High levels of low back disability, but not low back pain intensity, are associated with reduced physical activity: a community-based, cross-sectional study

Although clinical guidelines recommend physical activity for low back pain (LBP), our understanding of the physical activity undertaken by people with LBP is limited. We examined the amounts, intensities and types of physical activity performed by community-based women with different levels of pain and disability.


Introduction
Low back pain (LBP) is the leading cause of disability globally, with years lived with disability exceeding that of diabetes and cardiovascular disease (1). Despite the increased societal and nancial burden of LBP (2), effective treatments are limited. National and international clinical guidelines consistently recommend physical activity for the management of acute and chronic LBP (3). However, the condition of LBP itself, including the associated back disability, may signi cantly limit the physical activity performed by individuals, and in turn their involvement in family, work and social activities. This may lead to complex social and health issues, such as social isolation, psychological issues, including depression and/or anxiety, and addiction to prescribed medications, such as opioids (4).
A number of systematic reviews have examined the relationship between physical activity and LBP and disability, with a focus on determining whether reduced physical activity is a risk factor for the development and progression of LBP (5,6). In contrast, few studies have investigated the physical activity patterns of individuals with LBP. While a systematic review found individuals with high levels of low back disability performed less physical activity (7), the review did not examine the speci c amounts, types and intensities of physical activity performed. Moreover, recent cross-sectional studies found con icting results when examining the association between physical activity and LBP and disability (8, 9). While Sokunbi et al (2016) found no association between pain and disability and the amount of physical activity performed by nurses with chronic LBP(9), Ezeukwu et al (2019) found low levels of physical activity in patients presenting to a physiotherapy clinic with chronic, non-speci c LBP and disability (8). Thus, we are lacking a comprehensive understanding of the amount and nature of physical activity performed by community-based individuals with LBP and disability.
This study aimed to examine the amounts, intensities and types of physical activity performed by community-based individuals with different levels of LBP intensity and disability. We hypothesised that individuals with high levels of pain and disability would perform less physical activity of all types and intensities. An understanding of the physical activity patterns of individuals with LBP has the potential to provide important information about the burden of LBP and disability, the particular subgroups of individuals with LBP that perform reduced physical activity, as well as the types and intensities of physical activity that may be important to target in improving physical activity and reducing low back disability.

Study sample
Participants were recruited from a parent study of 1,423 community-dwelling women aged between 18 and 75 years. The sample was randomly recruited from a database that was derived from the Victorian Electoral Roll from 2002 to 2003. Details of the parent study methodology has been previously reported (10). Of the 1423 participants, 754 were willing to be contacted to participate in further research and 542 consented to take part in this study upon re-contact. Written informed consent was obtained from all participants and the research was approved by Monash University Human Research and Ethics Committee (CF07/2961-2006000113).

Demographics
Age, weight and height were self-reported by participants and body mass index (BMI; kg/m 2 ) was calculated. Participants were considered working if they had paid jobs, or were involved in volunteer work, course work and unpaid work outside the home. Depression was assessed based on the following question "Have you been diagnosed with depression in the past 2 years?".

Low back pain intensity and disability
The Chronic Pain Grade (CPG) questionnaire was used to assess LBP intensity and disability. It is a selfadministrated instrument that has been shown to be valid and reliable to assess chronic pain in community-based research (11). It consists of seven items that measure back pain intensity and degree of disability. Each item has an 11-point Likert scale that is summed to give a score ranging from 0 to 100.
High levels of pain intensity and disability were de ned as scores equal to or more than 50 and no or low pain intensity and disability were de ned as scores less than 50 according to the CPG (11).

