Taiwanese translation and psychometric testing of the revised illness perception questionnaire for patients with traumatic injury

METHODS

Taiwanese translation and psychometric testing of the revised illness

perception questionnaire for patients with traumatic injury

Wendy Chaboyer, Bih-O Lee, Marianne Wallis and Chi-Sheng Chien

Aims. The purpose of this study was to undertake a cross-cultural adaptation, translation and psychometric analysis of the

Illness Perception Questionnaire Revised for patients with traumatic injury in Taiwan.

Background. Illness representations are the cognitive understandings and emotional responses individuals develop, which help

to determine their responses to health threats.

Methods. This methodological study involved four phases. First, two subscales, identity and causes were modified. Second,

translation and back translation was undertaken by four translators, moderated by an expert committee. Third, ten patients

with traumatic injury pilot tested the feasibility and readability of the Chinese Illness Perception Questionnaire Revised

(Trauma), and three professionals assessed the scale for content validity, resulting in minor modifications. Finally, 173 patients

with traumatic injury were recruited to the main study and completed the Chinese Illness Perception Questionnaire Revised

(Trauma). Item analysis, factor analysis, Cronbach’s alpha and split-half reliability were used to the psychometric properties of

the Chinese Illness Perception Questionnaire Revised (Trauma).

Results. Eight items were removed from the scale as a result of the item analysis. The factor analysis demonstrated a six-factor

structure explained 60Æ3% of the total item variance in the scale, which was very similar to the original scale. The Cronbach’s

alphas ranged from 0Æ69–0Æ80 for each subscale, and the split-half reliability coefficients were from 0Æ70–0Æ82.

Conclusion. The Chinese Illness Perception Questionnaire Revised (Trauma) had good psychometric properties.

Relevance to clinical practice. Translation of the instrument into Chinese extends its utility to the traumatic injury population.

Awareness of patients’ illness representations can help clinicians provide appropriate interventions to patients.

Key words: illness representations, instrument development, nursing, traumatic injury

Accepted for publication: 31 August 2010

Introduction

Traumatic injury is associated with a wide range of both

physical and psychosocial sequela, which can have far-

reaching consequences for both those injured and their

families (Richmond et al. 1998, Davey et al. 2004, Aitken

et al. 2007). Recently, illness representation, or the under-

standing people develop about their illnesses, has been used

to better understand how those who have experienced a

traumatic injury perceive their injury (Lee et al. 2008, 2010),

Authors: Wendy Chaboyer, PhD, RN, Professor, Director of

NHMRC Centre of Research Excellence in Nursing Interventions

for Hospitalised Patients (NCREN), Griffith University, Queensland,

Australia; Bih-O Lee, PhD, RN, Associate Professor, Department of

Nursing, Chung Hwa University of Medical Technology, Tainan,

Taiwan, ROC; Marianne Wallis, PhD, RN, Professor, Research

Centre for Clinical Practice Innovation, Griffith University and Gold

Coast Health Service District, Queensland, Australia; Chi-Sheng

Chien, MD, Attending Physician, Department of Orthopaedic

Surgery, Chi Mei Medical Center, Taiwan, ROC

Correspondence: Bih-O Lee, Associate Professor, Department of

Nursing, Chung Hwa University of Medical Technology, No. 89,

Wen-Hwa 1st Street, Jen-Te Hsiang, Tainan County 71703, Taiwan,

ROC. Telephone: +886 6 2671214 ext. 521.

E-mail: [email protected]

� 2012 Blackwell Publishing Ltd

3466 Journal of Clinical Nursing, 21, 3466–3474, doi: 10.1111/j.1365-2702.2011.03964.x

with the hope that this understanding will help clinicians

devise more targeted treatments and support services. As part

of a larger study investigating recovery in the traumatic

injury population, this study reports on the translation and

testing of illness representation survey in Taiwan. Impor-

tantly, understanding patients’ illness representation may

help health professionals provide more appropriate interven-

tions and support, ultimately enabling patients to better

manage their recovery from traumatic injuries.

Illness representations as a construct emerged from psy-

chology, specifically the self-regulation school of thought,

which regards individuals as problem solvers who understand

and develop their own strategies to cope with their condition

(Leventhal et al. 2001). Conceptually, illness representations

are multidimensional, comprised of five cognitive and one

emotional components. The first dimension, identity, reflects

the physical symptoms individuals experience. Timeline, the

second dimension, reflects the perceived progress and dura-

tion of an illness (acute, chronic or cyclic). The third

component is consequences, both the anticipated and expe-

rienced expected experiences or results. Controllability, the

fourth dimension, is the perception of the responsiveness of

the condition to self-treatment and expert intervention (i.e.

whether the illness can be controlled). The fifth component is

causes, that is, the perceived cause of an illness including

external, internal and behavioural reasons (Leventhal et al.

2001). The emotional dimension reflects emotional responses

to the condition.

Given that illness representations can be modified over

time (Leventhal et al. 1984) and that they have been

associated with outcomes such as quality of life (Covic et al.

