Sensitivity to change of the Rheumatoid Arthritis Self-Efficacy scale (RASE) and predictors of...

49

Musculoskelet. Care 6: 49–67 (2008)Copyright © 2008 John Wiley & Sons, Ltd DOI: 10.1002/msc

Res

earc

h a

rtic

les

MUSCULOSKELETAL CAREMusculoskelet. Care 6(1): 49–67 (2008)Published online 25 January 2008 in Wiley InterScience(www.interscience.wiley.com) DOI: 10.1002/msc.125

Research article

Sensitivity to change of the Rheumatoid Arthritis Self-Effi cacy scale (RASE) and predictors of change in self-effi cacySarah Hewlett PhD MA RN,1 Zoë Cockshott MSc,2 Celia Almeida BA BSc MSc,3 Pam Richards,3 Rob Lowe PhD,4 Rosemary Greenwood MSc,5 John Kirwan BSc MD FRCD,3 and the RASE Study Group

1School of Nursing, University of the West of England, Bristol UK; 2Institute for Health Research, Lancaster University, Lancaster, UK; 3Academic Rheumatology, University of Bristol, Bristol, UK; 4Psychology Department, University of Wales, Swansea, UK; 5Research Development and Support Unit, Bristol Royal Infi rmary, Bristol, UK

RASE Study Group: Sue Benjamin, Bournemouth & Christchurch NHS Trust, UK

Kirsty Board, Hereford County Hospital, UK Anne Campbell, St Helens Hospital, Merseyside, UK Maureen Cox, Nuffi eld Orthopaedic Centre, Oxford, UK Janet Cushnaghan, Lymington Hospital, UK Maggie Hehir, Bristol Royal Infi rmary, UK Maggie Jolly, Derriford Hospital, Plymouth, UK Fiona Maggs, Selly Oak Hospital, Birmingham, UK Claire McEachern, Cheltenham General Hospital, UK Elizabeth Read, Cheltenham General Hospital, UK Nikki Seaman, North Devon District Hospital, Barnstaple,

UK Jill Spicer, Cannock Chase Hospital, UK Heather Unsworth, Lymington Hospital, UK

Abstract

Objectives: Patient education in rheumatoid arthritis (RA) aims to improve health out-comes by prompting people to adopt self-management behaviours. One precursor for initiating behaviour change is self-effi cacy (SE), a belief that you can do a task. This

Hewlett et al.50


Res

earc

h a

rtic

les

study tested the sensitivity to change of a new scale to measure SE for self-management in people with RA, the Rheumatoid Arthritis Self-Effi cacy scale (RASE). Exploratory analysis examined potential predictors of change in SE.Methods: People with RA at 11 rheumatology centres, who had accepted an education programme as part of clinical care, completed questionnaires at baseline, and two and eight weeks after their programme end. Programmes were not standardized, as this was a pragmatic study in clinical practice.Results: A total of 128 patients participated. After controlling for baseline scores, the RASE showed small but signifi cant improvements in SE from baseline (RASE 107.57, CI 105.42-109.72) to two weeks after programme end (RASE 110.80, CI 108.60-112.99), and eight weeks (RASE 110.62, CI 108.40-112.85, p < 0.001). Standardized response means, calculated both by absolute and percentage change, were 0.339 and 0.371 at two weeks after programme end, and 0.321 and 0.352 at eight weeks. Changes in the RASE were associated with behaviour initiation at two and eight weeks (r = 0.419, r = 0.342, p < 0.001). No substantial predictors of change in SE could be identifi ed.Conclusions: The RASE is sensitive to change in a cohort of people with RA in the UK receiving education programmes as routine clinical care. Exploratory analysis did not identify clinical or psychological factors that predict change in SE, suggesting that pro-grammes should not be restricted to particular patients. Copyright © 2008 John Wiley & Sons, Ltd.

Key words: Rheumatoid arthritis, self-effi cacy; RASE, scale; sensitivity; validation; predictors

Introduction

Patient education in rheumatoid arthritis (RA) aims to change knowledge, skills and attitudes and to improve patient outcome by initiating behaviour change. Behaviour change is infl uenced by both outcome expectancy (a belief that certain behaviours lead to certain outcomes) and self-effi cacy (SE), a belief in one’s ability to carry out a task with a desired outcome (Bandura, 1977). SE is associated with current and future health status, may be changed by appropriate interventions, and enhanced SE for certain behaviours is associated with improved health outcomes in those areas (Bandura, 1977; Cross et al., 2006; Lorig et al., 1989). If the patient develops the knowledge, skills and outcome expectancy for a self-management be-haviour, then SE is a major determinant of whether or not they initiate that be-haviour and persist in the face of diffi culty (Bandura, 1977; Maibach and Murphy, 1995). Thus, if education programmes are to be effective, they should enhance SE.

General measures of SE have been developed (General SE Scale, GSES) (Schwarzer, 1993), as well as disease-specifi c questionnaires (Bartholomew et al.,

Sensitivity to change of the RASE 51


Res

earc

h a

rtic

les

1993; Sallis et al., 1988; Stewart et al., 1994). The Arthritis SE Scale (ASES) was developed in the USA to measure SE for pain, function and other symptoms (Lorig et al., 1989; O’Leary et al., 1988). Measurement scales require revalidation and sometimes modifi cation prior to their reliable use in the UK (Hill et al., 1990; Kirwan and Reeback, 1986). The 20 visual analogue scales in the ASES include numbers, markers and words, which in clinical practice can cause confusion with scale completion. SE scales need careful phraseology if they are to refl ect beliefs rather than actual ability to perform the tasks being measured, and the wording of some existing SE scales give cause for reservation (Braden, 1991; Brady, 1997; Schiaffi no and Revenson, 1992).

