Please download to get full document.

View again

of 7

The Brief Illness Perception Questionnaire

The Brief Illness Perception Questionnaire
0 views7 pages
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Documenttranscript
  The Brief Illness Perception Questionnaire Elizabeth Broadbent  a, 4 , Keith J. Petrie a  , Jodie Main a  , John Weinman  b a   Department of Psychological Medicine, Faculty of Medical and Health Sciences, University of Auckland, Private Bag 92019, Auckland, New Zealand   b  Health Psychology Section, Department of Psychology (at Guy’s), Institute of Psychiatry, 5th floor Thomas Guy House, London Bridge, London SE1 9RT, UK  Received 12 June 2005; received in revised form 3 October 2005; accepted 18 October 2005 AbstractObjective:  This study evaluates the Brief Illness PerceptionQuestionnaire (Brief IPQ), a nine-item scale designed to rapidlyassess the cognitive and emotional representations of illness. Methods:  We assessed the test–retest reliability of the scale in132 renal outpatients. We assessed concurrent validity bycomparing the Brief IPQ with the Illness Perception Question-naire–Revised (IPQ-R) and other relevant measures in 309 asthma,132 renal, and 119 diabetes outpatients. Predictive validity wasestablished by examining the relationship of Brief IPQ scores tooutcomes in a sample of 103 myocardial infarction (MI) patients.Discriminant validity was examined by comparing scores on theBrief IPQ between five different illness groups.  Results:  TheBrief IPQ showed good test–retest reliability and concurrent validity with relevant measures. The scale also demonstratedgood predictive validity in patients recovering from MI withindividual items being related to mental and physical functioningat 3 months’ follow-up, cardiac rehabilitation class attendance,and speed of return to work. The discriminant validity of theBrief IPQ was supported by its ability to distinguish betweendifferent illnesses.  Conclusion:  The Brief IPQ provides a rapidassessment of illness perceptions, which could be particularlyhelpful in ill populations, large-scale studies, and in repeatedmeasures research designs. D  2006 Elsevier Inc. All rights reserved.  Keywords:  Illness perceptions; Questionnaire; Reliability; Validity; Brief IPQ; Chronic illness Introduction The study of individuals’ perceptions of illness stemmedfrom research into the communication of health threats in the1960s. Leventhal et al. [1] developed the self-regulatorymodel to describe the process by which individuals respondto a perceived health threat. The model proposes that situational stimuli (such as symptoms) generate bothcognitive and emotional representations of the illness or health threat. These representations are processed in parallelthrough three stages. The individual first forms the repre-sentation of the illness or health threat, next, they adopt  behaviours to cope with this, and, lastly, they appraise theefficacy of these behaviours. The model incorporates a con-tinuous feedback loop in which the results of the appraisal process are fed back into the formation of the illness/threat representation and the adoption of coping responses.Early research identified five dimensions within thecognitive representation of illness:  identity  —the label the person uses to describe the illness and the symptoms theyview as being part of the disease;  consequences  —theexpected effects and outcome of the illness;  cause  —  personal ideas about the cause of the illness;  timeline  — how long the patient believes the illness will last; and  cureor control   —the extent to which the patient believes that they can recover from or control the illness [1,2]. Theemotional representation incorporates negative reactionssuch as fear, anger, and distress. Ongoing research over the past 30 years has demonstrated the importance of illnessrepresentations to patient behaviour  [3]. Changing patients’illness perceptions has been shown to improve recoveryfollowing myocardial infarction (MI) [4], and other self-regulatory interventions in illnesses as diverse as diabetesand AIDS have also improved patient outcomes [5].Early research investigating the content of illnessrepresentations largely involved open-ended interviews. Asknowledge has grown and Leventhal’s self-regulatory model 0022-3999/06/$ – see front matter   D  2006 Elsevier Inc. All rights reserved.doi:10.1016/j.jpsychores.2005.10.020 4  Corresponding author. Tel.: +64 9 3737599x84511; fax: +64 9 3737013.  