2004, Rutter & Rutter 2007, Lee et al. 2008), complications

(Cherrington et al. 2004) and adherence to dietary regimes

(Coutu et al. 2003), researchers have focused on its measure-

ment. One early tool, the Illness Perception Questionnaire

(IPQ), only measured the cognitive dimensions (Weinman

et al. 1996), but the more widely used IPQ-revised (IPQ-R)

includes the emotional dimension (Moss-Morris et al. 2002).

The IPQ-R contains 70 items and eight subscales as the

timeline dimension has been divided into two subdimensions.

The eight subscales are as follows: (1) 14 common identity

symptoms that are answered by ticking yes or no, (2) A

38-item section that measures seven subscales including

timeline (acute/chronic) (six items), timeline cyclical (four

items), consequences (six items), personal control (six items),

treatment control (five items), illness coherence (five items)

and emotional representation (six items), and (3) A 18-item

section used to measure the causes subscale. The second and

third sections are scored on 5-point likert scales (Moss-

Morris et al. 2002). The IPQ-R is freely available and has

been used in a growing number of studies (Moss-Morris et al.

2002).

The IPQ-R can be modified for use for a variety of

populations, which makes it a particularly attractive instru-

ment. The recommended modification involves revising the

identity and causes subscales to reflect the frequent symptoms

and risk factors associated with the particular condition. This

study reports on the modification of the IPQ-R for the trauma

population [i.e. the IPQ-R (Trauma)] and then its translation

into Mandarin Chinese and testing in Taiwan. Because over

30 million people speak Mandarin Chinese (Gordon 2005),

a Chinese version of IPQ-R for traumatic injury population

may have substantial value in future. Further, in general,

translating instruments to other languages extends their

utility to different geographical regions and allows cross-

cultural comparisons of experience, extending understanding

of any given group (Beaton et al. 2000).

Methodology

A four-phased methodological study, which included modi-

fication, translation, pilot study and and a main study, was

undertaken from February–August 2006 in Taiwan. Each of

these steps is described, with the main study findings reported

in the results section.

Modifications

A review of the literature on the predictors and outcomes of

traumatic injury (Davey et al. 2004, Holbrook & Hoyt

2004) was undertaken to identify appropriate modifications

to the identity and causes subscales (as recommended).

Table 1 displays the original and modified subscales. To

note, four symptoms and eight causes were deleted from the

Trauma version, as they reflected chronic and infectious

diseases, and their retention was not supported in the

literature.

Translation

The cross-cultural translation process (Brislin 1986, Jones

et al. 2001) was used as it has been reported to enhance

the likelihood of achieving conceptual equivalence in trans-

lation (Beck et al. 2003). First, forward translation, which

involves one (Brislin 1986) or two (Jones et al. 2001)

bilingual persons translating the items in the original

language to the target language, was undertaken. To reduce

the likelihood of ambiguous translation, two bilingual

individuals, who had backgrounds as healthcare profession-

als, translated the IPQ-R (Trauma). As recommended, both

Methods Taiwanese translation and psychometric testing

� 2012 Blackwell Publishing Ltd

Journal of Clinical Nursing, 21, 3466–3474 3467

translators were blinded to the underlying construct to be

measured; one translator had experience in translating scales

and the second was experienced in Chinese composition

(Brislin 1986). To synthesise the two translated versions, an

expert committee including one of the researchers and the

two translators discussed the results of the forward transla-

tion versions. Four items were difficult to express in Chinese

because, in Chinese, they were more abstract statements.

Those items included item 12 ‘There is a lot which I can do to

control my symptoms’, item 19 ‘There is very little that can

be done to improve my injury’, item 24 ‘The symptoms of my

condition are puzzling to me’ and item 27 ‘My injury does

not make any sense to me’. After further discussion, the

expert committee and translators reached the consensus on

the Chinese wording of these items. Thus, the first Chinese

version of the IPQ-R (Trauma) was produced.

Next, two other individuals, blinded to the English version

and the underlying construct, translated the Chinese version

back into English. Back translation is one of the most

important steps to achieving semantic equivalence in trans-

lating scales (Brislin 1970). One back translator was a person

whose first language was English, and the second was an

experienced English teacher whose first language was Chi-

nese. Neither were health-related professionals. No major

difficulties were encountered in the back translation, except

two items had semantic discrepancies because they were

affective in nature and thus, their expression was somewhat

culture bound. When the two items were translated to

Chinese and back-translated to English, they were judged to

lack the original meaning by the expert committee. Item 27

‘My injury does not make any sense to me’ caused difficulties

because the term ‘make sense’ was an English language

idiom. Item 13 ‘What I do can determine whether my injury

gets better or worse’ was back-translated resulting in dissim-

ilar semantic meanings. Consensus for appropriate Chinese

wording of these items was reached after several discussions.

