The Health and Well-Being of Caregivers of Children With Cerebral Palsy

Parminder Raina, PhD*; Maureen ODonnell, MD; Peter Rosenbaum, MD#; Jamie Brehaut, PhD**;Stephen D. Walter, PhD*; Dianne Russell, MSc*#; Marilyn Swinton, BSc; Bin Zhu, MSc*; and

Ellen Wood, MD

ABSTRACT. Objective. Most children enjoy healthychildhoods with little need for specialized health careservices. However, some children experience difficultiesin early childhood and require access to and utilizationof considerable health care resources over time. Althoughimpaired motor function is the hallmark of the cerebralpalsy (CP) syndromes, many children with this develop-ment disorder also experience sensory, communicative,and intellectual impairments and may have complex lim-itations in self-care functions. Although caregiving is anormal part of being the parent of a young child, this roletakes on an entirely different significance when a childexperiences functional limitations and possible long-term dependence. One of the main challenges for parentsis to manage their childs chronic health problems effec-tively and juggle this role with the requirements of ev-eryday living. Consequently, the task of caring for a childwith complex disabilities at home might be somewhatdaunting for caregivers. The provision of such care mayprove detrimental to both the physical health and thepsychological well-being of parents of children withchronic disabilities. It is not fully understood why somecaregivers cope well and others do not. The approach ofestimating the independent or direct effects of thecare recipients disability on the caregivers health is oflimited value because (1) single-factor changes are rareoutside the context of constrained experimental situa-tions; (2) assumptions of additive relationships and per-fect measurements rarely hold; and (3) such approachesdo not provide a complete perspective, because they failto examine indirect pathways that occur between predic-tor variables and health outcomes. A more detailed ana-lytical approach is needed to understand both direct andindirect effects simultaneously. The primary objective ofthe current study was to examine, within a single theory-based multidimensional model, the determinants of

physical and psychological health of adult caregivers ofchildren with CP.Methods. We developed a stress process model and

applied structural equation modeling with data from alarge cohort of caregivers of children with CP. This de-sign allowed the examination of the direct and indirectrelationships between a childs health, behavior andfunctional status, caregiver characteristics, social sup-ports, and family functioning and the outcomes of care-givers physical and psychological health. Families (n 468) of children with CP were recruited from 19 regionalchildrens rehabilitation centers that provide outpatientdisability management and supports in Ontario, Canada.The current study drew on a population available to theinvestigators from a previous study, the Ontario MotorGrowth study, which explored patterns of gross motordevelopment in children with CP. Data on demographicvariables and caregivers physical and psychologicalhealth were assessed using standardized, self-completedparent questionnaires as well as a face-to-face home in-terview. Structural equation modeling was used to testspecific hypotheses outlined in our conceptual model.This analytic approach involved a 2-step process. In thefirst step, observed variables that were hypothesized tomeasure the underlying constructs were tested using con-firmatory factor analysis; this step led to the so-calledmeasurement model. The second step tested hypothesesabout relationships among the variables in the structuralmodel. All of the hypothesized paths in the conceptualmodel were tested and included in the structural model.However, only paths that were significant were shown inthe final results. The direct, indirect, and total effects oftheoretical constructs on physical and psychologicalhealth were calculated using the structural model.Results. The most important predictors of caregivers

well-being were child behavior, caregiving demands,and family function. A higher level of behavior problemswas associated with lower levels of both psychological ( .22) and physical health ( .18) of the caregivers,whereas fewer child behavior problems were associatedwith higher self-perception ( .37) and a greaterability to manage stress ( .18). Less caregiving de-mands were associated with better physical ( .23) andpsychological ( .12) well-being of caregivers, respec-tively. Similarly, higher reported family functioning wasassociated with better psychological health ( .33) andphysical health ( .33). Self-perception and stress man-agement were significant direct predictors of caregiverspsychological health but did not directly influence theirphysical well-being. Caregivers higher self-esteem andsense of mastery over the caregiving situation predictedbetter psychological health ( .23). The use of morestress management strategies was also associated withbetter psychological health of caregivers ( .11). Grossincome ( .08) and social support ( .06) had indirectoverall effects only on psychological health outcome,

