1-s2.0-S1356689X13000842-main (1)

at SciVerse ScienceDirect

Manual Therapy 18 (2013) 533e540

Contents lists available

Manual Therapy

journal homepage: www.elsevier .com/math

Original article

Outcomes of osteopathic manual treatment for chronic low back painaccording to baseline pain severity: Results from the OSTEOPATHICTrial

John C. Licciardone a,b,*, Cathleen M. Kearns a, Dennis E. Minotti a

a The Osteopathic Research Center, University of North Texas Health Science Center, 3500 Camp Bowie Boulevard, Fort Worth, TX 76107, USAbDepartment of Medical Education, Texas College of Osteopathic Medicine, University of North Texas Health Science Center, 3500 Camp Bowie Boulevard,Fort Worth, TX 76107, USA

a r t i c l e i n f o

Article history:Received 6 February 2013Received in revised form8 May 2013Accepted 13 May 2013

Keywords:Manual therapyOsteopathic medicineOsteopathyChronic low back pain

* Corresponding author. The Osteopathic ResearchTexas Health Science Center, 3500 Camp Bowie BoulUSA. Tel.: 1 817 735 2028; fax: 1 817 735 0157.

E-mail address: [email protected] (J.C.

1356-689X 2013 The Authors. Published by Elseviehttp://dx.doi.org/10.1016/j.math.2013.05.006

a b s t r a c t

Purpose: To assess response to osteopathic manual treatment (OMT) according to baseline severity ofchronic low back pain (LBP).Methods: The OSTEOPATHIC Trial used a randomized, double-blind, sham-controlled, 2 2 factorialdesign to study OMT for chronic LBP. A total of 269 (59%) patients reported low baseline pain severity(LBPS) (

J.C. Licciardone et al. / Manual Therapy 18 (2013) 533e540534

severity of chronic low back pain by comparing patient subgroupswithin the OSTEOPATHIC Trial.

2. Methods

2.1. Study overview

The OSTEOPATHIC Trial was approved by the Institutional Re-view Board at the University of North Texas Health Science Centerand registered with ClinicalTrials.gov (NCT00315120). Its method-ology has been previously described (Licciardone et al., 2008;Licciardone et al., 2013). The trial used a randomized, double-blind,sham-controlled, 2 2 factorial design (Fig. 1) to study OMT (factor1) and ultrasound therapy (factor 2) over 12 weeks in patients withnonspecific chronic LBP. Therein, OMT was shown to be safe, wellaccepted by patients, and associated with statistically significantand clinically relevant reduction in LBP (Licciardone et al., 2013).Consequently, the present study focused on comparing OMT vs.sham OMT in patient subgroups with low baseline pain severity(LBPS) and high baseline pain severity (HBPS). Ultrasound therapywas not studied herein because the OSTEOPATHIC Trial failed todemonstrate its efficacy.

2.2. Enrollment and randomization

Patients were recruited throughout DallaseFort Worth fromAugust 2006 to September 2010 through newspaper advertise-ments, community agencies, and medical clinics, including thoseaffiliated with the group practice of the University of North TexasHealth Science Center, exclusive of clinics that provided OMT spe-cialty services. The eligibility criteria were developed to includepatients with nonspecific chronic LBP and to exclude patients whorecently used manual therapy for LBP. Essentially, patients werethose 21e69 years of age who self-reported low back pain on mostdays during the past three months, but who were without any ofthe following: red flag conditions; history of recent low backsurgery, receipt of workers compensation benefits, or ongoinglitigation involving back problems; medical conditions that mightimpede OMT (or ultrasound therapy) protocol implementation;corticosteroid use in the past month; or clinical evidence of lumbarradiculopathy, as determined by the presence of ankle dorsiflexionweakness, great toe extensor weakness, impaired ankle reflexes,

Factor 1

Osteopathic manual treatment

Fact

or 2

U

ltras

ound

ther

apy

Active Sham

Sham

Ac

tive Sham OMT +

UST

(n=118)

