The Relationship Between Readiness to Change Pain-Related Exercise Participation and Perceived Work Function: A Cross-Sectional Study of Factory Workers


 Background: Healthy lifestyle behaviours are associated with protection against health disorders and pain. Exercise participation is one such behaviour, associated with improved outcomes in those experiencing pain. Musculoskeletal pain is highly prevalent in the workplace, particularly in factory workers and associated loss of work function is recognised as having a great impact on individuals, society and the economy. A worker’s ‘readiness to change pain behaviour’ is an important factor to consider in achieving a healthy lifestyle behaviour and potentially improved function. This study aimed to examine the relationship between a cohort of factory workers ‘readiness to change pain behaviour’ such as exercise and their ‘perceived work function’.Methods: A cross-sectional study design was used to establish the relationship between ‘readiness to change pain behaviours’ and ‘perceived work function’. The Multidimensional Pain Related Change Questionnaire 2 (MPRCQ2) was used to measure readiness to change various pain behaviours including exercise. The Work Ability Index (WAI) was used to assess ‘perceived work function’. 75 participants were recruited using convenience sampling. Correlation and multiple regression were used for statistical analysis.Results: Mean WAI, MPRCQ2 and MPRCQ2 exercise component were 41.89 (SD 5.28), 4.26 (SD 1.01) and 4.40 (SD 1.69). MPRCQ2 and MPRCQ2 exercise component were not significant predictors of WAI in factory workers (F (2, 72) = 2.17, p > 0.001). There was no significant relationship between MPRCQ2 and WAI (rs = .09, p > .05). However, there was a significant positive relationship between MPRCQ2 exercise component and WAI (rs = .23, p < .05). Conclusions: This study suggests that readiness to change pain-related exercise participation has a positive association with ‘perceived work function’. Further research should explore the causal relationship and consider strength training as a specific type of exercise.

their body is damaged or vulnerable because they were told this by health professionals or society and subsequently avoid activity and movement as a result. This is an understandable common-sense behaviour which may lead to inactivity [7].

Behaviour Change and Pain
A series of review papers were released by the Lancet journal in 2018 calling for a change in how we prevent and treat one of the most common pain areas, low back pain [8]. The main call for action was around changing the beliefs and behaviour of health professionals, patients and society to move away from a biological understanding of low back pain which creates fear and unhelpful behaviours, towards a holistic biopsychosocial understanding with person-centred care, focusing on self-management and healthy lifestyles [8, 9,10]. Other recent research has suggested a change in current musculoskeletal healthcare practice to focus on identifying lifestyle and behavioural risk factors, challenge unhelpful thoughts and beliefs that lead to unhealthy behaviours and provide support with adopting healthy lifestyle behaviours [3,4,11].
The Stages of Change Model [12] presents 6 stages of readiness to change behaviour: pre-contemplation, contemplation, preparation, action, maintenance and relapse. Described as a continuous cycle of an individual's state of readiness change. An individual's readiness to change and adopt self-management strategies is associated with improved coping [13]. A recent systematic review reported moderate-quality evidence supporting the use of behaviour change techniques to enhance exercise adherence in people with musculoskeletal pain [14].

Pain, Work Disability and Work Ability
Workplace disability is on the increase in the UK with more than 7 million working-age disabled people [15]. Pain has been identi ed as the leading cause of disability [15]. Work ability is de ned as the balance between an employee's individual resources and their work demands which can be used as a predictor of future sickness absence and work function [16]. Pain is recognised as a threat to work ability and function [17,18,19]. Poor self-e cacy [20], kinesiophobia (fear of movement) [21] and maladaptive pain behaviours [22] are all thought to reduce work ability in workers experiencing pain. Currently, factory workers have not been a focus of the research despite there being a high prevalence of musculoskeletal absence reported in the manufacturing and production industry [23].

Research Question
This study will explore the relationship between pain behaviour and function at work (work ability) in factory workers by assessing for the presence of key behavioural risk factors and readiness to change these behaviours. Focus will be given to readiness to change exercise participation.
Current research is now looking at prediction models for future sickness absence due to musculoskeletal disorders [24]. This current study will add to this body of research by identifying potential behavioural risk factors that need to be included in such prediction models for factory workers.