From the *Department of Clinical Epidemiology and Biostatistics, McMas-ter University Evidence-Based Practice Centre, and Department of Paedi-atrics, Faculty of Health Sciences, and #School of Rehabilitation Science,McMaster University, Hamilton, Ontario, Canada; Centre for CommunityChild Health Research, BC Research Institute for Childrens and WomensHealth, University of British Columbia, Vancouver, British Columbia, Can-ada; CanChild Centre for Childhood Disability Research, McMaster Uni-versity, Hamilton, Ontario, Canada; **Clinical Epidemiology Unit, OttawaHealth Research Institute, Ottawa, Ontario, Canada; and Department ofPediatrics, Dalhousie University, Halifax, Nova Scotia, CanadaAccepted for publication Dec 16, 2004.doi:10.1542/peds.2004-1689No conflict of interest declared.Reprint requests to (P.R.) McMaster Evidence-Based Practice Centre, De-partment of Clinical Epidemiology and Biostatistics, Faculty of HealthSciences, McMaster University, DTC, Room 306, 1280 Main St W, Hamilton,ON, Canada L8S 4L8. E-mail: praina@mcmaster.caPEDIATRICS (ISSN 0031 4005). Copyright 2005 by the American Acad-emy of Pediatrics.

e626 PEDIATRICS Vol. 115 No. 6 June 2005 www.pediatrics.org/cgi/doi/10.1542/peds.2004-1689 by guest on May 17, 2015pediatrics.aappublications.orgDownloaded from

whereas self-perception ( .22), stress management ( .09), gross income ( .07), and social support ( .06)had indirect total effects only on physical health out-comes.Conclusions. The psychological and physical health

of caregivers, who in this study were primarily mothers,was strongly influenced by child behavior and caregiv-ing demands. Child behavior problems were an impor-tant predictor of caregiver psychological well-being, bothdirectly and indirectly, through their effect on self-per-ception and family function. Caregiving demands con-tributed directly to both the psychological and the phys-ical health of the caregivers. The practical day-to-dayneeds of the child created challenges for parents. Theinfluence of social support provided by extended family,friends, and neighbors on health outcomes was second-ary to that of the immediate family working closely to-gether. Family function affected health directly and alsomediated the effects of self-perception, social support,and stress management. In families of children with CP,strategies for optimizing caregiver physical and psycho-logical health include supports for behavioral manage-ment and daily functional activities as well as stressmanagement and self-efficacy techniques. These datasupport clinical pathways that require biopsychosocialframeworks that are family centered, not simply techni-cal and short-term rehabilitation interventions that arefocused primarily on the child. In terms of prevention,providing parents with cognitive and behavioral strate-gies to manage their childs behaviors may have thepotential to change caregiver health outcomes. Thismodel also needs to be examined with caregivers of chil-dren with other disabilities. Pediatrics 2005;115:e626e636.URL: www.pediatrics.org/cgi/doi/10.1542/peds.2004-1689;caregiver, well-being, disability, cerebral palsy, informalsupport, stress process model, structural equation model-ing.

ABBREVIATIONS. CP, cerebral palsy; GMFCS, Gross MotorFunction Classification System; SES, socioeconomic status; OMG,Ontario Motor Growth; SEM, structural equation modeling; LLM,log linear modeling; RMSEA, root-mean-squared error of approx-imation; NNFI, Nonnormed Fit Index; CFI, Comparative Fit Index.

Most children enjoy healthy childhoods withlittle need for specialized services in thehealth care system. However 7.7% of chil-dren experience difficulties during their developingyears and require access to and utilization of exten-sive health care resources over time.1 Cerebral palsy(CP) is one such developmental disorder that beginsin early childhood as a set of functional limitationsthat stem from disorders of the developing centralnervous system.2 The current estimated incidence ofCP is 2.0 to 2.5 per 1000 live births in developedcountries.3 Although impaired motor function is thehallmark of the CP syndromes, many children alsoexperience sensory, communicative, and intellectualimpairments and may have complex limitations inself-care functions such as feeding, dressing, bathing,and mobility. These limitations can result in require-ments for long-term care that far exceed the usualneeds of children as they develop.4,5 We chose CP asa prototype condition to study the issues that parentswho care for a child with a disability face.Family caregivers often shoulder the principal,

multifaceted responsibilities of long-term disability

management.6 Although caregiving is a normal partof being the parent of a young child, this role takeson an entirely different significance when a childexperiences functional limitations and possible long-term dependence. One of the main challenges forparents is to manage their childs chronic healthproblems effectively while maintaining the require-ments of everyday living. In some cases, the provi-sion of such care can prove detrimental to both thephysical health and the psychological well-being ofparents of children with chronic disabilities and havean impact on family income, family functioning, andsibling adjustment.7In the past 2 decades, tremendous changes in