Accordingly, the RA SE scale (RASE) was developed and validated to measure SE for self-management in people with RA in the UK (Hewlett et al., 2001). The RASE measures beliefs about potential (rather than actual) ability to perform tasks and was developed using items initiated by people with RA and by rheumatology health professionals. The scale contains both general and task-specifi c items, and tasks have varying magnitudes of diffi culty. The 28 RASE items are rated from strongly disagree to strongly agree (1–5), giving a score of 28–140, where high scores refl ect high SE (Appendix 1). The RASE shows good face and content validity, construct validity (biological sense), criterion validity (comparison with other SE scales), reliability (internal consistency and stability), and is independent of mood, disease status and disability (Hewlett et al., 2001). Sensitivity to change was tested in a small study (n = 48) and suggested a signifi cant increase in RASE (+5.167, p < 0.025) and an association with behaviour initiation (r = 0.35, p < 0.01) (Hewlett et al., 2001), but needs to be tested in a larger study.

While the published evidence largely supports patient education leading to improved outcomes in RA (Lorig et al., 1993; Superio-Cabuslay et al., 1996), there are inconsistencies in the direction, degree and sustainability of change for a variety of outcomes (Hammond et al., 2004; Helliwell et al., 1999; Kirwan et al., 2005; Lindroth et al., 1989; Riemsma et al., 2004; Taal et al., 1997). One reason for this variation could be that SE – a principal mechanism through which patient educa-tion is believed to work – is not being altered. This could be because receptiveness to SE change and the timing of the education programme are not synchronized. For example, patients might be resistant to changing SE during early disease when illness beliefs that the timeline of RA will be short may limit the realization that behaviour change is necessary (Leventhal et al., 1984; Schiaffi no and Revenson, 1992), and education at this time may not result in improved outcomes (Freeman et al., 2002). Psychological distress, stress or active infl ammation may reduce the ability to concentrate on self-management. Personality traits, educational level, the impact of the disease or motivation to attend the programme might either reduce or enhance the ability to change SE. If patient attributes that either enhance or

Hewlett et al.52


Res

earc

h a

rtic

les

hinder the ability to change SE could be identifi ed, then patient education could be more specifi cally targeted and timed, just as drugs are specifi cally targeted to suit particular patients.

The primary aim of this study was therefore to examine the sensitivity to change of the RASE in a larger group of people with RA, with the secondary aim of confi rming the criterion and construct validity, and reliability (internal consis-tency) shown in the initial study. An exploratory analysis of potential predictors of change in SE was a subsidiary aim.

Methods

Patients

After obtaining ethics approval, people with confi rmed RA (Arnett et al., 1988) who had been offered an education programme as part of their normal clinical care were invited to complete a package of questionnaires one week before starting their programme, and then two and eight weeks after completion (baseline, End+2W and End+8W), and informed, written consent was obtained. There were no other entry criteria, as this was a pragmatic study to test the RASE during routine patient education in current clinical practice, rather than in a research situation where patient selection may be within narrow clinical parameters and where patient motivation may be to help with research, rather than a wish to improve self-man-agement of RA.

Education programmes

Patient education programmes were defi ned as existing, planned, structured RA patient education on self-management, either in groups or one-to-one settings, and delivered by a health care professional(s). The programmes were not standardized, and the 11 centres meeting the criteria delivered their existing, pre-determined education programmes. Information on the content and delivery of each pro-gramme was collected for descriptive purposes only.

Assessments

Sensitivity to change of the RASE was tested against two other SE scales, the ASES (Lorig et al., 1989) and the GSES (Schwarzer, 1993), and against behaviour initiation using a separate administration of the RASE questions with the request to tick each activity that had been tried in the previous two weeks. RASE validity,



Res

earc

h a

rtic

les

reliability and predictors of SE were assessed as follows. Clinical status: infl ammatory indices [C-reactive protein (CRP), plasma viscosity (PV)], articular index (Fuchs et al., 1989) and disability (Health Assessment Questionnaire, HAQ) (Fries et al., 1980). Two 10 cm visual analogue scales (VAS) assessed pain (‘How much pain have you had in the past 24 hours?’, from ‘No pain’ to ‘Severe pain’) and disease activity (‘Considering all the ways your arthritis affects you, how well are you doing?’, from ‘Very well’ to ‘Very badly’). Psychological assessment: anxiety and depression (Hospital Anxiety and Depression Scale, HAD) (Zigmond and Snaith, 1983), helplessness (Arthritis Helplessness Index, AHI) (Stein et al., 1988), coping strategies of confrontation, avoidance and acceptance-resignation (Medical Coping Modes Questionnaire, MCMQ) (Feifel et al., 1987a, 1987b) and ability to recognize and manage mood (Trait Meta-Mood Scale, mood clarity and management) (DeVellis et al., 1998). Personality traits: pessimism (Revised Life Orientation Test (LOTr)) (Scheier and Carver, 1985) and negative personality or neuroticism (Eysenck Personality Inventory (EPI)) (Eysenck and Eysenck, 1964). Educational level: age upon leaving full-time education. Motivation for attending the programme: 10 cm VAS (‘Thinking about your reasons for attending the education course, how much was this because you personally wanted to be there?’, from ‘Not at all’ to ‘Very much’). Engagement in the programme: as refl ected in a VAS for enjoyment (‘How much did you enjoy the programme?, from ‘Not at all’ to ‘Very much’). Disease impact: Personal Impact of disability (PI HAQ) (Hewlett et al., 2002). All measures were mailed at baseline, End+2W and End+8W apart from the articular index, infl ammatory indices, pessimism and neuroticism (baseline), and motivation and engagement (End+8W). Data were returned to researchers at the coordinating centre, who were not involved in programme delivery, and were analysed by the study statistician (RG).

Sample size

The analyses for the main aims (sensitivity to change, validity and reliability) included factor analysis, and using the guideline of fi ve subjects per variable for factor analysis, a sample size of 140 was required for the 28-item RASE. As the RASE was still under fi nal validation, the sample size for aim 3 (exploratory exami-nation for predictors of SE change) was based on a validated tool in a closely related domain (helplessness) (Stein et al., 1988). Using data from the RASE development studies (Hewlett et al., 2001), there would be an 80% chance of detecting a differ-ence of 2.2 on the helplessness scale between those who improve in SE and those who do not, using 150 patients (5% signifi cance level). This is equivalent to an effect size of 0.5.

