The prognostic value of catastrophizing for predicting the clinical evolution of low back pain patients: a study in routine clinical practice within the Spanish National Health Service

1 THE PROGNOSTIC VALUE OF CATASTROPHIZING FOR PREDICTING THE CLINICAL EVOLUTION OF LOW BACK PAIN PATIENTS. A s udy in ou ine clinical p ac ice wi hin he Spanish Na ional Heal h Se ice [NCT00502333]. F ancisco M. Ko acs, MD, PhD,1,2 Jesús Seco, MD, PhD,3,2 Ana Royuela, MSc,4,5,2 Josep Co coll,MD,6,2 And és Peña, MD, PhD,7,2 and he Spanish Back Pain Resea ch Ne wo k+ 1: Fundación Ko acs, Palma de Mallo ca, Spain. 2: Spanish Back Pain Resea ch Ne wo k. 3: Depa amen o de En e me ía y Fisio e apia. Ins i u o de Biomedicina. Uni e sidad de León, Pon e ada, Spain. 4: CIBER Epidemiología y Salud Pública (CIBERESP). Spain. 5: Unidad de Bioes adís ica Clínica. Hospi al Ramón y Cajal, IRYCIS, Mad id. Spain. 6: Cen o de Salud de T amun ana-Espo las, Mallo ca, Spain. 7: Se icio de Rehabili ación. Hospi al Ramón y Cajal. Mad id, Spain. +: O he membe s o he Spanish Back Pain Resea ch Ne wo k who au ho ed his s udy a e: Al onso Mu iel, MSc,4,5,2 Víc o Ab ai a,4,5,2 Ma ía Nie es Plana,4,5,2 An onio Pallice , Cen o de Salud San a Eulalia, Ibiza, Ma io Ges oso,1 Nicole Mu aggi,1 Ca los Isan a, Cen o de Salud San José No e, Za agoza, Monse a Núñez, Hospi al Clínic, Ba celona, José Luis Peña Sag edo, Hospi al Uni e si a io Ma qués de Valdecilla, San ande , Pila B ie a, Hospi al Uni e si a io Ma qués de Valdecilla, San ande , Luis Al a ez Galo ich, Fundación Jiménez Díaz, Mad id, Sal ado Fus e , Hospi al Clínic, Ba celona, Ja ie Zamo a,4,5,2 F ancisco Ja ie Cano-Ga cía, Depa amen o de Pe sonalidad, E aluación y T a amien o Psicológicos, Uni e sidad de Se illa, Daniel Bení ez Mele o, Cen o de Salud de Mon equin o, Se illa, F ancisco Ma ínez, Cen o de Salud Bembib e, León, Elena Rod íguez, Unidad Básica Bahía G and, Lluchmajo , Balea es, Rosa Vecino La o , Cen o de Salud San José No e, Za agoza, Dolo es Vázquez, Mu ua Asepeyo, Mad id, Ma iano O ega, Cen o de Salud Es T encado s, Lluchmayo , Balea es, Jenny Moix, Depa amen o de Psicología Básica, E olu i a y de la Educación, Uni e sidad Au ónoma de Ba celona, Bella e a, Spain, E a All é, Cen o de Salud San José No e, Za agoza, Te esa López- Melus, Cen o de Salud San José No e, Za agoza, Pa icia G acia, Cen o de Salud San José No e, Za agoza, Mª Dolo es Vicen e, Cen o de Salud San José No e, Za agoza, Mª Ángeles Gay, Cen o de Salud San José No e, Za agoza, Isabel To es, Cen o de Salud San José No e, Za agoza, Pila Bo az, Cen o de Salud San José No e, Za agoza, Emilio Suá ez Sánchez, Cen o de Salud de Mon equin o, Se illa, Ana Expósi o, Cen o de Salud San José No e, Za agoza, Ana He nández, Cen o de Salud San José No e, Za agoza, Glo ia Sanz, Cen o de Salud San José No e, Za agoza, Yolanda Sánchez, Cen o de Salud San José No e, Za agoza, Pila Melón Juncosa, Cen o de Salud San José No e, Za agoza, Mi eia Ba celó Cas elló, Cen o de Salud San José No e, Za agoza, Milag os Bue o Gallego, Depa amen o de Pe sonalidad, E aluación y T a amien o Psicológicos, Uni e sidad de Se illa, José Sánchez Blanco, Dis i o Sani a io Se illa Su Co esponding au ho : F ancisco M. Ko acs, MD, PhD. Depa amen o Cien í ico, Fundación Ko acs, Paseo de Mallo ca 36, 07012 Palma de Mallo ca, Spain. Tel: +34 971 720809; Fax: +34 971 720774; ko acs@ko acs.o g. All he au ho s con ibu ed subs an ially o he concep ion and design o his s udy, and o he e ision o his a icle, ake public esponsibili y o he app op ia eness o he design and me hod, and he collec ion, analysis and in e p e a ion o he da a. All he au ho s ha e e iewed and app o ed he inal e sion o his manusc ip o submission and/o publica ion. F ancisco Ko acs, Jesús Seco, Ana Royuela and Josep Co coll designed he o ms used o collec he da a, and he da abase in which hey we e in oduced o analysis. F ancisco 2 Ko acs, Jesús Seco, Josep Co coll, An onio Pallice , Ma io Ges oso, Nicole Mu aggi, Ca los Isan a, Monse a Núñez, José Luis Peña Sag edo, Pila B ie a, Luis Al a ez Galo ich, Sal ado Fus e , Daniel Bení ez Mele o, F ancisco Ma ínez, Elena Rod íguez, Rosa Vecino La o , Dolo es Vázquez, Ma iano O ega, E a All é, Te esa López-Melus, Pa icia G acia, Mª Dolo es Vicen e, Mª Ángeles Gay, Isabel To es, Pila Bo az, Emilio Suá ez Sánchez, Ana Expósi o, Ana He nández, Glo ia Sanz, Yolanda Sánchez, Pila Melón Juncosa, Mi eia Ba celó Cas elló, José Sánchez Blanco sc eened hei pa ien s, ec ui ed hose complying wi h inclusion c i e ia and ga he ed da a om hei assessmen s. F ancisco Ko acs, F ancisco Ja ie Cano-Ga cía, Jenny Moix and Milag os Bue o Gallego, ensu ed ha da a included in he da abase we e consis en wi h hose included in he o ms illed by he pa ien s. Ana Royuela, Víc o Ab ai a, Ma ía Nie es Plana, Al onso Mu iel and Ja ie Zamo a analyzed he da a. F ancisco M. Ko acs and Ana Royuela ake esponsabili y o he accu acy o da a anlaysis. F ancisco M. Ko acs w o e his manusc ip , had ull access o all o he da a in he s udy, and akes ull esponsibili y o he in eg i y o he da a and he decision o submi i o publica ion. Acknolwedgemen s: This s udy was join ly unded by he Spanish Minis y o Heal h’s Agency o Quali y and he Ko acs Founda ion, a Spanish no o p o i Ins i u ion wi h no links o he heal h indus y. The unding ins i u ions had no ole in he design and conduc o he s udy; da a collec ion; managemen , analysis and in e p e a ion o he da a; p epa a ion, e iew and app o al o he manusc ip ; o he decision o submi he a icle o publica ion. No bene i s in any o m ha e been o will be ecei ed om a comme cial pa y ela ed di ec ly o indi ec ly o he subjec o his a icle. The au ho s do no ha e con lic s o in e es o epo . 