E-mail address:  e.broadbent@auckland.ac.nz (E. Broadbent).Journal of Psychosomatic Research 60 (2006) 631–637  has become more widely used, more objective measureshave been developed. The Illness Perception Questionnaire(IPQ) [6] is a widely used multifactorial pencil-and-paper questionnaire which assesses the five cognitive illnessrepresentations on a five-point Likert scale. A revisedversion of this scale, the Illness Perception Questionnaire– Revised (IPQ-R), extended the original scale by addingmore items, splitting the control dimension into personalcontrol and treatment control, and incorporating a cyclicaltimeline dimension, an overall comprehension of illnessfactor, and an emotional representation [7].The IPQ-R has over 80 items, and in some situationssuch a long questionnaire is prohibitive. This is particularlythe case when patients are very ill or when there is limitedtime available for assessment. A shorter questionnairewould be more suitable for patients who are very ill or elderly because it would be less taxing and much quicker tocomplete. It may also be more acceptable to those who arelimited in their reading and writing ability. The shorter questionnaire offers the potential for illness percept ions to be investigated in a wider range of patient groups [8] andwould be especially useful when illness perceptions aremeasured as only one part of a larger set of psychologicalconstructs, in large population-based studies, and whenrepeated measures are taken on a frequent basis.This research aimed not only to construct a very short and simple measure of illness perceptions, but also toconstruct a measure with an alternative format to themultifactorial Likert scale approach used in the IPQ andIPQ-R. The Brief Illness Perception Questionnaire (Brief IPQ) uses a single-item scale approach to assess perceptionson a continuous linear scale. This paper assesses the psychometric properties of the Brief IPQ using samplesfrom several illness groups and investigates the value of a brief scale to assess illness perceptions. Method Scale development  The Brief IPQ has nine items and is shown in AppendixA. The items were developed by forming one question that  best summarised the items contained in each subscale of theIPQ-R. The Brief IPQ therefore has eight new items plus part of the causal scale previously used in the IPQ-R. All of the items except the causal question are rated using a 0-to-10 response scale. Five of the items assess cognitive illnessrepresentations: consequences (Item 1), timeline (Item 2), personal control (Item 3), treatment control (Item 4), andidentity (Item 5). Two of the items assess emotionalrepresentations: concern (Item 6) and emotions (Item 8).One item assesses illness comprehensibility (Item 7).Assessment of the causal representation is by an open-ended response item adapted from the IPQ-R, which asks patients to list the three most important causal factors intheir illness (Item 9). Responses to the causal item can begrouped into categories such as stress, lifestyle, hereditary,etc., determined by the particular illness studied, andcategorical analysis can then be performed.Like the IPQ and IPQ-R, the most general version of theBrief IPQ uses the word  d illness T , but it is possible to replacethis with the name of a particular illness such as diabetes or asthma. Similarly, the treatment control item uses the word d treatment  T , but this can be replaced by a particular treatment such as  d surgery T  or   d inhaler  T  if researchers areinterested in a particular treatment.  