development and validation of the caregiver guilt questionnaire

International Psychogeriatrics (2010), 22:4, 650–660 C© International Psychogeriatric Association 2010doi:10.1017/S1041610210000074

Development and validation of the CaregiverGuilt Questionnaire

.........................................................................................................................................................................................................................................................................................................................................................................

Andres Losada,1 Marıa Marquez-Gonzalez,2 Cecilia Penacoba1

and Rosa Romero-Moreno1

1Psychology Department, Universidad Rey Juan Carlos, Madrid, Spain2Clinical and Health Psychology Department, Universidad Autonoma de Madrid, Madrid, Spain

ABSTRACT

Background: Family care of frail elderly people has been linked to significant negative consequences forcaregivers’ mental health. Although outcome variables such as burden and depression have been widelyanalyzed in this population, guilt, an emotion frequently observed in caregivers, has not received sufficientattention in the research literature.

Methods: Face-to-face interviews were carried out with 288 dementia caregivers. Guilt was measured usingthe Caregiver Guilt Questionnaire (CGQ).

Results: Using principal components analysis, 22 items were retained and five factors were obtained whichexplained 59.25% of the variance. These factors were labeled: guilt about doing wrong by the care recipient,guilt about not rising to the occasion as caregivers, guilt about self-care, guilt about neglecting other relatives,and guilt about having negative feelings towards other people. Acceptable reliability indexes were found, andsignificant associations between the CGQ and its factors and the Zarit Burden Interview guilt factor werealso found. Caregivers with higher scores on the CGQ also scored higher in depression, anxiety, frequencyand appraisal of behavioral problems. Negative associations between the CGQ and its factors and frequencyof/and satisfaction with leisure and social support were also found. Being female and caring for a parent wereassociated with higher scores on the CGQ.

Conclusions: Feelings of guilt are significantly related to caregiver distress. The CGQ may be a useful measurefor acknowledging feelings of guilt in caregivers; moreover, it can be used as an outcome variable forpsychoeducational interventions aimed at reducing caregiver distress.

Key words: anxiety, assessment, behavioral problems, burden, dementia, depression, leisure, social support

Introduction

Family care of frail elderly people has been linkedto significant negative consequences for caregivers’mental health. Outcome variables such as burdenand depression have been widely analyzed, withmany research studies demonstrating that caregiversas a group, especially if they are caring for relativeswith Alzheimer’s disease and other dementias, showhigher scores on depression and stress than non-caregivers (e.g. Pinquart and Sorensen, 2003).Other negative emotional outcomes in caregiving,such as anger or anxiety, have received moreattention in recent years (Coon et al., 2003;

Correspondence should be addressed to: Andres Losada-Baltar, Facultad deCiencias de la Salud. Departamento de Psicologıa, Edificio Departamental II.Avda. de Atenas, s/n. 28922 Alcorcon, Madrid, Spain. Phone: + 34 914888941.Email: [email protected]. Received 30 Sep 2009; revision requested 26 Oct2009; revised version received 4 Jan 2010; accepted 5 Jan 2010. First publishedonline 22 February 2010.

Cooper et al., 2008). There remain, however, someemotions, such as guilt, which have not receivedadequate attention in the research literature, eventhough they are frequently observed in caregivers(Yaffe, 1988).

Guilt has been described as “the dysphoricfeeling associated with the recognition that onehas violated a personally relevant moral orsocial standard” (Kugler and Jones, 1992). Guilthas also been suggested as a factor potentiallycontributing to depression and distress in non-dementia caregivers (Boye et al., 2002; Spillerset al., 2008) and non-caregiving samples (Ghataviet al., 2002) and some research considers it as amain emotion for caregivers, which may exacerbatetheir burden (e.g. Brodaty, 2007). The associationbetween guilt and burden or depression may beeither direct or indirect, that is, mediated ormoderated by other relevant variables in caregiving,

Caregiver Guilt Questionnaire 651

such as social support or frequency of andsatisfaction with leisure or pleasant activities.

Despite the fact that researchers have called forthe analysis of guilt in both community and long-term care (e.g. Hamel, 1995; Ankri et al., 2005),the number of studies analyzing guilt in dementiacaregiving remains small.

Gonyea et al. (2008) analyzed guilt by ratinganswers to an open-ended question posed to 66caregivers of elderly relatives (“Some women tellus that they feel guilty about their caregiving: howguilty would you say you feel about it?”). Drawingupon the stress and coping model, these authorsfound that guilt contributed to the explanation of asignificant proportion of burden, even when othersignificant variables were controlled for (e.g. basicand instrumental activities of daily living).

Guilt has also been analyzed in dementiacaregiving as one of the components (factors) ofburden. Through principal component analysis ofthe Zarit Burden Interview (ZBI; Zarit et al.,1980), Ankri et al. (2005) obtained a factor madeup of four items, which they called “guilt”, asmost of these items assessed a general sense ofinadequacy as a caregiver (e.g. “Do you feel thatyou should be doing more for your relative?”or “Do you feel you could do a better jobin caring for your relative”). They found thissubscale to be significantly related to the patient’sbehavioral problems (e.g. verbal aggressiveness andsadness). In a study with caregivers of childrenwith intellectual disabilities, Gallagher et al. (2008)found that guilt, as assessed with the ZBI, was thestrongest predictor of caregiver anxiety. However,even though the ZBI’s guilt factor may functionas a good indicator of a general sense of guiltin caregivers, we consider this feeling to be arather complex one, since as well as this generalfeeling of inadequacy as a caregiver there maybe different types of guilt specifically associatedwith different sources or situations potentiallyaffecting a caregiver’s emotional state; for example,interactions with the care recipient, role conflict,devoting time to themselves, etc. The analysis ofthe specific type of guilt or the diverse sourcescontributing to the generation of this emotion maybe very helpful for clinicians working with thepopulation in question.