3 Abs ac Backg ound/con ex : Expe imen al s udies sugges ha ca as ophizing may wo sen he p ognosis o low back pain (LBP) and LBP- ela ed disabili y, and inc ease he isk o ch onici y. Pu pose: To assess: a) The p ognos ic alue o baseline ca as ophizing o p edic ing he clinical e olu ion o LBP pa ien s in ou ine clinical p ac ice b) The associa ion be ween he e olu ion o pain and ca as ophizing. S udy design/se ing: P ospec i e s udy in ou ine clinical p ac ice o he Spanish Na ional Heal h Se ice. Pa ien sample: 1,422 acu e and ch onic adul LBP pa ien s ea ed in p ima y and hospi al ca e. Ou come measu es: pain, disabili y and ca as ophizing, measu ed h ough alida ed ins umen s. Me hods: Pa ien s we e managed acco ding o ou ine clinical p ac ice. Ou come measu es we e assessed a baseline and 3 mon hs la e . Logis ic eg ession models we e de eloped o es ima e he associa ion be ween baseline ca as ophizing sco e and he imp o emen o LBP and disabili y, adjus ing o baseline LBP and leg pain se e i y, disabili y, du a ion o he pain episode, wo ke s’ compensa ion co e age, adiological indings, ailed back su ge y, diagnos ic p ocedu es and ea men s unde aken h oughou he s udy. Ano he model was de eloped o es ima e he associa ion be ween he e olu ion o LBP and he change in ca as ophizing, adjus ing o he same possible con ounde s plus he e olu ion o leg pain and disabili y. Models we e epea ed excluding he ea men s unde gone a e he baseline assessmen . Resul s: Reg ession models showed ha he deg ee o baseline ca as ophizing does no p edic he e olu ion o LBP and disabili y. Con e sely, as he deg ee o pain imp o emen inc eases, so does he OR o imp o emen in ca as ophizing, anging om 3 (CI 95% 2.00; 4.50, p<0.001) o imp o emen s in pain be ween 1.1 and 4 VAS poin s, o 7.3 (CI 95% 3.49; 15.36, p<0.001) o imp o emen s in pain > 6.1. Simila esul s we e ob ained when ea men s we e excluded om he models. Conclusions. In ou ine p ac ice, assessing he baseline sco e o ca as ophizing does no help clinicians o p edic he e olu ion o LBP and disabili y a 3 mon hs. Key wo ds: Low back pain, disabili y, p edic ion, ca as ophizing, ou ine clinical p ac ice. 4 In oduc ion Nonspeci ic o common low back pain (LBP) is de ined as pain be ween he cos al ma gins and he in e io glu eal olds, which may be associa ed wi h pain e e ed down o he leg (“leg pain”), and is usually accompanied by pain ul limi a ion o mo emen . Diagnosing common LBP implies ha he pain is no ela ed o condi ions such as ac u es, spondyli is, di ec auma, o neoplas ic, in ec ious, ascula , me abolic, o endoc ine- ela ed p ocesses.1 Two o he main psychological ac o s which ha e been conside ed o nega i ely in luence he p ognosis o pain and disabili y in LBP pa ien s, a e ea -a oidance belie s (FAB) and ca as ophizing.1-9 FAB e e o he ea -induced a oidance o mo emen s o ac i i ies which a e expec ed o be pain ul, whe eas ca as ophizing is de ined as an exagge a ed nega i e men al s a e ela ed o an ac ual o an icipa ed pain ul expe ience.1-9 In he Spanish cul u al en i onmen , FAB ha e shown o ha e an ei he negligible o non- exis en in luence on LBP among elde ly popula ions and among acu e, subacu e and ch onic LBP pa ien s ea ed in ou ine p ac ice,10-13 whe eas ca as ophizing co ela es wi h disabili y and explains app oxima ely one ou h o i s a iance, 13,14 sugges ing ha i may ha e an in luence on he p ognosis o LBP pa ien s. F om a heo e ical poin o iew, p e-exis ing ca as ophizing hough s may hampe pa ien s’ clinical e olu ion. Con e sely, i could also be hypo hesized ha ca as ophizing would appea o be ein o ced in pa ien s who expe ience a disappoin ing clinical e olu ion, successi e ailed ea men s and con inued pain and disabili y. This poses a “chicken and egg” dilemma, on he po en ial ecip ocal in luence be ween ca as ophizing and lack o clinical imp o emen .3,7-9 In ac , p e ious c oss-sec ional s udies ha e shown ha ca as ophizing, pain and disabili y co ela e wi h each o he ,13-18 bu esul s om p ospec i e s udies a e inconsis en . Some andomized con olled ials and small s udies in ou ine p ac ice sugges ha baseline ca as ophizing is associa ed wi h he e olu ion o pain and disabili y, some sugges he con a y, and o he s conclude ha ca as ophizing p edic s he ou come o acu e LBP, bu no be o e 6 weeks a e he onse o pain.2,4,7,19-31 Resul s om he only la ge p ospec i e s udy conduc ed in ou ine p ac ice, sugges ha baseline ca as ophizing does no p edic he e olu ion o LBP- ela ed disabili y.32 I ca as ophizing we e o ac ually ha e a nega i e in luence on p ognosis, i would ollow ha , in ou ine p ac ice, clinicians should iden i y pa ien s in whom psychological ea men o add ess ca as ophizing should be conside ed. To his end, a cu -o alue o baseline ca as ophizing, abo e which educing i would be equi ed o ea LBP success ully, should be iden i ied. The e o e, he objec i es o his s udy we e o: a) De e mine whe he assessing