Spine

SPINE Volume 39, Number 3, pp 263-273 ©2014, Lippincott Williams & Wilkins

LITERATURE REVIEW

Influence of Catastrophizing on Treatment Outcome in Patients With Nonspecific Low Back Pain A Systematic Review Maria M. Wertii, MD,*+ Jakob M. Burgstaller, DMD, MD,* Sherri Weiser, PhD,t Johann Steurer, MD,* Reto Kofmehl, BSc,* and Ulrike Held, PhD*

Study Design. Systematic review. Objective. The aim of this study was to assess the effect of catastrophizing on treatment efficacy and outcome in patients :reated for low back pain. Summary of Baciiground Data. Psychological factors including ;atastrophizing thoughts are thought to increase the risk for chronic ow back pain. The influence of catastrophizing is debated. Methods. In September 2012, the following databases were searched: BIOSIS, CINAHL, Cochrane Library, EMBASE, OTseeker, 'eDRO, PsyclNFO, MEDLINE, Scopus, and Web of Science. For 50 3f 706 references, full text was assessed. Results based on 11 studies ivere included in this analysis. Results. In the 11 studies, a total of 2269 patients were included, even studies were of good and 4 of moderate methodological quality. Heterogeneity in study settings, treatments, outcomes, înd patient populations impeded meta-analysis. Catastrophizing ît baseline was predictive for disability at follow-up in 4 studies ind for pain in 2 studies. Three studies found no predictive effect 3f catastrophizing. A mediating effect was found in all studies n = 5) assessing the impact of a decrease in catastrophizing during reatment. A greater decrease was associated with better outcome, vlost studies that investigated the moderating effects on treatment

=rom the *Horten Centre for Patient Oriented Research and Knowledge Transfer, Department of Internal Medicine, University of Zurich, Pestalozzisrasse, Zurich, Switzerland; and tNYU Hospital for loint Diseases, Dccupational and Industrial Orthopaedic Center (OIOC), New York Jniversity, New York, NY. \cknowledgment date: August 27, 2013. First revision date: October 1, 2013. Acceptance date: October 28, 2013. The manuscript submitted does not contain information about medical levice(s)/drug(s). \lo funds were received in support of this work. relevant financial activities outside the submitted work. \ddress correspondence and reprint requests to Jakob M. Burgstaller, DMD, •AD, Horten Centre for Patient Oriented Research and Knowledge Transfer, Department of Internal Medicine, university of Zurich, Pestalozzistrasse 24, Í091 Zurich, Switzerland; E-mail: [email protected] DOI: 10.1097/BRS.0000000000000110 ipine

efficacy found no effect (n = 5). However, most studies did not look for a direct interaction between the treatment and catastrophizing thoughts. No study investigated the influence of catastrophizing on work-related oufcomes including return to work. Conclusion. Catastrophizing predicted degree of pain and disability and mediated treatment efficacy in most studies. The presence of catastrophizing should be considered in patients with persisting back pain. Limited evidence was found for the moderating effects on treatment efficacy. Future research should aim to clarify the role of catastrophizing as a moderator of outcome and investigate its importance for work-related outcomes. Key words: low back pain, back pain, catasfrophizing, fear avoidance, fear avoidance beliefs, fear avoidance model, prognosis, outcome, treatment outcome, mediator, moderator, predictor.

Level of Evidence: 1 Spine 2014;39:263-273

P

atients' attitudes and coping mechanisms have been shown to play a causal role in the chronification of low back pain (LBP). Almost all adults once in their lifetime complain about LBP, but only 10% to 15% develop chronic LBP.' This small percentage of patients accounts for threequarters of the costs of medical care and lost productivity associated with LBP.-'' There is consensus among experts to avoid unnecessary investigation and overtreatment of patients with acute LBP by treating symptomatically with encouragement to return to normal activity.'' Persisting pain for several weeks strongly predicts the development of chronic LBP, a condition where complete recovery and return to full physical function are often difficult to achieve."^ Current research aims to identify risk indicators for delayed recovery in patients with subacute LBP to optimize treatment and avoid chronification. Targeted and timely interventions in patients at risk for chronic pain facilitate recovery and may reduce health care costs.'' The fear avoidance model is a theoretical model that describes how psychological factors affect the experience of pain and the development of chronic pain and disability.^ www.spinejournal.com