To our knowledge, there is only one instrumentthat has been specifically developed for measuringguilt in dementia caregivers, namely, the CaringGuilt Scale developed by Martin et al. (2006).This scale focuses on feelings of regret and senseof responsibility (e.g. “If I did not spend mytime caring for my relative I know I would feeldeep regret”). Martin et al. (2006) analyzed therelationship between guilt and depression in a

sample of 70 dementia caregivers, finding nosignificant association between the two variables. Itis important to note, however, that the six itemsmaking up the Caring Guilt Scale are written inconditional terms (e.g., “I would worry about myrelative if I did not care for them as I do”), so that theanswers may be reflecting not the current situation,but rather an imagined one. Also, the psychometricdata the authors provided for the scale were scarce.

Considering that (a) guilt is an important feelingfor caregivers, (b) improving our knowledge aboutthis variable will increase our ability to understandthe factors contributing to caregiver distress, and(c) there are few instruments available formeasuring this construct, the aim of this study is todevelop a measure for assessing guilt in dementiacaregivers and to assess its psychometric properties.

Methods

PopulationParticipants were caregivers of relatives withdementia living in the community. All caregiverswere recruited through social and health carecenters in Madrid (Spain) between September 2005and May 2009. Once located, an initial interviewwas carried out by telephone in order to confirmthat volunteers were the primary caregivers of familymembers with dementia. In order to participatein the study, caregivers were required to identifythemselves as the primary source of help for theirrelatives, and to report devoting more than onehour per day to caregiving duties for more thanthree months during the assessment period. Two-hundred-and-ninety caregivers were called for theassessment interview. Two of them quitted theinterview referring to lack of time due to caregivingduties. The sample characteristics of the finalsample (288 caregivers) are shown in Table 1.Caregivers gave their consent to participate in thestudy, and approval for the research was obtainedfrom both the Spanish Ministry of Education andthe Ethics Committee at the Rey Juan CarlosUniversity (Madrid).

MeasuresFace-to-face interviews were carried out by trainedpsychologists with each of the caregivers atthe participating social and health centers. Theinterviews lasted approximately 90 minutes, andincluded the following variables (means, standarddeviations and range for the assessed variables areshown in Table 2):

Sociodemographic information. This comprised thecaregiver’s and care recipient’s age, gender and

652 A. Losada et al.

Table 1. Sample characteristics

N % M SD RANGE...........................................................................................................................................................................................................................................................................................................

Caregiver age (years) 288 59.63 12.60 29–87Caregiver sex

Female 228 79.2Male 60 20.8

Daily hours caring 10.91 7.84 1–24Time caring (in months) 53.15 46.10 3–312Co-residence

Yes 225 78.1No 63 21.9

Care recipientSpouse 107 37.2Parent 166 57.6Other relative (father-in-law, 15 5.2

mother-in-law,aunt, etc.)

Care recipient age 78.97 8.35 48–97Relative’s illness

Alzheimer’s disease 167 58.4Other dementia 121 41.6

Table 2. Means, standard deviations and range forthe assessed variables

M SD RANGE........................................................................................................................................................

Burden 30.50 15.90 1–69Frequency of behavioral 34.90 14.68 0–83

problemsAppraisal of problem 17.81 14.62 0–64

behaviorsCare recipient functional 66.34 31.01 0–100

statusFrequency of leisure 8.21 4.41 0–23

activitiesSatisfaction with leisure 6.57 3.56 0–19

activitiesSocial support 10.63 3.84 1–18Anxiety 16.14 8.91 0–36Depression 18.00 11.73 0–55

(45.49∗)

∗Percentage of caregivers above clinical cut-off for depression.

kinship, time spent caring (in months), and numberof hours per day devoted to care.

Guilt. Following a review of the literature, andbased on the clinical experience of the authors andthree other experts consulted for this purpose, wedeveloped an initial pool of 34 items measuringguilt. All the items were originally created for thedevelopment of the Caregiver Guilt Questionnaireand no initial domains or factors were hypothesized,given the limited literature in this area. Responseoptions range from 0 (“never”) to 4 (“always oralmost always”) (see Appendix).

Burden. The Zarit Burden Interview (Zarit et al.,1980) was used, and the ZBI guilt factor obtained byAnkri et al. (2005) was used to analyze convergentvalidity. Because the burden scale was included inthe latter stages of the study, data for this instrumentare only available for 166 caregivers. The internalconsistency found in this study for this scale was0.90 (Cronbach’s α).

Frequency and appraisal of behavioral problems.The Revised Memory and Behavior ProblemsChecklist (Teri et al., 1992) was used to assess thesevariables. It includes 24 items which, through twosubscales, assess both the frequency of behavioralproblems and the appraisal of these behaviors asbeing irritating or upsetting for the caregivers. Inthis study, the internal consistency found for thefrequency scale was 0.82, and for the appraisal scalewas 0.88 (Cronbach’s α).

Functional status. The care recipient’s functionalstatus was assessed with the Barthel Index(Mahoney and Barthel, 1965). Higher scores on thismeasure are indicative of better physical functioningof the relative. Internal consistency of this scale was0.92. (Cronbach’s α).

Frequency of and satisfaction with leisure.Satisfaction with leisure time was rated using theStevens et al. (2004) Leisure Time satisfactionmeasure. This instrument consists of six itemsmeasuring the level of satisfaction felt by caregiverswith regard to the amount of time spent on leisureactivities over the past month (quiet time foryourself, engaging in hobbies, etc.). It was measuredon a three-point Likert-type scale (0 = “not at


all”; 2 = “a lot”). Frequency of leisure activitieswas measured using an adaptation of the LeisureTime satisfaction measure. Caregivers were askedto report the extent to which they had participatedin the assessed activities over the previous month.This was measured on a five-point Likert-type scale(0 = “not at all”; 4 = “a lot”). Internal consistencyindices in the present study were 0.83 for thesatisfaction scale and 0.73 for the frequency scale(Cronbach’s α).