health care systems have exerted a shift toward out-patient community and home-based settings, whichin turn have increased the responsibilities of infor-mal caregivers.8 In addition, several factors that maycontribute to the perceived burden and stress expe-rienced by parents of children with disabilities exist.These factors include smaller family units, increasedrate of marital breakdown,8 technologic innovations,and pharmacologic advancements in medicine.9Consequently, the task of caring for a child withcomplex disabilities at home might be somewhatdaunting for caregivers.The notion of caregiving as a career connotes a

dynamic process, whereby an individual movesthrough a series of stages that require adaptation andrestructuring of responsibilities over time.4,10,11These stages might include (1) anticipation for andacquisition of the caregiver role, (2) performance oftasks and responsibilities, and (3) eventual exit fromthe role.4,10,11 Unlike a conventional career, however,the caregiver role is usually not planned or chosenand is generally not seen as an appealing pursuit forthe future.It is not fully understood why some caregivers

cope well and others do not. Stress has been con-ceived as the balance between external environmen-tal demands and the perceived internal ability torespond and may occur when the demands preventthe pursuit of other life objectives.4,8,9 Modifyingfactors of caregiver stress include (1) the characteris-tics of the caregiver (eg, age, marital status, copingability),12,13 (2) characteristics of the recipient (eg, thedegree of disability),12,14 (3) the shared history be-tween the caregiver and the person being cared for,4(4) social factors (eg, access to social networks andsocial support),8,12 (5) economic factors (eg socioeco-nomic status [SES], ability to access formal care, em-ployment),4,12 and (6) cultural context.4 Each of thesefactors may influence the outcome of the caregivingsituation; together they suggest that stress occurs in abroader context than simply the provision of care fora child with a physical disability.Several conceptual models describe the impact of

stress on caregivers.9,15,16 These models have typi-cally used traditional analytic approaches (eg, linearregression) to examine the relationship between afactor and the outcome after adjusting for other vari-ables. The approach of estimating the independentor direct effects of the care recipients disability onthe caregivers health is of limited value because (1)

www.pediatrics.org/cgi/doi/10.1542/peds.2004-1689 e627 by guest on May 17, 2015pediatrics.aappublications.orgDownloaded from

single-factor changes are rare outside the context ofconstrained experimental situations; (2) assumptionsof additive relationships and perfect measurementsrarely hold; and (3) such approaches do not providea complete perspective, because they fail to examinedirect and indirect pathways that occur between pre-dictor variables and health outcomes. A more de-tailed analytical approach is needed to understandboth direct and indirect effects simultaneouslywithin a theory-based multidimensional model.Our primary objective was to examine, within a

single multidimensional model, a comprehensive setof factors that are relevant to the caregiving situation.The conceptual model (Fig 1) that guided this re-search is described in detail by Raina et al.17 The 5constructs in the proposed model include (1) back-ground and context, (2) child characteristics, (3) care-giver strain, (4) intrapsychic factors, and (5) coping/supportive factors. This research examines the directand indirect associations between caregiver charac-teristics, sources of caregiver stressors, family func-tioning, and informal social support on the well-being of the caregivers of children with CP.Specifically, we hypothesized that an increase in achilds disability as measured by the Gross MotorFunction Classification System (GMFCS)18; the Pedi-

atric Evaluation of Disability Inventory, Part 119; andthe Health Utilities Index, selected questions,20 andbehavioral problems would be associated directlywith poorer physical and psychological well-being ofprimary caregivers.7 However, we also hypothesizedthat the direct relationship between a childs disabil-ity or his or her behavioral problems and parentalwell-being would be mediated by intrapsychic andcoping/supportive factors as described in the pro-posed conceptual model (Fig 1).

