration of absenteeism. Questions on impact of LBP in terms of change of job, cost of work days lost, cost of treatment and compensation for work days lost showed moderate correlation. The question enquiring age at the initial onset of LBP showed low test retest correlation [Table 6].

Test retest correlation for questions on duration of the job was high for the number of years in to job and low for hours/day, hours/week and days/week. Work parameters such as inadequate rest intervals, shortage of staff and income adequacy showed moderate test re-test correlation [Table 6]. Test retest correlation was found to be moderate for activity limitation due to current LBP episode [Table 2] and activity limitation due to LBP in the past 12 months [Table 3].

Table 3

Activity limitation due to LBP in past 12 months

Question enquiring “whether physical factors at work are dangerous for back?” showed moderate test retest correlation. Physical factors such as frequent bending and whole body vibration showed high; lifting heavy objects, sustained sitting and sustained standing moderate; twisting, working with back in awkward position and lifting light objects showed low correlation. Kneeling or squatting and sitting cross legged showed moderately negative test re-test correlation [Table 4].

Table 4

Physical factors at work

Test re-test correlation for job satisfaction was found to be low. Psychosocial factors such as mental stress at job showed moderate, fear of loss of job due to back pain and poor relations with employer showed high test re-test correlation. Ability to take decisions on your own, co-worker support and monotonous, repetitive work showed low correlation. Test re-test correlation for lack of recognition at work couldn’t be computed as the variables were constant [Table 5].

Table 5

Psychosocial factors at work

The test re-tests correlation for remedial measures such as consultation sought and by whom was found to be moderate. Question on self-rating of general health status and physical activity showed moderate correlation but life-style factors such as past history of smoking, drinking alcohol and sleep disturbance due to LBP showed low test retest correlation. Correlation coefficient for questions on current smoking, number of years smoked and cigarettes per day could not be computed because the variable was constant [Table 6].

DISCUSSION

A majority of workers were male, young in age, migrant and belonged to lower middle and upper lower categories of socio-economic status. Nearly 6.89% were illiterate and the commonest language spoken was Hindi. The average duration of job was 5.59 years and they worked for an average of 51.03 h/week [Table 1]. Saha et al.[7] in their study on assessment of health status of workers in garment industry of Kolkata observed that most of the workers were males (76.79%) in the age group of 15-45 years (80.36%). Nearly 23.21% of the workers were illiterate and most of them belonged to poor socio-economic status.

Table 1

Demographic features of the workers in garment industry

Internal consistency and test retest reliability have been used to measure reliability in present study. Test re-test reliability has been considered to be more relevant in clinical medicine because the constructs to be measured are heterogeneous in nature as different constructs (activity limitation, physical factors, and psychosocial factors) in this study. Thus one may expect clinically heterogeneous scales to have poor internal consistency. Thus, internal consistency of individual constructs has been established in the present study. The problem of testing reliability with test retest method is that there is potential of learning, carry over or recall effects.[8] Most of the investigators have chosen an interval ranging from 2 days to 2 weeks. This time frame is generally believed to be a reasonable compromise between recollection bias and unwanted clinical change and the questionnaires repeated after 2 days or 2 weeks have shown no statistically significant differences,[9] thus an interval of 10 days for test retest reliability had been chosen for the present study. It took 20 min to complete an interval schedule in Hindi, at a pace understandable to them, as most of the workers were migrant and the commonest language known to them was Hindi.

The items on activity limitation due to current LBP episode and back pain in the past 12 months have been found to have excellent internal consistency and moderate test retest reliability. The probable reason for this is that questions on activity limitation due to current LBP episode have been referred from SF 36 physical functioning scale.[10] The internal consistency of SF 36 physical functioning scale has been found to be excellent (Cronbach's alpha = 0.93) in general and working age (Cronbach's alpha = 0.90) subjects of British population.[11] The items on activity limitation due to back pain in the past 12 months had been adopted from questionnaire for identification of back pain for epidemiological purposes[4] where the sensitivity of items on interference of back pain with normal work has been found to be moderate (64%).

The construct of physical and psychosocial factors at work were found to have unacceptable internal consistency. The items found to have some difficulty in interpretation, double barreled or leading questions were reworded and redrafted to improve internal consistency. The physical and psychosocial factors eliciting low test retest reliability or poor differentiation were deleted from the final questionnaire.

Items such as socio-economic status, cigarette smoking and alcohol consumption showed low test retest correlation or poor differentiation; a probable reason for this variability could be that the worker may not be willing to reveal his income, social status, education and/or addiction. Thus standardization of above mentioned environmental factors such as the absence of the industry owner/supervisor or co-workers at the time of interview and interview during leisure hours needs to be done at the time of interview.

During the interview, it was observed that there was difficulty in interpreting the response options for the question “how badly is sleep affected by back pain?” (Responses: [a] Sleep of <6 hours/day [b] Sleep of <4 h/day [c] Sleep of <2 h/day [d] Pain prevents sleeping at all) therefore the responses were changed to a likert scale (never, rarely, sometimes, often, all of the time).

Upon interview some modifications were felt to be made to the questionnaire such as an addition of option of self-medication and massage in health care services utilized, tobacco chewing in life-style factors, any time in question on when pain is felt during the day, cannot say in perceived cause of back pain and option of gym in question on the type of exercise.

Thus, the final questionnaire has been found to have acceptable internal consistency and moderate to good test re-test reliability and is ready for use.