Social support. The Psychosocial SupportQuestionnaire (PSQ; Reig et al., 1991) was used.This scale has six items (e.g. “My friends and/orrelatives pay me visits at home”) that were used toassess caregivers’ perceptions of the frequency ofsupport they receive. Response options range from0 (“never”) to 3 (“very often”). In the present studywe found an internal consistency index for this scaleof 0.78 (Cronbach’s α).

Anxiety. The Tension-Anxiety subscale from theProfile of Mood States (POMS; McNair et al.,1971) was used. It consists of nine items (e.g.“tense”), and caregivers are asked to rate on aLikert-type scale (0 = “not at all”; 4 = “extremely”)how they felt during the previous week. Internalconsistency for this scale in this study was 0.85(Cronbach’s α).

Depression. Depression was assessed by means ofthe Center for Epidemiological Studies-DepressionScale (CES-D; Radloff, 1977), a 20-item measurethat assesses depressive symptomatology (e.g. “I feltsad”), with response options ranging from 0 (“rarelyor none of the time [less than one day per week]”)to 3 (“most or all of the time [5–7 days perweek]”).Scores of 16 or higher may be indicative of clinicaldepression (Radloff and Teri, 1986). Cronbach’s α

coefficient for this scale was 0.90.

Results

Factorial validity of the Caregiver GuiltQuestionnaireFollowing the usual procedures for the developmentof assessment instruments, a principal componentsanalysis of the 34 initial items of the scale wascarried out using varimax rotation, in order tomaximize the orthogonality of the factors. Fivefactors were obtained with eigenvalues higher than1.0. The ten items showing the lowest factorloadings (<0.45) on the corresponding factorwere removed. In subsequent analysis, five factorsexplained 56.58% of the variance. Next, anothertwo items with the lowest factor loadings onthe corresponding factor (<0.45) were deleted.The resulting 22 items loaded on five factors

(eigenvalues from 4.08 for the first factor to 1.77 forthe fifth factor) explained 59.25% of the variance.Factor loadings are shown in Table 3.

The first factor includes seven items referring tocaregivers’ feelings of guilt associated with negativefeelings, emotions or acts towards their relative(e.g. “I have felt bad about getting angry with theperson I’m caring for”). This factor, of which allitems are positively scored, was labeled guilt aboutdoing wrong by the care recipient. The six items ofthe second factor mainly reflect caregivers’ feelingthat, in general, they could be doing a better jobas a caregiver (e.g. “I have thought that perhapsI’m not caring well for my relative”). This factorwas labeled guilt about failing to meet the challengesof caregiving. All items are positively scored exceptitem number 6, which is inversely scored. The thirdfactor contains four items regarding caregivers’negative feelings about looking after themselvesand taking part in activities other than caring fortheir relative (e.g. “I have felt bad for leaving myrelative in the care of someone else while I hadfun”). This factor was labeled guilt about self-care,and all its items are positively scored. The fourthfactor is made up of two items reflecting caregivers’negative feelings associated with the fact of notbeing able to devote as much time as they wouldwish to their other relatives (e.g. “I have felt badfor not looking after my other relatives [husband,wife, children, etc.] as I should”). This factor waslabeled guilt about neglecting other relatives, and all itsitems are positively scored. The fifth and final factorincludes three items about caregivers’ negativefeelings related to having had negative emotions orfeelings towards other people who do not devotetime and effort to caregiving duties as they should,or who do not have such responsibilities. Thisfactor was labeled guilt about having negative feelingstowards other people, and all its items are positivelyscored.

Scores for the resulting guilt subscales aredetermined by computing the sum of the scoresof the items belonging to the corresponding scale,with item number 6 being first reversed. The totalscore on the Caregiver Guilt Questionnaire (CGQ)consists of the sum of the scores of all the subscales.A higher score reflects greater guilt.

ReliabilityAcceptable to good reliability indices were obtainedfor the CGQ factors and the full scale. Specifically,a Cronbach’s α of 0.88 was obtained for the totalscale, 0.89 for the guilt about doing wrong by thecare recipient factor; 0.76 for the guilt about failingto meet the challenges of caregiving factor; 0.69 forthe guilt about self-care factor; 0.86 for the guilt


Table 3. Factor loadings of the CGQ items

FACTO R S

C G Q I T E M S 1 2 3 4 5...........................................................................................................................................................................................................................................................................................................................

11. I have felt bad about telling off the person I’m caring for, for somereason

0.849

10. I have felt bad about getting angry with the person I’m caring for 0.83214. I have felt bad about not having more patience with the person I’m

caring for0.754 0.375

2. I have felt guilty about the way I’ve sometimes behaved with myrelative

0.753

12. I’ve got angry with myself for having negative feelings toward theperson I’m caring for

0.693

8. I have felt bad about things I may have done wrong with the personI’m caring for

0.602 0.340

20. I have felt guilty about having so many negative emotions in relation tocaring

0.513 0.341

22. I have felt guilty thinking that my lack of information andpreparedness might mean that I’m not handling the care of my relativein the best way possible

0.771

9. I have thought that perhaps I’m not caring well for my relative 0.317 0.71121. I have thought that the way I care for my relative may not be

appropriate and may make his/her problem get worse0.636

13. I’ve found myself thinking that I’m not up to the job 0.5885. I have thought that I’m not doing things right with the person I’m

caring for0.545

6. I have thought that, given the circumstances, I’m doing a good job as acaregiver

−0.510 −0.330

16. I have felt bad for leaving my relative in the care of someone else whileI had fun

0.774

15. I have felt bad about leaving my relative in the care of someone elsewhile I do my own things (e.g. work, shopping, going to the doctor)