Hewlett et al.54


Res

earc

h a

rtic

les

Statistical approach

Demographic and baseline characteristics are described using means and SD, or medians, and fi rst and third inter-quartile range (IQR), where data are not normally distributed. Criterion validity was examined using Pearson’s correlation coeffi cient to investigate how the RASE score compared with other SE measures (ASES and GSES). Reliability was explored using Cronbach’s α to examine overall internal consistency, repeated with each item omitted in turn to explore potential redun-dancy, and also by inter-item correlation. Proposed acceptable standards for internal consistency are within the range of 0.7–0.95 (Bland and Altman, 1997). Preliminary analysis of factor structure in the development of the RASE suggested eight factors (Hewlett et al., 2001), and confi rmatory factor analysis was performed in this larger study using principal components analysis with a cut-off of an eigenvalue of 1 for assigning factors. Construct validity was tested using Pearson’s correlation coeffi -cient to test for association with mood variables. Sensitivity to change can be examined using absolute change scores, but this could be confounded by baseline scores (e.g. people who score low at baseline have the potential for greater change than those who score high). Therefore, sensitivity was examined using repeated linear regression (analysis of variance, ANOVA), which uses the residual compo-nent of the RASE score that remains after controlling for the baseline score. Two other methods of sensitivity to change were explored, the standardized response mean (SRM, mean change divided by the standard deviation of change scores) and Cohen’s effect size (mean change divided by the standard deviation of the baseline score) (Sim et al., 2006). It has been suggested that 0.2–0.49 refl ects low sensitivity, 0.5–0.79 moderate sensitivity and ≥0.8 high sensitivity (Cohen, 1988). Bonferroni’s correction to 0.004 was used in view of the number of tests performed. Variables that potentially might predict change in the RASE included demographic vari-ables, infl ammatory activity, clinical status, mood, helplessness, coping style, per-sonality and intervention (total hours attended). These were investigated using ANOVA to model the RASE at End+2W using the baseline RASE as a covariate, and then repeated for the ASES and GSES. Modelling was carried out using a forwards stepwise procedure and then confi rmatory backwards stepwise procedure using 5% inclusion and exclusion criteria.

Results

Demographic data

A total of 148 people with RA participated in the study, of whom 128 completed. Participants had a mean age of 56 years (SD 11.6 years), median disease duration



Res

earc

h a

rtic

les

of one year, but with a wide range (IQR 0.5, 47 years) and 80% were female (Table 1). Those who did not complete the study had lower baseline RASE scores (100.04 versus 107.48, p = 0.036) and were using fewer of the 28 self-management behaviours being measured (8.00 versus 13.79, p < 0.001) but were otherwise no different from

TABLE 1. Demographic data at entry for those completing or withdrawing from the study

Baseline variable1 Completed (n = 128)

Did not complete (n = 20)

p value2

Mean (SD) or Median (IQR)

n Mean (SD) or Median (IQR)

n

Age (years) 56.2 (11.62) 127 59.72 (14.72) 18 0.246Duration (years) 1 (0.5, 47) 126 2 (0.5, 44) 18 0.411Age left full-time education 16 (10, 25) 125 15 (14, 19) 15 0.167PV (1.5–2.7 mPa) 1.73 (0.16) 93 1.76 (0.15 13 0.612CRP (<10–200 mg/L) 9.95 (0, 189) 110 17 (5, 84) 16 0.061Tender joints (0–28) 6 (0, 28) 127 5.5 (0, 18) 18 0.862Swollen joints (0–28) 6 (0, 22) 127 7.5 (0, 19) 18 0.247Patient opinion of disease activity (0–10) 3.98 (2.48) 127 4.43 (2.35) 18 0.472Pain (0–10) 4.65 (2.73) 127 4.42 (2.97) 18 0.745Disability (HAQ, 0–3) 1.34 (0.7) 123 1.51 (0.82) 18 0.349Disability impact (PIHAQ, 0–9) 3.45 (1.82) 123 3.87 (2.13) 17 0.382Self-effi cacy: RASE (28–140) 107.48 (11.94) 128 100.04 (7.7) 12 0.036

ASES Pain (10–100) 55.25 (18.27) 127 51.64 (17.13) 11 0.528ASES Other Symptoms (10–100)

58.23 (19.53) 126 56.95 (17.98) 12 0.827

GSES (10–40) 29.01 (5.64) 125 25.91 (6.47) 11 0.087Coping (MCMQ) Confrontation (0–32) 20.13 (3.61) 122 19.4 (3.24) 10 0.537

Avoidance (0–28) 16.42 (3.58) 125 16.64 (3.56) 11 0.845Accept–Resign (0–16) 7.93 (2.58) 125 8 (2.57) 11 0.929

Mood (TMMS): Clarity (7–42) 27.89 (6.07) 123 28.2 (5.14) 10 0.877Management (4–24) 17.31 (3.24) 121 17 (2.67) 10 0.766

Anxiety (HAD, 0–21) 7.46 (4.17) 128 5.17 (3.3) 12 0.067Depression (HAD, 0–21) 5.56 (3.62) 128 6.58 (3.8) 12 0.354Helplessness (AHI, 5–30) 17.22 (4.58) 127 17.25 (4.56) 12 0.983Optimism (LOTr, 0–24)3 14.85 (5.15) 128 15.5 (4.25) 12 0.674Pessimism (EPI, 0–24) 11.45 (5.18) 128 8.75 (6.97) 12 0.096Behaviour initiation (RASE items, 0–28) 13.79 (4.73) 126 8 (5.46) 12 <0.001

1 For all variables, higher scores refl ect greater symptoms, feelings or greater use of the strategies measured.2 Unpaired t-test, chi-squared test or Mann Whitney test, as appropriate.3 LOTr, ??; EPI, ??.

Hewlett et al.56


Res

earc

h a

rtic

les

those who completed the study. At baseline, participants had moderate levels of pain and disability, little depression but slightly greater anxiety and helplessness. SE was relatively high, as measured by the RASE and GSES, but moderate as measured by the ASES (Table 1).