OMT +

UST

(n=115)

OMT +

Sham UST

(n=115)

Sham U

ST m

ain effects (n=222)

UST

main

effects (n=233)

Sham OMT

main effects

(n=225)

Sham OMT +

Sham UST

(n=107)

OMT

main effects

(n=230)

Fig. 1. Overview of the OSTEOPATHIC Trials 2 2 factorial design. OMT denotesosteopathic manual treatment; UST, ultrasound therapy. As indicated by the shadedbox, the present study focuses on OMT (factor 1) because it was found to be efficaciousin reducing low back pain in the OSTEOPATHIC Trial, whereas UST (factor 2) was notefficacious.

loss of light touch sensation in the medial, dorsal, and lateral as-pects of the foot, or shooting posterior leg pain or foot pain uponipsilateral or contralateral straight leg raising (Bigos et al., 1994).Patients who had received manual therapy in the past threemonths, or more than three times in the past year, were alsoexcluded. Patients were randomly allocated to either OMT or shamOMT by a computer-based process. These assignments wereconveyed to treatment providers via opaque sealed envelopes.Randomization was not stratified according to baseline painseverity. Patients and outcome assessors were not informed oftreatment group assignments.

2.3. Patient subgroups

Low back painwas measured with a 100-mmvisual analog scale(VAS) at baseline, before each treatment session, and at week 12.We dichotomized patients into two subgroups defined as havingLBPS (VAS< 50 mm/100 mm) or HBPS (VAS 50 mm/100 mm) forthree reasons. First, dichotomization yielded relatively larger sub-groups than would have been obtained with other polychotomouscategorizations (e.g., trichotomization as mild, moderate, orsevere). Second, it was intuitively appealing to simply bisect the100-mm VAS. Third, the 50-mm cutpoint would facilitate extrap-olation of our LBP results to numerical and other rating scales usedin research settings or clinical practice.

2.4. Treatment protocols

Treatment fidelity methods (Bellg et al., 2004) were usedto train 15 treatment providers to deliver the OMT and shamOMT protocols. Both protocols consisted of 15-min treatmentsessions at weeks 0e2, 4, 6, and 8, delivered by the same providerto a given patient unless there was a scheduling conflict. Osteo-pathic manual treatment included high-velocity, low-amplitudethrusts; moderate-velocity, moderate amplitude thrusts; softtissue stretching, kneading, and pressure; myofascial stretchingand release; positional treatment of myofascial tender points;and muscle energy techniques. These techniques were aimedprimarily at the lumbosacral, iliac, and pubic regions. Otherosteopathic techniques were allowed only if the treatment pro-vider judged a designated technique to be contraindicated orineffective for a given patient. The sham OMT protocol was basedon that developed in the North Texas Clinical Trial (Licciardoneet al., 2003) and subsequently determined to provide a robustresponse in comparison with other placebo treatments for pain(Hrobjartsson and Gotzsche, 2001). The sham methods includedhand contact, active and passive range of motion, and techniquesthat simulated OMT, but that utilized such maneuvers as lighttouch, improper patient positioning, purposely misdirectedmovements, and diminished treatment provider force. Patientswere allowed to receive their usual LBP care and other co-treatments during the study with the exception of manualtherapies.

2.5. Outcomes

2.5.1. Substantial low back pain improvementSubstantial LBP improvement was based on the Initiative on

Methods, Measurement, and Pain Assessment in Clinical Trials(IMMPACT) consensus statement recommendations (Dworkinet al., 2008). We used the relative threshold of 50% pain reduc-tion to determine substantial improvement at week 12, rather thanthe absolute threshold of 40 mm pain reduction, to minimizefloor effects in assessing OMT efficacy in patients with LBPS.This threshold is highly sensitive and specific in predicting global