0.706

7. When I’ve gone out to do some pleasant activity (e.g. eating out in arestaurant), I’ve felt guilty and unable to stop thinking that I should becaring for my relative

0.684

1. I have felt bad about having made some plan or done some activitywithout taking my relative into account

0.661

3. I have felt bad for not looking after my other relatives (husband, wife,children . . .) as I should, due to my caregiving

0.873

4. I have felt bad about not being able to devote more time to my family(husband, wife, children . . .), due to my caregiving

0.844

18. I have felt like a bad person for hating and/or envying other relativeswho could have taken responsibility for some caring and do not do so

0.833

19. I have felt bad for having negative feelings (e.g. hate, anger orresentment) toward some relatives

0.766

17. I have felt guilty about having wished that others “could have thisburden” or suffer as I do

0.590

Note: Numbers in bold in each column represent items clustered to the corresponding factor. Items <0.30 are not shown.

about neglecting other relatives factor; and 0.61 forthe guilt about having negative feelings towards otherpeople factor.

Associations with demographic variablesMean differences in the CGQ factors and totalscore across sex and kinship (spouses and children)

were analyzed through t tests. The results of theseanalyses are shown in Table 4. Female caregiversscored significantly higher than male caregivers onthe factors guilt about neglecting other relatives (t =−2.94; p < 0.01) and guilt about having negativethoughts towards other people (t = −2.15; p < 0.05).Female caregivers also obtained higher CGQ totalscores (t = −2.16; p < 0.05). Significant differences


Table 4 Guilt mean scores (and standard deviations) across sex and kinship

G U I LTA B O U TDOINGWRONG BYT H E CA R ERECIPIENT

G U I LTA B O U TFA ILING TOMEET THECHALLENGESO FCAREGIVING

G U I LTA B O U TSELF-CARE

G U I LT A B O U TNEGLECTINGOT HERRELATIVES

G U I LTA B O U THAVINGNEGATIVEFEELINGSTOWARDSOT HERPEOPLE

CGQ TOTALSCORE

.............................................................................................................................................................................................................................................................................................................................

SexFemale 8.83 (6.45) 5.14 (4.66) 1.98 (2.74) 1.06 (1.87) 1.64 (2.23) 18.69 (12.83)Male 7. 51 (6.11) 4.55 (4.62) 1.24 (2.39) 0.50 (1.13) 1.10 (1.57) 14.86 (11.81)

Care recipientSpouse 7.33 (6.27) 4.03 (4.07) 1.34 (2.50) 0.61 (1.27) 0.89 (1.46) 13.71 (10.95)Parent 9.70 (6.38) 5.65 (5.01) 2.13 (2.79) 1.17 (2.02) 1.99 (2.42) 20.85 (10.08)Other relative 7.56 (6.20) 4.96 (4.35) 1.81 (2.63) 0.89 (1.62) 1.35 (1.92) 14.40 (11.38)

Table 5. Correlations and intercorrelations between guilt and other variables

G U I LTA B O U TDOINGWRONGBY THEC A R E R E-CIPIENT

G U I LTA B O U TFA ILINGTO MEETTHE CHAL-LENGES O FC A R E G IV-ING

G U I LTA B O U TSELF-CARE

G U I LTA B O U TNEGLECT-INGOT HERRELATIVES

G U I LTA B O U THAVINGNEGATIVEFEELINGSTOWARDSOT HERPEOPLE

CGQTOTA LSCORE

G U I LT(ZBIFACTO R)

.........................................................................................................................................................................................................................................................................................................................

Guilt (ZBI factor) 0.401∗∗ 0.352∗∗ 0.182∗ 0.136 0.177∗ 0.455∗∗ –Frequency of

behavioral problems0.334∗∗ 0.331∗∗ 0.324∗∗ 0.102 0.230∗∗ 0.418∗∗ 0.338∗∗

Appraisal of behavioralproblems

0.456∗∗ 0.457∗∗ 0.228∗∗ 0.232∗∗ 0.158∗ 0.509∗∗ 0.337∗∗

Functional status 0.091 0.149∗ −0.055 −0.090 −0.031 0.057 −0.070Frequency of leisure −0.145∗ −0.113 −0.171∗∗ −0.075 −0.096 −0.192∗∗ −0.195∗

Satisfaction withleisure

−0.248∗∗ −0.165∗∗ −0.151∗ −0.097 −0.140∗ −0.286∗∗ −0.271∗∗

Social support −0.168∗∗ −0.157∗ −0.034 −0.073 −0.061 −0.188∗∗ −0.305∗∗

Depression 0.388∗∗ 0.364∗∗ 0.248∗∗ 0.212∗∗ 0.120∗ 0.462∗∗ 0.308∗∗

Anxiety 0.440∗∗ 0.368∗∗ 0.173∗∗ 0.177∗∗ 0.198∗∗ 0.462∗∗ 0.346∗∗

Guilt about doingwrong by the carerecipient

– 0.559∗∗ 0.352∗∗ 0.218∗∗ 0.339∗∗ 0.854∗∗ 0.401∗∗

Guilt about not risingto the occasion ascaregivers

– 0.364∗∗ 0.159∗∗ 0.255∗∗ 0.765∗∗ 0.352∗∗

Guilt about self-care – 0.210∗∗ 0.283∗∗ 0.585∗∗ 0.182∗

Guilt about neglectingother relatives

– 0.142∗ 0.383∗∗ 0.136

Guilt about havingnegative feelingstowards other people

– 0.507∗∗ 0.177∗

p < 0.05; ∗∗ p < 0.01

in CGQ score and all the factors were also found onconsidering kinship. Spouses obtained lower CGQtotal scores than those caring for a parent, andscored lower in all CGQ factors: guilt about doingwrong by the care recipient (F = 5.44; p < 0.01), guilt

about failing to meet the challenges of caregiving (F =5.51; p < 0.01), guilt about self-care (F = 5.84; p <

0.01), guilt about neglecting other relatives (F = 3.02;p < 0.05) and guilt about having negative thoughtstowards other people (F = 12.43; p < 0.01).