Page 4/19
The aim of this study was to examine the statistical relationship between a cohort of factory workers 'readiness to change pain behaviours' and their 'perceived work function'. With particular focus on readiness to change 'exercise participation'. The secondary aim was to compare this relationship in workers that were experiencing musculoskeletal pain and those that were not experiencing musculoskeletal pain.

Study Design
A correlational design was used to assess the relationship between the variables of 'readiness to change pain behaviour' and 'perceived work function' in a cohort of factory workers and establish if 'readiness to change pain behaviour' was associated with 'perceived work function'. This research used quantitative data to provide statistical analysis in the form of correlation and multiple regression. A cross-sectional study design was used for estimating the prevalence of readiness to change multiple pain behaviours, as the exercise component was considered independently.
The Multidimensional Pain Readiness to Change Questionnaire 2 (MPRCQ2) was used to measure readiness to change various pain management behaviours. This questionnaire has good validity and reliability [25,26]. The MPRCQ2 gives a score from 1-7 using 69 questions covering 9 behaviour components; use of relaxation, cognitive control, assertive communication, exercise participation, avoid resting due to pain, avoid regularly asking for assistance, task persistence, pacing and use of taught body mechanics. For the rst of two sections, 1 represents 'I am not doing this now, and am not interested in ever doing it.' and 7 represents 'I have been doing this for a long time (at least 6 months.)'. The second section inverts the scores, so 1 represents 'I am doing this now and am not interested in ever stopping' and 7 represents 'I have not done this for a long time (at least 6 months)'. An average score for each of the 9 behaviour components is calculated 1-7 and an overall score 9-63 [26]. The overall score and the 'exercise participation' behaviour component score were used in this study.
Participants reported if they were currently experiencing musculoskeletal pain or not, to allow for comparison between these two groups. Other participant characteristics measured included; gender, age range, shift pattern, job role, time in current job role, current or recent pain experience and other current health issues (see a full list of health conditions in appendix 1).

Participants
The study population consisted of 75 factory workers from a factory in the UK which manufactures home and personal care products and consisted of 98 total factory workers. Convenience sampling was used to recruit participants between September-November 2019. Participants gave consent for use of their data in group analysis. Participant characteristics were collected for population, but not individual analysis. No other personal information was collected to maintain anonymity.
The 75 participants (66 male, 9 female) modal age range was 55-64. Inclusion criteria were; aged over 18, employed within the factory for more than 3 months, English speaking. Exclusion criteria also included people registered as vulnerable. See Table 1 for participant characteristics.

Materials
The various questionnaires were inputted to Qualtrics software along with the consent form, debrief statement, participant information sheet (see appendix 1) and inclusion/exclusion criteria questions. A tablet computer was used to administer. SPSS (IBM, Armonk, NY) was used for statistical analysis.

Procedure
Invitation to take part in this research (see appendix 1) was displayed on a notice board and tables in a canteen area. Participants were able to sign up via their line manager or drop into a session during their working day, where a researcher was present.
After an initial explanation (using the participant information sheet) and a chance to ask questions, the participant was left alone in a room for 20 minutes, to complete all questionnaires. The link to the Qualtrics questionnaire was distributed using a tablet computer.
Qualtrics guided the participants through the consent process and various inclusion/exclusion criteria. If participants did not meet the criteria or consent, the questionnaire was stopped. Participants that did meet inclusion criteria and consented were taken through the WAI and MPRCQ2 questionnaires followed by a nal study debrief.
The details and answers for all participants were stored securely by Qualtrics. This data was cleansed and exported to SPSS for statistical analysis. SPSS analysis was completed on a secure passwordprotected laptop computer.