baseline ca as ophizing would help clinicians o p edic he clinical e olu ion o low back pain pa ien s, in ou ine clinical p ac ice, while es ablishing he cu -o poin o iden i y subjec s in whom ca as ophizing may hinde eco e y and should he e o e be ea ed, b) Assess he associa ion be ween imp o emen in pain and he e olu ion o ca as ophizing. Me hods 5 Se ing This s udy was pe o med in 14 Heal h Ca e Cen e s om 7 di e en egions in Spain. Twel e belonged o he Spanish Na ional Heal h Se ice (SNHS), and 2 o no - o -p o i Founda ions wo king o he SNHS. Pa icipa ing cen e s included 6 p ima y ca e cen e s and 8 special y cen e s in ehabili a ion, neu o e lexo he apy, o hopedic su ge y, and heuma ology. Subjec s Inclusion c i e ia we e: seeking ca e in a pa icipa ing cen e o LBP wi h o wi hou leg pain, no caused by di ec auma o sys emic diseases, no complying wi h c i e ia o e e al o su ge y, and being able o ead in Spanish. Pain no caused by sys emic diseases was de ined as pain in pa ien s who had no been diagnosed wi h cance , ib omyalgia o in lamma o y diseases, such as heuma oid a h i is o Bech e ew’s disease (Ankylosing Spondyli is), and who did no show signs sugges ing ib omyalgia o “ ed lags” o po en ial unde lying sys emic diseases. “Signs sugges ing ib omyalgia” we e de ined as di use pain wi h unexplained a igue o sleep dis u bances, and “ ed lags” o po en ial unde lying sys emic diseases we e de ined as oncologic disease du ing he p e ious 5 yea s, cons i u ional symp oms (unexplained weigh loss, e e , chills), his o y o in a enous d ug use, o immunocomp omised hos .1,33-35 C i e ia o e e al o su ge y we e de ined as signs sugges ing cauda equina synd ome, o ne e oo comp ession due o disk he nia ion o spinal s enosis po en ially quali ying o su ge y. Rele an o p og essi e pa esia, loss o sphinc e con ol o saddle anes hesia, we e conside ed as signs sugges ing cauda equina synd ome. Po en ial su gical c i e ia o disk he nia ion we e de ined as disabling scia ic pain las ing 6 weeks o mo e, caused by a comp omised ne e oo demons a ed by magne ic esonance (MRI). Po en ial su gical c i e ia o symp oma ic lumba spinal s enosis we e de ined as adicula pain las ing 3 o mo e mon hs, o claudica ion un ela ed o pe iphe al ascula disease, wi h e idence o s enosis on MRI o CT scans.1 Pa ien s who had unde gone unsuccess ul spine su ge y (“ ailed back su ge y”) and hose wi h “ ed lags” in which app op ia e es p ocedu es had uled ou sys emic diseases, we e in i ed o pa icipa e in he s udy. Exclusion c i e ia we e: ea ed o un ea ed cen al ne ous sys em impai men , e usal o sign he in o med consen , and lea ing a ques ionnai e assessing any o he a iables unanswe ed. The design o his s udy did no imply any a ia ion in he pa ien s’ clinical managemen . The e o e, as opposed o andomized clinical ials, he e we e no e hical easons o keeping he sample size as small as possible. As a esul , sample size o his s udy was es ablished a 1,500, in o de o ensu e enough s a is ical powe , gi en ha : a) p e ious s udies had sugges ed ha he po en ial e ec o ca as ophizing in Spanish 6 subjec s could be small,10,11,16 b) app oxima ely 80% o LBP pa ien s ea ed in ou ine p ac ice wi hin he Spanish Na ional Heal h Se ice, epo a clinically ele an imp o emen a 3 mon hs,36,37 and p e ious s udies ha e shown ha imp o emen s a e no no mally dis ibu ed,36,38 which excludes linea eg ession analysis and implies he need o dicho omize con inuous a iables o logis ic eg ession analyses, which in u n may educe s a is ical powe , d) his sample size would allow he in oduc ion o up o 30 a iables in he eg ession models as po en ial con ounde s.39 P ocedu e The s udy p o ocol was app o ed by he E hical Commi ees o he pa icipa ing Hospi als and ins i u ions. All pa ien s seeking ca e o LBP who we e ea ed by physicians pa icipa ing in his s udy, we e sc eened o inclusion and exclusion c i e ia. The physicians explained he s udy’s cha ac e is ics o eligible pa ien s, as well as how impo an i was o hem o ully and accu a ely answe he ques ionnai es. They inally in i ed pa ien s o sign he co esponding in o med consen . The pa ien s who signed i we e included in he s udy. Nei he pa ien s no ec ui ing physicians ecei ed any compensa ion o hei pa icipa ion in his s udy. Pa ien s we e assessed upon ec ui men and h ee mon hs la e . A bo h assessmen s, pa ien s comple ed all he sel -adminis e ed ques ionnai es by hemsel es, in p i a e. The only ins uc ions hey ecei ed, we e hose included in he s anda d alida ed e sions o he sel -adminis e ed ques ionnai es. They ecei ed no help o u he di ec ions om heal h ca e pe sonnel, esea ch s a o o he hi d pa ies. Once comple ed, he ques ionnai es we e collec ed by auxilia y pe sonnel no ela ed o he s udy. Da a we e in oduced in o a da abase a a cen al coo dina ion o ice by wo adminis a i e assis an s, who double-checked ha he da a in oduced coincided wi h a ings on he ques ionnai es. Following ou ine p ac ice condi ions, all decisions on clinical managemen , including he p esc ip ion o any kind o diagnos ic es s o ea men s, we e le up o he ea ing clinicians, and no measu es we e aken o homogenize hei c i e ia. Clinicians had access o he sco es o pain and disabili y, since hese da a can in luence hei clinical ecommenda ions, bu no o he sco e o ca as ophizing. Va