METHODS

SettingCanChild studies are made possible through a partnership

between the provincial government-funded CanChild Centre forChildhood Disability Research at McMaster University and the 19publicly funded regional ambulatory childrens rehabilitation cen-ters in Ontario, Canada. These regional centers deliver a range ofdevelopmental therapies and services (predominantly physical,occupational, speech-language, and recreational therapies) pro-vided by developmental professionals who are trained and expe-rienced in both the assessment and the management of childhooddisability. Each center serves the majority of eligible children intheir area. The current study drew on a population that wasavailable to the investigators from a previous study, the OntarioMotor Growth (OMG) study, which explored patterns of grossmotor development in children with CP and where more details ofthe sampling process are described.21

Fig 1. Conceptual model of the care-giving process among caregivers of apediatric population.

e628 HEALTH OF CAREGIVERS OF CHILDREN WITH CEREBRAL PALSY by guest on May 17, 2015pediatrics.aappublications.orgDownloaded from

SampleCaregivers were recruited for the present study from a sam-

pling frame, originally created in early 1996, of families who hadparticipated in the OMG study. For the OMG study, children hadbeen randomly selected in a stratified sampling procedure on thebasis of age and level of motor function using the GMFCS18 from18 of 19 regional centers and 1 hospital-based therapy program ina community without a regional center. A total of 657 children andtheir families participated in the OMG study, 632 of whom werestill involved at the end of OMG project; these 632 families wereinvited to participate in the current caregiver study.

Selection CriteriaOne primary caregiver per household was selected for this

study. The primary caregiver was defined as the person who ismost responsible for the day-to-day decision making and care ofthe child; the family determined who was best considered theprimary caregiver. Caregivers who were asked to participate inthis study had to meet the following criteria: (1) have a child whohad participated in the OMG study, (2) identify themselves as aprimary caregiver whose child lived with them, (3) give writtenconsent to participate, and (4) reside in Ontario. Initial recruitmentinvolved mailing families a package that contained a letter, aconsent form, and a brochure describing the study, along with alottery ticket as an incentive to participate. Telephone follow-upwith families was done by a person who was trained to answerany questions and obtain verbal consent.

Data CollectionThe data were collected in 2 steps. First, for minimizing the

time burden on the caregivers, a package that included an intro-ductory letter, a consent form, and a questionnaire was mailed tothem for completion before a face-to-face interview. The self-report questionnaire collected demographic information about thecaregiver, the child, and the family; the childs ability to performactivities of daily living; the childs day-to-day health; the childsbehavior; caregiver stress management strategies; caregiver per-ceptions of formal care for the child within the last 12 months; andthe caregivers perception of his or her own general health andwell-being (Table 1 summarizes the measures used).

The second step consisted of a home-based interview with theprimary caregiver of the child with CP. Interviewers who werehired and trained specifically for this study conducted the inter-views. The structured, face-to-face interview collected informationabout caregiving assistance provided to the child, the caregiversperceptions of his or her own physical and mental health andemotional well-being, mastery and self-esteem, informal socialsupport, family functioning, job-caregiving conflict issues, andgross income status (Table 2 briefly describes the measures used).The questionnaires were pretested with nonstudy families of chil-

dren with disabilities; the average length of each interview was1hour.

Statistical AnalysisDescriptive statistics were calculated for all variables in the

data set. structural equation modeling (SEM) was used to testspecific hypotheses outlined in our conceptual model. This ana-lytic approach involves a 2-step process. In the first step, observedvariables are hypothesized to measure the underlying constructsand tested using confirmatory factor analysis; this step leads to theso-called measurement model. The second step focuses on testinghypotheses about relationships among the variables in the struc-tural model. Several model diagnostic approaches were used toassess the integrity of each phase of the SEM development and theappropriateness of variables included in the model.7,2224 We usedthe PROC CALIS procedure in SAS version 8.2 (SAS Institute,Cary, NC), using covariance matrices and the maximum-likeli-hood estimation method to assess model fit. All of the hypothe-sized paths in the conceptual model (Fig 1) were tested andincluded in the structural model. However, only paths that weresignificant (P .05) are shown in the final results.

The direct, indirect, and total effects of factors on health werecalculated using the structural model. For example, the impact ofstress management on psychological health involves 1 direct path(14) and 1 indirect path (13 16). The total effect (T) wasestimated by summating the direct effect and the indirect effects[14 (13 16) T].

The process of developing and testing a structural equationmodel is theory and data driven. Nonconvergence is not uncom-mon in the process of parameter estimation in such models. There-fore, a log linear modeling (LLM) analysis was conducted as anadjunct to the SEM analysis to assess the robustness of SEManalysis. LLM requires discrete measurements. Instead of latentcontinuous variables, categorized data were used in LLM analysisto specify the log linear model. The robustness of the SEM modelwas comparable to the LLM. However, SEM analyses identifiedmore relationships than the LLM (which are not reported in thisarticle).

