lable at ScienceDirect

Manual Therapy 15 (2010) 382e387

Contents lists avai

Manual Therapy

journal homepage: www.elsevier .com/math

Original article

Pain biology education and exercise classes compared to pain biology educationalone for individuals with chronic low back pain: A pilot randomisedcontrolled trial

Cormac G. Ryan*, Heather G. Gray, Mary Newton, Malcolm H. GranatSchool of Health, Glasgow Caledonian University, Scotland G4 0BA, UK

a r t i c l e i n f o

Article history:Received 14 July 2009Received in revised form22 February 2010Accepted 2 March 2010

Keywords:Low back painEducationExerciseRandomised controlled trial

* Corresponding author. Tel.: þ44 141 331 3327; faE-mail address: cormac.ryan@gcal.ac.uk (C.G. Ryan

1356-689X/$ e see front matter � 2010 Elsevier Ltd.doi:10.1016/j.math.2010.03.003

a b s t r a c t

The aim of this single-blind pilot RCT was to investigate the effect of pain biology education and groupexercise classes compared to pain biology education alone for individuals with chronic low back pain(CLBP). Participants with CLBP were randomised to a pain biology education and group exercise classesgroup (EDEX) [n ¼ 20] or a pain biology education only group (ED) [n ¼ 18]. The primary outcome waspain (0e100 numerical rating scale), and self-reported function assessed using the Roland MorrisDisability Questionnaire, measured at pre-intervention, post-intervention and three month follow up.Secondary outcome measures were pain self-efficacy, pain related fear, physical performance testing andfree-living activity monitoring. Using a linear mixed model analysis, there was a statistically significantinteraction effect between time and intervention for both pain (F[2,49] ¼ 3.975, p < 0.05) and pain self-efficacy (F[2,51] ¼ 4.011, p < 0.05) with more favourable results for the ED group. The effects levelled offat the three month follow up point. In the short term, pain biology education alone was more effectivefor pain and pain self-efficacy than a combination of pain biology education and group exercise classes.This pilot study highlights the need to investigate the combined effects of different interventions.

� 2010 Elsevier Ltd. All rights reserved.

1. Introduction

Chronic low back pain (CLBP) is a complex condition for whichmany different interventions exists. A number of different treat-ments have been shown to be effective including; education, exer-cise, manual therapy, multidisciplinary and cognitive behaviouralinterventions (Hilde and Bo, 1998; Abenhaim et al., 2000; Tugwell,2001; Guzman et al., 2002; Moffett and Mannion, 2005; Waddelland Burton, 2005; Airaksinen et al., 2006). Clinical management ofpatients with CLBP often comprises of two or more differentmanagement strategies delivered simultaneously. This can occurwithout evidence that the two management strategies havea synergistic effect. The interaction effects of different managementstrategies need to be further investigated so that appropriatecombinations of interventions can be delivered for each patient.

Group, aerobic based, exercise classes are a common manage-ment strategy for individuals with CLBP. A number of randomisedcontrolled trials have found such classes to be beneficial for thispatient group (Frost et al., 1995, 1998; Moffett et al., 1999; Klaber

x: þ44 141 331 8112.).

All rights reserved.

Moffett et al., 2004; UK BEAM Trial Team, 2004). The exact mech-anisms by which group exercise classes bring about a therapeuticeffect are not fully understood. It may be that psychosocial, ratherthan physical, mechanisms bring about the effect considering thatas few as four to five classes can bring about a therapeutic effectdespite this being an insufficient training stimulus to bring aboutphysiological training effects (Klaber Moffett et al., 2004).

Pain biology education is a relatively new intervention for themanagement of CLBP. It is a cognitive behavioural based interven-tionwhich attempts to reduce pain and disability by explaining thebiology of the pain to the patient (Butler and Moseley, 2003). Thistype of education has been shown to be superior to morebiomedical forms of education for this patient group (Moseley et al.,2004). Pain biology education has also been found to be usefulwhen delivered in combination with usual care physiotherapy(Moseley, 2002).