Fig. 2. Schematic representation of the multidimensional approach for assessing theefficacy of osteopathic manual treatment in patients with chronic low back pain. Pa-tient improvement in low back pain is based on the Initiative on Methods, Measure-ment, and Pain Assessment in Clinical Trials (IMMPACT) consensus statementrecommendations (Dworkin, et al., 2008). The relevant thresholds are 50% painreduction for substantial improvement; 30% pain reduction for moderate improve-ment; and 10% pain reduction for minimal improvement. For simplicity, however, anytreatment effect that fails to reach the thresholds for moderate or substantialimprovement may be considered minimal. Population effect size is estimated using theCochrane Back Review Group criteria (Furlan, et al., 2009). The relevant criteria areresponse ratio (RR) > 2 for a large effect size; 1.25 RR 2 for a medium effect size;and RR < 1.25 for a small effect size. The cells within the 3 3 2 matrix are uniquelyidentified by a three-digit code sequentially based on patient improvement (1, mini-mal; 2, moderate; 3, substantial), population effect size (1, small; 2, medium; 3, large),and baseline pain severity (1, low; 2, high). This study focused on assessing the efficacyof osteopathic manual treatment in providing substantial low back pain improvementin patients with low baseline pain severity ( 2) effect sizes (Furlan et al.,2009). Our multidimensional approach for assessing OMT efficacyis illustrated by a 3 3 2 matrix based on three levels of patientimprovement (minimal, moderate, substantial), three levels ofpopulation effect size (small, medium, large), and two subgroups ofbaseline pain severity (LBPS, HBPS) (Fig. 2). We focused on the sixcells (Fig. 2, cells 311, 312, 321, 322, 331, 332) that potentiallycharacterize OMT efficacy in providing substantial LBP improve-ment. We took this highly specific analytical approach of assessingsubstantial improvement at the patient level as opposed to moresensitive assessments of minimal or moderate patient improve-ment to ensure that any positive study findings would have clinicalsignificance. Previous reviews have attributed only small and notapparently clinically relevant effects (Rubinstein et al., 2011) or, atbest, moderate efficacy (Chou and Huffman, 2007) to spinalmanipulative therapy.

Responder analysis was also used to assess secondary outcomes.Satisfaction with back care was dichotomized by combining verysatisfied and satisfied responses vs. all others. For work disabilityand use of LBP co-treatments, the Cochrane Back Review Groupcriteria are reversed such that medium and large effect sizes arerepresented by 0.5 RR 0.8 and RR< 0.5, respectively. Patientflow, treatment provider assignment, treatment adherence, andsafety were assessed by contingency table methods.

Hypothesis testing was by intention-to-treat with a two-sideda 0.05. Rothmans T statistic (Hogan et al., 1978) was used totest for statistical interaction between OMTand ultrasound therapy

in assessing substantial LBP improvement. Missing data generallywere imputed using the last observation carried forward. However,because relevant baseline data were not available for workdisability and were not feasible for satisfaction with back care, weused multivariate regression to impute missing data for thesevariables. Work disability was regressed on age, sex, and baselinework status, whereas satisfaction with back care was regressedexclusively on age and sex. Per-protocol analyses were conductedto assess the impact of treatment non-adherence and robustness ofour data imputation methods.

There were 269 (59%) patients in the LBPS subgroup and 186(41%) patients in the HBPS subgroup. Post-hoc subgroup-specificestimates of statistical power in detecting medium and large ef-fect sizes (Furlan et al., 2009) were computed under the assump-tion of a common sham OMT response across subgroups (Table 1).Statistical power in detecting large effect sizes exceeded 0.80 forthe primary outcome and four secondary outcome measures inboth subgroups. Statistical power in detecting medium effect sizeswas low in both subgroups with the exception of satisfaction withback care. We used the P for interaction (Altman and Bland, 2003)to compare subgroup treatment effects for each outcome to mini-mize the likelihood of spurious results and invalid conclusions

Fig. 3. CONSORT diagram. OMT denotes osteopathic manual treatment. aFour patients weremeet the inclusion criteria. Two of these patients provided false information to initially qual

Table 1Post-hoc statistical power in detectingmedium and large effect sizes for primary andsecondary outcomes according to baseline pain severity.a