RESULTS

Description of the SampleOf the initial sampling frame that contained 632

families, 42 caregivers were lost to follow-up. A totalof 590 (93%) families were contacted, and 570 (90%)were eligible; 503 (88%) of the 570 consented. A finalsample of 468 primary caregivers (82% of the 570eligible families) provided data (Table 3). The resultsshow that the age, gender, and marital status of our

TABLE 1. Child and Caregiver Questionnaire Measures

Variable Measure Description of Measure

ChildMotor severity GMFCS18 An ordinal descriptive scale of the gross motor

function abilities of children with CPActivities of daily living Pediatric Evaluation of Disability Inventory:

Part 119Child is rated as capable/unable on 73 self-careand 54 mobility items

Cognitive function Health Utilities Index: selected questions20 Items about childs ability to learn, remember,think, and solve problems

Child behavior Survey Diagnostic Instrument48 (SDI) SDI is a 24-item subset of the Child BehaviorChecklist with 3 scales: conduct disorder,hyperactivity, and emotional disorder

CaregiverCaregiver health status Medical Outcomes Study: Short Form 36 Health

Survey49 (SF36)The SF36 is a generic measure of health conceptsrelated to functional status and well-being

Perception of formal care Measures of Processes of Care50 Caregivers perceptions of the extent to whichspecific behaviors of health professionals occur

Stress management Coping Health Inventory for Parents11 Caregivers appraisal of their coping responses tomanagement of family life when he or she hasa child who is seriously/chronically ill

SES of caregivers National Longitudinal Study of Children andYouth20

Items related to education level (EDUC-Q2,EDUC-Q3), occupational status (LFS-Q1, LFS-Q2), and income (INCOMQ3B)

www.pediatrics.org/cgi/doi/10.1542/peds.2004-1689 e629 by guest on May 17, 2015pediatrics.aappublications.orgDownloaded from

caregiver sample was well matched to the corre-sponding OMG study. In addition, we compared ourstudy sample with a national population sample interms of age; gender distribution, with both samplesbeing primarily female; and martial status. Our care-giver sample differed from the national populationin the distribution of the educational categories. Bothsamples showed similar proportions at the lower andhigher ends of the educational continuum (Fig 2).25The mean age of the children was 10.6 (SD: 2.69)

years, 56% of whom were boys. Half of the childrenwith CP (49.8%) were first-born. The mean age ofcaregivers was 40.3 (SD: 6.72) years, and 94.4% werefemale, 89.7% of which were birth mothers. Lan-guage preference of caregivers was English (98.5%).The GMFCS levels indicated that of the children, 28%walked without restrictions (level I), 11% walkedwithout assistive devices (level II), 19% walked withassistive mobility devices (level III), 21% had nonam-bulatory self-mobility with limitations (level IV), and20% were severely limited even with the use of as-sistive technology (level V). The range, mean, and SDof the observed variables for caregivers are shown inTable 4. The correlations between latent constructsare shown in Table 5.

SEM

Measurement ModelInitially, the hypothesized model was created with

the predicted paths among the latent structural vari-ables predicted from the literature (see Fig 1). Theconfirmatory factor analysis was used to test thereliability of the hypothesized measurement model.Testing the fit of the measurement model led to

additional refinements to the conceptual model.First, we dropped 2 variables (chronicity of dis-tress and reported health transitions from SF-36)from the psychological health construct because ofsmall factor loadings (0.4) from the exploratoryfactor analyses with the psychological health factor.Second, the perception of formal care (Measures ofProcesses of Care) construct was dropped because ofmissing and nonapplicable responses to the ques-tionnaire. Finally, the disability construct wasdropped because of its strong correlation with thecaregiving demands construct. The caregiving de-mands construct was collapsed into a single ob-served variable because the measured variablescaregiver assistance-self care and caregiver assis-tance mobility were highly correlated. Two vari-ables were relocated in the model: vitality fromSF-36 was relocated to the psychological health con-struct from physical health construct, and socialprovision scale was relocated to the family functionconstruct instead of the social support construct.After respecifying and reestimating the model,

income was found to have more effect than edu-cation, and subsequently the latter variable wasdropped from the SES construct. We also modifiedthe model by dropping the social functioning vari-able from the social support construct to improve thegoodness of fit. The final measurement model of 23observed variables indicated an acceptable fit.Several statistical tests and goodness-of-fit indices

are used to assess the adequacy of model fit. In thisstudy, the root-mean-squared error of approxima-tion (RMSEA), Bentler and Bonetts Nonnormed FitIndex (NNFI), and Bentlers Comparative Fit Index