Education programmes

The 11 programmes being delivered in clinical practice varied widely in delivery, ranging from one to six sessions, from one to 12 participants and providing 2.5–12.0 education hours (Table 2). However, there was considerable similarity in content, with all courses covering the disease process, fl ares, rheumatology services, joint protection, relaxation, pacing and pain management. There was greater variation in coverage of footwear (82%), mood, stress management, employment and comple-mentary therapies (73%), utilizing clinic consultations (64%) and sexual issues (55%). Programmes were multi-disciplinary, including physiotherapists and occupa-tional therapists (91%), nurses (82%), podiatrists and doctors (36%), and psycholo-gists (18%). Half of centres reported using goal-setting strategies (55%).

Criterion validity

Criterion validity or accuracy of the RASE was assessed by association with other SE scales. At baseline, the RASE was moderately associated with ASES Pain and

TABLE 2. Description of education programmes

Centre No of sessions

Total hours

Patients per group

Programme offered to

Offered to newly diagnosed patients

A 1 3 1 RA patients YesE 1 2.5 2 Any diagnosis NoD 2 5 6 Infl amm1 arthritis YesG 3 6 10 Infl amm arthritis YesC 4 8 5 Infl amm arthritis NoH 6 9 10 Infl amm arthritis NoF 5 10 8 Infl amm arthritis YesK 5 10 8 Infl amm arthritis NoI 6 12 7 Infl amm arthritis NoJ 6 12 8 Infl amm arthritis NoB 6 12 12 Infl amm arthritis No

1 Infl amm, Infl ammatory.



Res

earc

h a

rtic

les

ASES Other Symptoms (r = 0.352, r = 0.362, p < 0.05). Both the RASE and ASES Pain were only weakly associated with the GSES (r = 0.186, r = 0.239, p < 0.05). ASES Other Symptoms was more strongly associated with the GSES (r = 0.398, p < 0.05) and the two ASES scales were highly correlated (r = 0.677).

Reliability

Cronbach’s α showed the internal consistency of the total RASE score at baseline to be 0.89. Internal consistency was further examined using inter-item correlation of the 28-item scale, which gave 378 possible correlations (Spearman’s rank). Of these, only 15 were greater than r = 0.5, including only one at r = 0.74, suggesting that there are no redundant items (Table A available online). Item 7 (‘I believe I could have a hot drink before bed, to improve my sleep’) had negative inter-item correlations on 6/27 occasions. Correlation between each of the 28 RASE items and the total RASE score was moderate to strong (r = 0.341 to r = 0.667), with the exception of item 7 (r = 0.267). Recalculation of Cronbach’s α for the total RASE with each item omitted in turn did not improve on the overall result of 0.89 (omit-ting item 7 made only a nominal difference to 0.892).

Confi rmatory factor analysis was performed using principal components anal-ysis (Table 3), which showed similar loadings on the eight factors identifi ed in the initial validation study (Hewlett et al., 2001), supporting the concept that SE for self-management of RA is multi-dimensional.

Construct validity

The RASE was moderately associated with psychological constructs, with lower SE associated with greater helplessness and depression (r = −0.224, p = 0.009; r = −0.304, p < 0.0001). There was no association between anxiety and SE as measured by the RASE (r = −0.075). By contrast, the ASES Pain and ASES Other Symptom scales were more strongly associated with helplessness, depression and anxiety (r = −0.431 to r = −0.631).

Sensitivity to change

Following their education programmes, patients made improvements in SE, as well as the use of the coping strategies of illness confrontation and resignation (Table 4). Using ANOVA to control for baseline scores, the RASE showed small but sig-nifi cant improvements in SE from baseline (RASE 107.57, SD 12.07) to End+2W (RASE 110.80, SD 12.35, p < 0.0010). This was maintained at End+8W (RASE 110.62, SD 12.46, p < 0.001) (Figure 1). The SRMs showed sensitivity to change of

Hewlett et al.58


Res

earc

h a

rtic

les

TA

BLE

3.

Fact

or lo

adin

gs fo

r th

e R

ASE

usi

ng c

onfi r

mat

ory

fact

or a

naly

sis

Item

Fact

or lo

adin

g (p

revi

ous

stud

y fi n

ding

s)

12

34

56

78

1 U

se r

elax

atio

n to

hel

p w

ith

pain

0.75

(0.

76)

21

Use

rel

axat

ion

to d

eal w

ith

wor

ries

0.79

(0.

75)

23

Use

rel

axat

ion

tape

to

help

me

rela

x0.

72 (

0.72

)22

A

lloca

te t

ime

for

rela

xati

on0.

79 (

0.71

)15

Ed

ucat

e fa

mily

/frie

nds

to h

elp

rela

tion

ship

s0.

80 (

0.87

)16

Ex

plai

n to

fri

ends

whe

n I

do/d

on’t

need

hel

p0.

85 (

0.86

)17

D

iscu

ss p

robl

ems

wit

h m

y pa

rtne

r or

fam

ily0.

71 (

0.86

)12

A

sk fo

r he

lp t

o de

al w

ith

ever

yday

tas

ks0.

64 (

0.82

)5

Avo

id d

oing

thi

ngs

that

cau

se p

ain

0.82

(0.

82)

14

Plan

/pri

orit

ise

to d

eal w

ith

doin

g ev

eryd

ay t

asks

0.82

(0.

77)

4 T

hink

pos

itiv

ely

to h

elp

wit

h pa

in0.

44 (

0.73

)18

M

ake

tim

e fo

r le

isur

e ac

tivi

ties

0.89

(0.

73)

19

Save

ene

rgy

for

leis

ure

acti

viti

es0.

90 (

0.67

)0.

63 (

0.51

)11

U

se g

adge

ts t

o he

lp w

ith

mob

ility

/tas

ks0.

17 (

0.60

)24

U

se r

egul

ar e

xerc

ise

0.87

(0.

80)

13

Do

exer

cise

s to

dea

l wit

h di

ffi cu

ltie

s of

tas

ks0.

79 (

0.67

)28

Se

ek h

elp

wit

h pe

rsis

tent

sid

e-ef

fect

s0.