Outcome measure LBPS (0.99 0.23 >0.99

Secondary outcomesRolandeMorris disability score 0.16 0.95 0.13 0.84SF-36 general health score 0.20 0.99 0.15 0.94Work disabilityb 0.11 0.50 0.09 0.31Satisfaction with back care 0.93 >0.99 0.82 >0.99Use of co-treatments for LBP during the trialExercise programming 0.13 0.62 0.11 0.47Non-prescription medication 0.33 0.98 0.24 0.91Prescription medication 0.14 0.64 0.11 0.49Physical therapy 0.08 0.27 0.07 0.20CAM therapies 0.12 0.55 0.10 0.41

CAM denotes complementary and alternative medicine; HBPS, high baseline painseverity; LBPS, low baseline pain severity; SF-36, Medical Outcomes Study ShortForm-36 Health Survey.

a Statistical power was computed under the assumption of a common sham OMTresponse across both subgroups. The thresholds for medium and large effect sizeswere based on the Cochrane Back Review Group criteria (Furlan, et al., 2009).

b Work disability analyses included only patients whowere employed full-time atbaseline.

J.C. Licciardone et al. / Manual Therapy 18 (2013) 533e540536

(Brookes et al., 2004). Statistical analyses were performed with theSPSS Statistics version 20 software (IBM Corporation, Armonk, NY).

3. Results

3.1. Patient flow and characteristics

The CONSORT diagram illustrates patient flow (Fig. 3). It dem-onstrates similar allocation to treatment providers, treatmentadherence, and follow-up in the OMTand sham OMTgroups withinboth subgroups. Patients in the HBPS subgroup reported signifi-cantly poorer back-specific functioning and general health thanpatients in the LBPS subgroup (Table 2). Patients in the HBPS sub-group were also more likely to have been taking prescriptionmedicine for LBP prior to randomization and were more oftenhospitalized for LBP than patients in the LBPS subgroup. Co-morbiddiabetes mellitus and depression were also more common in theHBPS subgroup. Patient characteristics in the OMT and sham OMTgroups were comparable within each subgroup (Table 3).

3.2. Substantial low back pain improvement

There was no statistical interaction between OMT and ultra-sound therapy in assessing substantial LBP improvement (T, 0.05;95% CI, 0.23 to 0.13; P 0.61). In the LBPS subgroup, 65 (48%)patients who received OMT vs. 56 (42%) patients who received

excluded for cause post-hoc because it was subsequently discovered that they did notify for the study. bRandomization was not stratified according to baseline pain severity.

Table 2Patient characteristics according to baseline pain severity.

Characteristic Total LBPS (

Table 4Outcomes of osteopathic manual treatment at week 12 according to baseline pain severity.a

Outcomes LBPS (

J.C. Licciardone et al. / Manual Therapy 18 (2013) 533e540 539

(Costa et al., 2013). Our dichotomization strategy may be easilyapplied in clinical practice to target patients with chronic LBP ofhigh severity for a short course of OMT to reduce pain and improveback-specific functioning, as demonstrated herein. Our OMTregimen involving six treatment sessions was parsimonious, beingwell within the recommended guideline of nine treatment sessionsover 12 weeks for persistent LBP (National Institute for Health andClinical Excellence, 2009). By contrast, a typical initial trial ofchiropractic care would have entailed six to 12 treatment sessionsover twoefour weeks, with potentially up to 36 treatment sessionsover 12 weeks depending on patient progress and prognosis (Globeet al., 2008).