Statistical Analysis
Descriptive statistics were completed on 'perceived work function' and 'readiness to change pain behaviours'. Average scores with standard deviations and distribution of data were reported. A correlational design was chosen to assess the relationship between two variables, both 'readiness to change pain behaviour' with 'perceived work function' and 'readiness to change pain behaviour speci c to exercise' with 'perceived work function'. A correlational design was also used to assess the variables for 2 groups; participants reporting a pain experience in the past 7 days (n = 52) and those that had not reported a pain experience in the past 7 days (n = 23).
A multiple regression analysis was completed to test both 'readiness to change pain behaviour' and 'readiness to change pain behaviour speci c to exercise' as predictors of 'perceived work function'.

Results
Two potential participants were excluded because they were employed within the factory for less than 3 months and 2 potential participants did not consent after reading the participant information sheet. Giving a total of 75 participants participated in this study. Descriptive details of the participants are presented in Table 1. Table 2 presents descriptive stats for WAI, MPRCQ2 and MPRCQ2 exercise component. Mean average WAI score was 41.89 (SD 5.28) which is categorised as a 'good' level of work ability [27] with a range of 28 to 49. Mean average MPRCQ2 score was 4.26 (SD 1.01) with a range of 2.1 to 6.6 and MPRCQ2 exercise component score was 4.40 (SD 1.69) with a range of 1 to 7.
Skewness scores for WAI did not fall within the +/-1.96 Z score range to meet assumptions of normality of data, Z > 3 (see Table 2). Non-parametric testing was used.
Correlations between the variables are shown in Table 3. The results showed that there was a positive relationship between MPRCQ2 and WAI, though this was not signi cant (rs = .09, p > .05). There was a signi cant positive relationship between MPRCQ2 exercise component and WAI (rs = .23, p < .05). This shows that 'readiness to change pain behaviour speci c to exercise', has a positive association with 'perceived work function'.
Tests to see if the data met the assumption of collinearity indicated that multicollinearity was not a concern (MPRCQ2 VIF = 1.49; MPRCQ2 exercise component, VIF = 1.49). The data met the assumption of independent errors (Durbin-Watson value = 1.89).
Data were analysed using a Multiple Regression using the Enter Method. The regression equation produced a small effect size (R 2 = .06, R 2 Adj = .03), indicating that 'readiness to change pain behaviour' (when all components are measured together) was not a signi cant predictor of 'perceived work function' (F (2, 72) = 2.17, p > 0.001). 'Readiness to change pain behaviour' was not a signi cant predictor of 'perceived work function' (t = .29, df = 74, p = 0.776). 'readiness to change pain behaviour related to exercise participation' participation was not a signi cant predictor of 'perceived work function' (t = 1.53, df = 74, p = 0.131).
Data were split into 2 groups; participants reporting a pain experience in the past 7 days (n = 52) and those that hadn't reported a pain experience in the past 7 days (n = 23). For participants in pain, correlations between the variables are shown in Table 4. The results showed that there was a positive relationship between MPRCQ2 and WAI, though this was not signi cant (rs = .12, p > .05). There was a positive relationship between MPRCQ2 exercise component and WAI, though this was not signi cant (rs = .18, p > .05). For participants not in pain, correlations between the variables are shown in Table 5. The results showed that there was a positive relationship between MPRCQ2 and WAI, though this was not signi cant (rs = .10, p > .05). There was a positive relationship between MPRCQ2 exercise participation