55 (

0.56

)7

Hav

e a

hot

drin

k to

im

prov

e m

y sl

eep

0.77

(0.

73)

6 W

ind

dow

n an

d re

lax

to i

mpr

ove

my

slee

p0.

85 (

0.71

)3

Use

my

join

ts c

aref

ully

to

help

wit

h pa

in0.

46 (

0.52

)26

M

anag

e m

y m

edic

atio

n0.

85 (

0.79

)27

Lo

ok o

ut fo

r &

avo

id s

ide-

effe

cts

of m

edic

atio

n0.

85 (

0.75

)9

Pace

mys

elf t

o he

lp d

eal w

ith

tire

dnes

s0.

86 (

0.72

)10

A

ccep

t fa

tigu

e as

par

t of

my

arth

riti

s0.

64 (

0.68

)20

Fo

cus

on t

he p

osit

ive

whe

n I’m

feel

ing

dow

n0.

63 (

0.59

)2

Thi

nk a

bout

som

ethi

ng e

lse

to h

elp

wit

h pa

in8

Use

rel

axat

ion

befo

re b

ed, t

o im

prov

e m

y sl

eep

25

Be

awar

e of

my

limit

s in

exe

rcis

e



Res

earc

h a

rtic

les

the RASE, whether calculated as absolute change or percentage change, and were similar in both the short term (baseline to End+2W: 0.34, 0.37) and longer term (baseline to End+8W: 0.32, 0.35). The SRMs were of a similar order to those for the ASES (0.32–0.53) and slightly greater than the GSES (0.23–0.3). There was little change in SE after the education programme had fi nished (i.e. between End+2W and End+8W), with low scores for all three scales (SRM range −0.02 to 0.14), suggesting that changes in SE mainly occur during programmes. Cohen’s effect size also supported sensitivity to change in the RASE, with baseline to End+2W at 0.27 and baseline to End+8W at 0.26.

After controlling for baseline scores, changes in the RASE were associated with self-management behaviour initiation at both End+2W and End+8W (r = 0.370, r = 0.397, p < 0.05). Change in the RASE was also associated with an increase in the coping strategy of illness confrontation (End+2W r = 0.228, End+8W r = 0.241, both p < 0.05). However, changes in ASES and GSES were not associated with changes in behaviour or illness confrontation. A reduction in illness resigna-tion was associated with an increase in the RASE after programme completion (between End+2W and End+8W, r = −0.206, p < 0.05), and with an increase in ASES and GSES between baseline and End+2W (r = −0.293 to r = −0.307, p < 0.05) and baseline and End+8W (r = −0.0247 to r = −0.343, p < 0.05).

Predictors of change in SE

An exploratory examination for predictors of change in the RASE at End+2W was made using linear regression, controlling for baseline RASE. There were no clini-

100

102

104

106

108

110

112

114

Baseline End+2W End+8W

RA

SE s

core

28–

140

FIGURE 1. Change in RASE following education programmes.Mean and 95% confi dence interval (n = 121)

Hewlett et al.60


Res

earc

h a

rtic

les

TA

BLE

4.

Cha

nges

follo

win

g ed

ucat

ion

prog

ram

mes

(n

= 12

1 pa

tien

ts w

ho c

ompl

eted

all

thre

e R

ASE

sca

les)

Var

iabl

e1B

asel

ine

End+

2WEn

d+8W

p va

lue2

Mea

n (S

D)

Mea

n (S

D)

Mea

n (S

D)

Pain

(0–

10)

4.54

(2.

92)

4.12

(2.

86)

4.16

(2.

69)

0.11

7Pa

tien

t op

inio

n of

dis

ease

act

ivit

y (0

–10)

3.98

(2.

64)

3.51

(2.

30)

3.68

(2.

41)

0.13

7D

isab

ility

(H

AQ

, 0–3

)1.

33 (

0.73

)1.

34 (

0.67

)1.

32 (

0.67

)0.

858

Dis

abili

ty im

pact

(PI

HA

Q, 0

–9)

3.43

(1.

91)

3.39

(1.

85)

3.39

(1.

85)

0.93

2Se

lf-ef

fi cac

y:R

ASE

(28

–140

) 10

7.57

(12

.07)

110.

80 (

12.3

5)11

0.62

(12

.46)

<0.0

01A

SES

Pain

(10

–100

) 55

.7 (

17.9

6)61

.9 (

19.0

8)62

.1 (

19.0

8)<0

.001

ASE

S O

ther

Sym

ptom

s (1

0–10

0)58

.23

(20.

37)

65.0

(17

.40)

64.7

(17

.96)

<0.0

01G

SES

(10–

40)

29.0

1 (3

.98)

30.7

(5.

05)

30.9

(5.

05)

0.00

2C

opin

g (M

CM

Q)

Con

fron

tati

on (

0–32

) 20

.13

(9.7

1)21

.4 (

3.37

)21

.1 (

3.37

)<0

.001

Avo

idan

ce (

0–28

) 16

.32

(3.7

6)16

.10

(3.6

5)16

.21

(3.3

2)0.

656

Acc

ept–

Res

ign

(0–1

6)

7.93

(2.

97)

7.4

(2.4

5)7.

6 (2

.24)

0.02

9M

ood

(TM

MS)

:C

lari

ty (

7–42

) 27

.85

(6.6

2)28

.45

(6.7

3)28

.36

(6.6

8)0.

378

Man

agem

ent

(4–2

4)

17.3

8 (3

.54)

17.3

5 (3

.76)

17.6

4 (3

.48)

0.62

5A

nxie

ty (

HA

D, 0

–21)

7.35

(3.

93)

7.53

(4.

15)

7.22

(4.

43)

0.43

6D

epre

ssio

n (H

AD

, 0–2

1)5.

54 (

3.54

)5.

26 (

3.37

)5.

31 (

3.54

)0.

386

Hel

ples

snes

s (A

HI,

5–30

)17

.10

(4.6

6)16

.39

(5.4

4)16

.63

(5.2

2)0.

