Abstract
Background Back pain beliefs (BPBs) are an important modifiable factor related to disability associated with low back pain (LBP). Back pain beliefs have not been characterized in baby boomers, a group at risk for decreased activity levels and reduced productivity.
Objective The aims of this study were: (1) to identify factors related to BPBs and (2) to evaluate the association between LBP disability and beliefs.
Design A cross-sectional survey of community-dwelling baby boomers (born 1946–1964) was conducted.
Methods Nine hundred fifty-eight baby boomers (mean age=56.2 years) participating in the Busselton Healthy Aging Study provided their history of LBP, BPBs, LBP behaviors related to care seeking (taking medication, seeking professional help) and activity modification (missing work, interference with normal activities, interference with recreational activities), LBP-related disability, and additional covariates with known associations with BPBs. Regression analyses were used to: (1) identify factors associated with more positive beliefs and (2) test the association between more positive BPBs and lower LBP disability, independent of other correlates of BPBs.
Results More positive BPBs were associated with younger age, better mental well-being, and higher income, whereas more negative BPBs were associated with receiving sickness or disability benefits and the experience of LBP in the previous month. In participants who reported experiencing LBP within the previous month, more positive BPBs were associated with lower disability scale scores and a decreased probability of interference with usual activities, independent of pain intensity, age, mental well-being, income, and employment status.
Limitations Cross-sectional analysis limits assessment of causality.
Conclusions Poorer BPBs were associated with greater disability. Characterization of the relationships between BPBs and LBP-associated behaviors and disability in baby boomers can assist in developing interventions to improve activity participation and productivity, potentially reducing the burden of LBP in this age group.
Globally, low back pain (LBP) has been the leading cause of disability for at least the last decade.1,2 Contemporary thinking suggests disability with LBP is related to a complex interaction of biopsychosocial factors,3–5 with cognitive factors such as beliefs identified as an important determinant of both disability and the experience of pain.
There is growing interest in the role of beliefs regarding LBP, as beliefs and expectations can modulate (facilitate or inhibit) pain responses in the brain6,7 and are central drivers of behavior8,9 and recovery.10 Beliefs and expectations are processed through interactions in multiple areas of the brain, including the dorsolateral prefrontal cortex, orbitofrontal cortex, insula, cingulate, and thalamus.6
There are multiple belief constructs potentially important for consideration in LBP. These constructs include beliefs related to: (1) the cause, nature, and meaning of the disorder; (2) the future course of the disorder; (3) the consequence, effects, and impact of the disorder; and (4) treatment, “cure,” and control.8,11,12 Beliefs in each of these constructs have the potential to positively or negatively influence pain, disability, and coping. Beliefs can be assessed at a clinical level by direct questioning or through the use of validated questionnaires. Perhaps the most recognized beliefs in musculoskeletal pain relate to fear avoidance or kinesiophobia.13–15 Higher levels of fear-avoidance beliefs are a negative prognostic factor for LBP outcomes16 and are an important consideration in the management of LBP.17,18
Another aspect of beliefs related to LBP is the future course and inevitability of back pain,19 commonly referred to as back pain beliefs (BPBs). The Back Beliefs Questionnaire (BBQ) was developed to assess these beliefs.19 The BBQ contains items such as “Back trouble may mean you end up in a wheelchair” and “Later in life, back trouble gets progressively worse.” Back pain beliefs have been shown to be influenced by the individual's sex, culture, familial environment, professional training, education level, mental health, perceived general health, work absenteeism history, pain experience, and LBP activity limitation.20–25 Additionally, increased interest in the role of BPBs has occurred following findings that more negative BPBs are linked to higher levels of disability,22,24–27 work absenteeism, and reduced productivity.19,28,29 There is also some evidence that interventions to improve BPBs may positively influence these outcomes,23,30 although this finding is not universal.28 Fear-related beliefs and BPBs have been identified as independent constructs in terms of influence on LBP prognosis.27