Physical activity
The long version of the International Physical Activity Questionnaire (IPAQ) was used to assess physical activity. The IPAQ consists of 31 items, which capture different domains of physical activity over the last 7 days and include physical activity associated with leisure-time, work, active transportation, and domestic and garden activities. Physical activity was also categorised according to intensity as vigorous, moderate and walking activities ( Table 1). Each physical activity domain was reported in terms of the number of days and hours/minutes per day that the activity was performed in the past 7 days. The sum of hours per week for each activity was weighed to an assigned energy expenditure equivalent, the metabolic equivalent task (MET), based on the intensity of this physical activity domain (MET x hours per day x number of days per week) (12,13). Based on the IPAQ guidelines, any participant reporting over 16 hours of physical activity per day or more than 7 days a week for activity was considered an outlier and omitted from the analysis (14). To determine the association between back pain and disability and the physical activity domains of domestic and garden activity, walking, moderate physical activity and total physical activity (MET, hours/week), estimated adjusted medians were obtained using quantile regression analysis and the 95% con dence intervals (CI) were obtained by bootstrapping with 1000 replicates from the quantile regression. There was insu cient participation in work-related activity, active transportation, vigorous and leisure physical activities and thus these activities were dichotomised as individuals participating or not. Binary logistic regression was used to examine the association between LBP and disability and these physical activity domains.
With regards to the data for sitting time, the discreteness of the data made quantile regression unreliable as a continuous endpoint distribution is required for this technique to be valid. We adopted a square root transformation of the sitting time endpoint to achieve approximate normality (up to rounding). Linear regression was then used to examine the association between LBP and disability and the transformed endpoint; the estimated marginal medians and 95% CIs were obtained by back-transformation of the adjusted means.
Multivariable analyses were adjusted for potential confounders in model 1; age, BMI and depression, in model 2 we added working status as it particularly showed in the univariate analysis that it was signi cantly different between women with high levels of pain intensity and disability and those with no or low levels of pain intensity and disability. Statistical signi cance was denoted by a p value of less than 0.05. Data were analysed using Stata 12.0 SE (StataCorp).

Results
Of the 542 participants who were recruited for this study, 506 (93.4%) returned the study questionnaire. 463 participants were included in the analyses after excluding 43, as they were considered outliers (as per the IPAQ guidelines) (14). However, there were no differences between the outliers and those included in the nal analyses in terms of age, BMI, depression, pain intensity and disability (p = 0.07-0.63), with the exception of a higher percentage of women employed in the outlier group (39 (90.7%) vs. 278 (61.2%), p < 0.001).
The characteristics of participants with no/low and high LBP intensity are presented in Table 2. High intensity back pain was reported in 71 (15.3%) women. Compared to women with no or low intensity back pain, those with high intensity back pain had a greater BMI (mean(SD) = 29.5 (5.3) vs. 27.0 (5.7), p = 0.001), were more likely to be depressed (16 (22.5%) vs. 46 (11.7%), p = 0.01) and less likely to be employed (33 (47.1%) vs. 245 (62.6%), p = 0.02). There were no differences between the pain intensity groups in their physical activity and sitting time, with the exception that fewer women with high pain intensity reported doing physical activity at work (16 (23.5%) vs. 142 (36.6%), p = 0.04).  Fig. 1(A) and (B)), these results remained unchanged.  Fig. 2).

Associations Between Low Back Disability And Physical Activity
The associations between physical activity (MET; hours/week) and low back disability are shown in Table 3 and Fig. 1 (Table 4, model 2; Fig. 2).