After the translation process was completed, the expert

committee reexamined the Chinese version of the IPQ-R

(Trauma). The panel used four-point scale from 1 ‘worst’

to 4 ‘best’ agreements to assess the items in the Chinese

version. Additionally, to ensure equivalence between the

English and Chinese versions, the discussion focused on the

meaning of illness representation in different cultures. As a

result of the ratings and discussion, several minor wording

changes were made to the Chinese version. This discussion

was vital to ensure the translation process maintained

concept integrity and that it could be used in different

cultural contexts (Brislin 1986).

Pilot study

In the pilot study, content validity, feasibility, readability and

the estimated time for completion were tested. The sample for

the test consisted of three health professionals (one physician,

one nurse and one rehabilitation practitioner) and 10 patients

with traumatic injury at a medical centre in Taiwan. The

professionals assessed the instrument for content validity,

readability and feasibility, focusing especially on the identity

and causes subscales. The 10 patients assessed the scale’s

readability and completed the scale to determine the com-

pletion time.

In the feasibility and readability testing of the instrument,

one of the professionals mentioned that some patients who

live in southern Taiwan use both Chinese and the Taiwanese

dialect, but a few people use the Taiwanese dialect only. The

researcher therefore excluded people who only used the

Taiwanese dialect, as they may not have completely under-

stood the questions. The experts supported the content

validity of the identity and causes of the instrument and did

not recommend any additional items. The Chinese IPQ-R

(Trauma) took the 10 patients about 10–15 minutes to

complete. Based on feedback from two patients, the Chinese

Table 1 Original items and modified items in subscale identity and subscale causes

Original items removed

from subscale 1 (Identity)

Items selected

for subscale 1 (Identity)

Original items removed

from subscale 9 (Causes)

Items selected for causes

subscale 8 (Causes)

1. Sore throat 1. Pain 1. Hereditary – it runs in my family 1. Stress or worry

2. Nausea 2. Loss of strength 2. A germ or virus 2. Chance or bad luck

3. Wheeziness 3. Breathlessness 3. Diet or eating habits 3. My own behaviour

4. Upset stomach 4. Weight loss 4. Poor medical care on my past 4. Mental attitude

5. Fatigue 5. Pollution in the environment 5. Family problems or worries caused

6. Stiff joints 5. Smoking 6. Overwork

7. Sore eyes 6. Accident or injury 7. My emotional state

8. Headaches 7. Altered immunity 8. Ageing

9. Sleep difficulties 9. Alcohol

10. Dizziness 10. My personality

W Chaboyer et al.

� 2012 Blackwell Publishing Ltd

3468 Journal of Clinical Nursing, 21, 3466–3474

characters in one item were revised because it was not easily

understood by patients.

Main study

A convenience sample of 232 patients from three hospitals

who had moderate to severe traumatic injuries, as indicated

by injury severity scores (ISS) of nine or greater, were invited

to complete the Chinese IPQ-R (Trauma). ISS is an interna-

tionally accepted method of measuring severity of injury

(Baker & O’Neill 1976). The ISS score ranges from a low of 0

to a high of 75. ISS scores from 9–14 are considered

moderate, and 15 or greater are considered severe trauma.

Study inclusion criteria were as follows: (1) 18 years of age or

over, (2) able to read, write and communicate in Mandarin,

and (3) ISS score of nine or greater. Patients were excluded if

they were unable to provide consents because of cognitive

impairment such as severe brain injuries, stroke or burn

injuries. After participants signed a consent form, they were

interviewed using the Chinese IPQ-R (Trauma) just prior to

hospital discharge. Institutional Review Board approvals

were given by three teaching hospitals in Taiwan and a

university in Australia. Data analysis involved assessing the

Chinese IPQ-R (Trauma) in two ways: first, item analysis and

then psychometric testing, which involved assessing the

construct validity and the reliability of the scale. Each is

described next.

Item analysis

The main purpose of item analysis is to find the best items

that can be used to measure the illness representations

(Murphy & Davidshofer 2005). Item analysis was used to

determine items in the Chinese IPQ-R (Trauma) that should

be removed because they did not add to understanding of

the subscale (DeVellis 2003). The two modified subscales,

identity and causes were excluded in the item analysis,

because they were lists of symptoms and possible reasons

for traumatic injury and were not appropriate for this type

of analysis. Thus, a total of 30 items underwent the item

analysis.

Item analysis involves several steps. First, the discrimina-

tion index was calculated, which tests the difference between

the proportions of high (the top 25%) and low scores (the

bottom 25%) (Murphy & Davidshofer 2005). In essence, this

index helps to determine whether those individuals who score

high on the various scales can be distinguished from those

who score low, with good scales detecting this difference. A

next step in the item analysis involves calculating item-total

correlations to evaluate the discriminative power of each item

(Munro & Page 2004) and how similar each individual item

is to the whole scale. A third step in the item analysis involved

calculating independent t-tests to examine both extremes of

the data (DeVellis 2003). That is, the top and bottom 25% of

scores were compared, with the expectation that there would

be significant differences between the two groups. Because

the Chinese IPQ-R (Trauma) was underpinned by a strong

theoretical framework, only items that performed poorly on

both t-tests (i.e. non-significant differences) and low cor-

rected item-total correlation were deleted. The criteria for

deleting items included both corrected item-total correlation

coefficients below 0Æ3 (Munro & Page 2004) and non-

significant t-test results (DeVellis 2003), demonstrating that

the items did not discriminate well.