iables A he i s assessmen , pa ien s we e asked o comple e ques ionnai es ga he ing da a on gende , age (da e o bi h), du a ion o cu en pain episode (days), and wo king s a us (classi ied as “no eligible”-i.e. s uden s, housewi es, unemployed, e i ed-, o “eligible o wo ke ’s compensa ion bene i s”, which in Spain can ep esen up o 100% o he sala y i espec i e o whe he he wo ke is wo king o no –i.e., wo king, on sick lea e o disabled-). In his s udy, LBP and leg pain se e i y, and LBP- ela ed disabili y we e conside ed he main indica o s o pa ien s’ clinical e olu ion.40 A bo h assessmen s, pa ien s we e asked o a e he in ensi y o low back pain (LBP), leg pain (LP), LBP- ela ed disabili y, and ca as ophizing. Pain in ensi y was measu ed wi h a 10-cm isual analog scale (VAS, o 7 which 0 = no pain and 10= wo s possible pain).41 Low back pain- ela ed unc ional disabili y was measu ed using he alida ed Spanish e sion o he Roland-Mo is ques ionnai e (RMQ),42 in which disabili y is sco ed om 0 o 24 poin s (be e o wo se). Ca as ophizing was measu ed using he ca as ophizing subscale o he alida ed Spanish e sion o he Coping S a egies Ques ionnai e, in which pa ien s’ use o ca as ophizing s a egies o cope wi h pain is sco ed om 0 (no use) o 36 (maximum possible use o hose s a egies).43 Rec ui ing physicians p o ided da a on pa ien s’ adiological indings (no indings, disc degene a ion, scoliosis, spondylolis hesis, spondylolysis, annula ea , disc p o usion, disc he nia ion, > 1 cm. di e ence in leg leng h, lumba iza ion o S1, sac aliza ion o L5, o he adiological indings), and his o y o ailed back su ge y ela ed o cu en episode (yes/no), as well as diagnos ic p ocedu es (X-Rays, scanne , MRI, EMG, blood analyses, scin ig aphy, o he ) and ea men s which he pa ien had unde gone h oughou he s udy (d ugs –NSAIDs, muscle elaxan s, o he d ugs-, physio he apy o ehabili a ion, neu o e lexo he apy (NRT) in e en ion, su ge y, o he ea men s). Analysis Absolu e and ela i e equencies we e calcula ed o ca ego ical a iables, and mean and s anda d de ia ion (SD) o con inuous ones. The cha ac e is ics o he pa ien s who imp o ed and did no imp o e we e compa ed using he Mann-Whi ney es o con inuous a iables. Ca ego ical a iables we e compa ed h ough he chi-squa ed es , o he Fishe ’s exac p obabili y es when chi-squa ed was no applicable. Imp o emen s in pain and disabili y we e de ined as any educ ion in he sco e o VAS o RMQ being highe han he minimal clinically impo an change (MCIC). P e ious s udies ha e es ablished MCIC o pain and disabili y a 30% o hei baseline sco e, wi h a minimum alue o 1.5 o VAS and 2.5 o RMQ.37 RMQ canno be sco ed wi h decimals so, in his s udy, imp o emen was de ined as “clinically ele an ” when ≥1.5 VAS poin s o ≥3 RMQ poin s. Simila ly, a “ ele an educ ion” in he CSQ sco e was de ined as any educ ion ≥ 30% o i s baseline alue.37,44 Since no da a on MCIC o CSQ a e a ailable, a sensi i i y analysis was conduc ed in which “change in he CSQ sco e” was de ined as any posi i e di e ence be ween assessmen s a baseline and 3 mon hs. Acco ding o hese de ini ions, imp o emen was impossible when baseline sco es we e ≤ 1,5 VAS poin s o pain, ≤ 3 RMQ poin s o disabili y, o ≤ 1 CSQ poin o ca as ophizing. The e o e, pa ien s wi h such a baseline sco e o a gi en a iable, we e excluded om he analysis which ocused on he imp o emen o ha a iable. Two logis ic eg ession models we e de eloped o es ima e he associa ion be ween baseline CSQ sco e and he imp o emen o LBP and LBP- ela ed disabili y du ing he s udy pe iod, adjus ing o o he possible con ounde s. In o de o elax he assump ion o linea i y among dependen a iables and baseline CSQ, baseline CSQ sco e was ca ego ized in qua iles and in oduced in o he models as dummy a iables, using he i s qua ile as he e e ence one. A he design phase o his s udy, i was decided ha all eco ded a iables ha migh exe an in luence on he e olu ion o pain and disabili y, would be included as po en ial con ounde s in he models. These included gende , age (in yea s), baseline alues o LBP (VAS poin s), leg pain (VAS poin s) and LBP- ela ed disabili y (RMQ poin s), du a ion o 8 he cu en episode (collec ed in days and classi ied as acu e –less han 14 days-, subacu e –14 o 90 days-, ch onic -91 o 365 days- and ex emely ch onic -o e 365 days-),38,45 wo ke s’ compensa ion co e age, diagnosis o “ ailed back su ge y”, adiological indings (disc degene a ion, spondylolis hesis/spondylolysis, spinal s enosis, disc p o usion/he nia, no indings), diagnos ic p ocedu es pe o med h oughou he s udy (X-Rays, MRI, CT scan, scin ig aphy, elec omyog aphy, blood analysis), and ea men s ecei ed be o e and du ing he s udy (NSAIDs, s e oids, muscle elaxan s, o he d ugs, physio he apy, ehabili a ion, NRT in e en ion, o su ge y). In o de o assess whe he ch onici y modi ied he associa ion be ween baseline CSQ sco e and he e olu ion o pain and disabili y, he in e ac ion be ween baseline CSQ sco e and ch onici y was also included in he maximal model. CSQ was o ced in o a non-au oma ic backwa d elimina ion s a egy o ien ed a p o iding a alid es ima e, so ha he a iable wi h he highes P alue ha was no a con ounde was excluded a e e y s ep.46,47 Va iables we e conside ed o be con ounde s i he es ima e o he coe icien o CSQ changed by mo e han 10% when ha a iable was