Considering both pain biology education and group exerciseclasses attempt to decrease fear of harm and increase physicalactivity (PA) in a paced manner it is logical to suggest that bothinterventions delivered together would have an added benefit toone another. However, the combined effect of these interven-tions has not been investigated. Such work is required to guideclinical practise.

C.G. Ryan et al. / Manual Therapy 15 (2010) 382e387 383

The aim of this study was to provide pilot data investigatingthe combined effect of pain biology education and group exerciseclasses compared to pain biology education alone for individualswith CLBP. The primary outcome measures were self-reportedpain and function. Secondary outcome measures were painrelated fear, pain self-efficacy, physical performance testing andfree-living physical activity (PA) monitoring.

2. Methods

2.1. Participants

All patients who met the inclusion/exclusion criteria wereprovided with the opportunity to opt into the study. The inclusioncriteria were: 18e65 years, non-specific low back pain >3 monthsduration and no history of surgery. The exclusion criteria were;physiotherapy within the past three months, involvement inregular sports activities twice per week for the past six months,constant or persistent pain adjudged clinically to be due to nerveroot irritation, fractures, non-back related musculoskeletal prob-lems which may affect ability to participate in the exercise classes,women who are or have been pregnant in the past year, anda positive response to red flag questions indicating a more seriouspathology such as malignancy. Recruitment occurred betweenSeptember-2005 and January-2007.

This study received ethical approval from the Greater GlasgowNational Health Service and the Glasgow Caledonian University,School of Health and Social Care, research and ethics committees.Written informed consent was obtained from all participants.

2.2. Experimental protocol

In this single-blinded RCT, participants were recruited from fivedifferent physiotherapy departments and randomised usinga random number generator to the education only group (ED) orthe education and exercise group (EDEX). At baseline, a set of self-report outcome measures were collected, three physical perfor-mance tests were completed and free-living step count wasrecorded. Demographic characteristics were also collected. Allpatients then participated in a two and a half hour session of painbiology education. Those in the ED group received no furtherintervention while those in the EDEX group were invited to takepart in six exercise classes over an eight week period. Eight weekslater, the post treatment assessment was performed, and alloutcome measures were collected. Three months later a postalfollow up was performed and the self-report outcome measureswere collected. The therapist delivering the education session andcollecting the outcome measures was blinded to treatment allo-cation. Blinding allocation was concealed using sealed envelopesand allocation patients were not allocated until after the educationhad been received.

2.3. Interventions

The pain biology education used in this study was based uponthat developed by Butler and Moseley (2003). The educationsession was delivered as a one off, two and a half hour cognitivebehavioural intervention focused on reshaping the participant'sbeliefs and attitudes about their back pain, attempting to decreasefear avoidance and harm beliefs, increase self-efficacy, and decreaseavoidance behaviour. This was achieved by providing informationon the biology of pain. The education was delivered using verbalcommunication, prepared diagrams and free-hand drawings. Thisform of intervention can decrease pain, pain related fear andbehaviour, and increase self-reported function when provided

alone or in conjunction with usual care physiotherapy (Moseley,2002, 2004, 2005; Moseley et al., 2004). Additionally, all partici-pants received “The Back Book”, a booklet which has been shown tobe beneficial for individuals with CLBP (Burton et al., 1999).

The exercise classes used in this study are knownwithin the UKas the “Back to Fitness exercise classes” (Moffett and Frost, 2000).These exercise classes were ongoing within the Greater GlasgowNational Health Service (NHS) at the time the study was beingundertaken, and participants randomised to the exercise groupjoined in with NHS patients. Each individual was invited to attendsix classes, one a week for six weeks. The classes involved circuitbased, graded, aerobic exercise with some core stability exercises.Three different exercise sites were used, involving differentinstructors. As the classes were organised and delivered by theNHS, making this a pragmatic RCT, it was not practical to change theclasses so that each site performed identical exercises. However allthree classes were similar and heavily based upon the guidelinesoutlined in the literature (Moffett and Frost, 2000). The classesinvolved awarm-up phase (10 min), an aerobic phase (20e30 min),and a warm-down phase (10e15 min). The aerobic phase involvedcircuit based exercise. For most exercises there was an easy,moderate, and hard version, and the participant could choosewhich version to perform. Participants were encouraged to work atan intensity considered “somewhat hard” for them.