Convergent and criterion validityCorrelations between CGQ factors and total scoreare shown in Table 5. Convergent validity wasassessed through the correlation between CGQfactors and total score, and guilt as measured bythe ZBI guilt factor (Ankri et al., 2005). As it can beseen, significant correlations were found betweenthe ZBI guilt factor and CGQ factors and totalscore, except in the case of the factor Guilt aboutneglecting other relatives. No significant differenceswere found in guilt scores between those caregiverswho completed the ZBI and those who did notexcept for the factor Guilt about neglecting otherrelatives. Caregivers who completed the ZBI hadhigher scores (M = 1.15; SD = 1.01) in this factorthan those who did not (M = 0.71; SD = 1.38)(t = −2.21; p < 0.05).

Significant associations between CGQ factorsand total score and appraisal of problem behaviors,depression and anxiety were also found. Caregiverswith higher CGQ scores also scored higher onappraisals, depression and anxiety. Similar findingswere obtained for frequency of behavioral problems,except that the association between this variable andthe CGQ factor guilt about neglecting other relativeswas not significant. Functional status was foundto be significantly associated only with guilt aboutfailing to meet the challenges of caregiving factor:caregivers who report higher functional status oftheir care recipients also score higher scores onthis CGQ factor. Caregivers who are less satisfiedwith their leisure time score higher on the CGQand all its factors, except for the factor guilt aboutneglecting other relatives. Similar findings were foundon analyzing frequency of leisure activities: thosecaregivers who report taking part in more leisureactivities score lower on guilt about doing wrongby the care recipient, guilt about self-care and CGQtotal score. Caregivers with higher scores on socialsupport present lower scores on guilt about doingwrong by the care recipient, guilt about failing to meetthe challenges of caregiving and CGQ total score.

Similar associations to those found with the CGQwere found between the ZBI guilt factor and allthe other variables assessed: all associations weresignificant, except that with functional status.

Discussion

The main aim of this study was to examinethe psychometric properties of a new scalefor measuring guilt among dementia caregivers,namely, the Caregiver Guilt Questionnaire (CGQ).Through principal component analysis, 22 itemswere found to be represented by a five-factor

structure. The factors were labeled guilt about doingwrong by the care recipient, guilt about failing to meetthe challenges of caregiving, guilt about self-care, guiltabout neglecting other relatives and guilt about havingnegative feelings towards other people. A significantproportion of variance of the guilt construct isexplained by these factors, and reasonable internalconsistency indices were found for all the factorsand for the total questionnaire.

The convergent validity of the CGQ wassupported by the substantial correlations foundbetween the CGQ (total score and some of itsfactors) and the ZBI guilt factor (Ankri et al., 2005),especially with the first two factors, guilt about doingwrong by the care recipient and guilt about failing tomeet the challenges of caregiving. This result is notsurprising if we consider that the content of theitems included in these two factors is similar to thatof the items included in the ZBI guilt factor: guiltassociated with a general sense of not being a goodcaregiver. The low, though significant, correlationsbetween the ZBI guilt factor and the CGQ factorsguilt about self-care and guilt about having negativefeelings towards other people, on the one hand, andthe non-significant association with the factor guiltabout neglecting other relatives, on the other, suggestthat the CGQ measures dimensions of guilt whichare not tapped by the ZBI guilt factor. Furthermore,the low to moderate intercorrelations between theCGQ factors, together with the different patternsof associations found between the CGQ factorsand the other variables analyzed in the study,reveal that guilt may be better conceived as amultidimensional and domain-specific experience.However, this should not necessarily be takenas an argument against the usefulness of a totalCGQ score, which is supported by the interestingcorrelations found between this global score and thevariables included in the study.

In this regard, similar results to those found byAnkri et al. (2005) with the ZBI guilt factor wereobtained. Caregivers looking after their parents hadhigher guilt scores than spouses in CGQ total scoreand all its factors. This result could be explained bythe fact that adult children giving care, especiallydaughters, may have more sources of guilt, as theyare more likely to experience conflict between theirmultiple roles (caregiver, mother or father, worker,etc.) and to perceive interference of their caregivingrole with work and family (Yee and Schulz, 2000).In this regard, adult children caregivers have beenfound in some studies to admit potentially harmfulbehaviors (e.g. verbal aggression, ignoring) towardstheir care recipients to a greater extent than spousecaregivers (Sasaki et al., 2007). Such behaviors arelikely to generate guilt in caregivers.


In this study, guilt has also been found to besignificantly associated with behavioral problems inthe care recipient. As Pagel et al. (1985) argue,caregivers facing more behavioral problems maymake stronger attempts to control their relative’sbehaviors. Given the difficulty of controlling suchbehaviors, their attempts may be frustrating and/orunsuccessful, thus generating feelings of guilt.