Participants Data were collected from six illness groups to evaluatethe psychometric properties of the scale: MI (  N  =103), renaldisease (  N  =132), type 2 diabetes (  N  =119), asthma(  N  =309), minor illnesses (allergies, colds, headaches)(  N  =166), and a group with chest pain undergoing stress-exercise testing prior to diagnosis (  N  =62). The MI groupwas involved in a psychological intervention trial at Auckland Hospital aimed at improving recovery, and inthese analyses only the control group was used. The renal,diabetes, and stress-exercise test groups were recruited fromoutpatient clinics at Auckland Hospital. The asthma patientswere recruited from general practitioner clinics around theUK by postal questionnaire. The minor illness group wasrecruited from undergraduate classes at The University of Texas who were asked to recall a recent illness. Patient characteristics of these samples are presented in Table 1. Results  Reliability The test–retest reliability of the Brief IPQ was assessedin renal patients attending outpatient clinics. The first questionnaire was filled in at the clinic and then follow-upquestionnaires were sent to half of the participants after 3weeksandtotheother half oftheparticipantsafter 6weeks. Table 1Characteristics of patient samples used in the validation of the Brief IPQIllness group  N  Gender (% male)Age mean(S.D.)Length of illnessmean (S.D.) yearsMyocardial infarction 103 88.3 54.7 (8.1) in hospital post-infarct Renal disease 132 70.7 58.0 (17.4) 8.8 (13.3)Type 2 diabetes 119 52.9 57.2 (13.2) 11.0 (11.1)Asthma 309 41.1 39.8 (10.1) 22.3 (13.4)Minor illnesses 166 39.8 18.4 (1.7)Allergies 65 53.3 18.75 (2.6)Colds 49 43.5 18.0 (0.5)Headaches 52 28.6 18.3 (0.9)Prediagnosis stressexercise testing62 54.8 52.3 (11.3)  E. Broadbent et al. / Journal of Psychosomatic Research 60 (2006) 631 – 637  632  Pearson correlations demonstrate that the items have goodtest–retest reliability over both time periods (see Table 2). Concurrent validity Illness Perception Questionnaire–Revised  To assess the concurrent validity of the Brief IPQ, weasked patients in the renal, diabetes, and asthma samples tocomplete both the Brief IPQ and IPQ-R (the questionnaireswere presented in alternate order between patients). Thecorrelations between the scales are presented in Table 3 andshow that the equivalent scales of the Brief IPQ and theIPQ-R are appropriately correlated.To establish the validity of the causal item, patients’answers to the Brief IPQ were compared with the causalfactors they endorsed in the IPQ-R list. The top four ratedcauses of asthma in the Brief IPQ were hereditary,respiratory virus, pollution, and allergies. These sameanswers were the most commonly endorsed causes of asthma in the IPQ-R (except allergies because it was not included in the IPQ-R). Of all causes given in response tothe Brief IPQ, 75% could be categorised within the 20 causalfactors listed in the asthma IPQ-R. In the renal sample, thetop four rated causes in the Brief IPQ were another medicalcondition or medication for it (e.g., lupus SLE or reaction toarthritis medication), diet, diabetes, and chance. These sameanswers were the highest rated causes in the IPQ-R (except other medical conditions or medication for them because it was not included in the IPQ-R), and 82% of Brief IPQanswers could be categorised into the 21 causal items listedin the IPQ-R. Self-efficacy Because the correlations between the Brief IPQ andIPQ-R personal control and treatment control subscaleswere comparatively low, further validity testing was performed on these dimensions. In social cognitive theory, perceived control is measured in terms of self-efficacy,which is an individual’s belief or level of confidence that they can successfully perform a particular task. Previousresearch with diabetes patients has shown significant moderat e correlations between self-efficacy and perceivedcontrol [9], and we expected to find similar correlations between the Brief IPQ personal control item and self-efficacy. We tested the association between the Brief IPQ personal control item and self-efficacy, using previouslyvalidated measures. These were the self-efficacy scales fromThe Knowledge, Attitude, and Self-Efficacy Asthma Ques-tionnaire [10] and The Multidimensional Diabetes Ques-tionnaire [11]. The Brief IPQ personal control item wassignificantly correlated with diabetes self-efficacy ( r  =.61,  P  b .001) and with asthma self-efficacy ( r  =.47,  P  b .001). Incomparison, the IPQ-R personal control item was not significantly correlated with diabetes self-efficacy ( r  =.26,  P  =.09) but was significantly correlated with asthma self-efficacy ( r  =.39,  P  b .001).  