The baby boomer generation refers to individuals born between 1946 and 1964 following increased birth rates after World War II, particularly in the so-called Western countries of Europe, North America, and Australia and New Zealand. Key strategies in meeting the societal needs of the baby boomer generation now and into the future include promotion of active lifestyles and community participation due to the positive effects these strategies have on health and well-being.31–33 As a specific example of community participation, the Australian government's strategy has focused on increasing workforce participation of baby boomers as part of an overall strategy to increase productivity.34 Given the high burden of LBP and the links among BPBs, activity, and productivity, addressing BPBs may be a useful strategy to improve activity and productivity in baby boomers. Before this strategy can be implemented, a characterization of baby boomer BPBs is needed. Characterization of this nature may assist in identifying vulnerable or at-risk individuals for targeted interventions.24,35
Characterization of BPBs in baby boomers (current age=49–67 years) is currently limited, with most studies assessing BPBs in younger cohorts, in cohorts with wide age ranges without consideration of specific age ranges,27 or in a single-sex cohort.25 Thus, the first aim of this study was to evaluate the BPBs (related to the future course and inevitability of back pain) of baby boomers and identify factors related to more negative beliefs of this nature. Consistent with previous findings,24 we hypothesized that more negative BPBs would be related to male sex, older age, higher body mass index, smoking, poorer mental well-being, lower income, not being employed, and experiencing LBP.
Understanding the relationship between BPBs and LBP disability in baby boomers with LBP could provide additional insight into keeping this age group active and productive.31–33 Therefore, the second aim of the study was to evaluate the association between BPBs and LBP disability, controlling for potential confounding variables. We hypothesized that more negative BPBs would be related to increased levels of LBP disability, consistent with associations between more negative BPBs and accumulated LBP-related activity modification behaviors we have previously found in 17-year-olds.24
Method
Participants
A cross-sectional evaluation of BPBs was performed using data from baby boomers in the Busselton Healthy Aging Study (BHAS).36 The BHAS is part of the broader Busselton Health Study that has successfully conducted repeated cross-sectional surveys of adults in the shire of Busselton, Western Australia, from 1966 to 2009, with participation rates ranging from 64% to 91%.37 The current survey targeted all noninstitutionalized adults born from 1946 to 1964 who resided in the shire and were listed on the electoral roll. Order of invitation to participate was randomized, with recruitment efforts focused on sequential, random 10% sample draws. Participants underwent a comprehensive clinical assessment and completed a self-administered questionnaire encompassing a wide range of health measures and conditions.36
Data from the first 1,004 participants who completed the survey between May 2010 and July 2011 were used in this analysis. This procedure was based on previous consideration of the power of a data set used for a similar study of 17-year-olds24 and the fact that these individuals represented a random sample.36 Furthermore, initial analysis of health risk factors of the first 300 participants demonstrated prevalence rates equal to those reported for the general Australian population, supporting the representativeness of the sample.36 By mid-2015, it is expected that data collection will be completed for more than 4,300 baby boomers (or 64% of the 6,690 individuals listed on the 2010 electoral roll), with longitudinal follow-up studies of this cohort planned to commence every 5 years.
Back Pain Beliefs Questionnaire
The BBQ is a 14-item questionnaire, with 9 items ranked on a 5-point scale and used to calculate a final score from 9 to 45.8,19 If there was one missing response, a score was imputed from the remaining responses. No score was calculated for individuals missing 2 or more responses. Higher scores are equated with more positive beliefs. There is no established cutoff score for what may equate to “good” or “poor” BPBs. However, a mean difference of 2 points can be considered clinically relevant,23 as a change of this amount is associated with reduced LBP work absenteeism and a reduction in workers' compensation costs associated with LBP.38 The BBQ has been reported to have good test-retest reliability (intraclass correlation coefficient=.87) and good internal consistency (Cronbach α=.7).19 These results have been supported by more recent findings.20
LBP Prevalence
The prevalence of LBP was based on questions modified from the Nordic questionnaire.24,39 Participants were asked, “Have you ever had low back pain?” This question was accompanied by a diagram with the low back indicated by a shaded area from the level of the 12th thoracic vertebra to the inferior gluteal folds. Participants also were asked, “Has your low back been painful at any time in the last month?” From these 2 questions, 3 mutually exclusive groups were formulated: (1) no LBP ever, (2) past history of LBP but no LBP in the previous month, and (3) LBP in the previous month. If participants indicated having had LBP in the previous month, they were asked to rate the intensity of pain experienced in the last week on a numerical rating scale from 0 to 10.