In a similar way to what has been found for otherdistress-related variables in caregiving research (e.g.Pinquart and Sorensen, 2003), female caregiversscored significantly higher than male caregivers,specifically in the factors guilt about neglecting otherrelatives and guilt about having negative feelingstowards other people. Caring for others is a morecentral element in their identity than it is for men,and female caregivers are usually more involvedthan male caregivers in the roles of assisting andnurturing all family members, besides the carerecipient. Given that women usually perceive moreresponsibility for caring for all family members, theyare likely to be more vulnerable to guilt associatedwith the perception that they are neglecting otherrelatives due to caregiving. In fact, female caregivershave been found to report more role conflict, morecaregiving costs and more interference with familyand leisure time than do men (Ingersoll-Daytonand Raschick, 2004). The finding that femalecaregivers report more guilt than men about havingnegative feelings towards other people could beexplained by the fact that female caregivers reportgreater role strain than male caregivers (Ingersoll-Dayton et al., 1996) and receive less formal andinformal support (Yee and Schulz, 2000). Thesedifferences between male and female caregiversmay be related to different coping styles, butalso to different internalized norms about caringand nurturing responsibilities, with women moreliable to assume caregiving as a moral (filial)responsibility they have to bear, that is, as “thenormal way” they should behave according tosocial and family expectations (Ingersoll-Daytonand Raschick, 2004). As suggested by Gonyeaet al., (2008), role theory can provide interestinginsights for understanding the emergence of guiltfeelings among adult children caregivers, especiallywomen. These gender differences may explain whyfemale caregivers have been systematically foundin the caregiving literature to report more burden,anxiety, depression and health problems than malecaregivers (Pinquart and Sorensen, 2003; Yee andSchulz, 2000). These negative emotional outcomesmay facilitate the generation of negative thoughts orfeelings towards other people, such as irritability orresentment.

The results of this study are consistent withthose of previous studies suggesting the relevance

of guilt in caregiving outcomes. Strong associationswere found between guilt and important outcomevariables such as depression or anxiety, suggestinga significant relationship between guilt feelings andcaregiver distress. Similarly, Gallagher et al. (2008)found guilt, as measured with the ZBI guilt factor,to be associated with greater anxiety in caregiversof children with intellectual disabilities. It is ourhypothesis that caregiver guilt contributes to thegeneration and exacerbation of depression andburden. However, the cross-sectional nature of thisstudy precludes us from drawing any conclusionsabout causal links between guilt and other negativeemotional outcomes in caregiving.

Another relevant finding of this study thatendorses the need to take guilt feelings into accounton analyzing caregiver distress are the significantassociations between guilt and other variablespertaining to key dimensions in theoretical modelssuch as the stress and coping model (e.g. Knightet al., 2000). For example, caregivers who reporthaving guilt feelings about doing wrong by their carerecipient or failing to meet the challenges of care-giving report low scores on leisure (frequency and,to a greater extent, satisfaction with leisure) and so-cial support. Once again, the cross-sectional designof the study does not permit us to make causalinferences in this regard, but it would be reallyinteresting for future studies to analyze whetherguilt feelings have any influence on the reduction oftime devoted to leisure or of social support levels.

Some other limitations of this study shouldbe acknowledged. With regard to random error,although internal consistency of the scale (andits factors) has been found to be reasonable, notest-retest assessment of guilt has been made, sodata on the stability over time of CGQ scores arenot available. With regard to potential sources ofsystematic error, some of these can be addressed.For example, regarding content validity, otherdomains or experiences that may be sources of guiltfor caregivers (e.g. thoughts about placing the carerecipient in a nursing home) are not tapped by theCGQ, even though they may influence caregiverdistress. Furthermore, as several empirical studieshave shown in recent years, there are substantialdifferences in caregivers’ experience of distressacross cultures (e.g. Knight et al., 2000), and guiltis considered to be a culturally sensitive emotion,strongly influenced by cultural perspectives andsocial norms (Bierbrauer, 1992). In order to test thecross-cultural content validity of the CGQ, it seemsnecessary to apply this instrument in other culturesso as to confirm the findings of the present work.Finally, the convenience-based nature of the samplelimits its representativeness and, in consequence,the generalizability of the results.


As has been already noted, the construct ofguilt has not yet been adequately integrated intoempirically based models of personality and/oremotion (Kugler and Jones, 1992). Currently, solidtheoretical accounts of this emotional experienceare still lacking, especially in caregiving research,where this variable has started to be analyzed onlyvery recently. Hence, there is a clear need forempirical studies, such as that presented here, whichmay contribute to improve our understanding of thepsychological and behavioral implications of guilt,and the dynamic of its relationships with caregiversdistress and coping behaviors. These studies mayeventually help us to develop more comprehensivecaregiving models which include this variable andthus describe more accurately caregivers’ subjectiveexperiences. The main contributions of the presentstudy to the literature on guilt may be summarizedas follows: it provides evidence on (a) the multi-dimensional and domain-specific nature of guilt;(b) the association between guilt and caregivers’distress; and (c) potential behavioral implicationsof guilt (frequency of and satisfaction with leisure),which may mediate its relationship with distress.It is clear that there are many questions regardingguilt which deserve further attention, such as therelationship between guilt and caregivers’ values(e.g. familism), moral or religious beliefs, ordysfunctional thoughts about caregiving. Thesequestions could be explored in future studies whichwould provide important insights into the nature ofguilt and its role in the caregiving stress process.

The practical implications of our findings areevident: they indicate the importance of addressingguilt feelings in interventions and helping caregiversto acknowledge and manage them, as a way ofreducing their distress and of promoting moreadaptive ways of coping with caregiving. In thisregard, helping caregivers to review and adjusttheir expectations and standards in relation to theirobligations and capacities, to accept their limitationsand weaknesses, and to acknowledge their needs,might facilitate a reduction in their guilt feelingswhich, in turn, could lead to the development ofmore adaptive behaviours, such as increasing theirleisure activities or seeking social and instrumentalsupport, which constitute important targets incognitive-behavioral interventions (e.g. Coon et al.,2003). Even when this study has focused onthe measurement of guilt in dementia caregivers,this emotional phenomenon is considered to bea key experience in many different caregivingexperiences, and it has been found to be associatedwith psychosocial maladjustment in caregivers (e.g.Spillers et al., 2008). In this sense, it might beinteresting to explore the usefulness of the CGQ toassess guilt in caregivers of people suffering from

other chronic conditions (e.g. terminal illnessessuch as cancer).