Psychometric testing

Construct validity was tested using an exploratory principle

component factor analysis (Pett et al. 2003, Hair et al. 2006,

Watson & Thompson 2006). Because of their composition

and scaling, two subscales, identity and causes, could not be

entered into the factor analysis. The Kaiser–Meyer–Olkin

(KMO) value was 0Æ763, which reflects items for this sample

were adequate for factor analysis, as the KMO exceeded the

recommended value of 0Æ60. The Bartlett’s tests of sphericity

were statistically significant in this sample (Chi-Square

2013Æ400, df 435, p < 0Æ001), suggesting all correlations in

the correlation matrix were greater than zero, as recom-

mended for factor analysis. That is, using a factor analysis on

the Chinese IRQ-R (Trauma) was deemed to be appropriate

(Hair et al. 2006). First, Eigenvalues greater than one and an

inspection of the scree plot were used to determine the factors

(Watson & Thompson 2006). The orthogonal rotation was

used to assist in factor interpretation because there were no

relationships amongst the factor in the correlation matrix

(Watson & Thompson 2006). Nine factors were extracted,

but because using eigenvalues greater than one can overes-

timate the number of factors in the solution and because

illness representations had a strong theoretical foundation

(Kline 1994), a second factor analysis was undertaken,

requesting a forced seven-factor solution.

Reliability of the Chinese IPQ-R (Trauma) was assessed

using Cronbach’s alphas and split-half reliability. All statis-

tical analysis was undertaken in SPSSSPSS (version 13.01, SPSS

Inc., Chicago, IL, USA).

Results

A total 173 (74Æ6%) of the invited patients consented to

participate in this study. The average age of the participants

Methods Taiwanese translation and psychometric testing

� 2012 Blackwell Publishing Ltd

Journal of Clinical Nursing, 21, 3466–3474 3469

was 37Æ2 (SD 14Æ8) years. A total of 122 (69%) were

men and 51 (31%) were women. In addition, the ISS ranged

from 9–29 with a mean of 13Æ5 (SD 4Æ5). Eight items were

removed after the item analysis, as they met the deletion

criteria (DeVellis 2003, Munro & Page 2004) including

two items each from the original personal control and

illness coherence subscales and four items from the

original treatment control subscale. Thus, the psychometric

properties of seven subscales with a total of 30 items were

assessed.

The results of the forced seven-factor solution were very

similar to the seven subscales of the IPQ-R with a few

differences. Two items, (2 and 5), loaded on both the second

and the seventh factors. Conceptually, it made sense for them

to be on the second factor reflecting the timeline (acute/

chronic) subscale as they had been classified as such in the

original IPQ-R which meant that the seven factors were

reduced to six. Only three other items loaded on factors that

were different from the original IPQ-R. Item 6 ‘My injury is a

serious condition’ originally belonged to the consequences

subscale, but loaded on the factor timeline (acute/chronic).

Item 9 ‘My injury strongly affects the way others see me’ was

from the original consequences subscale, but loaded on the

factor emotional representations. Finally, item 23 ‘There is

nothing which can help my condition’ was originally from the

treatment control subscale, but loaded on the factor illness

coherence. In addition, the original subscales personal control

and treatment control loaded on the same factor. Table 2

Table 2 Factors and factor loading and Cronbach’s alphas of the Chinese IPQ-R (Trauma) (n = 173)

No. Mean (SD)

Factors (Varimax rotated)