emo ed om he maximal model. In o de o a oid he loss o s a is ical powe , independen con inuous a iables we e no ca ego ized.48 I was hypo hesized ha , al hough clinicians did no ha e access o pa ien s’ baseline CSQ sco es, he deg ee o baseline ca as ophizing migh ha e been linked o some pa ien s’ cha ac e is ics which, in u n, could ha e in luenced clinicians’ decisions wi h ega ds o ea men . The e o e, analyses we e epea ed excluding ea men s ecei ed du ing he ollow-up pe iod. Ano he logis ic eg ession model was de eloped o es ima e he associa ion be ween he e olu ion o LBP (ca ego ized in qua iles and in oduced in he model as dummy a iables), and he change in CSQ (de ined as any change in he sco e being ≥ 30% o baseline alue), adjus ing o o he possible con ounde s. In addi ion o all he po en ial con ounde s lis ed in he models desc ibed abo e, he maximal model also included he e olu ion o LP and disabili y. E olu ion o LBP, LP and disabili y was de ined as he baseline sco e minus he inal one, so ha posi i e alues e lec imp o emen . C i e ia used o conside a a iable as a con ounde we e he same as desc ibed abo e. To alida e hese h ee models, he sample was andomly spli in o wo sub-samples o each model. The i s one ( aining sample) included 80% o pa ien s. The second one ( alida ion sample), included he emaining pa ien s (i.e., 20%).47 Collinea i y o he maximal model in aining samples was e alua ed using he c i e ia p oposed by Belsley.49 The SPSS 17 (SPSS Inc, Chicago, IL) was used o s a is ical analysis. Resul s Thi y h ee clinicians sc eened 1,565 pa ien s. Six y- i e pa ien s declined o commi o he ollow-up isi and did no sign he in o med consen . The emaining 1,500 (95.67%) pa ien s we e included. The e we e no losses o ollow-up, bu 78 subjec s (5.2%) we e excluded a he analysis phase o ha ing le ques ionnai es on LBP (32 pa ien s), LP (47), disabili y (39), ca as ophizing (34), o wo ke s’ compensa ion co e age (4), unanswe ed. Six y- wo ou o hose 78 pa ien s le wo o mo e ques ionnai es unanswe ed. The e o e, 1,422 pa ien s we e included in he analysis. Thei mean age was 52.6 yea s, and mos we e women (62.6%) wi h ch onic LBP (58.4%). A baseline, hei mean CSQ sco e was 9 15 poin s, al hough 143 pa ien s (10.1%) had a CSQ sco e o e 75% o he maximum possible. Pa ien s’ baseline cha ac e is ics a e shown in Table 1. A baseline, 24 pa ien s had a pain sco e ≤ 1,5 VAS poin s, 77 had a disabili y sco e ≤ 3 RMQ poin s, and 6 had a sco e on he CSQ ≤ 1. The e o e, hey we e excluded om he analysis ocusing on he imp o emen o he co esponding a iable (Table 2). Pain wo sened in only 36 pa ien s (2.5% o he sample), disabili y in 40 (2.8%) and ca as ophizing in 91 (6.4%). The e o e, no sepa a e analyses we e made o hese pa ien s and hey we e included in he “did no imp o e” ca ego y o he co esponding a iable. The e we e some di e ences be ween pa ien s who expe ienced imp o emen s in pain, disabili y and ca as ophizing, and hose who did no (Table 2). Among pa ien s who imp o ed, ailed back su ge y was less common, diagnos ic p ocedu es we e less equen ly pe o med, and he ea men s p esc ibed di e ed om hose unde gone by pa ien s who did no imp o e. In addi ion, among he 1,043 pa ien s in whom pain imp o ed h oughou he s udy pe iod, baseline pain se e i y was highe , ewe we e ch onic, and ewe showed disk p o usion o he nia on MRI. Among he 879 in whom disabili y imp o ed, baseline RMQ sco es we e highe and ewe showed no adiological indings in imaging. Among he 1,010 in whom ca as ophizing imp o ed, baseline CSQ sco es we e highe , leg pain was mo e common, and ewe showed no adiological indings in imaging (Table 2). The model analyzing he associa ion be ween baseline ca as ophizing and he e olu ion o low back pain had o be adjus ed by baseline se e i y o low back pain, baseline deg ee o disabili y, and whe he he pa ien ecei ed NRT in e en ion. The aining sample included 1,103 pa ien s, and he alida ion one, 295. The e we e no collinea i y p oblems. No associa ion was ound be ween baseline CSQ sco e and he e olu ion o LBP (Table 3). This was he case o bo h he aining and alida ion samples. When ea men s we e emo ed om he model, changes in esul s we e mino (Table 3). The model analyzing he associa ion be ween baseline ca as ophizing and he e olu ion o disabili y, had o be adjus ed by baseline sco e o disabili y, and whe he he pa ien ecei ed NRT in e en ion. The aining sample included 1,088 pa ien s, and he alida ion one, 257. The e we e no collinea i y p oblems. No associa ion was ound be ween baseline CSQ and he e olu ion o disabili y (Table 4). This was he case o bo h he aining and alida ion samples. When ea men s we e emo ed om he model, changes in esul s we e mino (Table 4). The las eg ession model showed ha he e olu ion o ca as ophizing h oughou he s udy pe iod was associa ed wi h he e olu ion o LBP (Table 5). The aining sample included 1,124 pa ien s and he alida ion one, 292. As he deg ee o pain imp o emen inc eased, so did he OR o imp o emen in ca as ophizing, anging om 3 (CI 95% 2.00; 4.50, p<0.001) o imp o emen s in pain be ween 1.1 and 4 VAS poin s, o 7.3 (CI 95% 3.49; 15.36, p<0.001) o imp o emen s in pain > 6.1 VAS poin s (Table 5). The ORs (95% CI) in he alida ion sample we e simila o he ones ound in he aining sample (da a no shown). When ea men s we e emo ed om he model, changes in esul s we e mino (Table 5). The e we e no collinea i y p oblems. In he sensi i i y analysis in which “change” in he CSQ sco e was de ined as “any change”, ins ead o a a ia ion ≥ 30% o i s baseline sco e, he esul s we e i ually iden ical (Table 5). 