Disability and LBP-Related Behaviors
The Oswestry Disability Index (ODI)40 was included as a specific measure of low back–related disability. The ODI is one of the most commonly used disability outcome measures in LBP and has been recommended for assessing function in people with LBP based on its established validity and reliability.41 Scores range from 0 (no disability) to 100 (maximum disability).
Low back pain–related care-seeking behaviors were ascertained from participants' responses to 2 questions24,39,42: (1) “Do you usually seek health professional advice or treatment for your low back pain?” and (2) “Do you usually take medication to relieve your low back pain?” Low back pain–related activity modification behaviors were ascertained from their responses to 3 questions24,39: (1) “Do you usually miss work due to your low back pain?” (2) “Does your low back pain usually interfere with your normal activities?” and (3) “Does your low back pain usually interfere with recreational physical activities (eg, sports, walking, cycling)?” Responses to these questions were either “yes” or “no.”
BPB Correlates
Factors investigated as correlates of BPB were selected based on previously determined constructs related to BPB23–25 and availability in the BHAS data set. We identified associations among male sex, higher body mass index, and lower family income and more negative BPB in young adults and considered them important constructs to include in the current analysis.24 We included age as a correlate given the range available in the sample of baby boomers. The Mental Health scale of the 12-Item Short-Form Health Survey (SF-12), version 2,43 was used as a measure of mental well-being, which has previously been correlated with BPBs.24 The scale was normalized to have a mean of 50 and a standard deviation of 10 based on 1998 general US population scores.44 Smoking also has been identified as a correlate of poorer BPBs in young adults24 and might relate to underlying constructs of beliefs and negative health behaviors. Workforce participation (employment status) was included given the relationships between beliefs and work productivity19,28,29 coupled with the notion of the positive effect of work on overall well-being and outlook.45,46
Data Analysis
Descriptive statistics were calculated for all measures used in the study. For the first aim (identifying factors associated with more negative BPBs), a series of univariable linear regressions with BBQ scores as the dependent variable were used to quantify univariable associations between the BBQ scores and each potential correlate. The independent association of each factor with BPBs was estimated by including all variables displaying evidence of an association with BPBs at P<.01 as independent variables in a multivariable linear regression model with BBQ scores as the dependent variable.
For the second aim (evaluating the association between BPBs and LBP disability), analyses were conducted using data only from participants who reported experiencing LBP in the previous month. A series of unadjusted (6) and adjusted (6) regressions were performed with ODI and each measure of LBP behavior as the dependent variable, with adjusted models including all variables identified as independent correlates of BPB from the previous analyses, plus pain intensity in the last week, as potential confounders. As ODI scores were non-normally distributed and displayed positive skew, quantile regression was used to estimate change in the median ODI score, rather than change in the mean, as a function of change in BBQ scores. Logistic regression was used to estimate the change in odds for each of the 5 LBP behaviors as a function of change in BBQ scores. Logistic regression for work absenteeism was performed in the subsample of employed participants. Standard diagnostics were used to ensure linearity of effects and absence of influential outliers in all models. Statistical analysis was performed with Stata/IC version 12.1 for Windows (StataCorp LP, College Station, Texas).
Role of the Funding Source
The Busselton Healthy Ageing Study is supported by grants from the Government of Western Australia (Department of Commerce and Industry, Department of Health) and the City of Busselton and from private donations to the Busselton Population Medical Research Institute. Fellowships provided funding for Dr Beales (National Health and Medical Research Council of Australia), Dr Smith (Curtin University), and Professor Straker (National Health and Medical Research Council of Australia).