2.4. Primary outcome measures

Self-reported functional ability was assessed using the RolandMorris Disability Questionnaire (RMDQ). The RMDQ consists of 24dichotomous items, is widely used in the literature and has demon-strated validity, reliability, and responsiveness to change (Roland andFairbank, 2000; Peat, 2004). A change of 4 or more points wasconsidered clinically important (Roland and Fairbank, 2000).

Pain was assessed using a numerical rating scale (NRS) goingfrom 0 to 100 rated upon pain experienced on the day of assess-ment. The scale was anchored using the following terms 0 ¼ nopain and 100 ¼ pain as bad as it could be. The pain NRS has beenshown to be a valid measure of pain demonstrating convergentvalidity (r¼ 0.65e0.88, p< 0.001) with other pain assessment tools(Jensen et al., 1986; Von Korff et al., 2000).

2.5. Secondary outcome measures

Physical performance was assessed using the repeated sit-to-stand test, the fifty-foot walk test and the 5-min walk test(Simmonds et al., 1998). During the repeated sit-to-stand test theparticipant was required to sit-to-stand five times from a standardchair. The shorter the time taken to complete the better the perfor-mance. The fifty-foot walk test required the participant to walka distance of fifty feet. The shorter the time taken to complete thetask the better the performance. The 5-min walk test required theparticipant to walk as far as a possible in a period of 5 min betweentwo markers 30 m apart. The three tests have demonstrated validityand reliability as performance measures (Simmonds et al., 1998).

Pain related fear was assessed using the Tampa Scale of Kine-siophobia-13 (TSK-13). The TSKe13 is a modified version of theoriginal Tampa Scale of Kinesiophobia. The questionnaire consistsof 13 items, on a four point scale. Higher scores indicate greaterlevels of pain related fear. The TSK-13 has demonstrated a goodlevel of internal consistency in a CLBP population (Cronbach's alpha0.82) (Goubert et al., 2004).

Pain self-efficacy, which was assessed using the pain self-effi-cacy questionnaire (PSEQ) (Nicholas, 1989), is a measure of anindividual's belief that they can carry out activities and functionsdespite their pain (Nicholas et al., 1992). The questionnaire contains

C.G. Ryan et al. / Manual Therapy 15 (2010) 382e387384

10 items (0e6 scale) where participants are asked to rate howconfident they are that they can perform a certain activity. Thehigher the score the greater the individuals pain self-efficacy. Thequestionnaire has demonstrated a high level of internal consistency(Cronbach's alpha ¼ 0.92), and a test-retest reliability of r ¼ 0.73(p < 0.001), for this patient group (Nicholas, 2007).

Objectively measured free-living PA has been shown to bea useful outcome measure for individuals with musculoskeletalconditions (Bussmann and Stam, 1998; Walker et al., 1998, 1999).Step-count was measured over a one week period using theactivPAL� activity monitor (PAL Technologies Ltd; Glasgow, Scot-land). The activPAL� has been shown to be a valid monitor formeasuring free-living PA in healthy adults (Grant et al., 2006; Ryanet al., 2006; Godfrey et al., 2007) and individuals with CLBP (Ryanet al., 2008).

2.6. Data analysis

An a-priori sample size calculation using the pain outcomeresults of Frost et al., (1995) estimated that 41 participants would berequired in both groups to identify an effect size of 0.63 with analpha level of 0.05 and a power of 80%. This was used to guideparticipant recruitment numbers. A post hoc power calculationwasalso performed to inform a future large scale trial based upon theprimary outcome measures. Power calculations were carried outusing the statistical package G*Power 3 (Faul et al., 2007).

Those allocated to a specific group at the beginning of the studywere in the same group at the end of the study, regardless of theirbehaviour during the study (e.g. those randomised to the EDEXgroup were analysed as part of the EDEX group even if they did notattend any exercise classes).