HbA 1c To further validate the Brief IPQ scale in type 2 diabetes patients, we tested its associations with HbA 1c , an estimateof blood glucose control over the past 3 months. Higher HbA 1c  indicates poorer metabolic control. The closest regular blood test to the date of questionnaire completionwas chosen for each patient. Previous research hasfound higher perceived control beliefs and self-efficacy to be related to better self-reported adherence to diet,medication, and exercise, as well as better metabolic control[9]. We therefore expected that higher personal control Table 2Test–retest reliability of the Brief IPQItemRenal sample3 weeks 6 weeksConsequences .70 44  .71 44 Timeline .67 44  .73 44 Personal control .63 44  .42 4 Treatment control .55 44  .70 44 Identity .65 44  .75 44 Concern .66 44  .66 44 Understanding .48 44  .61 44 Emotional response .65 44  .72 444  P  b .01. 44  P  b .001.Table 3Pearson correlations between the Brief IPQ and the IPQ-R Brief IPQIPQ-R Identity Timeline Consequences Personal control Treatment control Concern Emotional response CoherenceIdentity .48 444  .10 .46 444   .01 .01 .31 444  .29 444  .08Timeline .19 444  . 53 444  .30 444   .18 444  .06 .24 444  .10 4  .12 4 Timeline cyclical .34 444  .01 .17 444   .03   .02 .07 .21 444   .10 4 Consequences .40 444  .18 444  .62 444   .22 444   .08 4  .54 444  .47 444  .08Personal control   .08 .03   .06 .33 444  .22 444  .01   .07 .14 444 Treatment control   .14 444   .08   .18 444  .34 444  .32 444   .16 444   .16 444  .10 4 Emotional representation .27 444  .03 .42 444   .24 444   .12 4  .49 444  .63 444   .02Illness coherence   .04 .12 44   .04 .23 444  .24 444   .05   .13 44  . 46 4444  P  b .05. 44  P  b .01. 444  P  b .001.  E. Broadbent et al. / Journal of Psychosomatic Research 60 (2006) 631 – 637   633   beliefs measured by the Brief IPQ would also be associatedwith better metabolic control. Because HbA 1c  was signifi-cantly correlated with the duration of diabetes ( r  =.33,  P  =.001), we conducted partial correlations controlling for length of illness.The partial correlations indicate that, as hypothesised,higher personal control measured by the Brief IPQ wasassociated with lower HbA 1c , which indicates better metabolic control ( r  =  .30,  P  b .01). Also in line with previous research [9], higher identity beliefs measured bythe Brief IPQ were associated with poorer metabolic control( r  =.25,  P  b .05). In addition, higher treatment control beliefswere associated with poorer metabolic control ( r  =.21,  P  b .05). There were no significant correlations with theother items. In contrast, HbA 1c  was not significantlycorrelated with the IPQ-R identity ( r  =.10,  P  =.37), treatment ( r  =.18,  P  =.09), or personal control ( r  =.02,  P  =.86) scales.  Asthma morbidity and beliefs about medication In the asthma sample, we investigated how the Brief  IPQwas related to the Jones Asthma Morbidity Index [12] and the Beliefs about Medicines Questionnaire [13]. In previousstudies, we have found that illness r epresentations areassociated with medication beliefs [14] and we expectedthat the Brief IPQ would display similar patterns to thosefound previously with the IPQ-R. We expected that poorer  perceptions would be associated with higher asthma morbid-ity. The correlations with the Brief IPQ displayed logicalrelationships that confirmed these hypotheses and weresimilartothecorrelationsfoundwiththeIPQ-R(seeTable4).  Predictive validity We investigated whether the Brief IPQ predicted anumber of key outcomes following MI. A multivariateanalysis of variance found that those who attendedrehabilitation classes had a higher identity score at hospitaldischarge (mean=3.37, S.E.