In conclusion, the CGQ presents acceptablepsychometric properties, and may contribute tothe development of sensitive and conceptually validoutcome measures for psychosocial interventionresearch (Moniz-Cook et al., 2008). It is our hopethat future studies using this instrument in ourcontext and in other countries will permit furtheranalysis of its utility for getting closer to our sharedobjective of improving our ability to help caregivers.

Conflict of interest

None.

Description of authors’ roles

Andres Losada and Marıa Marquez-Gonzalez bothworked on the design of the study and its imple-mentation, the development of the CGQ items, thedata analysis and the writing of the paper. CeciliaPenacoba collaborated on the development of theCGQ items. Rosa Romero-Moreno worked on thestudy implementation and the writing of the paper.

Acknowledgments

The authors would like to thank Nancy Pachana,Ph.D., from the University of Queensland, for herhelp with the CGQ back-translation. This studywas partially funded by grants from the SpanishMinistry of Education (SEJ2006–02489/PSIC)and from the Spanish Ministry of Science andInnovation (PSI2009-08132/PSIC).

References

Ankri, J., Andrieu, S., Beaufils, B., Grand, A. andHenrad, J. C. (2005). Beyond the global score of the ZaritBurden Interview: useful dimensions for clinicians.International Journal of Geriatric Psychiatry, 20, 254–260.

Bierbrauer, G. (1992). Reactions to violation of normativestandards: a cross-cultural analysis of shame and guilt.International Journal of Psychology, 27, 181–193.

Boye, B., Bentsen, H. and Malt, U. F. (2002). Does guiltproneness predict acute and long-term distress in relativesof patients with schizophrenia? Acta PsychiatricaScandinavica, 106, 51–57.

Brodaty, H. (2007). Meaning and measurement or caregiveroutcomes. International Psychogeriatrics, 19, 363–381.

Coon, D. W., Thompson, L., Steffen, A., Sorocco, K. andGallagher-Thompson, D. (2003). Anger and depressionmanagement: psychoeducational skill training interventionsfor women caregivers of a relative with dementia. TheGerontologist, 43, 678–689.


Cooper, C., Katona, C., Orrell, M. and Livingston, G.(2008). Coping strategies, anxiety and depression incaregivers of people with Alzheimer’s disease. InternationalJournal of Geriatric Psychiatry, 23, 929–936.

Gallagher, S., Phillips, A. C., Oliver, C. and Carroll, D.(2008). Predictors of psychological morbidity in parents ofchildren with intellectual disabilities. Journal of PediatricPsychology, 33, 1129–1136;

Ghatavi, K., Nicolson, R., MacDonald, C., Osher, S. andLevitt, A. (2002). Defining guilt in depression: acomparison of subjects with major depression, chronicmedical illness and healthy controls. Journal of AffectiveDisorders, 68, 307–315.

Gonyea, J. G., Paris, R. and Zerden, L. (2008). Adultdaughters and aging mothers: the role of guilt in theexperience of caregiver burden. Aging and Mental Health,12, 559–567.

Hamel, G. (1995). Guilt: a hidden problem in caregiving.Provider, 21, 59–60.

Ingersoll-Dayton, B. and Raschick, M. (2004). Therelationship between care-recipient behaviors and spousalcaregiving stress. The Gerontologist, 44, 318–327.

Ingersoll-Dayton, B., Starrels, M. E. and Dowler, D.(1996). Caregiving for parents and parents-in-law: isgender important?. The Gerontologist, 36, 483–491.

Knight, B. G., Silverstein, M., McCallum, T. J. and Fox,L. S. (2000). A sociocultural stress and coping model formental health outcomes among African Americancaregivers in southern California. Journal of Gerontology:Psychological Sciences, 55, 142–150.

Kugler, K. and Jones, W. H. (1992). On conceptualizing andassessing guilt. Journal of Personality and Social Psychology,62, 318–327.

Mahoney, F. L. and Barthel, D. W. (1965). Functionalevaluation: the Barthel Index. State Journal, 4, 61–65.

Martin, Y., Gilbert, P., McEwan, K. and Irons, C. (2006).The relation of entrapment, shame and guilt to depression,in carers of people with dementia. Aging and Mental Health,10, 101–106.

McNair, D., Lorr, M. and Droppleman, L. (1971). Profileof Mood States. Manual. San Diego, CA: Educational andIndustrial Testing Service.

Moniz-Cook, E. et al. and the INTERDEM Group(2008). A European consensus on outcome measures forpsychosocial intervention research in dementia care. Agingand Mental Health, 12, 14–29.

Pagel, M. D., Becker, J. and Coppel, D. B. (1985). Loss ofcontrol, self-blame, and depression: an investigation ofspouse caregivers of Alzheimer’s disease patients. Journal ofAbnormal Psychology, 94, 169–182.

Pinquart, M. and Sorensen, S. (2003). Differences betweencaregivers and noncarevigers in psychological health andphysical health: a meta-analysis. Psychology and Aging, 18,250–267.

Radloff, L. S. (1977). The CES-D scale: a self-reportdepression scale for research in the general population.Applied Psychological Measurement, 1, 385–401.

Radloff, L. S. and Teri, L. (1986). Use of the Center forEpidemiological Studies-Depression Scale with olderadults. In T. L. Brink (ed.), Clinical Gerontology: A Guide toAssessment and Intervention (pp. 119–136). New York:Haworth Press.

Reig, A., Ribera, D. and Miguel, J. (1991). Psychologicalsupport and daily stress in non-institutionalized elderly.European Journal of Psychological Assessment, 7, 191–200.

Sasaki, M., Arai, Y., Kumamoto, K., Abe, K., Arai, A.and Mizuno, Y. (2007). Factors related to potentiallyharmful behaviors towards disabled older people by familycaregivers in Japan. International Journal of GeriatricPsychiatry, 22, 250–257.