201

Beh

avio

urs

(RA

SE it

ems,

0–28

)13

.88

(4.7

7)15

.90

(5.7

2)16

.07

(5.7

2)<0

.001

1 For

all

vari

able

s, hi

gher

sco

res

refl e

ct g

reat

er s

ympt

oms,

feel

ings

or

grea

ter

use

of t

he s

trat

egie

s m

easu

red.

2 Rep

eate

d m

easu

res

AN

OV

A.



Res

earc

h a

rtic

les

cal, psychological, demographic or programme variables that predicted change in RASE score. Several variables were signifi cant for change in ASES and GSES scales, but when entered into a stepwise model, only baseline anxiety remained as contributing a small explanation for the variation in change in SE: combined anxiety and mood management explained 12% of the change in ASES Pain; com-bined anxiety and illness resignation explained 9% of the variation in change in ASES Other Symptoms; and anxiety accounted for 10% of the variation in change in GSES.

Discussion

This study showed that the RASE is sensitive to change when used in non-stan-dardized, routine patient education programmes already in clinical use across the UK. It also confi rmed the initial construct and criterion validity, and reliability in a larger population. Other clinical research studies have shown the RASE to be sensitive to change (Hammond et al., 2005) and the RASE performed well in a review of arthritis scales (Katz, 2003).

The RASE was tested in education programmes in clinical practice, and although these showed some diversity in construction and delivery, there is no reason to suppose that these are anything other than typical of many such courses being offered in routine care, as the RASE study group centres covered many geo-graphical areas, and both teaching and non-teaching, town and rural rheumatology units. In addition, the range of patient characteristics was large, with both early and late disease duration, varying levels of infl ammation, age, education, and clini-cal and psychological status. Nonetheless, these clinical programmes, offered in response to perceived clinical need, produced small but signifi cant overall changes in SE. It would be interesting to test the RASE further in cognitive-behavioural programmes in RA, rather than educational programmes. The RASE, unlike the other SE scales, showed an association with the initiation of changes in behaviour. Although the questions on behaviour change were those in the RASE itself, it was expected that the other SE scales would also refl ect change in such self-management behaviours. The RASE refl ected that an improvement in SE was associated with greater confrontation of one’s problems – that is, greater belief in an ability to make a difference to one’s condition is associated with greater willing-ness to tackle problems, a reasonable expectation that was not detected by the other SE scales.

The arthritis-specifi c RASE and ASES scales are moderately correlated, but both show only weak association with the GSES, suggesting, in line with SE theory, that SE can be task specifi c and that a disease-specifi c scale may be more appropri-ate in RA. The RASE was moderately associated with related psychological con-

Hewlett et al.62


Res

earc

h a

rtic

les

structs of helplessness and depression, but the ASES was fairly strongly associated with these concepts and also with anxiety. While these psychological constructs are related to SE, strong correlation might indicate potential diffi culty for the scales to differentiate between SE and mood. Factor analysis confi rms that SE, as mea-sured by the RASE, is multi-dimensional when applied to the self-management of RA, refl ecting the many facets of RA self-management. However, this was not an attempt to validate subscales within the RASE, which remains a single scale of 28 items. These sensitivity, validity and reliability data together provide support-ing evidence that the RASE is a valid and sensitive tool to assess SE for self-management behaviours in people with RA in the UK.

Participants had been offered, and accepted, a patient education programme on the grounds of clinical need, which in theory should increase the likelihood of higher motivation for self-management (rather than motivation to help the researcher). Even so, exploratory analyses for predictors of change in SE, including clinical and psychological status, disease duration, age, personality and motivation, did not reveal any strong predictors of improvement in SE, using any of the SE scales (with the exception of a small contribution from anxiety). If no predictors are readily apparent but overall these clinical programmes enhance SE, then there is no reason as yet, to limit their availability to particular patients (e.g. defi ned by disease duration or clinical status).

Study limitations

The mode of education delivery may be a factor in enhancing SE. However, this was not a randomized controlled trial of patient education but a pragmatic study, examining patients undergoing routine care; therefore, the programmes were not standardized but were those used in normal practice. Standardizing programmes and designing them specifi cally to enhance SE might have produced greater change in SE, but would not refl ect routine practice. There was no intention to examine the outcomes from the programmes of individual centres, as centres contributed insuffi cient numbers (e.g. for comparison of programme duration).

Those who did not complete the study had signifi cantly lower baseline SE than those who completed and were using fewer of the 28 self-management behav-iours being measured. These differences might refl ect patients who were not ready to make changes in their lives and who would not have improved in SE even had they completed the courses (Prochaska and Diclemente, 1983). Although these patients were no different in any other respect to those who completed the study, had they remained in the study, it might have diluted the overall improvement in SE. Low SE at the time of invitation to join an education programme might explain the anecdotal low uptake of education programmes in clinical practice. This sug-



Res

earc

h a

rtic

les

gests that SE, or motivation for attending a course, may need to be addressed at the time of invitation to participate in an education programme, so that the invita-tion might be taken up.

Change in the RASE was associated with behaviour initiation, although other SE scales were not. SE is important in relation to initiating behaviours that are valued, and it is probable that not all of the behaviours on the RASE were desired by all patients. However, the scale consists of items proposed by patients (Hewlett et al., 2001); therefore, it is anticipated that most of them would be valued by RA patients.

The analyses on potential predictors of outcome were intended to be explor-atory; however, the reduced sample size might further limit their interpretation (n = 128 instead of n = 150). While a range of clinical, psychological and per -sonality variables were assessed as predictors of change in SE, there may be other candidate variables that were not measured. For example, problematic social support (e.g. the critical spouse) may affect SE (Riemsma et al., 1998) and the concept of readiness to change behaviour (Hammond, 2003), but these were not measured.

Future research might usefully explore whether the RASE, although generated by people with RA, is equally valid in people with other types of arthritis. Further, more defi nitive research into potential predictors of change in SE, and in particular the role of anxiety, might be warranted. A third issue raised by this study is the level of both diversity and commonality in the RA education programmes being offered in clinical practice. Unlike medication or exercise therapy, which are gener-ally identifi able and quantifi able, the delivery of educational interventions in common practice currently remains diverse, non-standardized and therefore diffi -cult to replicate. The development of consensus and standards on core topics and common approaches to professionally led RA education programmes, such as have been developed for lay-led self-management programmes (Lorig et al., 1993), might be potentially benefi cial.