263

Spine

LITERATURE REVIEW

Within this theoretical model, the presence of catastrophizing thoughts or behavior is a prerequisite for poor outcome and is defined as "an exaggerated negative mental set brought to bear during actual or anticipated painful experience."•* It is theorized that negative beliefs about pain and/or negative illness information leads to a catastrophizing response in which patients imagine the worst possible outcome. This leads to fear of activity and avoidance that in turn causes disuse and resultant distress, reinforcing the original negative appraisal in a deleterious cycle.^ In chronic cases, catastrophizing may become a cognitive coping strategy based on the patient's characteristic coping style or because catastrophizing is thought to have prevented severe pain or other aversive outcomes in the past.' The fear avoidance model suggests that patients without catastrophizing and fear avoidance beliefs (FAB) are more likely to confront pain problems and are more active in the coping process. This type of "good" coping has been used to develop interventions for those high in catastrophizing and FAB. Although there is some empirical support for the fear avoidance model, it is a matter of debate as to how and when to best assess catastrophizing in clinical practice. Current treatment guidelines for LBP recommend the timely identification and initiation of multidisciplinary treatment for other psychological factors {e.g., depression, distress, job dissatisfaction) associated with increased risk for delayed recovery.•'''°'" Whether catastrophizing influences treatment outcome in patients with LBP remains unclear. To date, the role of catastrophizing on treatment efficacy in LBP has not been reviewed systematically. The aim of this review is to assess the influence of catastrophizing on treatment response in randomized controlled trials (RCTs) in patients with LBP.

Influence of Catastrophizing • Wertii et al

Eligibility Criteria

AH RCTs were considered eligible that met the following criteria: they reported results concerning patients seeking care for LBP, they assessed the influence of catastrophizing on treatment outcome, and they were published between January 1980 and September 2012. We focused on RCTs with at least 30 patients per group because of a concern about sample size. Assuming a reduction in perceived disability that was onethird greater in the treatment group than the reference group, a sample size of 37 patients per group would be sufficient to detect the difference in allowing a drop-out rate of 15% (a, 0.80, P = 0.05). No limits for the study setting or language of the publication were applied. Excluded were the reports from conference proceedings.

Study Selection, Data Extraction, and Synthesis

The bibliographic details of all retrieved articles were stored. Two reviewers (M.W. and J.B.) independently screened all references by title and abstract and reviewed full texts in all studies that met the predefined eligibility criteria. Disagreements were discussed and resolved by consensus or by thirdparty arbitration (S.W.). Alternative researchers with specific language proficiencies were approached for non-English language references.

Outcome Definition

All investigated outcomes were extracted and categorized into work-related {e.g., sick days, employment) and nonwork-related outcomes {e.g., pain, perceived disability). Each method of outcome measurement was appraised with regard to their validity and reliability and was operationalized {e.g. perceived disability measured by Oswestry Disability Index).

MATERIALS AND METHODS

Quality Assessment

This systematic review follows the recommendation of the preferred reporting items for systematic reviews and metaanalyses statement (Figure 1) on conducting systematic reviews of RCTs.'-

The internal validity of each study was assessed using the Scottish Intercollegiate Guidelines Network (SIGN) Methodology checklist for RCTs by the 2 reviewers independently (M.W. and J.B.).'^ Quality was rated as follows: "High (-1- -1-)" denoted that most of the criteria have been fulfilled, If not fulfilled, the conclusions of the study are very unlikely to alter. "Moderate (-1-)" denoted that some criteria were fulfilled. Criteria not adequately described are unlikely to altei the conclusions. "Low ( —)" denoted that few or no criteria were fulfilled. The conclusions are likely to alter. As recommended by Scottish Intercollegiate Guidelines Network, studies rated by both reviewers as low quality were excluded from further analysis.