16 back pain in gene al p ac ice. A c i e ia-based e iew o he li e a u e. Spine 1995; 20: 318-27. 36. Ko acs FM, Fe nandez C, Co de o A, Mu iel A, González-Luján L and he Spanish Back Pain Resea ch Ne wo k. Non-speci ic low back pain in p ima y ca e in he Spanish Na ional Heal h Se ice: a p ospec i e s udy on clinical ou comes and de e minan s o managemen . Biomed Cen al Heal h Se Res 2006 URL: h p://www.biomedcen al.com/1472-6963/6/57 37. Ko acs FM, Ab ai a V, Royuela A, Co coll J, Aleg e L, Cano A, Mu iel A, Zamo a J, Gil del Real MT, Ges oso M, Mu aggi N, Spanish Back Pain Resea ch Ne wo k. Minimal clinically impo an change o pain in ensi y and disabili y in pa ien s wi h nonspeci ic low back pain. Spine 2007;32:2915–2920 38. Ko acs FM, Ab ai a V, Zamo a J, Fe nández C, and he Spanish Back Pain Resea ch Ne wo k. The ansi ion om acu e o subacu e ch onic low back pain. A S udy based on de e minan s o quali y o li e and p edic ion o ch onic disabili y. Spine 2005; 30: 1786-92. 39. Ha ell FE. Mul i a iable Modeling S a egies. In: Sp inge , edi o . Reg ession Modeling S a egies. New Yo k: 2002. 60-61. 40. Deyo RA, Ba ie M, Beu skens AJ, Bomba die C, C o P, Koes B, Malmi aa a A, Roland M, Von Ko M, Waddell G. Ou come measu es o low back pain esea ch. A p oposal o s anda dized use. Spine 1998;23(18):2003-13. 41. Huskinson EC. Measu emen o pain. Lance 1974; 2: 1127-1131 42. Ko acs FM, Llobe a J, Gil del Real MT, e al. Valida ion o he Spanish e sion o he Roland Mo is Ques ionnai e. Spine 2002; 27 (5): 538-542 43. Rod íguez L, Cano FJ, Blanco A. E aluación de las es a egias de a on amien o del dolo c ónico. Ac as Esp Psiquia 2004; 32(2):82-91 44. Os elo RW, Deyo RA, S a o d P, Waddell G, C o P, Von Ko M, Bou e LM, de Ve HC. In e p e ing change sco es o pain and unc ional s a us in low back pain: owa ds in e na ional consensus ega ding minimal impo an change. Spine 2008;33(1):90-4. 45. Me skey H, Bogduk N. Desc ip ion o ch onic pain synd omes and de ini ions o pain e ms. In: Classi ica ion o Ch onic Pain, 2nd ed. Sea le: IASP P ess, 1994. 39-58 46. Kleinbaum DG, Kuppe LL, Mo gens e n H. Epidemiologic Resea ch: P inciples and Quan i a i e Me hods. New Yo k: Van Nos and Reinhold; 1982. 47. Kleinbaum DG, Kuppe LL, Mulle KE. Applied Reg ession Analysis and O he Mul i a iable Me hods, 2nd ed. Bos on: PWS-Ken ; 1988. 48. Al man DG, Roys on P. The cos o dicho omising con inuous a iables. BMJ 2006; 332(7549):1080. 49. Belsley DA. Condi ioning Diagnos ics: Collinea i y and Weak Da a in Reg ession. New Yo k: John Wiley & Sons, 1991. 50. Moix J, Ko acs FM, Ma ín A, Plana MN, Royuela A, and he Spanish Back Pain Resea ch Ne wo k. Ca as ophizing, s a e-anxie y, ange and dep essi e symp oms do no co ela e wi h disabili y when a ia ions o ai -anxie y a e aken in o accoun . A s udy o ch onic low back pain pa ien s ea ed in Spanish pain uni s [NCT00360802]. Pain Med 2011;12:1008–1017 51. an Tulde M, Becke A, Bekke ing T, B een A, Gil del Real MT, Hu chinson A, Koes B, Lae um E, Malmi aa a A, On behal o he COST B13 Wo king G oup on Guidelines o he Managemen o Acu e Low Back Pain in P ima y Ca e. Eu opean guidelines o he managemen o acu e nonspeci ic low back pain in p ima y ca e. Eu Spine J 2006;15:S169-191. 17 52. Ai aksinen O, B ox JI, Ced aschi C, Hildeb and J, Klabe -Mo e J, Ko acs FM, Mannion AF, Reis S, S aal JB, U sin H, Zanoli G, On behal o he COST B13 Wo king G oup on Guidelines o Ch onic Low Back Pain. Eu opean guidelines o he managemen o ch onic nonspeci ic low back pain. Eu Spine J 2006;15:S192-300 53. Leeuw M, Goossens ME, Lin on SJ, e al. The ea -a oidance model o musculoskele al pain: cu en s a e o scien i ic e idence. J Beha Med 2007;30(1):77-94 54. Rosens iel AK, Kee e FJ. The use o coping s a egies in ch onic low back pain pa ien s: ela ionship o pa ien cha ac e is ics and cu en adjus men . Pain 1983; 17:33-44. 55. El ing B, Ande sson T, G oo en WJ. Low le els o physical ac i i y in back pain pa ien s a e associa ed wi h high le els o ea -a oidance belie s and pain ca as ophizing. Physio he Res In 2007;12(1):14-24 56. Lin on SJ. The ole o psychological ac o s in back pain and i s emedia ion. Pain Re iews 1994;1:231–43. 57. Ande sson GBJ. Epidemiologic ea u es o ch onic low-back pain. Lance 1999; 354: 581-585. 58. an Tulde MW, Koes BW, Bou e LM. A cos -o -illness s udy o back pain in The Ne he lands. Pain 1995, 62:233–240. 59. Heymans MW, an Buu en S, Knol DL, Anema JR, an Mechelen W, de Ve HCW. The p ognosis o ch onic low back pain is de e mined by changes in pain and disabili y in he ini ial pe iod. Spine J 2010;10:847-856 60. Tu ne JA, Hol zman S, Mancl L. Media o s, mode a o s, and p edic o s o he apeu ic change in cogni i e-beha io al he apy o ch onic pain. Pain 2007;127(3):276-86 61. Vlaeyen JW, Lin on SJ. Fea -a oidance and i s consequences in ch onic musculoskele al pain: a s a e o he a . Pain 2000;85(3):317-32 62. McC acken LM, G oss RT. Does anxie y a ec coping wi h ch onic pain? Clin J Pain 1993;9(4):253-9. 