Discussion
This study hypothesised that 'readiness to change pain behaviour' has a statistically signi cant positive relationship with 'perceived work function' in factory workers. The results indicated that this positive relationship was not statistically signi cant. Therefore, we accept the null hypothesis. A similarly unsupportive association was reported between readiness to self-manage pain and both physical and psychological functioning [13]. Although similar, the measure used was different, using the pain stages of change questionnaire (PSOCQ) instead of MPRCQ2 [13] which focuses on readiness to adopt a selfmanagement approach rather than focusing on speci c behavioural components like MPRCQ2. Patient disability and depression were the functional focus [13] opposed to 'perceived work function' as in this current study. One study [28] used the MPRCQ2 to measure readiness to change and assess the association with pain-related function (rather than 'perceived work function'). No signi cant association was reported with the MPRCQ2 or any subcomponents (including exercise) with pain-related functioning. The cohort recruited [28] had some key differences with this current study populations as they all had low back pain and were general population rather than factory workers.
In contrast to this current study's results data reported by Pitt-Catsouphes [29] in support of a statistically signi cant relationship between health behaviour changes in uencing physical health and work ability (measured by WAI). Various in uential review papers are supportive of behaviour change, exercise and self-management strategies for managing pain and improving function at work [4,8,9,10,30]. This body of research is speci c to low back pain rather than musculoskeletal pain as a whole and of these 5 studies, only NICE [30] completed a robust systematic literature review. Behaviour change and 'readiness to change pain behaviour' may be more important for people experiencing low back pain than other musculoskeletal pain. More primary research and systematic reviews are recommended focusing on musculoskeletal pain, other than low back pain.
This study hypothesised that 'readiness to change pain behaviour related to exercise participation' has a statistically signi cant positive relationship with 'perceived work function' in factory workers. The results indicated that there was a statistically signi cant positive relationship. Therefore, this hypothesis was accepted. This nding builds on the already established research base linking higher levels of physical activity and exercise as a key lifestyle behaviour change for improving pain and function associated with musculoskeletal pain [4,6,8,9,10,14,30]. 'Readiness to change' is a new area of focus and the speci c relationship between 'readiness to change pain behaviour related to exercise participation' and 'perceived work function' was not reported in these previous review papers, highlighting a new nding to add to the research base and investigate further. Applying these results practically, readiness to participate in exercise seems to be a positive behaviour and higher perceived work function seems to be a positive perception. Both of which should be encouraged.
The mean average WAI score was categorised as a 'good' level of work ability [27] and seemed generally high for this cohort of factory workers, with no workers scoring in the lowest category. This study hypothesised that 'readiness to change pain behaviour' and 'readiness to change pain behaviour related to exercise participation' are good predictors of 'perceived work function' in factory workers. The results indicate that neither 'readiness to change pain behaviour' and 'readiness to change pain behaviour related to exercise participation' are good predictors of 'perceived work function'. Therefore, we accept the null hypothesis. No other research presents 'readiness to change pain behaviour' as a predictor of 'perceived work function'. Several studies have linked exercise participation (as a behaviour, but not readiness to change) with work ability. A systematic review and meta-analysis on workplace health promotion on wellbeing and work ability was completed in 2008 [33]. Exercise was reported to increase both wellbeing and work ability of workers. A similar systematic review conducted more recently [34] reported moderate-quality evidence for exercise and lifestyle education as a workplace intervention. Research considering behaviour change speci c to exercise and workability was reported on [29] in support of a predictor relationship between 'behaviour change related to exercise participation' and 'perceived work function'. With changes in health behaviour speci c to exercise associated with a statistically signi cant improvement in physical health and work ability [29]. 'Readiness' to change was not considered [29].
Exercise as a health behaviour has been considered as part of a paradigm for health performance [35], where exercise was 1 of 5 health behaviours considered to contribute to health performance. The authors [35] discuss attitudes as a key in uencer of behaviour. They do not consider where readiness to change ts into this paradigm. The results from this current study suggest 'readiness to change pain behaviour related to exercise participation' should at least be considered. The exact nature of this association is still unclear and there does not appear to be a predictor relationship between 'readiness to change pain behaviour' and perceived work function, based on the results of this current study.
The tertiary aim of this study was to compare the statistical relationship between factory workers 'readiness to change pain behaviour' and their 'perceived work function' in workers that were experiencing musculoskeletal pain and those that were not experiencing musculoskeletal pain. The results reported no statistically signi cant relationship for either group. A previous study with similarly unsupportive ndings was a randomised controlled trial (RCT) of 66 slaughterhouse workers with upper limb pain and work disability [32]. Workers had either 10 weeks of strength exercise or ergonomic training. Strength exercise was reported as superior for preventing deterioration of work ability but not improving work ability. This was an insigni cant nding with low effect size, although this may be due to the small sample size. The researchers [32] also focused on strength training as a particular type of exercise in contrast to any exercise participation in this current study. One previous study did report con icting results [36] using a subset of 80 participants from a larger cohort study [37]