Data were analysed using SPSS (version 16.0) and a significancelevel was set at p � 0.05. The normality of the data was assessedusing the One-Sample KolmogoroveSmirnov test. All data wasfound to me normally distributed. Comparisons between groupcharacteristics (age, height, weight, BMI and pain duration) weremade using unpaired t-tests. The intervention effects were assessedusing a linear mixed model analysis (interaction between time and

Fig. 1. The figure shows the participant pathway. F ¼ females, EDEX

intervention group) which accounted for any differences in base-line values within participants and accounted for missing data. Thismodel considered repeated measures over three time periods andtwo groups entered as fixed factors. Additionally, the crossovereffect of group and time period was entered as an interaction term.Within the model duration of pain was entered as a covariate toaccount for differences between groups for this participant char-acteristic at baseline. This was the only significantly differentcharacteristic between groups.

3. Results

3.1. Participants

Seventy individuals expressed an interest in taking part in thisstudy. The participant pathway is shown in Fig. 1. Twenty-fiveindividuals initially agreed to be contacted by the researcher butthen decided not to participate in the study for the followingreasons; insufficient time, failed to opt in, not interested, no reasongiven. Seven individuals consented to take part but on initialassessment did not meet the inclusion/exclusion criteria. Theparticipant characteristics for the 38 individuals who provideda full set of baseline data are presented in Table 1. There was nosignificant difference between the EDEX group and the ED groupfor any of the participant characteristics except for pain durationwhich was significantly longer in the ED group.

A further seven individuals dropped out before the post treat-ment free-living PA data could be collected. Reasons for droppingout at this point included insufficient time, family commitments,one lady became pregnant and another broke her ankle, One maleparticipant could not be contacted. At the 3 month assessment 27participants completed the study and 11 did not. Non-completerswere significantly younger than completers (47.8 � 9.4 yrs vs.39.3 � 11.8 yrs, 95% CI �15.8 to �1.2, p < 0.05). There was nosignificant difference between completers and non-completers forgender, height, weight, BMI, duration of pain or employment status.

Exercise class attendance ranged from 0 to 6 sessions. Of the 20participants randomised to the EDEX group, two dropped-out

¼ Education and exercise group, ED ¼ Education only group.

Table 1Participant Characteristics This table shows the participant characteristics of the twogroups who provided baseline data. Data are presented as mean (SD) and groupcomparisons were performed using student independent t-tests.

EDEX(n ¼ 20)

ED(n ¼ 18)

p-value Mean diff(95% CI)

Gender 14F 6M 11F 7MAge 45.2 (11.9) 45.5 (9.5) 0.92 �0.4 (�7.5 to 6.8)Height 1.66 (0.09) 1.72 (0.12) 0.10 �0.06 (�0.13 to 0.01)Weight 79.2 (15.1) 77.5 (10.7) 0.70 1.7 (�7.0 to 10.4)BMI 28.7 (5.6) 26.2 (3.5) 0.11 2.5 (�0.6 to 5.6)Duration of pain 7.6 (7.0) 13.7 (10.2) 0.04 �6.1 (�11.8 to �0.4)

C.G. Ryan et al. / Manual Therapy 15 (2010) 382e387 385

prior to beginning the exercise classes. Of the 18 who providedpost treatment data, three attended zero classes and six attendedsix classes. The median attendance was 4.5 classes, and the mode6.0 classes. Twelve (60%) of the participants attended at least halfof the classes.

3.2. Primary outcome measures

The mean (SD) data for the two primary outcome measures arepresented in Table 2 and in Figs. 2 and 3. There was a statisticallysignificant interaction effect between time and group interventionfor pain with more favourable results for the ED group (F[2,49]¼ 3.975, p< 0.05) (Fig. 3). Therewas a similar non-significanttrend for a more favourable functional outcome in the ED group (F[2,51] ¼ 2.152, p ¼ 0.127) (Fig. 2). The effect for pain and functionlevelled off at the three month follow-up point.

3.3. Secondary outcome measures

The mean (SD) data for the secondary outcome measures arepresented Table 2. There was a statistically significant interactioneffect between time and group intervention for pain self-efficacywith more favourable results for the ED group (F[2,51] ¼ 4.011,p < 0.05). The effect for pain self-efficacy levelled off at the threemonth follow-up point.