18 Table 1. Baseline cha ac e is ics o pa ien s included in he s udy Va iables All pa ien s (n = 1422) Gende (males) * Age (yea s) ¤ Eligible o wo ke s’ compensa ion Du a ion o pain * Acu e (< 14 days) Subacu e (14-90 days) Ch onic: 91-365 days > 365 days Failed back synd ome * Findings on imaging es s * No indings Disc degene a ion Spondylolis hesis/spondylolysis Spinal s enosis Disc p o usion/he nia ion Annula ea > 1cm di e ence in leg leng h Lumba iza ion o S1 Sac aliza ion o L5 O he adiological indings Diagnos ic p ocedu es * X-Rays Magne ic Resonance CT-scan Elec omiog aphy Blood analysis Scin ig aphy D ug ea men * NSAIDs Co icoids Muscle elaxan s O he Physio he apy o ehabili a ion * Neu o- e lexo he apy * Su ge y * O he ea men s* LBP (VAS) ¤ LP (VAS) (n= 1098) ¤ Disabili y (RMQ) ¤ Ca as ophizing (CSQ) ¤ 532 (37.4) 52.6 (15.0) 824 (58.0) 113 (7.9) 479 (33.7) 553 (38.9) 277 (19.5) 21 (1.5) 255 (17.9) 890 (62.6) 91 (6.4) 118 (8.3) 485 (34.1) 3 (0.2) 12 (0.8) 14 (1.0) 9 (0.6) 15 (1.1) 359 (25.2) 355 (25.0) 50 (3.5) 58 (4.1) 53 (3.7) 3 (0.2) 806 (56.7) 96 (6.8) 274 (19.3) 139 (9.8) 176 (12.4) 1242 (87.3) 8 (0.6) 15 (1.1) 6.7 (2.1) 6.2 (2.4) 12.8 (5.6) 15.0 (8.5) * F equency (%) ¤ Mean (SD) LBP: Se e i y o low back pain LP: Se e i y o e e ed pain down in o he leg (leg pain) (in he 1098 pa ien s who had i ) RMQ: Sco e in he Roland-Mo is Ques ionnai e CSQ: Sco e in he Coping S a egies Ques ionnai e 19 Table 2. Baseline cha ac e is ics o pa ien s in whom low back pain, disabili y and ca as ophizing imp o ed and did no imp o e h oughou he s udy. Va iable LBP imp o ed (n =1043) † LBP did no imp o e (n =355) † p Disabili y imp o ed (n =879) ‡ Disabili y did no imp o e (n =466) ‡ p Ca as ophizing imp o ed (n =1010) ¥ Ca as ophizing did no imp o e (n =406) ¥ p Gende (males) * Age (yea s) ¤ Eligible o wo ke s’ compensa ion Ch onici y * < 14 days 14-90 days 91-365 days > 365 days Failed back su ge y * Findings on imaging es s * Disc degene a ion Spondylolis hesis/spondylolysis Spinal s enosis Disc p o usion/he nia Annula ea > 1cm di e ence in leg leng h Lumba iza ion o S1 Sac aliza ion o L5 No inding Diagnos ic p ocedu es * X-Rays Magne ic Resonance CT scan Elec omiog aphy Blood analysis Scin ig aphy D ug ea men * NSAIDs Co icoids Muscle elaxan s O he Physio he apy o ehabili a ion * Neu o- e lexo he apy * Su ge y * O he ea men s* Se e i y o LBP (VAS) ¤ Pa ien s wi h LP* Se e i y o LP (VAS)¤ Disabili y (RMQ) ¤ Ca as ophizing (CSQ) ¤ 391 (37.5) 52 (41;63) 598 (57.3) 91 (8.7) 362 (34.7) 403 (38.6) 187 (17.9) 8 (0.8) 662 (63.5) 71 (6.8) 79 (7.6) 330 (31.6) 2 (0.2) 8 (0.8) 9 (0.9) 6 (0.6) 177 (17.0) 223 (21.4) 230 (22.1) 23 (2.2) 25 (2.4) 22 (2.1) 2 (0.2) 584 (56.0) 51 (4.9) 183 (17.5) 100 (9.6) 101 (9.7) 975 (93.5) 3 (0.3) 7 (0.7) 7 (6;8) 775 (74.4) 7 (5;8) 14 (0;18) 16 (9;22) 126 (37.2) 52 (41;64) 207 (58.6) 18 (5.1) 105 (29.6) 146 (41.1) 86 (24.2) 13 (3.7) 221 (62.3) 19 (5.4) 38 (10.7) 145 (40.8) 0 (0) 3 (0.8) 4 (1.1) 2 (0.6) 70 (19.7) 127 (35.8) 114 (32.1) 26 (7.3) 28 (7.9) 28 (7.9) 0 (0) 209 (58.9) 41 (11.5) 84 (23.7) 39 (11.0) 70 (19.7) 258 (72.7) 5 (1.4) 6 (1.7) 6 (4;8) 247 (69.6) 1 (0;.6) 12 (8;16) 13 (7;19) 0.501 0.983 0.668 0.006 <0.001 0.681 0.335 0.066 0.002 1.000 1.000 0.749 1.000 0.241 <0.001 <0.001 <0.001 <0.001 <0.001 1.000 0.344 <0.001 0.011 0.447 <0.001 <0.001 0.029 0.106 <0.001 0.078 0.154 0.451 0.443 333 (37.9) 52 (41;63) 519 (59.0) 71 (8.1) 307 (34.9) 341 (38.8) 160 (18.2) 8 (0.9) 546 (62.1) 55 (6.3) 74 (8.4) 298 (33.9) 2 (0.2) 7 (0.8) 8 (0.9) 4 (0.5) 142 (16.2) 178 (20.3) 206 (23.4) 18 (2.0) 21 (2.4) 17 (1.9) 2 (0.2) 492 (56.0) 42 (4.8) 169 (19.2) 84 (9.6) 86 (9.8) 832 (94.7) 3 (0.3) 5 (0.6) 7 (5;8) 647 (73.7) 6 (2;8) 14 (10;18) 15 (9;21) 167 (35.8) 53 (43;65) 255 (55.0) 30 (6.4) 143 (30.7) 193 (41.4) 100 (21.5) 13 (2.8) 314 (67.4) 33 (7.1) 41 (8.8) 169 (36.3) 1 (0.2) 3 (0.6) 6 (1.3) 5 (1.1) 88 (18.9) 157 (33.7) 134 (28.8) 28 (6.0) 32 (6.9) 31 (6.7) 1 (0.2) 278 (59.7) 47 (10.1) 94 (20.2) 54 (11.6) 78 (16.7) 365 (78.3) 5 (1.1) 9 (1.9) 7 (5;8) 348 (74.7) 5 (2;8) 12 (8;16) 14 (8;20) 0.460 0.071 0.149 0.177 0.008 0.056 0.561 0.813 0.386 1.000 1.000 0.576 0.290 0.026 <0.001 0.033 <0.001 <0.001 <0.001 1.000 0.194 <0.001 0.677 0.243 <0.001 <0.001 0.134 0.025 0.290 0.694 0.481 <0.001 0.028 378 (37.4) 53 (42;64) 585 (57.9) 82 (8.1) 341 (33.8) 394 (39.0) 193 (19.1) 8 (0.8) 651 (64.5) 63 (6.2) 84 (8.3) 333 (33.0) 1 (0.1) 9 (0.9) 7 (0.7) 6 (0.6) 160 (15.8) 213 (21.1) 226 (22.4) 21 (2.1) 21 (2.1) 17 (1.7) 1 (0.1) 550 (54.5) 44 (4.4) 163 (16.1) 93 (9.2) 92 (9.1) 947 (93.8) 2 (0.2) 6 (0.6) 7 (5;8) 754 (74.7) 5 (1;8) 13 (9;17) 15 (8;21) 152 (37.4) 50 (40;62) 235 (58.2) 30 (7.4) 137 (33.7) 157 (38.7) 82 (20.2) 12 (3.0) 238 (58.6) 28 (6.9) 34 (8.4) 151 (36.9) 2 (0.5) 2 (0.5) 4 (1.0) 2 (0.5) 94 (23.2) 141 (34.7) 126 (31.0) 28 (6.9) 35 (8.6) 34 (8.4) 2 (0.5) 254 (62.6) 51 (12.6) 109 (26.8) 46 (11.3) 81 (20.0) 295 (72.7) 6 (1.5) 9 (2.2) 7 (5;8) 277 (78.2) 5 (2;7) 13 (8;17) 14 (8;20) 0.996 0.035 0.932 0.944 0.002 0.040 0.647 0.972 0.130 0.199 0.738 0.522 0.818 0.001 <0.001 <0.001 <0.001 <0.001 <0.001 0.199 0.005 <0.001 <0.001 0.225 <0.001 <0.001 0.009 0.017 0.009 0.013 0.911 0.970 <0.001 * F equency (%) ¤ Median (p25;p75) LBP: Low back pain LP: Re e ed pain down in o he leg (in he 1098 pa ien s who had i ) RMQ: Sco e in he Roland-Mo is Ques ionnai e CSQ: Sco e in he Coping S a egies Ques ionnai e † Only includes pa ien s whose pain se e i y a baseline was high enough o allow o a clinically ele an imp o emen (i.e., baseline VAS > 1.5). ‡ Only includes pa ien s whose disabili y a baseline was high enough o allow o a clinically ele an imp o emen (i.e., baseline RMQ > 3). ¥ Only includes pa ien s whose ca as ophizing a baseline was high enough o allow o imp o emen (i.e., baseline CSQ ≥ 1). 