TABLE 2. Caregiver Interview Measures

Variable Measure Description of Measure

Distress National Population Health Survey52(NPHS)

Subset of items from the Composite InternationalDiagnostic Interview (CIDI) related to distress(MH Q1A to Q1F) and chronicity of distress (MHQ1GQ1L)

Depression NPHS52 Subset of items from the CIDI related to majordepressive episodes (MHLTH Q2 to Q28)

Mastery NPHS53 Scale that measures caregiver self-control and self-concept (MAST-Q1)

Self-esteem NPHS52 Six-item scale that measures caregiver self-esteem(ESTEEM-Q1)

Chronic health conditions National Longitudinal Study ofChildren and Youth51 (NLSCY)

Chronic health conditions that last 6 mo or moreand are diagnosed by a health care professional(CHRON-Q1)

Health status McMaster Health Utility Index (HUI)in NPHS52

Items that measure health status, health-relatedquality of life, and producing utility scores(HSTATQ1 to HSTQ30)

Caregiving assistance Pediatric Evaluation of DisabilityInventory (PEDI): Parts II and III19

Measures amount of caregiver assistance providedto a child during basic functional activities ofdaily living

Job-caregiving conflict Pearlins Scale54 Five-item scale related to job-caregiving conflictsInformal social support Social Network and Frequency of

Contact Index in NPHS52Items summarize possible people in the caregiverssocial network and the average number ofcaregiver contacts in the past 12 mo with family,friends, and neighbors (SUP-Q7A to SUP-Q7H)

Social Provision Scale (SPS) inNLSCY51

Short version of the Social Provision Scale55 thatmeasures perceived social support from familyand friends

Family functioning Family Assessment Device (FAD) inNLSCY51

A global assessment of family functioning and thequality of relationships between parents orpartners (FNC-Q1A to Q1M)

e630 HEALTH OF CAREGIVERS OF CHILDREN WITH CEREBRAL PALSY by guest on May 17, 2015pediatrics.aappublications.orgDownloaded from

(CFI) were used to assess model fit. Typically, anRMSEA value of0.05 and an NNFI value and a CFIvalue close to 1 are indicative of good fit. However,it is not necessary for a model to display all of thesecharacteristics to be considered acceptable.26 In thisstudy, the RMSEA value was 0.06, the NNFI valuewas 0.90, and the CFI value was 0.92. All factorloadings were substantial in magnitude, and signif-icantly different from 0, indicating that the latentconstructs were adequately operationalized by theobserved variables. The final standardized loadingsof observed variables on the latent constructs areshown in Table 6.

The Structural ModelA path diagram of the structural model of factors

that influence the health of caregivers is shown in Fig3. The latent constructs are denoted by ellipses. Twoconstructs are each represented by a single observedvariable and are denoted by rectangles. Only pathsthat were significant (P .05) are shown in thediagram.The RMSEAwas 0.06, NNFI was 0.90, and CFI was

0.91, which indicated an acceptable fit. The 2 for the

structural model was 583.3 with degrees of freedom215 (P .0001). The standardized regression coeffi-cients and the corresponding R2 statistics for each ofthe significant hypothesized paths (P .05) areshown in Fig 3. Five constructs influenced caregiv-ers psychological health, whereas 3 constructs influ-enced their physical health. Thus, for example, psy-chological health has an R2 of 0.50, indicating thatchild behavior, self-perception, family function, care-giving demands, and stress management togetheraccounted for 50% of the variation in psychologicalhealth.Our conceptual model permits the estimation of

both direct and indirect effects of constructs on care-giver health. With respect to specific hypothesizedrelationships in this analysis, the most importantpredictors of caregivers well-being were child be-havior, caregiving demands, and family function. Anincrease in reported child behavior problems wasassociated with decrease in both psychological (3 .22) and physical (4 .18) health of the caregiv-ers, whereas fewer child behavior problems wereassociated higher self-perception (2 .37) and agreater ability to manage stress (5 .18; Fig 3).Decreased caregiving demands were associated withan increase in physical (7 .23) and psychological(6 .12) well-being of caregivers. Similarly, higherreported family functioning was associated with im-provements in both psychological health (16 .33)and physical health (15 .33).