1 2 3 4 5 6

37 3Æ0 ± 1Æ1 0Æ82 0Æ03 0Æ02 0Æ05 0Æ07 0Æ12

34 3Æ2 ± 1Æ1 0Æ78 �0Æ07 �0Æ09 0Æ01 0Æ27 0Æ18

35 2Æ7 ± 1Æ0 0Æ72 0Æ02 �0Æ05 0Æ02 0Æ12 0Æ00

38 2Æ8 ± 1Æ1 0Æ72 0Æ11 �0Æ02 �0Æ15 �0Æ05 0Æ11

33 2Æ9 ± 1Æ1 0Æ67 �0Æ07 �0Æ15 0Æ08 0Æ27 0Æ17

36 3Æ3 ± 1Æ1 0Æ49 0Æ36 �0Æ01 0Æ00 �0Æ16 0Æ11

9 2Æ7 ± 1Æ0 0Æ42 0Æ15 �0Æ06 �0Æ12 0Æ08 0Æ09

3 3Æ0 ± 1Æ1 0Æ02 0Æ83 �0Æ02 0Æ02 0Æ17 0Æ00

4 2Æ7 ± 1Æ1 �0Æ01 0Æ80 �0Æ08 0Æ13 0Æ16 0Æ06

1 2Æ8 ± 1Æ1 �0Æ01 0Æ75 �0Æ08 0Æ17 0Æ18 0Æ11

6 3Æ4 ± 1Æ0 0Æ38 0Æ49 0Æ09 �0Æ20 0Æ26 0Æ01

5 2Æ5 ± 1Æ1 0Æ14 0Æ41 �0Æ10 0Æ03 0Æ13 �0Æ05

2 2Æ3 ± 0Æ9 0Æ19 0Æ30 �0Æ03 0Æ00 0Æ08 0Æ06

19 3Æ5 ± 0Æ9 �0Æ21 0Æ04 0Æ75 �0Æ01 0Æ05 �0Æ03

15 3Æ5 ± 0Æ9 �0Æ16 �0Æ06 0Æ71 0Æ11 0Æ10 �0Æ06

12 3Æ5 ± 0Æ8 0Æ15 0Æ10 0Æ69 �0Æ04 �0Æ01 0Æ01

17 3Æ1 ± 1Æ0 �0Æ05 �0Æ17 0Æ63 0Æ20 �0Æ11 �0Æ05

14 3Æ3 ± 0Æ9 0Æ05 �0Æ19 0Æ56 0Æ05 0Æ03 �0Æ02

27 3Æ2 ± 1Æ1 0Æ05 0Æ11 0Æ07 0Æ84 �0Æ07 �0Æ12

25 3Æ2 ± 1Æ1 �0Æ02 �0Æ03 0Æ03 0Æ82 �0Æ11 �0Æ03

26 3Æ3 ± 1Æ1 �0Æ05 0Æ17 0Æ12 0Æ81 �0Æ10 �0Æ08

23 3Æ6 ± 0Æ9 �0Æ25 �0Æ15 0Æ31 0Æ34 0Æ11 �0Æ07

8 3Æ6 ± 1Æ0 �0Æ02 0Æ19 0Æ04 �0Æ03 0Æ85 0Æ03

7 3Æ7 ± 1Æ1 0Æ05 0Æ20 0Æ07 �0Æ04 0Æ82 0Æ09

11 3Æ7 ± 0Æ9 0Æ29 0Æ09 0Æ07 �0Æ11 0Æ61 0Æ02

10 3Æ2 ± 1Æ2 0Æ21 0Æ12 �0Æ11 �0Æ14 0Æ42 0Æ01

30 2Æ6 ± 0Æ9 0Æ08 0Æ13 �0Æ02 0Æ02 0Æ04 0Æ87

29 2Æ4 ± 0Æ7 0Æ09 �0Æ07 �0Æ05 �0Æ09 0Æ10 0Æ81

31 2Æ4 ± 0Æ8 0Æ33 0Æ28 0Æ02 �0Æ18 �0Æ03 0Æ59

32 2Æ3 ± 0Æ8 0Æ25 �0Æ06 �0Æ14 �0Æ13 0Æ01 0Æ40

Variance explained 19Æ90 10Æ17 8Æ09 7Æ54 5Æ71 4Æ94

Mean (SD) of factor 2Æ9 ± 1Æ1 2Æ8 ± 1Æ1 3Æ4 ± 0Æ9 3Æ3 ± 1Æ1 3Æ6 ± 1Æ1 2Æ4 ± 0Æ8Cronbach’s a 0Æ83 0Æ80 0Æ71 0Æ76 0Æ71 0Æ69

Split-half reliability 0Æ82 0Æ80 0Æ70 0Æ75 0Æ70 0Æ70

W Chaboyer et al.

� 2012 Blackwell Publishing Ltd

3470 Journal of Clinical Nursing, 21, 3466–3474

shows the factors, factor loading, mean, standard deviation

(SD), Cronbach’s alpha coefficients and the split-half reli-

ability for the Chinese IPQ-R (Trauma) items. The Cron-

bach’s alpha coefficients for the six subscales ranged from

0Æ69–0Æ80, and the split-half reliability coefficients ranged

from 0Æ70–0Æ82.

Following this analysis, six factors extracted for the

Chinese IPQ-R-Trauma were named emotional representa-

tions, timeline (acute/chronic), controllability, illness coher-

ence, consequences and timeline cyclical. All items have a

loading score of >0Æ40 except two. Item 2 and item 23 had

loading levels of 0Æ30 and 0Æ34. The six-factor structure

explained 60Æ31% of the total item variance. To ensure the

change and combination of the items were appropriate, the

researchers went back to review the illness representation

construct to check that the changes did not deviate from the

conceptualisation of illness representations. It did not appear

that combining the personal control and treatment control

scales was an appropriate revision for the Chinese culture.

Interestingly, the original IPQ only had one-scale measuring

control, whereas the IPQ-R split this scale into two.

In summary, after modification and testing, the Chinese

IPQ-R (Trauma) contained 50 items grouped into eight

subscales including the following: (1) 10 commonly experi-

enced symptoms answered by ticking yes or no to understand

the identity of traumatic injuries, (2) 30 items grouped into

six subscales, used to measure the patients’ views of their

traumatic injury, and (3) 10 items used to measure possible

causes of a traumatic injury. Scores range from 1–5 for all

subscales except identity. The identity subscale ranged from

1–10 with higher scores indicating more symptoms. Table 3

presents the Chinese IPQ-R (Trauma) and its scoring on each

subscale.