Literature Search We identified all RCTs meeting our eligibility criteria published between January 1980 and September 2012. The following databases were search by an experienced librarian (M.G.): BIOSIS, CINAHL, Cochrane Library, FMBASF, OTseeker, PeDRO, PsycINFO, MEDLINE, Scopus, and Web of Science. Search terms for catastrophizing were identified in the literature {e.g.., catastrophising, catastrophization, catastrophisation). Two detailed search strategies are depicted in Supplemental Digital Content, Appendix 1 available at http://links.lww.com/BRS/A846. To ensure the completeness of the literature search, 1 reviewer (J.B.) conducted an electronic hand search of the 6 most often retrieved journals and added all potentially eligible references not retrieved by the systematic search. In addition, bibliographies of included studies relevant to the research question were searched and potential eligible references included in the full text review (inclusion and exclusion criteria applied). 264

www.spinejournal.com

Operationalization of Catastrophizing as Predictor, Mediator, and Moderator

The definitions for predictor, mediator, and moderator were adopted from the study by Pincus et al,^'* which are listed i the following text.

• Predictor: Baseline catastrophizing affects outcome but does not interact with the allocated treatmeni intervention.

February 201':

LITERATURE REVIEW

Influence of Catastrophizing • Wertli et al

Number of records identified through database search:

Hand search ofthe journals most often publishing about the topic: n = 35 Screen of Bibliographies: n = 20

n = 1473 n = 55

Number of records after removal of duplicates: BIOSIS CINHAHL Chochrane Library EMBASE OT Seeker PeDRO PsyclNEO PubMed/MEDLINE Scopus Web of Science

n= 2 n =8 n =3 n = 51 n= 2 n =2 n = 35 n = 174 n= 8 n = 366

n = 651

Number of records screened for title and abstract: n = 706

Number of full-text articles exiuded: Number of fijll-text articles assessed for eligibility: n = 50

No influence of catastrophizing Study design other than RCT Patients otherthanLBP Patients < 30

n = 14 n = 12 n=10 n =1

n = 37

Number of RCTs included in the analysis:

Figure 1. Exclusion criteria.

13 publications based on 11 RCTs

Mediator: Change in catastrophizing during treatment impacts outcome, with or without interacting with allocated treatment. Moderator: Catastrophizing at haseline interacts with treatment. The quality of the moderator analysis was assessed for ;ach study hy 2 reviewers (M.W. and J.B.) and discussed with m experienced statistician (U.H.). The following factors were :onsidered: (1) Was the analysis a priori defined; (2) Was the selection of factors for the analysis clinically plausihle; ¡3) Were moderators measured prior to randomization; and [4) Was there an adequate quality of measurement of haseine factors, that contains an explicit test of the interaction between moderator and treatment?

Psychometric Properties and Description of the Questionnaires The pain catastrophizing scale (PCS) consists of 13 questions."*"' The score is a sum of all 13 items (each item on a scale of 0 ^ , range, 0-52). The higher the score, the more the presence of catastrophizing thoughts. The internal consistency is high (Cronhach a = 0.87-0.95).'*""* The 3 catastrophizing subscales are: rumination (sum of items 8, 9, 10, 11; range, 0-16), magnification (sum of items 6, 7, 13; range, 0-12), and helplessness (sum of items 1, 2, 3, 4, 5, 12; range, 0-24). The internal consistency is moderate to high (Cronbach a: rumination 0.87-0.95, magnification 0.66-0.88, helplessness The coping strategies questionnaire (CSQ) consists of a 48-item checklist assessing 6 cognitive and 2 behavioral www.spinejournal.com