Fig 2. Caregiver sample selection and recruitment.

TABLE 3. Demography of Caregivers

Variable Caregiver(n 468)

Gender, n (%)Female 442 (94.4)Male 26 (5.6)

Age, meanFemale 40.06 6.54Male 44.42 8.32

Marital status, n (%)Never married 25 (5.3)Married or living with a partner 379 (81.0)Separated 34 (7.3)Divorced 19 (4.1)Widowed 11 (2.4)

Caregiver relation with child with CP, n (%)Mother 420 (89.7)Father 25 (5.3)Foster 15 (3.2)Other as caregiver 8 (1.6)

Education level, n (%)Elementary school 10 (2.1)Secondary school 145 (31.0)Trade, technical, or vocational school 79 (16.9)Community college 125 (26.7)University 28 (6.0)Undergraduate degree 63 (13.5)Postgraduate degree 18 (15.6)

Caregivers main activity, n (%)Caring for family and working for pay/profit 272 (58.1)Caring for family 173 (37.0)Working for pay/profit 5 (1.1)Going to school 6 (1.3)Recovering from illness/on disability 3 (0.6)Other 6 (1.3)

Hours usually worked, n (%)Full-time (30 h/wk) 208 (44.4)Part-time (30 h/wk) 100 (21.4)Did not work for pay 160 (34.1)

Gross household income, n (%)Less than $29 999 114 (24.3)$30 000$59 999 156 (33.3)$60 000 or more 187 (40.0)

Variables with missing data: caregivers main activity 3 (0.6%);gross household income 11 (2.4%).

www.pediatrics.org/cgi/doi/10.1542/peds.2004-1689 e631 by guest on May 17, 2015pediatrics.aappublications.orgDownloaded from

Self-perception and stress management were sig-nificant direct predictors of caregivers psychologicalhealth but did not directly influence their physicalwell-being. Caregivers higher self-esteem and senseof mastery over the caregiving situation predictedbetter psychological health (9 .23). The use ofmore stress management strategies was also associ-ated with better psychological health of caregivers(14 .11). In addition to direct relationships, weobserved the important mediating effect of familyfunctioning on several latent constructs. For exam-ple, higher levels of self-perception (10 .56), socialsupport (12 .18), and stress management (13 .27) all were associated with better family function-ing.The total effect of hypothesized relationships be-

tween the latent constructs on caregiver health out-come can also be calculated using the direct andindirect pathways in the conceptual model (Fig 3).Thus, the total effect of child behavior (T.38) onpsychological health outcome was the sum of onedirect pathway (3 .22) and 7 indirect pathways( .16), whereas the total effect of child behavioron physical health was T.28. Similarly, the totaleffect of stress management on psychological health

was T .20. Gross income ( .08) and socialsupport ( .06) had indirect overall effects on onlypsychological health outcome, whereas self-percep-tion ( .22), stress management ( .09), grossincome ( .07), and social support ( .06) hadindirect total effects only on physical health out-comes.

DISCUSSIONSeveral factors are known or thought to influence

the health outcomes of parents who raise a child witha developmental disability. The direct and indirectrelationships among these variables were examinedusing a single comprehensive structural equationmodel.The psychological and physical health of caregiv-

ers, who were primarily mothers, was strongly influ-enced by child behavior and caregiving demands.These results reiterate associations made betweencaregiving and negative health outcomes in previousstudies.24,2730 Our study corroborated earlier find-ings that child behavior problems are the single mostimportant child characteristic that predicts caregiverpsychological well-being.7 However, as with theKing et al study,7 the childrens behavioral issues