Directions For Further Research
The concept of readiness to change was rst discussed in 1983 [12], yet it has not been studied in great depth. Its importance was highlighted for adopting self-management strategies for managing chronic pain [13]. In contrast, Byrka & Kaiser [35] discuss health attitudes as a key in uencer of health behaviour rather than readiness to change.
Consensus is yet to be agreed on which individual characteristics (readiness, attitudes, motivation, beliefs and expectations) are most important for health and pain behaviour change. The results from this current study suggest 'readiness to change pain behaviour related to exercise participation' should at least be considered although the exact nature of this association is still unclear. The range of potential behaviour in uencing characteristics may be the reason that readiness to change has not been studied in great depth since it was rst proposed [12]. All are likely to play some role and further research is needed to clarify how these characteristics interact.
The statistically signi cant positive relationship between 'readiness to change pain behaviour related to exercise participation' and 'perceived work function' in factory workers reported in the results, is a new nding to add to the current research base which already recognises exercise as a positive behaviour for improving pain and function associated with musculoskeletal pain [4, 6, 8, 9, 10, 14; 30]. It may be useful to attain from future research, which speci c type of exercise is most bene cial for workers for improving pain and work function.
This research area has already received attention in the literature with strength training emerging as the most effective exercise improving pain in the workplace. No less than 4 systematic reviews [38, 39; 40, 41] have been completed in recent years to evaluate the evidence of workplace strength training for managing musculoskeletal pain amongst workers. The rst of these systematic revies [41] assessed the effectiveness of a range of workplace interventions in the prevention of upper limb pain. Strong evidence was reported for the effectiveness of workplace-based resistance training and moderate evidence for stretching programmes. A robust systematic review was completed [41], although they do not consider unpublished or grey literature and ndings may be subject to publication bias. One systematic review [39] assessed the effectiveness of a range of workplace interventions for reducing chronic musculoskeletal pain and symptoms. The authors reported consistency in results for high-intensity exercises for decreasing pain and symptoms in employees with chronic musculoskeletal pain [39]. Another systematic review [38] focused on workplace-based rehabilitation of upper limb pain in their systematic review.
Strong evidence was again reported for workplace exercise programs over ergonomic controls and training. The exercise programs were a mix of resistance training, stretching and conditioning exercise.
The most recent systematic review [40] focusing on workplace interventions to rehabilitate musculoskeletal disorders among employees with physically demanding work, report further strong evidence for the positive effect or workplace strength training for rehabilitation of workers with musculoskeletal pain. The systematic review [40] was extremely robust, with multiple reviewers involved in the quality review process.
None of the systematic reviews discussed concerning exercise and strength training for managing musculoskeletal pain in the workplace [38,39,40,41] consider 'perceived work function' and this was recommended as a sensible focus for future research [40]. In contrast, a prediction model was developed [24], for future sick leave and loss of work function due to musculoskeletal pain. They use the occupational health check questionnaire to measure predictor variables such as psychological distress, work pace and presence of musculoskeletal complaints, but this questionnaire does not consider strength training or any other exercise, physical activity, lifestyle behaviours or readiness to change. These variables are subsequently missing from the prediction model as they have not been considered. The ndings of this current study would suggest that these two research areas would bene t from coming together; readiness to change exercise behaviour (speci cally strength training) and 'perceived work function' in workers with musculoskeletal pain.