An unanswered question is why OMT yielded a large effect sizein our HBPS subgroupwhile a Cochrane Collaboration review foundspinal manipulative therapy to be no more effective than shamspinal manipulative therapy in providing short-term pain relief orimprovement in functional status (Rubinstein et al., 2011). Onepossible explanation is that our a-posterioriHBPS subgroup analysiswas biased by confounders that were no longer distributed atrandom in this subgroup (Hennekens and Demets, 2009). However,patients appeared to be adequately balanced on sociodemographic,clinical, and baseline outcome characteristics in each subgroup(Table 3). Another possibility is that previous studies (Waagen et al.,1986; Licciardone et al., 2003; Ghroubi et al., 2007) suffered fromhigh risk of bias (Rubinstein et al., 2011), thereby reducing theirlikelihood of detecting significant treatment effects. A thirdpossible explanation involves the high prevalence of dysfunction inthe lumbar, sacral, pelvic, and innominate regions of patients withchronic LBP (Licciardone and Kearns, 2012). Our multimodal OMTregimen included six techniques in a comprehensive approach fortreating the dysfunctions underlying LBP severity and associatedwith deficits in back-specific functioning. Previous trials involvingunimodal approaches (e.g., high-velocity, low-amplitude thrustingin the lumbar region) may not have adequately addressed multi-focal dysfunction in patients with chronic LBP. The association oflumbar dysfunction with higher baseline pain severity in our pa-tients (Licciardone and Kearns, 2012) also helps to explain thegreater response to OMT within the HBPS subgroup. A fourthpossible explanation for our results is that the majority of treat-ments was provided by osteopathic physicians who received fi-delity training in implementing the study protocol. Nevertheless,we believe that our results may be generalizable to other manualtherapy practitioners because several OMT techniques in our pro-tocol were accepted for LBP treatment by professional associationsrepresenting chiropractors and physiotherapists (Harvey et al.,2003).

The overarching strengths and limitations of the OSTEOPATHICTrial have been previously described (Licciardone et al., 2008,2013). Essentially, strengths included allocation concealment,blinding of outcome assessors, high levels of treatment adherenceand outcomes reporting, and intention-to-treat analysis. We alsopragmatically assessed OMT as practiced in real-life settings tocomplement usual care and self-care for chronic LBP. Limitationsincluded patient self-reporting of co-morbid conditions, workdisability, and LBP co-treatments. Additionally, despite our effortsto maintain patient blinding during the study, it is possible thatsome degree of unblinding may have occurred.

To our knowledge, the OSTEOPATHIC Trial is the largest OMTtrial to date. Consequently, its sample size facilitated the perfor-mance of selected subgroup analyses. Statistical power exceeded0.80 for analyses aimed at detecting large effect sizes in both LBPSand HBPS subgroups for the primary outcome variable. Both sub-group analyses were also adequately powered for detecting largeeffect sizes for clinically relevant improvements in back-specificfunctioning and general health, and for satisfaction with back

care and use of non-prescription medication as a LBP co-treatment.Thus, our analyses were powered to detect the most important andclinically relevant treatment effects of OMT, such as those observedfor substantial LBP improvement on the VAS (Fig. 2, cell 332) andchange in back-specific functioning on the RMDQ. In the HBPSsubgroup, and to a lesser degree in the LBPS subgroup, otherpotentially important benefits of OMT could not be ruled outbecause of low statistical power. We were also unable to defini-tively classify OMT effects in reducing LBP in the LBPS subgroupbecause of inadequate statistical power. However, based on the RRsand 95% CIs for patients with LBPS, OMT effects in this subgroup liein one of six possible cells (Fig. 2, cells 111, 121, 211, 221, 311, 321).

5. Conclusions

The large effect size for OMT in providing substantial LBPimprovement in patients with HBPS was associated with clinicallyimportant improvement in back-specific functioning. Based onthese results, and on safety and satisfaction with back care, OMTappears to be an attractive option in patients with chronic LBP ofhigh severity before proceeding to more invasive and costly treat-ments. Our results and conclusions should be interpreted in lightof the usual caveats that accompany subgroup analyses, includingstatistical power limitations and the potential for unknownconfounding.

Acknowledgments

This study was funded by grants to JCL from the National In-stitutes of HealtheNational Center for Complementary and Alter-native Medicine (K24-AT002422) and the Osteopathic HeritageFoundation. The funders had no role in the design and conduct ofthe study; collection, management, analysis, and interpretation ofthe data; and preparation, review, or approval of the manuscript.We express our gratitude to the patients and personnel at TheOsteopathic Research Center for their contributions to this study.