=.47) than nonattendees(mean=1.67, S.E.=.59) [  F  (39,1)=5.11,  P  =.03]. We alsofound that slower return to work was significantly asso-ciated with higher concern ( r  =.43;  P  =.03) and with higher treatment control beliefs ( r  =.44;  P  =.03). The Brief IPQ at discharge also predicted cardiac anxiety measured by theCardiac Anxiety Questionnaire [15] and quality of life measured by the Seattle Angina Questionnaire [16] and the SF-36 vitality and mental health scales [17] (only these partsof the SF-36 were used) 3 months after the MI. Theseassociations are shown in Table 5. Table 4Correlations between Illness Perception Scales, the Jones Asthma Morbidity Index, and the Beliefs About Medication Questionnaire (BMQ) in an asthmasampleJones Asthma Morbidity Index BMQ necessity BMQ concernsBrief IPQ IPQ-R Brief IPQ IPQ-R Brief IPQ IPQ-R Consequences .39 444  .37 444  .46 444  .46 444  .13 4  .28 444 Timeline .02 .11 .24 444  .39 444   .09   .11Personal control   .18 44   .10   .13 4   .11   .24 444   .08Treatment control   .13 4   .16 44  .12 4   .03   .35 444   .33 444 Identity .42 444  .28 444  .34 444  .26 444  .08 .25 444 Concern .32 444  .40 444  .29 444 Understanding   .09   .08 .13 4  .02   .26 444   .35 444 Emotional response .25 444  .32 444  .34 444  .34 444  .26 444  .35 4444  P  b .05. 44  P  b .01. 444  P  b .001.Table 5Associations between the Brief IPQ and 3-month outcomes in myocardial infarction patientsCardiac AnxietyQuestionnairetotal scoreSF36 Seattle Angina Questionnaire a  VitalityMentalhealthPhysicallimitationAnginafrequencyAnginastabilityTreatment satisfactionDisease perceptionConsequences .33 4   .52 444   .58 444   .11 .09   .13   .20   .27Timeline   .08   .09 .10 .24 .18   .30   .04 .04Personal control   .09 .12 .11 .07   .05 .36 4  .08 .01Treatment control   .02 .20 .24 .17   .21 .08 .25 .09Identity .36 4   .45 44   .45 44   .50 444   .36 4   .08   .05   .34 4 Concern .36 4   .32 4   .46 44   .36 4   .24 .29   .04   .38 4 Understanding   .21 .39 4  .20 .25 .03 .16 .40 4  .33 4 Emotional response .47 44   .21   .45 44   .32 4   .06 .01   .05   .45 44 a  Higher scores on the Seattle Angina Questionnaire indicate better functioning. 4  P  b .05. 44  P  b .01. 444  P  b .001.  E. Broadbent et al. / Journal of Psychosomatic Research 60 (2006) 631 – 637  634   Discriminant validity To assess the extent to which the Brief IPQ coulddistinguish between different illnesses, we compared meanscores across people with diabetes, asthma, colds, MI patients prior to discharge, and prediagnosis chest pain patientsawaiting stress-exercise testing. Each of these illnesses variesin presentation, chronicity, effects on patients lives, andmanageability, and the chest pain group has no formaldiagnosis. We were interested in whether the Brief IPQcould identify distinct patient beliefs in these groups. A seriesof one-way ANOVA with Scheffe post hoc tests showedsignificant differences between illnesses as indicated inTable 6. The differences were in line with expectations. For example, those with chronic illnesses (asthma and diabetes)had much longer timeline perceptions than all of those in theother illness groups, and MI patients had longer timeline perceptions than the colds and prediagnosis groups. Patientswith the greatest control beliefs both in terms of personalcontrolandtreatmentcontrolwerethehospitalisedMIpatientswho were at the time receiving new medical and surgicaltreatments as well as lifestyle advice. Patients with the lowest control beliefs were those with colds (a virus for whichantibiotics are ineffective) and those who were not yet diag-nosed and therefore had no information on appropriate be-haviours or treatments. In terms of emotional representations,as would be expected, people with colds were the least con-cerned, while those with diabetes, an illness with potentiallysevere long-term complications, were the most concerned. Discussion This paper reports the psychometric properties of a newnine-item scale, the Brief IPQ. The scale measures patients’cognitive