265

Spine

LITERATURE REVIEW

coping strategies.' Six questions assess catastrophizing (items 5, 12, 14, 28, 38, 42). The score is computed by summing responses to the 6 items (each item is scored 0-6 points, range 0-36). Internal consistency and reliability in a LBP population was good in all subscales (Cronbach a, between 0.71 and 0.85).' The Cronbach a of the catastrophizing subscale was between 0.78' and 0.84." The pain-related self-statements scale (PRSS) is intended to assess situation-specific cognitions that either promote or hinder attempts to cope with pain.'" Catastrophizing is assessed with the items 2, 4, 7, 9, 10, 13, 15, 16 and has been shown to be reliable and valid (Cronbach a, 0.83).'" Items are scored on a Likert scale (0-5 points); the score is the average of all items (range, 0-5). Higher values indicate more catastrophizing. The pain cognition list (PCL) is a 50-item scale that measures a verbal-cognitive response system of chronic pain.^' Catastrophizing is measured by 17 items (each item scored on a 5-point Likert scale; 1: highly disagree to 5: totally agree). A sum score is obtained per subscale for each patient. The Catastrophizing subscale (range, 17-85) has been shown to be reliable and valid (Cronbach a 0.88).^''2^ The pain coping and cognition list is a 42-item self-report questionnaire, developed on the basis of the PCL, CSQ, and MPLC (multidimensional pain locus of control questionnaire) covering attributions, expectancies, and cognitive coping strategies. Each item is scored on a 6-point Likert scale (1: totally disagree to 6: totally agree). Catastrophizing is covered by 1 of the 4 subscales (12 items). The internal consistencies of the catastrophizing subscale proved to be good (Cronbach a O.S5):-^'^^ The PCS and the CSQ are considered to be equally reliable and valid for the measurement of catastrophizing thoughts.^^'^* It has therefore been proposed to use the PCS in research that aims to explore catastrophizing.'"^ The PRSS is considered to be more pain-specific when compared with the CSQ. A direct comparison of the PRSS and the CSQ showed a moderately strong linear relationship between the 2 scales (r = 0.56).'" The correlation between the Catastrophizing subscale of the PCL and the CSQ or PCS was high (r = 0.70).^^ Statistical Analysis Because of heterogeneous study populations, measurements, and scales used as well as outcomes investigated, only descriptive statistics were used to summarize findings across all cohort studies. Forest plots were generated on the basis of values reported using R statistical software for Windows (Statistics Department of the University of Auckland, Auckland, New Zealand).^^

Influence of Catastrophizing • Wertli et al

are summarized in Figure 1. In total, 13 publications based on 11 RCTs were included in the analysis.

Study Characteristics RCTs conducted in a general practitioner setting (n = 3), in rehabilitation clinics (n = 3), hospitals/specialists (n = 3), and physical therapy outpatient clinics (n = 2) (baseline characteristics in Table 1). The study quality was good in 7 and moderate in 6 studies (Supplemental Digital Content, Appendix 2 available at http://links.lww.com/BRS/A846). The primary outcome in most RCTs was self-report measurements (i.e.. pain, disability, change in pain, or disability). No RCT investigated return to work or other work-related outcomes. Five publications (4 RCTs) used the CSQ for assessing catastrophizing, ^«-^^ 3 the PCS (2 RCTs),'-'-'-' 2 the PCL,^''-'^ 2 the PRSS Catastrophizing subscale,^^-^' and 1 the PCCL.^'' Cutoff values were only applied once (median split > 1 1 , 0)

R¿

fO Q;

S.

A

Z

z

,t^ ' ^ "D ro

Cii

Q



C

c •^

"D

ro c

ro

ro aJ O O

û_

o

i/i

1

_aj_

ro

ëu

eu

I

c rc

eji

"rô 'S 3; eu 00 OOO

s° u

ro

OJ _^ fN

^

ro cb

S



if

200

_C

'S cto

^

fO

"rô S m

?nec 201

ro

S _^

s meef; 200

eu

rO

Hem

c

Hem 200

udy

"rô

OJ

CO

esse 200

ro

E

002

9 (49)

49.9 (

3^

ro

ro

yr (9.4

PL

m II

£-

y

D.

01

ed via 9 nt faciliti newspap id, freafe 4 outpat ient rehabilit centers. the 1Nefherla

— O' Z)

ln Q



a o "ô-— _ •— o

eN

i!P!

'c

U

.SE/ 65)

"rô

LO

eN O

II c

fena

tme

u ro 'c "Q. II ro ¿J

2^ D .