Spillers, R. L., Wellisch, D. K., Kim, Y., Matthews, B. A.and Baker, F. (2008). Family caregivers and guilt in thecontext of cancer care. Psychosomatics, 49, 511–519.

Stevens, A. B. et al. (2004). Measurement of leisure timesatisfaction in family caregivers. Aging and Mental Health,8, 450–459.

Teri, L., Truax, P., Logsdon, R., Uomoto, J., Zarit, S.and Vitaliano, P. P. (1992). Assessment of behavioralproblems in dementia: the revised memory and behaviorproblems checklist. Psychology and Aging, 7, 622–631.

Yaffe, M. J. (1988). Implications of caring for an aging parent.Canadian Medical Association Journal, 138, 231–235.

Yee, J. L. and Schulz, R. (2000). Gender differences inpsychiatric morbidity among family caregivers: a reviewand analysis. The Gerontologist, 40, 147–164.

Zarit, S. H., Reever, K. E. and Bach-Peterson, J. (1980).Relatives of the impaired elderly: correlates of feelings ofburden. The Gerontologist, 20, 649–655.

Appendix: Caregiver Guilt Questionnaire (CGQ)

Below is a list of some feelings and thoughts.Please tell us how often you have had these feelingsand thoughts during the past few weeks using thefollowing options:

0 = never; 1 = rarely; 2 = sometimes; 3 = severaltimes; 4 = always or almost always

1. I have felt bad about having made some plan ordone some activity without taking my relative intoaccount [Me he sentido mal por haber hecho algunplan o actividad sin contar con mi familiar].

2. I have felt guilty about the way I’ve sometimesbehaved with my relative [Me he sentido culpablepor la forma en que me he comportado en ocasionescon mi familiar].

3. I have felt bad for not looking after my other relatives(husband, wife, children . . .) as I should, due to mycaregiving [Me he sentido mal por no atender amis otros familiares (marido, mujer, hijos. . .) comodeberıa, debido al cuidado].

4. I have felt bad about not being able to devote moretime to my family (husband, wife, children. . .), dueto my caregiving [Me he sentido mal por no poderdedicar mas tiempo a mi familia (marido, mujer,hijos . . .) como deberıa, debido al cuidado].

5. I have thought that I’m not doing things right withthe person I’m caring for [He pensado que no estoyhaciendo las cosas bien con la persona a la que estoycuidando].

6. I have thought that, given the circumstances, I’mdoing a good job as a caregiver [He pensado


que, teniendo en cuenta las circunstancias, estoyhaciendo bien mi tarea como cuidador/a].

7. When I’ve gone out to do some pleasant activity (e.g.eating out in a restaurant), I’ve felt guilty and unableto stop thinking that I should be caring for myrelative [Cuando he salido a hacer alguna actividadagradable (p.ej., ir a cenar a un restaurante), mehe sentido culpable y no he dejado de pensarque deberıa estar cuidando o atendiendo a mifamiliar].

8. I have felt bad about things I may have done wrongwith the person I’m caring for [Me he sentido malpor cosas que quiza habıa hecho mal con la personaa la que estoy cuidando].

9. I have thought that perhaps I’m not caring wellfor my relative [He pensado que quiza no estoycuidando bien].

10. I have felt bad about getting angry with the personI’m caring for [Me he sentido mal por habermeenfadado con la persona a la que estoy cuidando].

11. I have felt bad about telling off the person I’m caringfor, for some reason [Me he sentido mal por haberreganado por alguna razon a la persona a la queestoy cuidando].

12. I’ve got angry with myself for having negativefeelings toward the person I’m caring for [Me heenfadado conmigo mismo/a por tener sentimientosnegativos hacia la persona a la que cuido].

13. I’ve found myself thinking that I’m not up to thejob [Me he encontrado pensando que no estoy a laaltura de las circunstancias].

14. I have felt bad about not having more patience withthe person I’m caring for [Me he sentido mal porno tener mas paciencia con la persona a la que estoycuidando].

15. I have felt bad about leaving my relative in the careof someone else while I do my own things (e.g. work,shopping, going to the doctor) [Me he sentido malpor dejar a mi familiar al cuidado de otra persona

mientras yo hacıa mis tareas (p.ej., trabajo, compra,ir al medico, etc.)].

16. I have felt bad for leaving my relative in the care ofsomeone else while I had fun [Me he sentido malpor dejar a mi familiar al cuidado de otra personamientras yo me divertıa].

17. I have felt guilty about having wished that others“could have this burden” or suffer as I do [Me hesentido culpable por desear que a otros “les toqueesta cruz” o sufran lo mismo que yo].

18. I have felt like a bad person for hating and/orenvying other relatives who could have takenresponsibility for some caring and do not do so [Mehe sentido mala persona por odiar y/o envidiar aotros familiares que podrıan hacerse responsablesparcialmente del cuidado y no lo hacen].

19. I have felt bad for having negative feelings (e.g., hate,anger or resentment) toward some relatives [Me hesentido mal por tener sentimientos negativos (p.ej.,odio, enfado o rencor) hacia algunos familiares].

20. I have felt guilty about having so many negativeemotions in relation to caring [Me he sentidoculpable por el hecho de tener tantas emocionesnegativas en relacion al cuidado].

21. I have thought that the way I care for my relative maynot be appropriate and may make his/her problemget worse [He pensado que la forma en la que cuidode mi familiar podrıa no ser adecuada y contribuira que su problema vaya a peor].

22. I have felt guilty thinking that my lack of informationand preparedness might mean that I’m not handlingthe care of my relative in the best way possible[Me he sentido culpable al pensar que mi faltade informacion y preparacion podrıa hacer que noestuviera manejando el cuidado de mi familiar de lamejor forma possible].

Note: The Spanish items are shown in squarebrackets.

development and validation of the caregiver guilt questionnaire

Documents