There was no statistically significant effect for the remainingsecondary outcome measures.

Table 2Group comparisons. Data are presented as mean (SD). * Significant at p � 0.05. F and p-v

Pre Post

Primary outcome measuresFunction (0e24)Exercise & Education 9.4 (4.2) 5.6 (3.9)Education 10.8 (5.2) 3.3 (3.0)Pain (0e100)Exercise & Education 28.1 (20.4) 23.9 (23.3)Education 39.3 (26.2) 8.4 (7.5)

Secondary outcome measuresPain related fear (13e52)Exercise & Education 25.8 (7.4) 21.9 (8.2)Education 28.4 (8.2) 21.3 (6.5)Pain self-efficacy (0e60)Exercise & Education 50.0 (11.4) 48.8 (12.2)Education 41.9 (12.5) 55.1 (4.7)50 ft walk (sec)Exercise & Education 11.42 (3.82) 10.76 (2.76)Education 9.79 (2.56) 9.00 (1.68)5 min walk (m)Exercise & Education 390.2 (89.8) 433.9 (81.42)Education 439.1 (86.6) 490.6 (85.1)Free-living step count (steps)Exercise & Education 8284 (3725) 8927 (3932)Education 8001 (2071) 9166 (2774)

4. Discussion

The aim of this study was to provide pilot data investigating thecombined effect of pain biology education and group exerciseclasses compared to pain biology education alone for individualswith CLBP. The findings suggest that for pain and pain self-efficacy,in the short term, the education only intervention had a greatereffect than the education and exercise group.

Frost et al. (1995) performed an RCT similar in design to thecurrent study, where one group received an education session andone group received education as well as group exercise classes.Frost et al. reported significantly greater improvement in thecombined education and exercise group. In the Frost et al. study theeducation was based on the medical model of back pain andincluded information on anatomy, biomechanics and posture. Thepain biology education session used in the current study wasa cognitive behavioural intervention based on the biopsychosocialmodel of back pain attempting to educate patients on the biology ofpain aiming to decrease fear and distress, and encourage a return tousual daily activity (Butler and Moseley, 2003). Previous researchhas directly compared these two forms of education and found painbiology education to be superior and has suggested thata biomedical back education may actually have a negative effect onpatient outcome (Moseley et al., 2004). Thus, in the study by Frostet al., (1995) the back to fitness classes may have shown addedbenefit to the education session because of the lack of effect, orindeed negative effect of such a biomedical education.

It is unclear as to why the education only group had better shortterm outcomes in this study. However there are a number ofpossible explanations. Among other things, the purpose of theeducation session was to “demedicalise” the person's condition,shift attitudes towards a more biopsychosocial self-managementapproach, and highlight that “hurt does not equal harm”. Attendingthe exercise classes in a “clinical setting” with a clinical classinstructor (physiotherapist) may have detracted from that messageand reinforced the concept of the participants being patients, withsomething medically/structurally wrong, requiring medical treat-ment. It may have reinforced the individual's perception of them-selves as a patient, the patient persona. The minimally invasivesingle cognitive behavioural education session may have

alues were calculated using a mixed model analysis.

FU F-value p-value

6.4 (5.1) 2.152 0.1274.3 (4.2)

19.1 (18.9) 3.975 0.025*22.6 (30.8)

21.5 (7.5) 0.440 0.64623.7 (6.6)

49.5 (13.1) 4.011 0.024*49.5 (9.8)

0.009 0.924

0.037 0.848

0.111 0.740

Fig. 2. The Function scores for both groups, pre-intervention, post-intervention and atthe three month follow up stage. Data are presented as mean (SD). RMDQ ¼ RolandMorris Disability Questionnaire.

C.G. Ryan et al. / Manual Therapy 15 (2010) 382e387386

contributed to a speedier exit from the person's patient-personastate, resulting in a faster improvement in outcome.