Results
Data collected from the first 1,004 participants were considered for this study. One thousand participants had valid BBQ scores (4 had missing BBQ scores). A further 42 participants were missing 1 (40 participants) or 2 (2 participants) of the other variables considered in this study and were excluded from analyses, leaving 958 cases for analysis. Descriptive statistics for all variables considered in the study are provided in Table 1. The mean age of these 958 participants (53.6% female) was 56.1 years (SD=5.4), and the mean BBQ score was 29.1 (SD=6.4). Half of the participants reported experiencing LBP in the previous month (486 of 958 [50.7%]), and the median ODI score for these participants was 12 (interquartile range=6–20). The proportion of participants with LBP in the previous month reporting each of the LBP-related behaviors is shown in Table 1, with “seeking professional care” the most common, reported by 270 (55.6%) of the 486 participants, followed by “interference with physical activities,” reported by 225 (46.3%) of the 486 participants.
Descriptive Statistics for All Study Measuresa
Table 2 displays the univariable associations between BBQ scores and each factor considered as a potential correlate of BPB in this study. Mean (standard deviation) BBQ scores in categories of continuous variables are presented for additional interpretative purposes only, with continuous measures used in regression analyses. Statistically significant, univariable associations for BBQ scores were identified for age, body mass index, smoking status, mental well-being, income, employment, and LBP status, but not sex.
Univariable and Adjusted Associations With Back Beliefs Questionnaire (BBQ) Scoresa
In the adjusted model, independent and significant associations were found between BBQ scores and age, SF-12 mental component summary (MCS) score, income, employment, and LBP status. Body mass index and smoking were no longer significant after accounting for other variables. Lower BBQ scores were associated with increasing age, with an increase in age of 5 years associated with an estimated decrease in BBQ scores of −0.5 points (95% confidence interval [95% CI]=−0.9, −0.6; P<.001). Higher BBQ scores were associated with better mental well-being, with an increase of 10 points in SF-12 MCS score associated with an estimated increase in BBQ scores of 1.1 points (95% CI=0.6, 1.5; P<.001). Higher BBQ scores were associated with higher income, with each increment of $20,000 associated with an estimated increase in BBQ scores of 0.5 points (95% CI=0.2, 0.8; P=.001). Lower BBQ scores were found in participants on sickness or disability benefits, with the estimated difference between this group and working participants being −3.3 (95% CI=−5.5, −1.2; P=.002).
Finally, higher BBQ scores were found in participants with a history of LBP but not reporting LBP in the previous month, with the 1.4-point difference between this group and those with LBP in the previous month being statistically significant (95% CI=0.5, 2.2; P=.002). However, the 1.1-point difference between this group and those with no LBP ever was not statistically significant (95% CI=−0.1, 2.4; P=.078). The final adjusted R2 for this model was .078, with the majority of this explained variance in BBQ scores due to SF-12 MCS scores.
Table 3 presents adjusted and unadjusted associations between BBQ scores and LBP disability (ODI scores) in the subgroup of participants reporting LBP in the previous month. In the unadjusted model, for each 1-point increase in BBQ score, the median ODI score was estimated to decrease by 0.7 points, and although this association was attenuated by adjustment for covariates (current pain level, age, SF-12 MCS score, income, and employment), it remained significant in the multivariable model, with a 1-point increase in BBQ score estimated to be associated with a decrease of 0.4 points in the median ODI score (95% CI=−0.5, −0.2; P<.001). To provide context, from this adjusted model, there is an estimated difference in median ODI score of 4.4 points between people with a BBQ score of 23 points (approximately 1 standard deviation below the total sample mean) and people with a score of BBQ score of 35 points (approximately 1 standard deviation above the total sample mean).