Discussion

This methodological study involved modifying, translating

and pilot testing and then assessing the psychometric prop-

erties of the Chinese IPQ-R (Trauma). In translation, four

items were difficult to translate because they reflected

abstract concepts in the Chinese culture context and two

items had semantic discrepancies because they were affective

in nature and thus, their expression was somewhat culture

bound. This highlights a serious issue, that is, items in a

questionnaire may be needed to be revised several times to

obtain semantic and idiomatic equivalence between two

languages. Accordingly, use of an expert committee can assist

this process, as we demonstrated in this study. Also, it is

essential to carefully select appropriate translators to assist in

both the translation and back translation process to ensure

the meaning of words is maintained.

The Chinese IPQ-R (Trauma) was pilot tested by three

professionals and 10 patients. It was advantageous to have

trauma experts’ participation at this stage of the process

because professionals who care for patients with traumatic

injury bring a certain level of understanding about their

patients’ condition in a particular cultural context. In this

study, the doctor identified the issue of Chinese and

Taiwanese dialects and confirmed the lists of symptoms for

identity and causes.

Another 173 patients with traumatic injury were inter-

viewed to test the psychometric properties of the Chinese

IPQ-R (Trauma) in the main study. Theoretically, the ideal

number recruited to test a new scale is at least five times,

which was included the number of items (Davis &

Robinson 1995). While the original IPQ-R has 38 items

that would have been assessed, because of the item

analysis, only 30 items were retained. Thus, 173 partici-

pants appear to be a sufficient sample size for the factor

analysis if using the rule of five participants per item (Davis

& Robinson 1995).

Seven factors were reduced to six factors after a factor

analysis. The five cognitive factors were very similar to the

original IPQ, and the sixth was consistent with the emotional

subscale of the IPQ-R. Given these results, the Chinese IPQ-R

(Trauma) appears to be both valid and reliable. That is, five

of the original seven factors extracted from the factor

analyses were named using exactly same names as the IPQ-R

English version. The original two factors personal control

and treatment control were combined into one factor

following the factor analysis and named controllability, as

they were in the IPQ. The result shows that the personal

control and treatment control could not be separately

measured in this population. Further understanding of this

Table 3 The Chinese IPQ-R (Trauma)

Subscales

IPQ-R (Trauma)

Number of items Scoring

Identity 10 1 less to 10 more

symptoms

Causes 10 1 less to 5 more causes

Emotional

representation

7 1 best to 5 worse

Timeline

(acute/chronic)

6 1 best to 5 worse

Consequences 4 1 best to 5 worse

Timeline cyclical 4 1 best to 5 worse

Controllability 5 1 worst to 5 best

Illness coherence 4 1 worst to 5 best

Total items 50

IPQ-R, Illness Perception Questionnaire Revised.

Methods Taiwanese translation and psychometric testing

� 2012 Blackwell Publishing Ltd

Journal of Clinical Nursing, 21, 3466–3474 3471

is needed to identify if this finding is related to the Chinese

culture on the patient group.

Interestingly, in the new-named factor controllability, only

item 19, ‘There is very little that can be done to improve my

injury’ was retrieved from the original factor treatment

control, with all the rest of the items from the personal

control subscale. Another item ‘There is nothing which can

help my condition’ from treatment control was classified into

illness coherence. The results may indicate that treatment

control is not an appropriate concept for patients with

traumatic injury in Taiwan for two possible reasons. First,

with sudden unexpected moderate to severe injury, survival

may be of paramount importance to this group. That is, in

the acute phase, treatment is controlled by clinical profes-

sionals not by patients. Second, patients with traumatic

injury in Taiwan may be more passive compared with

Western patients. Traditionally, physicians are regarded as

authorised professionals in Taiwan. Accordingly, sharing in

decision making and treatment control may be not part of

patients’ illness representations. This difference between

Eastern and Western conceptualisations of the doctor and

patient relationship requires further investigation because it

may have implications for other related concepts such as self

management and self care.

Item 9, ‘My injury strongly affects the way others see me’

was attributed to the factor consequences originally. How-

ever, the item was classified into an emotional factor but it

loaded on the factor emotional representation with a loading

of 0Æ42. This may be because the patients worried that their

traumatic injuries might bring disabilities and cause others to

discriminate against them when they went back to work.

Item 6, ‘My injury is a serious condition’ was originally from

the subscale consequences. The item was classified into the

timeline factor with loading factor 0Æ49. It may be injury was

an accident event, so the relatively young patient group

estimated a period to recover from the injury. The estimation

may be related to when they could back to work. This result

was correspondent to Taiwan’s culture. Thus, the researchers

speculated that the two item results are both culture bound

and injury bound.

The six-factor structure entered into the factor analy-

sis explained about 60% of the total item variance. Addi-

tionally, most items demonstrated satisfactory loading levels

despite two items having a loading level of 0Æ30 and 0Æ34.