References

Altman DG, Bland JM. Interaction revisited: the difference between two estimates.BMJ 2003;326:219.

Barnes PM, Bloom B, Nahin RL. Complementary and alternative medicine useamong adults and children: United States, 2007. Natl Health Stat Report 2008:1e23.

Becker A, Held H, Redaelli M, Strauch K, Chenot JF, Leonhardt C, et al. Low back painin primary care: costs of care and prediction of future health care utilization.Spine (Phila Pa 1976) 2010;35:1714e20.

Bellg AJ, Borrelli B, Resnick B, Hecht J, Minicucci DS, Ory M, et al. Enhancingtreatment fidelity in health behavior change studies: best practices and rec-ommendations from the NIH Behavior Change Consortium. Health Psychol2004;23:443e51.

Bigos SJ, Bowyer OR, Braen GR, Brown K, Deyo R, Haldeman S, et al. Acute low backproblems in adults. Clinical Practice Guideline No. 14. Rockville, MD: Agency forHealthcare Research and Quality, Public Health Service, U.S. Department ofHealth and Human Services; 1994.

Bombardier C. Outcome assessments in the evaluation of treatment of spinal dis-orders: summary and general recommendations. Spine (Phila Pa 1976)2000;25:3100e3.

Bono CM, Harris MB, Warholic N, Katz JN, Carreras E, White A, et al. Pain intensityand patients acceptance of surgical complication risks with lumbar fusion.Spine (Phila Pa 1976) 2013;38:140e7.

Brookes ST, Whitely E, Egger M, Smith GD, Mulheran PA, Peters TJ. Subgroupanalyses in randomized trials: risks of subgroup-specific analyses; power andsample size for the interaction test. J Clin Epidemiol 2004;57:229e36.

Chou R, Huffman LH. Nonpharmacologic therapies for acute and chronic low backpain: a review of the evidence for an American Pain Society/American Collegeof Physicians clinical practice guideline. Ann Intern Med 2007;147:492e504.

Chou R, Qaseem A, Snow V, Casey D, Cross Jr JT, Shekelle P, et al. Diagnosis andtreatment of low back pain: a joint clinical practice guideline from the Amer-ican College of Physicians and the American Pain Society. Ann Intern Med2007;147:478e91.

J.C. Licciardone et al. / Manual Therapy 18 (2013) 533e540540

Clinical Guideline Subcommittee on Low Back Pain. American Osteopathic Associ-ation guidelines for osteopathic manipulative treatment (OMT) for patientswith low back pain. J Am Osteopath Assoc 2010;110:653e66.

Costa LD, Koes BW, Pransky G, Borkan J, Maher CG, Smeets RJ. Primary careresearch priorities in low back pain: an update. Spine (Phila Pa 1976) 2013;38:148e56.

Deyo RA, Mirza SK, Turner JA, Martin BI. Overtreating chronic back pain: time toback off? J Am Board Fam Med 2009;22:62e8.

Dworkin RH, Turk DC, Wyrwich KW, Beaton D, Cleeland CS, Farrar JT, et al. Inter-preting the clinical importance of treatment outcomes in chronic pain clinicaltrials: IMMPACT recommendations. J Pain 2008;9:105e21.

Emshoff R, Bertram S, Emshoff I. Clinically important difference thresholds of thevisual analog scale: a conceptual model for identifying meaningful intra-individual changes for pain intensity. Pain 2011;152:2277e82.

Farrar JT, Portenoy RK, Berlin JA, Kinman JL, Strom BL. Defining the clinicallyimportant difference in pain outcome measures. Pain 2000;88:287e94.

Furlan AD, Pennick V, Bombardier C, van Tulder M. 2009 updated method guide-lines for systematic reviews in the Cochrane Back Review Group. Spine (Phila Pa1976) 2009;34:1929e41.