Conclusion

Assuming change in SE to be a necessary precursor to the adoption of self-manage-ment behaviours, the RASE provides a useful assessment tool for examining SE in people with RA in the UK.

Acknowledgements

The authors would like to thank the Arthritis Research Campaign for funding the study and Professor Hewlett’s academic post, the people with RA who participated

Hewlett et al.64


Res

earc

h a

rtic

les

and Vanessa Lock (initial research sister). SH, ZC, RL and JK designed the study; SH, ZC and CA managed the study; SH, ZC, RL and RG analysed the data; and all authors and all members of the RASE study group contributed to data inter-pretation and drafts of this paper. The authors have declared no confl icts of interest.

References

Arnett FC, Edworthy SM, Bloch DA et al. (1988). The American Rheumatism Association 1987 revised criteria for the classifi cation of rheumatoid arthritis. Arthritis and Rheumatism; 31: 315–24.

Bandura A (1977). Self-effi cacy: Toward a unifying theory of behavioural change. Psychological Review 84: 191–215.

Bartholomew LK, Parcel GS, Swank PR, Czyzewski DI (1993). Measuring self-effi cacy expectations for the self-management of cystic fi brosis. Chest 103: 1524–30.

Bland JM, Altman DG (1997). Statistics notes: Cronbach’s alpha. British Medical Journal 314: 527.

Braden CJ (1991). Patterns of change over time in learned response to chronic illness among participants in a systemic lupus erythematosus self-help course. Arthritis Care and Research 4: 158–67.

Brady TJ (1997). Do common arthritis self-effi cacy measures really measure self-effi cacy? Arthritis Care and Research 10: 1–8.

Cohen J (1988). Statistical power analysis for the behavioral sciences (2nd edn). Hillsdale, NJ, Lawrence Erlbaum.

Cross MJ, March LM, Lapsley HM, Byrne E, Brooks PM (2006). Patient self-effi cacy and health locus of control: Relationship with health status and arthritis-related expenditure. Rheumatology 45: 92–6.

DeVellis RF, Carl KL, DeVellis BM, Blalock SJ, Patterson CC (1998). Correlates of changes in mood following a mood induction in OA patients. Arthritis Care and Research 4: 234–42.

Eysenck HJ, Eysenck SBG (1964). Manual of the Eysenck Personality Inventory. London, Hodder & Stoughton.

Feifel H, Strack S, Nagy VT (1987a). Degree of life-threat and differential use of coping modes. Journal of Psychosomatic Research 31: 91–9.

Feifel H, Strack S, Nagy VT (1987b). Coping strategies and associated features of medically ill patients. Psychosomatic Medicine 49: 616–25.

Freeman K, Hammond A, Lincoln NB (2002). Use of cognitive-behavioural arthritis education programmes in newly diagnosed rheumatoid arthritis. Clinical Rehabilitation 16: 828–36.

Fries JF, Spitz P, Kraines RG, Holman HR (1980). Measurement of patient outcome in arthritis. Arthritis and Rheumatism 23: 137–45.

Fuchs HA, Brooks RH, Callahan LF, Pincus T (1989). A simplifi ed twenty-eight joint quantitative articular index in rheumatoid arthritis. Arthritis and Rheumatism 32: 531–7.

Hammond A (2003). Why do people choose whether to attend arthritis education programmes? Annals of the Rheumatic Diseases 62: 547 (abstract).

Hammond A, Bryan J, Hardy A (2005). One-year outcomes of the ‘Lifestyle management for arthritis’ programme for people with infl ammatory arthritis. Rheumatology 44 (Suppl. 1): OP26.



Res

earc

h a

rtic

les

Hammond A, Young A, Kidao R (2004). A randomised controlled trial of occupational therapy for people with early rheumatoid arthritis. Annals of the Rheumatic Diseases 63: 23–30.

Helliwell PS, O’Hara M, Holdsworth J, Hesselden A, King T, Evans P (1999). A 12-month randomised controlled trial of patient education on radiographic changes and quality of life in early RA. Rheumatology 38: 303–8.

Hewlett S, Cockshott ZC, Kirwan JR, Barrett J, Stamp J, Haslock I (2001). Development and validation of a self-effi cacy scale for use in British patients with rheumatoid arthritis (RASE). Rheumatology 40: 1221–30.

Hewlett S, Smith AP, Kirwan JR (2002). Measuring the meaning of disability in rheumatoid arthritis: The Personal Impact Health Assessment Questionnaire (PI HAQ). Annals of the Rheumatic Diseases 61: 986–93.

Hill J, Bird JA, Lawson CW, Wright V (1990). The Arthritis Impact Measurement Scales: an anglicized version to assess the outcome of British patients with RA. British Journal of Rheumatology 29: 193–6.

Katz P (Ed.) (2003). Patient outcomes in rheumatology: A review of measures. Arthritis and Rheumatism (Arthritis Care and Research); 49: 5 (Suppl.): S159–S161.

Kirwan JR, Reeback JS (1986). Stanford health assessment questionnaire modifi ed to assess disability in British patients with rheumatoid arthritis. British Journal of Rheumatology 25: 206–9.

Kirwan JR, Hewlett S, Cockshott Z, Barrett J (2005). Clinical and psychological outcomes in a randomised controlled clinical trial of patient education in rheumatoid arthritis. Musculoskeletal Care 3: 1–16.

Leventhal H, Nerenz DR, Steel DJ (1984). Illness representations and coping with health threats. In Baum A, Taylor S, Singer J (Eds). Handbook of psychology and health vol IV: Social psychological aspects of health. London, Lawrence Erblaum Associates.

Lindroth Y, Bauman A, Barnes C, McCredie M, Brooks PM (1989). A controlled evaluation of arthritis education. British Journal of Rheumatology 28: 70–82.