However, those results meet the recommended criteria for

loading factors in a study (Munro & Page 2004). In

addition, acceptable Cronbach’s alpha coefficients and split-

half reliability coefficients were obtained, based on DeVellis’s

(2003) suggestions, and support the claime that the subscales

of the Chinese IPQ-R (Trauma) are reasonably reliable.

Although the number of Chinese IPQ-R (Trauma) was

reduced after factor analysis, a total of 50 items may be

still too many for patients with traumatic injury. When

instruments have more items, participants may refuse or

be unwilling to complete them, and the quality of the research

may be jeopardised (See et al. 2007). Therefore, it may be

beneficial to construct a brief version to understand traumatic

injury patients’ illness representations in the future.

Limitations

This study has several limitations. First, while a literature

review was used to develop the identity and causes scale (as

recommended), it may have been better to combine this

review with qualitative interviews with Taiwanese individu-

als who experienced a traumatic injury. However, a strong

theoretical model and research underpinned the original

IPQ-R, and thus, for the most part, the Chinese IPQ-R

(Trauma) is grounded in previous empirical work. Second, it

would have been ideal to compare the scale with other similar

questionnaires that are well constructed and tested. However,

no other scales are available. Thus, the further research will

be needed to establish concurrent validity for the Chinese

IPQ-R (Trauma).

Conclusion

This study reports on the translation and testing of the IPQ-R

from English to Chinese. After adapting the instrument for

the particular population, a structured process for transla-

tion, back translation and assessment of feasibility, readabil-

ity, content and construct validity and reliability was

undertaken. Transcultural translation of an instrument may

have difficulties in relation to semantic equivalence such as

cultural, conceptual and semantic equivalence; however, the

use of experts and pretesting are two strategies can address

this issue. Validity and reliability of the Chinese IPQ-R

(Trauma) have been demonstrated. Further studies will need

to be undertaken to extend the application of the Chinese

IPQ-R (Trauma) in Chinese culture context.

Relevance to clinical practice

Findings from this study provide evidence to indicate that

people with traumatic injury perceived and behaved in

individual ways to self-regulate their injuries. Specifically,

the underlying beliefs have the potential to influence how

individuals view their injuries on controllability in Taiwan.

There may be a window of opportunity to adapt the IPQ-R

for the injured population in Chinese culture.

W Chaboyer et al.

� 2012 Blackwell Publishing Ltd

3472 Journal of Clinical Nursing, 21, 3466–3474

Contributions

Study design: BOL, WC, MW; data collection: BOL, WC,

CSC; data analysis: BOL, WC, CSC and manuscript prepa-

ration: BOL, WC.

Conflict of interest

There is no conflict of interest in this paper.

References

Aitken LM, Davey TM, Ambrose JE,

Connelly LB, Swanson C & Bellamy N

(2007) Health outcomes of adults

3 months after injury. Injury 38, 19–26.

Baker SP & O’Neill B (1976) The injury

severity score: an update. Journal of

Trauma 16, 882–885.

Beaton DE, Bombardier C, Guillemin F &

Ferraz MB (2000) Guidelines for the

process of cross-cultural adaptation of

self-report measures. Spine 25, 3186–

3191.

Beck CT, Bernal H & Froman RD (2003)

Methods to document semantic equiv-

alence of a translated scale. Research in

Nursing & Health 26, 64–73.

Brislin RW (1970) Back-translation for

cross-cultural research. Journal of

Cross-Cultural Psychology 1, 185–216.

Brislin RW (1986) The wording and trans-

lation of research instruments. In

Field Methods in Cross-Cultural Re-

search (Lonner WL & Berry JW eds).

Sage, Newbury Park, CA, pp. 137–164.

Cherrington CC, Moser DK, Lennie TA &

Kennedy CW (2004) Illness represen-

tation after acute myocardial infarc-

tion: impact on in-hospital recover.

American Journal of Critical Care 13,

136–145.

Coutu MF, Dupuis G, D’Antono B &

Rochon-Goyer L (2003) Illness repre-

sentation and change in dietary habits

in hypercholesterolemic patients. Jour-

nal of Behavioral Medicine 26,

133–152.

Covic A, Seica A, Gusbeth-Tatomir P,

Gavrilovici O & Goldsmith JA (2004)

Illness representations and quality of

life scores in haemodialysis patients.

Nephrology Dialysis Transplantation

19, 2078–2083.

Davey T, Aitken LM, Ambrose JE, Connelly

LB, Swanson C, Gee T & Bellamy N

(2004) Study 4b-audult injury outcome

study. Queensland Trauma Plan Pro-

ject. Conrod, Herson.

Davis GC & Robinson MC (1995)

Advanced statistical methods. In

Principles and Practices of Nursing

Research (Talbot LA ed.). Mosby, New

York, pp. 373–383.

DeVellis RF (2003) Scale Development:

Theory and Applications. Sage, Thou-

sand Oaks, CA.