Ghroubi S, Elleuch H, Baklouti S, Elleuch MH. [Chronic low back pain and vertebralmanipulation]. Ann Readapt Med Phys 2007;50:570e6.

Globe GA, Morris CE, WhalenWM, Farabaugh RJ, Hawk C. Chiropractic managementof low back disorders: report from a consensus process. J Manipulative PhysiolTher 2008;31:651e8.

Harvey E, Burton AK, Moffett JK, Breen A. Spinal manipulation for low-back pain: atreatment package agreed to by the UK chiropractic, osteopathy and physio-therapy professional associations. Man Ther 2003;8:46e51.

Hennekens CH, Demets D. The need for large-scale randomized evidence withoutundue emphasis on small trials, meta-analyses, or subgroup analyses. JAMA2009;302:2361e2.

Hogan MD, Kupper LL, Most BM, Haseman JK. Alternatives to Rothmans approachfor assessing synergism (or antagonism) in cohort studies. Am J Epidemiol1978;108:60e7.

Hrobjartsson A, Gotzsche PC. Is the placebo powerless? An analysis of clinicaltrials comparing placebo with no treatment. N Engl J Med 2001;344:1594e602.

Licciardone JC, Kearns CM. Somatic dysfunction and its association with chronic lowback pain, back-specific functioning, and general health: results from theOSTEOPATHIC Trial. J Am Osteopath Assoc 2012;112:420e8.

Licciardone JC, King HH, Hensel KL, Williams DG. OSTEOPAThic health outcomes inchronic low back pain: the OSTEOPATHIC Trial. Osteopath Med Prim Care2008;2:5.

Licciardone JC, Minotti DE, Gatchel RJ, Kearns CM, Singh KP. Osteopathic manualtreatment and ultrasound therapy for chronic low back pain: a randomizedcontrolled trial. Ann Fam Med 2013;11:122e9.

Licciardone JC, Stoll ST, Fulda KG, Russo DP, Siu J, Winn W, et al. Osteopathicmanipulative treatment for chronic low back pain: a randomized controlledtrial. Spine (Phila Pa 1976) 2003;28:1355e62.

National Institute for Health and Clinical Excellence. Low back pain: early man-agement of persistent non-specific low back pain. National Health Service; 2009.

Ostelo RW, Deyo RA, Stratford P, Waddell G, Croft P, Von Korff M, et al. Interpretingchange scores for pain and functional status in low back pain: towards inter-national consensus regarding minimal important change. Spine (Phila Pa 1976)2008;33:90e4.

Roland M, Morris R. A study of the natural history of back pain. Part I: Developmentof a reliable and sensitive measure of disability in low-back pain. Spine (Phila Pa1976) 1983;8:141e4.

Rubinstein SM, van Middelkoop M, Assendelft WJ, de Boer MR, van Tulder MW.Spinal manipulative therapy for chronic low-back pain: an update of a Cochranereview. Spine (Phila Pa 1976) 2011;36:E825e46.

Von Korff M, Ormel J, Keefe FJ, Dworkin SF. Grading the severity of chronic pain.Pain 1992;50:133e49.

Vos T, Flaxman AD, Naghavi M, Lozano R, Michaud C, Ezzati M, et al. Years lived withdisability (YLDs) for 1160 sequelae of 289 diseases and injuries 1990e2010: asystematic analysis for the Global Burden of Disease Study 2010. Lancet2013;380:2163e96.

Waagen GN, Haldeman S, Cook G, Lopez D, DeBoer KF. Short term trial of chiropracticadjustments for the relief of chronic low back pain. Man Med 1986;2:63e7.

Yarlas A, Miller K, Wen W, Dain B, Lynch SY, Pergolizzi JV, et al. A randomized,placebo-controlled study of the impact of the 7-day buprenorphine trans-dermal system on health-related quality of life in opioid-naive patients withmoderate-to-severe chronic low back pain. J Pain 2013;14:14e23.