and emotional representations of their illnessincludingconsequences,timeline,personalcontrol,treatment control, identity, coherence, concern, emotional response,and causes. The Brief IPQ allows very simple interpretationof scores: increases in item scores represent linear increasesin the dimension measured. Results indicate that the Brief IPQ has good test–retest reliability, and there are moderateto good associations between the Brief IPQ and the IPQ-R on all the equivalent dimensions. The lowest associationsare between the control dimensions. Measuring control perceptions on a single-item scale corresponds with thetraditional measurement of self-efficacy strength (percent-age confidence that one can perform a behaviour) [18].Support for the validity of the Brief IPQ personal controlitem is provided by its association with self-efficacy.Furthermore, diabetes patients’ blood glucose control wasassociated with the Brief IPQ personal control, treatment control, and identity items. Overall, the pattern between theBrief IPQ and the IPQ-R is fairly comparable, but perhaps inthe control area the more direct and straightforwardapproach of the Brief IPQ may have an advantage.The causal question in the Brief IPQ identified thesame top-rating causal factors as did the IPQ-R in bothasthma and renal samples. This is in line with a recent systematic review that found no differences betweenstudies that measured experimenter-generated causal beliefsandstudiesthatmeasuredrespondent-generatedcausalbeliefs[19]. It is of note that not all the responses to the Brief IPQcausal question could be categorised into the items listed inthe IPQ-R. This highlights the advantage of the open-endedcausal question in the Brief IPQ to identify causal beliefs that are not listed, for example, allergies in the asthma population.The method of analysing the causal dimension in theBrief IPQ is likely to depend on the aims of the study. Insome cases, it may be best to analyse only the first-rankedcause, and in other cases it may be better to include all threeof the causes generated by patients. Another method may beto categorise answers into groups that fit the particular illness, such as risk factors for MI that cannot be changed(e.g., hereditary, ageing) and risk factors that can bechanged (e.g., diet, lack of exercise). In large datasets, wesuggest that researchers first look at a sample of responses tothe causal question to work out the appropriate range of causal categories. Data can be coded into these categories,which, if required, can be later collapsed into smaller clusters of causal beliefs.The Brief IPQ demonstrated good predictive validity in patients recovering from MI. The consequences, identity,concern, understanding, and emotional response at dischargewere all fairly consistently related to mental and physicalfunctioning at 3 months’ follow-up. Identity also predicted Table 6Brief IPQ mean scores (S.D.) in diabetes, asthma, colds, prediagnosis, and MIDiabetes Asthma Colds MI (discharge) Prediagnosis  F  Consequences 4.7 (2.9) a  3.5 (2.3) a,b 3.8 (2.2) 4.1 (2.8) 4.6 (2.6)  b 6.6 4 Timeline 9.2 (1.9) a,b,c 8.8 (2.2) d,e,f  5.4 (3.1) a,d,g 7.2 (3.1) c,f,g,h 4.5 (3.0)  b,e,h 67.7 4 Personal control 6.7 (2.3) a,b 6.7 (2.4) c,d 4.7 (2.5) a,c,e 7.7 (1.7) e,f  5.2 (2.8)  b,d,f  15.2 4 Treatment control 8.0 (2.3) a,b 7.9 (2.0) c,d 5.5 (2.9) a,c,e 8.8 (1.2) e,f  5.3 (2.8)  b,d,f  32.7 4 Identity 4.6 (2.8) a  4.5 (2.3)  b 4.5 (2.4) 3.1 (2.6) a,b,c 5.1 (2.5) c 4.7 4 Concern 7.0 (3.1) a,b 4.6 (2.8) a,c,d,e 2.5 (2.5)  b,c,f,g 6.2 (3.4) e,g 6.0 (3.0) d,f  28.3 4 Understanding 7.9 (2.3) a,b,c 6.5 (2.6) a,d 6.4 (2.7)  b 8.0 (2.2) d,e 6.1 (2.9) c,e 11.6 4 Emotional response 4.3 (3.3) a  3.3 (2.9) a,b 3.8 (2.9) 4.2 (3.1) 5.2 (2.8)  b 7.4 4 Superscripts (a, b, etc.) denote pairs of groups different at .05 level Scheffe test. 4  P  b .001.  E. Broadbent et al. / Journal of Psychosomatic Research 60 (2006) 631 – 637   635
Advertisement
MostRelated
View more
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks
SAVE OUR EARTH

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!

x