The EDEX group had increased therapistepatient interaction. Asa result, patients may have attributed improvements to the thera-pist and not themselves, decreasing feelings of accomplishmentand self-efficacy, which may have undermined the education(Dolce, 1987; Klaber Moffett and Richardson, 1997). There is someevidence that attending such a class may inadvertently reinforceany negative pre-existing beliefs, either through the therapists useof language and terminology or through certain behaviours andinteractions (Morris, 2004). This may have been even more of anissue if the therapist providing the classes had beliefs and attitudesdeeply rooted in the biomedical model subscribing to a morestructural-pathology understanding of the condition, which theymay have been passed onto the participants in their class. This mayhave diluted many of the messages put forward in the pain biologyeducation. There is a growing body of evidence that the attitudesand beliefs of the therapist affect the information they provide theirpatients (Rainville et al., 2000; Linton et al., 2002). However, theattitudes and beliefs of the therapists delivering the exercise classeswere not assessed in this study.

Participants randomised to the EDEX group participated inexercise classes whichwere also attended by low back pain patientswhowere not participating in this study. It is possible that this mayhave had an effect on the outcomes for the EDEX patients. Forexample, the non-participant patients are unlikely to have had thesame level of pain biology education, and indeed may have

Fig. 3. The pain scores for both groups, pre-intervention, post-intervention and atthe three month follow up stage. Data are presented as mean (SD). NRS ¼ numericalrating scale.

possessed a strong structural pathological view point of painwhichthey may have communicated to the participants in this studyattending the same exercise class. This interactionmay have dilutedthe message provided in the pain biology lecture and may partiallyexplain the superior outcomes for the ED group.

Greater improvement in the ED groupmay have been associatedwith “mixed information” in the EDEX group. Providing two sets ofinformation, using different amounts and formats of informationcan lead to poorer outcomes than using one set of informationalone (Little et al., 2001). The information and advice provided bythe therapists who supervised the exercise classes was notcontrolled and was left to the physiotherapists' professionaljudgement. This was done to create a realistic NHS based “back tofitness” exercise class environment. It is likely that the information/advice differed somewhat from the information provided in thepain biology education session. This may have led to confusion, andeven frustration, on the part of the patient, which could have hada negative impact on patient improvement.

This pilot study had a number of strengths, including randomisedassignment and investigator blinding. The study had three types ofoutcome measure, self-report, PPT and objectively measured free-living activity. The agreement infindings between the different typesof outcomemeasure increases the confidence in the results. The studyalso had a number of weaknesses; the study was not double blinded,participants knew to which group they had been allocated. This mayhave resulted in a placebo effect, although it could be argued that anyplacebo affect would have produced better outcomes in the EDEXgroup. The sample size in this study was small, increasing the likeli-hood of a type II error. Based upon the three month follow up dataa power calculation, based upon a simple t-test, has shown that for80% power at an alpha level of 0.05, a sample size of 66 participantswould be needed in each group to detect a significant difference inpain levels and 162 participants in each group to detect a significantdifference in self-reported function.

The study would have benefited from a group-exercise-class-only group, to better understand if the poorer outcomes in theEDEX group were due to the interaction between pain biologyeducation and exercise or if it was due to the exercise alone. Whileparticipants were requested not to seek co-interventions beyondtheir GP during the course of the study, one participant from the EDgroup reported that she received osteopathic manipulations. Noneof the remaining participants reported receiving co-interventions.Finally, three of the participants attended no exercise classes andonly six attended all six classes, the results of this study may havebeen considerably different if all participants had attended all sixclasses. However this was a pragmatic trial and artificially enforcingfull attendance would have reduced the generalisability of theresults. Other studies which have investigated the efficacy of the“back to fitness” exercise classes have had similar issues withattendance (UK BEAM Trial Team, 2004).

5. Conclusion

In conclusion, pain biology education was more effective forpain, and pain self-efficacy than a combination of pain biologyeducation and group exercise classes, for individuals with CLBP, inthe short term. The rationale for this finding is not fully understood.This pilot study highlights the need to investigate the combinedeffects of different interventions, as it cannot be assumed thata synergistic effect will occur, and a negative interaction is possible.