Unadjusted and Adjusted Associations Between the Back Beliefs Questionnaire (BBQ) Score and Measures of Disability and Low Back Pain–Related Behaviorsa
The adjusted and unadjusted associations between BBQ scores and LBP-related behaviors in the subgroup of participants reporting LBP in the previous month are presented in Table 3. In unadjusted models, lower BBQ scores were associated with medication use, interference with normal and physical activities, and work absenteeism (in working participants). After adjustment for covariates (current pain level, age, SF-12 MCS score, income, and employment), BBQ scores remained statistically significantly associated with interference with normal activities only, with the major confounding factor in other models being current pain level. A 1-point increase in BBQ score was estimated to be associated with a 5% reduction in the odds of interference with normal activities (95% CI=2%, 9%; P=.001). This reduction corresponds to a 0.13 decrease in the predicted probability of interference with normal activities from 0.41 for people with a BBQ score of 23 (approximately 1 standard deviation below the total sample mean) to 0.30 for people with a BBQ score of 35 (approximately 1 standard deviation above the total sample mean).
Discussion
Low back pain had been experienced by the majority of the BHAS cohort and resulted in impacts on care seeking and activity levels (Tab. 1). These findings are consistent with international findings for adolescents and adults.1,47,48 The results of this study provide the first detailed characterization of BPB in baby boomers. The overall mean BBQ score for the BHAS cohort was 29.1. The score is similar to that reported for Australian 17-year-olds in the Raine Study cohort (29.3)24 and female adults from a wide age range (30.7).25 In contrast, the overall BBQ score of the cohort was higher than several large general population cohort studies of adults in Australia (26.5 and 26.3),30 Canada (26.1),20 and the United States (27.7),21 indicating more positive BPBs in the baby boomers of the BHAS. These differences may reflect regional or cultural differences in BPBs or differences in the age range of the cohorts.22,49,50
BPB Correlates
In line with the first aim of this study, this study demonstrated independent and significant associations of more negative BPBs (ie, lower BBQ scores) with increasing age, poorer mental health status (lower SF-12 MCS score), lower income, lack of workforce participation due to sickness or disability, and experiencing LBP in the previous month. A strong association between more negative BPBs and poorer mental health status in this group of baby boomers was similar to the association we found in 17-year-olds.24 This finding is consistent with the strong relationship between LBP and poorer mental well-being (negative psychological affect) considered in the contemporary biopsychosocial understanding of pain.3–5,51 The finding of more negative BPBs with lower socioeconomic status (measured via income) is consistent with previous findings21,24 and is potentially related to broader associations between income and health.52 More negative BPBs in the group not participating in the workforce due to sickness or disability are potentially consistent with more negative BPBs leading to reduced work productivity.19,28,29 Workforce participation is an important factor in promoting health and well-being,45,46 and beliefs may be part of an important cognitive interplay in this relationship, a premise worthy of further investigation.
Baby boomers who had experienced LBP but not in the previous month had more positive BPBs than the other participants (Tab. 3). We have documented a similar finding in 17-year-olds,24 and other authors have documented this finding in adults.22,26,28 This finding, however, is not universal21,25 and may depend on pain intensity.20 It appears that recovery from an episode of LBP may improve BPBs.53 Overall, the relationships found in the current study support the importance of considering BPBs in baby boomers.
BPBs, Disability, and LBP-Related Behaviors
In line with the second aim, there was a significant association between more negative BPBs and higher levels of disability (higher ODI score) in baby boomers. This association remained significant after adjusting for age, SF-12 MCS score, income, employment, and current pain level, and it is consistent with findings from other age groups22,24–27 and highlights that BPBs are an important construct as a determinant of back pain–related disability in baby boomers. However, as the association between BPB and disability was weak, there are clearly other factors important to the reduction of LBP disability, and BPBs should be seen as just one important factor in preventing and managing LBP disability effectively.
The activity modification behaviors of “interfering with normal activities” and “interfering with physical activities” are similar conceptually to individual items in the ODI. Interestingly, BPBs were significantly associated with interference in normal activity but not physical activities after adjustment for covariates (Tab. 3). This finding could have been due to the slightly weaker observed relationship and the limitations of self-reported physical activity. Investigation of BPBs with objective measures of physical activity may enhance understanding of this finding. Care-seeking behaviors and the “missing work” question are not conceptually related to specific items in the ODI. Although the various disability-related behaviors were correlated, the use of these questions in addition to the ODI in the current study was potentially useful, as seeking care and missing work arguably add to the economic burden of LBP and are specific concerns for the societal impact of aging baby boomers.31–33
Implications
Given the cross-sectional nature of this study, the potential direction of the relationships between BPBs and the factors we have investigated cannot be ascertained. Such relationships are likely to be complex and bidirectional. Although our similar findings in 17-year-olds24 suggest that BPBs are well established at least by the age of 17 years, they are likely to be molded through one's life course. Further longitudinal research is needed to track how BPBs change over time.