13 n D . eu

— , ro

o XI

o

U) O,

lafio

C

co

63)

la-1-

(12

(10

eN

y

e3i

ô

LO rvj

ro

ro

c n

o

1

52 wk

LO

C; c

|i

E •a ro

1

"1

Posttreatmei (13 wk

o

:^

ro *--

m

cy OD O O 01 CN 01 1

www,spinejournal,com

267

LITERATURE REVIEW

u ro

o ^

Al 5 ro

ui Z

Z

g/-e

0;

o AI D.

m1

m

Z

• ^

ti. t :

268

www.spinejournal.com

land

tation Is

S5/on;

elect

eatr

trans erani de;rs; n

tic rce

eri( ills tes e indi nitive •r se

tition er, lera

•3 0

eu

•5

1'1t

c TO

+

0

u

u

•g

s'

i 'S0 "0

u QJ

Ci,

/¡I

"o

j ;

1^

.0 ^^

Q

r-

o o

QJ CN

0 3 — •

.^

ra

a

fo

eu

CT*

1

Cr

0

E

IQR, indic

re

"S

AM,

Oos

Stu
g

tñ -C eu n

;rred

'S

.S



c/) re - ^

tients medic Hoen Cente

"rr ni ^

CO

o o

ecrijitr

.£3

îrred speci; jtpatii ers, tl^

5^ z

,0)

£ DJO

'c ::^ D Z

c

1-

U

lients

C

for or thePt

1

1^ n ^?

ntre

iali S t f o

îrred

Setti

(D

E

B

eu Oi

c

'(M)l

'S

c

tu —

D pasi

ia

ra

1

QJ

_^ c

IB,

n

eu D.

DC

c

pnoq



eu

m

•- t§ "? c Z

exCQ

-S

e stimula-

1, pain •eale; )roi

^ pan

•cor ris ( Q

tj

c"

u

.0

//egí

ra

7;

OJ

%

Uj

11

WLC 1 f/on;

O

^ -c ^ . ^ .S

'£ ^

ica

¡cal

ion

.E c

CD

m

"S s 2' j ; LBR, specifi

Q. CO

Í7Í Z

ej

., c o î! O 're re TO g Q. cjo

"S

0

uó u_ >:,CQ

0

S

S

•i

•¿

re¿

AI

m Al

5c

pai

Ó

l -

•Q S r~ :a

,

eu

mo, eau; rk or solic •ntion

o

i_

P ± referí

Hit

O

ra

5t 100 m '

Oiagno

o

CD

1

pu.

U

_y ^-^ e/5

terr

ria Oj

u

..J"

S a ~'_

1

"S -5

101

îO

'S E ï

2

.r-

0/:

c

1

'~ tu



inter :riteri



(valu

ro ro

DJO

c ^

eu ' ^

q

O

EX.

lizir

posur

c ro oc

^• S -2

is ^ 1

vi

.^ ai



C 'isi (—

^

Q (—^

|do.

More c:ata with PDS PSC

"LO

associa d disab ups (AP OJO C

OJO ' ¡ ~ :

/ii

Q UJ

C

c ro ^ ^ Ñ c eN

C CU

f^

tx

C;



Qj

.C

i£ S

Ö

1¿ _o ^

CL

o o eN

CM

ro

lore t

-C

ro "oj

Leeu'

^ C

.^

o a; o

o o

O

"ro

"TO

01

eu

.2 ^ ro >—

ll cai ultidi iin-n QJ

o c

a;

a;

1

> eN ti

^ u Oj ' ^

VIVO 1

ro

t investigatec n of the Wes t Have dence intervi ll; PRS nt; n.r., not i

o

5 ^•g

Ü

1 /nd/cate

P

CO O O

'CS 0-52, CS:QO-3

1/5

3

PCC

PCS:

u

,E|

0..

etal,

eN

(S

SpinI

o^ PCL:

(12 ro

PCL:



¿

g

d

Q>

tudy

atí

5



CO

fi

K

C

LU

~ ^

1

Û



f^

u^

N

o-°"E'i

Catastr treat

E

ro (/>

O

8

in Inventor :ements sa

o_

1



??

-.35, PCCL

D.

^^

u 0) c

Influence of catastrophizing on treatment outcome in patients with nonspecific low back pain: a systematic review.

Systematic review...
11MB Sizes 0 Downloads 0 Views