Acknowledgements

This study was funded by the School of Health and Social Care ofGlasgow Caledonian University, and no financial support was

C.G. Ryan et al. / Manual Therapy 15 (2010) 382e387 387

received from any commercial company. One of the authors is a co-inventor of the activPAL� physical activity monitor and a director ofPAL technologies Ltd. The remaining authors declare no competinginterests. The authors would like to acknowledge the significantstatistical input of Dr. Sebastien Chastin and Dr. Jon Godwin of Glas-gow Caledonian University.

References

Abenhaim L, Rossignol M, Valat JP, Nordin M, Avouac B, Blotman F, et al. The role ofactivity in the therapeutic management of back pain. Report of the internationalParis task force on back pain. Spine 2000;25:1Se33S.

Airaksinen O, Brox JI, Cedraschi C, Hildebrandt J, Klabber-Moffett J, Kovacs F, et al.Chapter 4 European guidelines for the management of chronic nonspecific lowback pain. European Spine Journal 2006;15(Suppl. 2):S192e300.

Bussmann JB, Stam HJ. Techniques for measurement and assessment of mobility inrehabilitation: a theoretical approach. Clinical Rehabilitation 1998;12:455e64.

Burton AK, Waddell G, Tilloston KM, Summerton N. Information and advice topatients with back pain can have a positive effect. A randomized controlled trialof a novel educational booklet in primary care. Spine 1999;23:2484e91.

Butler DS, Moseley L. Explain pain. Adelaide: Noigroup; 2003.Dolce JJ. Self-efficacy and disability beliefs in behavioral treatment of pain.

Behaviour Research and Therapy 1987;25:289e99.Faul F, Erdfelder E, Lang AG, Buchner A. G*Power 3: a flexible statistical power

analysis program for the social, behavioural, and biomedical sciences. BehaviorResearch Methods 2007;39:175e91.

Frost H, Klaber Moffett JA, Moser JS, Fairbank JC. Randomised controlled trial forevaluation of fitness programme for patients with chronic low back pain. BritishMedical Journal 1995;310:151e4.

Frost H, Lamb SE, Klaber Moffett JA, Fairbank JC, Moser JS. A fitness programme forpatients with chronic low back pain: 2-year follow-up of a randomisedcontrolled trial. Pain 1998;75:273e9.

Godfrey A, Culhane KM, Lyons GM. Comparison of the performance of the activPALProfessional physical activity logger to a discrete accelerometer-based activitymonitor. Medical Engineering & Physics 2007;29:930e4.

Goubert L, Crombez G, Van Damme S, Vlaeyen JW, Bijttebier P, Roelofs J. Confir-matory factor analysis of the Tampa Scale for Kinesiophobia: invariant two-factor model across low back pain patients and fibromyalgia patients. TheClinical Journal of Pain 2004;20:103e10.

Grant PM, Ryan CG, Tigbe WW, Granat MH. The validation of a novel activitymonitor in the measurement of posture and motion during everyday activities.British Journal of Sports Medicine 2006;40:992e7.

Guzman J, Esmail R, Karjalainen K, Malmivaara A, Irvin E, Bombardier C. Multidis-ciplinary bio-psycho-social rehabilitation for chronic low back pain. CochraneDatabase of Systematic Reviews (Online) 2002;1(1). CD000963.

Hilde G, Bo K. Effect of exercise in the treatment of chronic low back pain:a systematic review emphasizing type and dose of exercise. Physical TherapyReviews 1998;3:107e17.

Jensen MP, Karoly P, Braver S. The measurement of clinical pain intensity:a comparison of six methods. Pain 1986;27(1):117e26.

Klaber Moffett JA, Carr J, Howarth E. High fear-avoiders of physical activity benefitfrom an exercise program for patients with back pain. Spine 2004;29:1167e73.

Klaber Moffett JA, Richardson PH. The influence of the physiotherapist-patientrelationship on pain and disability. Physiotherapy Theory and Practice1997;13:89e96.

Linton SJ, Vlaeyen J, Ostelo R. The back pain beliefs of health care providers: are wefear-avoidant? Journal of Occupational Rehabilitation 2002;12:223e32.

Little P, Roberts L, Blowers H, Garwood J, Cantrell T, Langridge J, et al. Should wegive detailed advice and information booklets to patients with back pain? A

randomized controlled factorial trial of a self-management booklet and doctoradvice to take exercise for back pain. Spine 2001;26:2065e72.