The observed relationship between BPBs and disability underscores the importance of considering BPBs to address the societal burden related to LBP. The BBQ has been used to track changes in BPBs following targeted education of the general population,30 specific occupational groups,23,54 and health care practitioners.49,50 These changes in BPBs have been associated with positive shifts in LBP-related disability behaviors.23,30,54 However, not all results have been positive,35 and some forms of education have resulted in only a short-term shift in BBQ scores, highlighting the difficulty in changing BPBs.55 The improved understanding of the BPBs of baby boomers and factors that influence these beliefs suggests 2 approaches to intervention may be required.
At a population level, the results from this study of baby boomers, where more negative BPBs were significantly associated with higher levels of disability even after adjustment of pain levels, age, mental health status, income, and workforce participation, suggest that universal interventions as previously conducted19 should be continued. However, the findings also supported refining interventions to target groups of individuals with LBP, as participants with LBP in the previous month had more negative BPBs. Further research should explore optimal methods for identifying at-risk groups and tailoring interventions to these groups. Consideration of other factors such as motivation, readiness to change, opportunity to change, and ability to change also may be important.35 Due to the association between BPBs and activity modification in this study, BPBs also should be considered for public health initiatives directed at keeping baby boomer populations active to prevent chronic diseases.56
Back pain beliefs also are important to consider at a clinical level when addressing the burden of LBP for a particular individual. Interventions that target beliefs within the context of the biopsychosocial profile of an individual demonstrate better outcomes.57,58 Screening questionnaires such as the BBQ, the STarTBack Screening Tool,59 and the Örebro Musculoskeletal Pain Questionnaire60,61 can be incorporated into clinical practice to identify people with negative BPBs and high levels of distress. The approach of using screening tools can direct clinicians to further investigate both the nature of the beliefs (ie, the cause and future course of their disorder) and the biopsychosocial factors underlying these beliefs in order to facilitate targeted interventions.
In conclusion, this study has shown that various factors are associated with BPBs and that BPBs are associated with disability in baby boomers with LBP. Beliefs should be targeted at both an individual clinical level and a population level in attempts to reduce the burden of LBP. The characterization of BPBs in baby boomers provided in this study should assist in refinement of such interventions.
Footnotes
All authors provided concept/idea/research design. Dr Beales, Dr Smith, Dr O'Sullivan, and Professor Straker provided writing. Dr O'Sullivan, Dr Hunter, and Professor Straker provided data collection. Dr Beales, Dr Smith, and Dr O'Sullivan provided data analysis. Dr Beales, Dr O'Sullivan, Dr Hunter, and Professor Straker provided project management. Dr Hunter and Professor Straker provided fund procurement. Dr Hunter provided participants, facilities/equipment, institutional liaisons, and consultation (including review of the manuscript before submission). The authors thank the community of Busselton for their ongoing support and participation.
Ethics approval for this study was granted from Curtin University Human Research Ethics Committee and the West Australian Department of Health Ethics Committee in accordance with the Australian National Health and Medical Research Council National Statement on Ethical Conduct in Human Research. The study also received ethics approval from the University of Western Australia Human Research Ethics Committee (number RA/4/1/2203).
The Busselton Healthy Ageing Study is supported by grants from the Government of Western Australia (Department of Commerce and Industry, Department of Health) and the City of Busselton and from private donations to the Busselton Population Medical Research Institute. Fellowships provided funding for Dr Beales (National Health and Medical Research Council of Australia), Dr Smith (Curtin University), and Professor Straker (National Health and Medical Research Council of Australia).
- Received February 25, 2014.
- Accepted September 13, 2014.
- © 2015 American Physical Therapy Association
References
Issue highlights