Discussion
In this study we found that community-based women with high levels of back disability, not back pain intensity, performed less overall physical activity compared to those with no or low levels of back pain and disability. Moreover, those with high levels of back disability differed in the intensities and types of physical activity they reported, performing less moderate and vigorous activities, as well as less discretionary physical activity, including leisure, domestic and gardening activities. These ndings not only identify an important subgroup of individuals in the community that have high levels of back disability in combination with reduced physical activity, but also highlight the need for further investigation to examine the bi-directional nature of this relationship and identify potential targets for increasing physical activity and reducing disabling LBP.
We found that women with high levels of back disability performed only 50% of the physical activity of those with no or low levels of back disability. In addition, only 46% of the women with high disability performed leisure activity and as few as 8% were involved with vigorous physical activity. The amount of moderate and vigorous physical activity reported by these women was lower than that recommended by the World Health Organisation Guidelines, which suggest that individuals should perform 20 METS (hours/week) of moderate or vigorous physical activity per week, or an equivalent combination to gain health bene ts (15). While it may be unrealistic for some women with high back disability to achieve these recommended activity levels, it is important to recognise that this may have major implications for their general health, including the development and management of co-morbidities, such as obesity and psychological conditions. Our data support this theory, as the women in our cohort with high disability had a mean BMI of 30 kg/m 2 with 50% classi ed as obese, and almost 16% reported depression.
Understanding this burden of low back disability with regard to physical activity levels is important, given low back disability is the leading cause of disability globally (1), the majority of the socioeconomic burden associated with LBP results from those with high levels of disability (16), and obesity is in epidemic proportions (17,18). Overall, these ndings highlight the poor health and lack of daily physical activity performed by community-based women with high back disability.
In contrast to disability, we found no differences in physical activity in those with different levels of back pain. These results are novel, as previous systematic reviews and epidemiological studies have largely examined whether physical activity is a risk factor for LBP, rather than characterising the physical activity performed by individuals with different levels of pain and disability (5,6,19). Moreover, few epidemiological studies have examined the relationship between both LBP and disability and physical activity in community-based individuals (7). The current ndings suggest that an individual's physical activity is not linked to their level of pain, but rather to their level of function. This may be explained by current theoretical models, such as the Fear-avoidance model, which proposes that if an individual believes that their LBP is an indicator of severe tissue damage and certain activities will cause reinjury, they will display avoidance behaviours that will ultimately result in less physical activity and high levels of disability (20,21). Conversely, individuals who believe that they do not have a serious condition and movement will assist their recovery, will move and be physically active despite their pain levels (20,(22)(23)(24). Overall, the current results highlight the importance of targeting individuals with high levels of disability, rather than high levels of pain, when improving physical activity levels for LBP.
This study also found that individuals with high levels of disability differed in both the intensities and types of physical activity they performed compared to those with no or low levels of disability. Women who were highly disabled by their back pain reported less involvement in moderate and vigorous activity, as well as discretionary activities, including leisure, domestic and garden activities. These ndings have signi cant implications. Given there is growing evidence that moderate intensity exercise improves pain sensitivity for musculoskeletal pain (25), our ndings suggest that individuals with high disability may not be receiving the potential bene ts in pain relief associated with this intensity of activity. Individuals who are not participating in recommended levels of moderate or vigorous exercise will also not receive the important health bene ts of exercise, from improved mood and mental health, weight regulation and increased energy levels. Furthermore, the women who were undertaking reduced levels of discretional activities, including leisure, domestic and gardening activities, may experience a reduction in their psychosocial wellbeing as a result of not being able to ful ll their familial roles, which in turn may further impact on their back disability. While further longitudinal investigation is needed, these ndings suggest that there is a need to focus on speci c intensities and types of physical activity to improve the physical activity levels of individuals with high levels of disability.
There are some limitations in our study. Due to the cross-sectional design, a causal relationship between low back disability and physical activity could not be determined. While we had a modest sample size with a relatively small number of participants with high intensity LBP and disability, it was comparable to the prevalence of LBP and disability in the Australian population (26). Our study participants were recruited from a parent study of community-based women, who were originally recruited through random sampling from the state electoral roll, and without having LBP as a recruitment criterion. Our sample had comparable baseline demographics to those of the parent study (age mean (SD): 50 (14.4); BMI mean (SD): 27.8 (6.5) kg/m 2 ) (10). While the IPAQ is a self-reported questionnaire, which makes it susceptible to recall bias and participants may over-report physical activity (27), it has been shown to have good reliability and validity when compared to objective physical activity measures (13).
There are a number of strengths to our study. We used validated questionnaires to measure LBP intensity and disability. Our data allowed us to examine different levels of pain and disability (ie no/low versus high) and to investigate both participation and METs for different types and intensities of physical activity. Given we recruited our community-based women from the Victorian state electoral roll and without reference to the presence of back pain, it suggests that our results have high generalizability to the general community.

Conclusions
Our study found that high levels of low back disability, not back pain intensity, were associated with reduced physical activity, including less total activity, moderate and vigorous activities and discretionary types of activity. These ndings highlight the excessive burden of high levels of back disability with respect to physical activity and the speci c subgroups of individuals with low back disability who participate in less physical activity. While further investigation is needed to examine this likely bidirectional relationship, these data highlight the potential amounts, intensity and types of physical activity that need to be investigated further to increase physical activity and reduce disability in community-based women with LBP.  Association between levels of low back pain intensity and disability (no/low versus high) and participation in physical activity (%), adjusted for age, BMI, depression and working status.