20 Table 3. Associa ion be ween ca as ophizing a baseline, and he e olu ion o he se e i y o low back pain h oughou he s udy pe iod (3 mon hs). C ude analysis Adjus ed analysis** OR (CI 95%) p OR (CI 95%) P Ca as ophizing a baseline* Q1 (≤8) Q2 (9-15) Q3 (16-21) Q4 (≥22) Re . ca 1.42 (0.99; 2.02) 1.13 (0.79; 1.61) 1.28 (0.88; 1.85) 0.253 0.055 0.502 0.197 Re . ca 0.87 (0.58; 1.31) 0.66 (0.43; 1.01) 0.63 (0.39; 1.02) 0.163 0.509 0.056 0.060 *: Ca ego ized in qua iles. **: Va iables included in he model we e: gende , age, baseline alues o LBP, leg pain and LBP- ela ed disabili y, du a ion o he cu en episode (acu e, subacu e, ch onic and ex emely ch onic), he in e ac ion be ween baseline CSQ and ch onici y, wo ke s’ compensa ion co e age, diagnosis o “ ailed back su ge y”, adiological indings, diagnos ic p ocedu es pe o med h oughou he s udy, and ea men s ecei ed. Resul s had o be adjus ed only by baseline se e i y o low back pain, baseline deg ee o disabili y, and ha ing unde gone neu o- e lexo he apy in e en ion. OR’s o hese a iables a e no shown, since hey may ha e been con ounded by unknown a iables which we e no con olled o .24 In he model in which p esc ibed ea men s we e emo ed om he models, only baseline se e i y o low back pain showed a con ounding e ec , and changes in esul s we e mino [adjus ed global p: 0.080, and adjus ed OR (CI95%) o Q2, Q3 and Q4: 1.16 (0.80;1.69), 0.81 (0.55;1.18) and 0.71 (0.47;1.07), espec i ely]. 21 Table 4. Associa ion be ween ca as ophizing a baseline, and he e olu ion o disabili y h oughou he s udy pe iod (3 mon hs). C ude analysis Adjus ed analysis** OR (CI 95%) p OR (CI 95%) P Ca as ophizing a baseline* Q1 (≤8) Q2 (9-15) Q3 (16-20) Q4 (≥21) Re . ca 1.43 (1.04; 1.97) 1.71 (1.21; 2.41) 1.73 (1.24; 2.40) 0.003 0.028 0.002 0.001 Re . ca 0.87 (0.61; 1.23) 0.76 (0.51; 1.12) 0.64 (0.43; 0.96) 0.167 0.428 0.167 0.029 *: Ca ego ized in qua iles. **: Va iables included in he maximal model we e: gende , age, baseline alues o LBP, leg pain and LBP- ela ed disabili y, du a ion o he cu en episode (acu e, subacu e, ch onic and ex emely ch onic), he in e ac ion be ween baseline CSQ and ch onici y, wo ke s’ compensa ion co e age, diagnosis o “ ailed back su ge y”, adiological indings, diagnos ic p ocedu es pe o med h oughou he s udy, and ea men s ecei ed. Resul s had o be adjus ed only by baseline deg ee o disabili y, and ha ing unde gone neu o- e lexo he apy in e en ion. OR’s o hese a iables a e no shown, since hey may ha e been con ounded by unknown a iables which we e no con olled o .24 In he model in which p esc ibed ea men s we e emo ed om he models, only baseline se e i y o disabili y showed a con ounding e ec , and changes in esul s we e mino [adjus ed global p: 0.459, and adjus ed OR (CI95%) o Q2, Q3 and Q4: 1.14 (0.82;1.60), 1.01 (0.70;1.47) and 0.85 (0.58;1.25), espec i ely] 22 Table 5. Associa ion be ween he e olu ion o he se e i y o low back pain and he imp o emen o ca as ophizing.† C ude analysis Adjus ed analysis** OR (CI 95%) p OR (CI 95%) P E olu ion o se e i y o LBP* Q1 (≤1.0) Q2 (1.1-4.0) Q3 (4.1-6.0) Q4 (≥6.1) Re . ca 4.97 (3.54; 6.97) 10.61 (6.92; 16.27) 35.87 (18.65; 68.98) <0.001 <0.001 <0.001 <0.001 Re . ca 3.00 (2.00; 4.50) 3.18 (1.91; 5.29) 7.32 (3.49; 15.36) <0.001 <0.001 <0.001 <0.001 *: Ca ego ized in qua iles. The e olu ion o LBP was de ined as he baseline sco e minus he inal one, so ha posi i e alues e lec imp o emen . **: Va iables included in he maximal model we e: gende , age, baseline alues o LBP, leg pain and LBP- ela ed disabili y, du a ion o he cu en episode (acu e, subacu e, ch onic and ex emely ch onic), he in e ac ion be ween baseline CSQ and ch onici y, wo ke s’ compensa ion co e age, diagnosis o “ ailed back su ge y”, adiological indings, diagnos ic p ocedu es pe o med h oughou he s udy, ea men s ecei ed, he e olu ion o e e ed pain and he e olu ion o disabili y. Resul s had o be adjus ed only by baseline deg ee o disabili y, imp o emen in disabili y h oughou he s udy pe iod, and ha ing unde gone neu o- e lexo he apy in e en ion. OR’s o hese a iables a e no shown, since hey may ha e been con ounded by a iables which we e no con olled o .24 † De ining “imp o emen ” as any educ ion in he CSQ sco e being ≥ 30% o i s baseline alue. In he sensi i i y analyses in which “imp o emen ” was de ined as “any educ ion in he CSQ sco e” – ega dless o i s size-, esul s had o be adjus ed only by imp o emen in disabili y h oughou he s udy pe iod and adjus ed ORs (CI 95%) o Q2, Q3 and Q4 we e: 2.87 (1.91;4.32), 4.04 (2.28;7.16) and 5.49 (2.46;12.24), espec i ely. In he model in which p esc ibed ea men s we e emo ed om he models, only baseline deg ee o disabili y and imp o emen in disabili y h oughou he s udy pe iod showed a con ounding e ec , and changes in esul s we e mino [adjus ed global p: <0.001, and adjus ed OR (CI95%) o Q2, Q3 and Q4: 3.43 (2.30;5.10), 4.06 (2.48;6.66) and 9.58 (4.62;19.84), espec i ely]