Strengths And Limitations
The regression equation produced a small effect size. A sample size of 81 was needed to achieve a larger effect size at a 95% con dence interval and a 5% precision rate [42]. The smaller sample of 75 participants used in this study could have contributed to a lack of statistical association between MPRCQ2 exercise component as a predictor and WAI as an outcome. Convenience sampling was used due to the nature of the study design and the small study population available. Convenience sampling does expose the study ndings to selection bias and reduces population validity. For example, the researchers were unable to recruit any workers that were currently absent from work. All participants were currently at work and so the results of this study can only be applied to workers in work, not those absent. The controlled workplace environment where participants completed the survey, with the researcher partially present may have unwittingly led to researcher bias which is a threat to internal validity. A lack of normality of WAI data was identi ed for this sample of workers and so non-parametric testing was used. Non-parametric testing is less powerful for detecting variability in data [43].
The WAI was used to measure 'perceived work function' due to its high validity and reliability [27,44]. It is widely used in this eld of research [29,31,32,33). One alternative for the future may be the work ability survey (WAS-R). Early research [45] suggests the WAS-R extends the measurement of WAI to also re ect organisations work demands. Validity and reliability are yet to be reported on, but the WAS-R may be considered as an alternative to WAI in future studies measuring 'perceived work function'. WAI seems to be the best current option and is a strength of this study.
MPRCQ2 was used to measure readiness to change various pain management behaviours including exercise. MPRCQ2 has good validity and reliability [25,26]. Although this research focuses on patients with bromyalgia, arthritis, spinal cord injury and amputation [26]. Patients with musculoskeletal pain such as back pain and upper limb pain have not been studied. MPRCQ2 is not as widely used as the WAI. Partly because of alternatives such as the PSOCQ [13] and partly because 'readiness to change pain behaviour' is yet to be studied in great depth. The MPRCQ2 was chosen for this study over PSOCQ due to the additional focus on speci c behavioural components like exercise. Yet, the 9 behaviour components may need an update. 'Use of taught body mechanics' [26] for example has come under scrutiny as a risk factor for musculoskeletal pain. This change in understanding was well summarised by a recent systematic review [46] reporting that lumbar spine exion when lifting was not a risk factor for low back pain onset, persistence or a differentiator for people in pain. Other important risk factors for pain, such as diet [35], sleep [11], smoking [10,11] and obesity [10,11] have also emerged but are not considered by MPRCQ2. This could be considered a limitation of the MPRCQ2 and subsequently this study. Future development and inclusion of these risk factors are recommended for any tool measuring 'readiness to change pain behaviour'.

Conclusions
This research does not provide support for an association or predictor relationship between factory workers 'readiness to change pain behaviour' and their 'perceived work function'. Previous research suggestive of any kind of relationship seems to focus on behaviour change as an intervention rather than readiness to change or focuses on different study populations outside of the workplace. Suggesting that these current results may be determined by 'readiness to change' as a concept and/or the ndings may be speci c to this study population of factory workers or this speci c sample.
This research does provide support for a positive relationship between factory workers 'readiness to change pain behaviour related to exercise participation' and their 'perceived work function'. In a practical sense, readiness to participate in exercise seems to be a positive behaviour and may be constructive for improving 'perceived work function'. Equally, the higher perceived work function seems to be a positive perception and may impact on readiness to participate in exercise. Both behaviours and perceptions should be encouraged and supported in a clinical setting.
Although this nding doesn't represent causation, this gives a good starting point for further research. To enhance the internal validity and population validity of these ndings, future research should apply a methodology which reduces use the potential for selection and researcher bias and look to recruit a larger and more varied sample.
Further research should explore the causal relationship between readiness to change exercise participation as a pain behaviour and 'perceived work function'. This research should consider motivations, attitudes, beliefs and expectations in relation to exercise participation as well as readiness to change. Workers that are absent from work should also be included to compare differences. Given the emerging evidence base for strength training for rehabilitating workers with musculoskeletal pain, this speci c type of exercise should be applied to future research exploring an association with work function.

Declarations
Ethics approval and consent to participate: Relevant ethical considerations (consent, deception, debrie ng, withdrawal, data protection, con dentiality and anonymity) were addressed and approved following the University of Derby ethical framework. All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1964, as revised in 2000. Informed consent was obtained from all patients for being included in the study. Consent for publication: Not applicable for individual data. The organisational involved did provide consent using the university of Derby ethical consent form. This is available upon request.
Availability of data and materials: The raw data and datasets used and/or analysed during the current study are available from the corresponding authors upon reasonable request.