Lorig K, Chastain R, Ung E, Shoor S, Holman H (1989). Development and evaluation of a scale to measure perceived self-effi cacy in people with arthritis. Arthritis and Rheumatism 32: 37–44.

Lorig K, Mazonson P, Holman H (1993). Evidence suggesting that health education for self-management in patients with chronic arthritis has sustained health benefi ts while reducing health care costs. Arthritis and Rheumatism 36: 439–46.

Maibach E, Murphy D (1995). Self-effi cacy in health promotion research and practice: Conceptualization and measurement. Health Education and Research 10: 37–50.

O’Leary A, Shoor S, Lorig K, Holman H (1988). A cognitive-behavioural treatment for rheumatoid arthritis. Health Psychology 7: 527–44.

Proschaska JO, Diclemente CC (1983). Stages and processes of self-change of smoking. Toward an integrative model of behavior change. Journal of Consulting and Clinical Psychology 51: 390–5.

Riemsma R, Rasker JJ, Taal E, Griep EN, Wouters J, Wiegman O (1998). Fatigue in rheumatoid arthritis: The role of self-effi cacy and problematic social support. British Journal of Rheumatology 37: 1042–6.

Riemsma RP, Taal E, Kirwan JR, Rasker JJ (2004). Systematic review of rheumatoid arthritis patient education. Arthritis and Rheumatism 51: 1045–59.

Sallis JF, Pinski RB, Grossman RM, Patterson TL, Nader PR (1988). The development of self-effi cacy scales for health-related diet & exercise behaviour. Health Education & Research 3: 283–92.

Scheier MF, Carver CS (1985). Optimism, coping and health: Assessment and implications of generalised outcome expectancies. Health Psychology 4: 219–47.

Hewlett et al.66


Res

earc

h a

rtic

les

Schiaffi no KM, Revenson T (1992). The role of perceived self-effi cacy, perceived control, and causal attribution in adaptation to rheumatoid arthritis: Distinguishing mediator from moderator effects. Personality and Social Psychology Bulletin 18: 709–18.

Schwarzer R (1993). Measurement of perceived self-effi cacy. Psychometric scales for cross cultural research. Berlin, Freie Universität.

Sim J, Jordan K, Lewis M, Hill J, Hay EM, Kziedzic K (2006). Sensitivity to change and internal consistency of the Northwick Park Neck Pain Questionnaire and derivation of a minimal clinically important difference. Clinical Journal of Pain 22: 820–6.

Stein MJ, Wallston KA, Nicassio PM (1988). Factor structure of the Arthritis Helplessness Index. Journal of Rheumatology 15: 427–32.

Stewart KJ, Kelemen MH, Ewart CK (1994). Relationships between self-effi cacy and mood before and after exercise training. Journal of Cardiopulmonary Rehabilitation 14: 35–42.

Superio-Cabuslay E, Ward MM, Lorig K (1996). Patient education interventions in osteoarthritis and rheumatoid arthritis: A meta-analytic comparison with non-steroidal anti-infl ammatory drug treatment. Arthritis Care and Research 9: 292–301.

Taal E, Rasker JJ, Wiegman O (1997). Group education for rheumatoid arthritis patients. Seminars in Arthritis and Rheumatism 26: 805–16.

Zigmond AS, Snaith RP (1983). The Hospital Anxiety and Depression Scale. Acta Psychiatrica Scandinavica 67: 361–70.

Correspondence should be sent to Sarah Hewlett PhD MA RN, arc Professor of Rheumatology and Nursing, Academic Rheumatology, Bristol Royal Infi rmary, Bristol BS2 8HW, UK. Tel: +44 (0) 117 928 2903; Fax: +44 (0) 117 928 3841. E-mail: [email protected]

Appendix 1: The Rheumatoid Arthritis Self-Effi cacy Questionnaire (RASE)

We are interested in fi nding out what things you believe you could do to help you with your arthritis. We want to know what you think you could do, even if you are not actually doing it at the moment.

Please tick one column for each question.Do you believe you could do these things to help you with your arthritis? 1 I believe I could use relaxation techniques to help with pain 2 I believe I could think about something else to help with pain 3 I believe I could use my joints carefully (joint protection) to help with pain 4 I believe I could think positively to help with pain 5 I believe I could avoid doing things that cause pain 6 I believe I could wind down and relax before going to bed, to improve my

sleep 7 I believe I could have a hot drink before bed, to improve my sleep 8 I believe I could use relaxation before bed, to improve my sleep 9 I believe I could pace myself and take my arthritis into account to help deal

with tiredness10 I believe I could accept fatigue as part of my arthritis



Res

earc

h a

rtic

les

11 I believe I could use gadgets to help with mobility, household tasks or personal care

12 I believe I could ask for help to deal with the diffi culties of doing everyday tasks

13 I believe I could do exercises to deal with the diffi culties of doing everyday tasks

14 I believe I could plan or prioritise my day to deal with diffi culties of doing everyday tasks

15 I believe I could educate my family and friends about my arthritis to help with the strains that arthritis can make on relationships

16 I believe I could explain to friends and family when I do or don’t need help17 I believe I could discuss any problems with my partner or family18 I believe I could make time for leisure activities, hobbies or socializing19 I believe I could save energy for leisure activities, hobbies or socializing20 I believe I could focus on the positive when I’m feeling down21 I believe I could use relaxation to deal with worries22 I believe I could allocate time for relaxation23 I believe I could use a relaxation tape or instructions to help me relax24 I believe I could use regular exercise25 I believe I could be aware of my limits in exercise26 I believe I could manage my medication, knowing how and when to take it27 I believe I could look out for and avoid side-effects of my medication28 I believe I could seek help with persistent side-effects

Scoring:

Likert scoring, strongly disagree to strongly agree, 1–5Sum scores, range 28–140, high scores refl ect high self-effi cacy

Sensitivity to change of the Rheumatoid Arthritis Self-Efficacy scale (RASE) and predictors of...

Documents

Transcript of Sensitivity to change of the Rheumatoid Arthritis Self-Efficacy scale (RASE) and predictors of...