Gordon RGJ (2005) Ethnologue: Languages

of the World, 15th edn. Available at:

http://www.ethnologue.com/ (accessed

18 May 2007).

Hair J, Anderson E, Tathan R & Black W

(2006) Multivariate Data Analysis.

Prentice-Hall, Upper Saddle River, NJ.

Holbrook TL & Hoyt DB (2004) The im-

pact of major trauma: quality-of-life

outcomes are worse in women than in

men, independent of mechanism and

injury severity. The Journal of Trauma

Injury, Infection and Critical Care 56,

284–290.

Jones PS, Lee JW, Phillips LR, Zhang XE &

Jaceldo KB (2001) An adaptation of

Brislin’s translation model for cross-

cultural research. Nursing Research 50,

300–304.

Kline P (1994) An Easy Guide to Factor

Analysis. Routledge, London.

Lee BO, Chaboyer W & Marianne W

(2008) Predictors of health-related

quality of life three months after trau-

matic injury. Journal of Nursing

Scholarship 40, 83–90.

Lee BO, Chaboyer W & Marianne W (2010)

Illness representations in patients with

traumatic injury: a longitudinal study.

Journal of Clinical Nursing 19, 556–563.

Leventhal H, Nerenz D & Steele D (1984)

Illness representation and coping with

health threats. In A Handbook of Psy-

chology and Health (Baum A & Singers

J eds). Erlbaum, Hilladale, NJ, pp.

219–252.

Leventhal H, Leventhal EA & Cameron L

(2001) Representing, procedures and

affect in illness self-regulation: a per-

ceptual-cognitive model. In Handbook

of Health Psychology (Baum A,

Revenson TA & Singer JE eds). Lawence

Eelbaum, Mahwah, NJ, pp. 19–48.

Moss-Morris R, Weinman J, Petrie KJ,

Horne R, Cameron LD & Buick D

(2002) The revised Illness Perception

Questionnaire (IPQ-R). Psychology and

Health 17, 1–16.

Munro BH & Page EB (2004) Statistical

Methods for Health Care Research.

Lippincott, Philadelphia.

Murphy KR & Davidshofer CO (2005)

Psychological Testing: Principles and

Applications. Prentice Hall, Upper

Saddle River, NJ.

Pett MA, Lackey NR & Sullivan JJ (2003)

Making Sense of Factor Analysis: The

Use of Factor Analysis for Instrument

Development in Health Care Research.

Sage, Thousand Oaks, CA.

Richmond TS, Donald D & Schwab CS

(1998) A prospective study of predic-

tors of disability at 3 months after non-

central nervous system trauma. The

Journal of Trauma Injury, Infection

and Critical Care 44, 635–643.

Rutter CL & Rutter DR (2007) Longitudi-

nal analysis of the illness representation

model in patients with irritable bowel

syndrome (IBS). Journal of Health

Psychology 12, 141–148.

See LC, Chang HJ, Liu MJ & Cheng HK

(2007) Development and evaluation of

validity and reliability of a questionnaire

on Medical workers’ stress. Taiwan

Journal of Public Health 26, 452–461.

Watson R & Thompson DR (2006) Use of

factor analysis in Journal of Advanced

Nursing: Literature review. Journal of

Advanced Nursing 55, 330–341.

Weinman J, Petrie KJ, Moss-Morris R &

Horne R (1996) The Illness Percep-

tion Questionnaire: a new method for

assessing the cognitive representation of

illness. Psychology and Health 11,

431–445.

Methods Taiwanese translation and psychometric testing

� 2012 Blackwell Publishing Ltd

Journal of Clinical Nursing, 21, 3466–3474 3473

The Journal of Clinical Nursing (JCN) is an international, peer reviewed journal that aims to promote a high standard of

clinically related scholarship which supports the practice and discipline of nursing.

For further information and full author guidelines, please visit JCN on the Wiley Online Library website: http://

wileyonlinelibrary.com/journal/jocn

Reasons to submit your paper to JCN:High-impact forum: one of the world’s most cited nursing journals, with an impact factor of 1Æ118 – ranked 30/95 (Nursing

(Social Science)) and 34/97 Nursing (Science) in the 2011 Journal Citation Reports� (Thomson Reuters, 2011)

One of the most read nursing journals in the world: over 1Æ9 million full text accesses in 2011 and accessible in over 8000

libraries worldwide (including over 3500 in developing countries with free or low cost access).

Early View: fully citable online publication ahead of inclusion in an issue.

Fast and easy online submission: online submission at http://mc.manuscriptcentral.com/jcnur.

Positive publishing experience: rapid double-blind peer review with constructive feedback.

Online Open: the option to make your article freely and openly accessible to non-subscribers upon publication in Wiley

Online Library, as well as the option to deposit the article in your preferred archive.

W Chaboyer et al.

� 2012 Blackwell Publishing Ltd

3474 Journal of Clinical Nursing, 21, 3466–3474

This document is a scanned copy of a printed document. No warranty is given about the accuracy of the copy.

Users should refer to the original published version of the material.