Moffett JK, Frost H. Back to fitness programme: the manual for physiotherapists toset up the classes. Physiotherapy 2000;86:295e305.

Moffett JK, Mannion AF. What is the value of physical therapies for back pain? Bestpractice & research. Clinical Rheumatology 2005;19:623e38.

Moffett JK, Torgerson D, Bell-Syer S, Jackson D, Llewlyn-Phillips H, Farrin A, et al.Randomised controlled trial of exercise for low back pain: clinical outcomes,costs, and preferences. British Medical Journal 1999;319:279e83.

Morris AL. Patients' perspectives on self-management following a back rehabilita-tion programme. Musculoskeletal Care 2004;2:165e79.

Moseley GL. Widespread brain activity during an abdominal task markedlyreduced after pain physiology education: fMRI evaluation of a single patientwith chronic low back pain. The Australian Journal of Physiotherapy 2005;51:49e52.

Moseley GL. Evidence for a direct relationship between cognitive and physicalchange during an education intervention in people with chronic low back pain.European Journal of Pain 2004;8:39e45.

Moseley GL, Nicholas MK, Hodges PW. A randomized controlled trial of intensiveneurophysiology education in chronic low back pain. The Clinical Journal ofPain 2004;20:324e30.

Moseley L. Combined physiotherapy and education is efficacious for chronic lowback pain. The Australian Journal of Physiotherapy 2002;48:297e302.

Nicholas MK. The pain self-efficacy questionnaire: taking pain into account. Euro-pean Journal of Pain 2007;11:153e63.

Nicholas MK. Self-Efficacy and chronic pain. In: Annual Conference of the BritishPsychological Society British Psychological society, United Kingdom, 1989.

Nicholas MK, Wilson PH, Goyen J. Comparison of cognitive-behavioral grouptreatment and an alternative non-psychological treatment for chronic low backpain. Pain 1992;48:339e47.

Peat G. PPA recommendations for low back pain-related functional limitationoutcome measures. London: Chartered Society of Physiotherapy; 2004.

Rainville J, Carlson N, Polatin P, Gatchel RJ, Indahl A. Exploration of physicians'recommendations for activities in chronic low back pain. Spine 2000;25:2210e20.

Roland M, Fairbank J. The roland-morris disability questionnaire and the oswestrydisability questionnaire. Spine 2000;25:3115e24.

Ryan CG, Grant PM, Gray H, Newton M, Granat MH. Measuring postural physicalactivity in people with chronic low back pain. The Journal of Back andMusculoskeletal Rehabilitation 2008;21:43e50.

Ryan CG, Grant PM, Tigbe WW, Granat MH. The validity and reliability of a novelactivity monitor as a measure of walking. British Journal of Sports Medicine2006;40:779e84.

Simmonds MJ, Olson SL, Jones S, Hussein T, Lee CE, Novy D, et al. Psychometriccharacteristics and clinical usefulness of physical performance tests in patientswith low back pain. Spine 1998;23:2412e21.

Tugwell P. Philadelphia panel evidence-based clinical practice guidelines onselected rehabilitation interventions for low back pain. Physical Therapy2001;81:1641e74.

UK BEAM Trial Team. United Kingdom back pain exercise and manipulation (UKBEAM) randomised trial: effectiveness of physical treatments for back pain inprimary care. British Medical Journal 2004;329:1377.

Von Korff M, Jensen MP, Karoly P. Assessing global pain severity by self-report inclinical and health services research. Spine 2000;25:3140e51.

Waddell G, Burton AK. Concepts of rehabilitation for the management of low backpain. Best practice & research. Clinical Rheumatology 2005;19:655e70.

Walker DJ, Heslop PS, Kay LJ, Chandler C. Spontaneous ambulatory activity asa quantifiable outcome measure for osteoarthritis of the knee. British Journal ofRheumatology 1998;37:969e71.

Walker DJ, Kidd E, Heslop PS, Chandler C. Spontaneous ambulatory activity asa quantifiable outcome measure for rheumatoid arthritis. Rheumatology1999;38:1234e8.