TABLE 4. Range, Mean, SD, and Sample Size for Observed Variables of Caregivers

Variables Range of Score High Score Equivalency Mean SD Sample Size

Distress score 024 More distress 4.71 4.35 468Role: emotional 0100 Better health 63.46 18.34 468Mental health 0100 Better health 69.34 42.60 468Vitality 090 Better health 47.53 22.19 468General health 0100 Better health 67.90 23.04 468Physical functioning 0100 Better health 83.75 21.77 468Role: physical functioning 0100 Better health 68.75 39.02 468Bodily pain 0100 Better health 67.97 25.48 468Chronic 09 Worse health 2.03 1.89 467Netsize 28 More support 5.02 0.94 468Contacts 645 More support 22.31 5.36 468Social provision scale 018 More provision 14.51 3.40 468Family functioning 036 More dysfunction 8.59 5.64 468Integration 057 Better management 37.75 10.62 465Support, esteem 054 Better management 30.63 10.28 464Medical communication 024 Better management 13.66 5.78 462Mastery 328 Superior mastery 14.88 5.63 468Self-esteem 524 Greater self-esteem 19.25 3.24 468Caregiving demand 0100 Less demand 56.56 31.40 468Conduct disorder 1126 Worse behaviors 12.41 2.34 468Hyperactivity 618 Worse behaviors 9.39 2.88 468Emotional disorder 718 Worse behaviors 9.70 2.56 468Gross household income 111 Better SES 8.25 2.49 457

Chronic indicates number of chronic conditions; netsize, existence of possible people to be contacted; contacts, number of contacts for allcategories; integration, integration, cooperation, optimism; support, esteem, support, esteem, stability; medical communication, medicalcommunication and consultation.

TABLE 5. Correlations Among Latent Constructs in the Measurement Model

Income CaregivingDemand

ChildBehavior

Self-Perception SocialSupport

FamilyFunction

StressManagement

PsychologicalHealth

Caregiving demand 0.08Child behavior 0.17* 0.11*Self-perception 0.11* 0.01 0.36*Social support 0.02 0.09 0.05 0.18*Family function 0.22* 0.11* 0.30* 0.66* 0.32*Stress management 0.13* 0.08 0.28* 0.35* 0.17* 0.50*Psychological health 0.17* 0.15* 0.42* 0.59* 0.18* 0.62* 0.43*Physical health 0.14* 0.25* 0.24* 0.32* 0.10 0.41* 0.22* 0.71*

* Values were significant from 0 at P .05.

e632 HEALTH OF CAREGIVERS OF CHILDREN WITH CEREBRAL PALSY by guest on May 17, 2015pediatrics.aappublications.orgDownloaded from

were not assessed to be in the clinical ranges ofseverity. These nonclinical behavioral issues of achild with a disability also influenced caregiverspsychological health indirectly through their effecton self-perception and family function. Thus, inplanning interventions for the child and the family, itis important for service providers to consider chil-drens behavioral issues as an important determinantof the well-being of both the child and the caregiver.Clearly, it is important for health care providers toassess how caregivers are affected by behavioral aswell as functional aspects of the childs disabilityin the provision of comprehensive family-orientedservices. In terms of prevention, providing parentswith cognitive and behavioral strategies to managetheir childs behaviors may have the potential tochange caregiver health outcomes.Caregiving demands contributed directly to both

the psychological and the physical health of the care-givers. Despite being a strong predictor of physicalhealth of caregivers, there was no evidence to sup-port hypothesized relationships between caregivingdemands and self-perception31 or social support.3234This suggests that it is the practical day-to-day needsof the child that create challenges for parents but thatneither their sense of self nor their social supportsmediate the impact of their childs level of disabilityon health outcomes.The direct effect of self-perception and stress man-

agement on psychological health supports the cen-tral relationships of the proposed model. Previousresearch has shown that social support, family func-tioning, and stress management may constrain re-sources with regard to caregiver health out-comes.7,12,29,3541 The proposed direct link between

social support and health was not replicated in thisstudy. Clearly, the influence of social support pro-vided by extended family, friends, and neighbors onhealth outcomes was secondary to that of the imme-diate family working closely together.Family function played a central role in both the

physical and the psychological health of caregivers.This construct affected health directly and also me-diated the effects of self-perception, social support,and stress management. These findings suggest thathealth care providers who work with families ofchildren with long-term disabilities should developinterventions that support and nurture the family asa whole. Therefore, health care providers should beencouraged to value family functioning as much asthe developmental and technical aspects of theservices that are offered to children with complexdisabilities.The evidence in the literature supporting the link

between SES and health is equivocal.30,4244 It is in-teresting that in this study, there was little evidenceto support the proposed link between SES and care-giver health outcomes. Gross household income didnot directly influence caregiver health. Instead, thedirect effect of gross income suggests that higherincome is predictive of improved child behavior. It isof course possible that there was insufficient socio-economic variation across the respondents for an SESeffect to be detectable in this study.The SEM method involves testing a theoretically

derived model. Data-driven considerations may re-quire changes to the theoretical model that result ina model that does not match the original, intendedmodel.45 In our SEM model, the final measurementmodel was somewhat different from the hypothe-

TABLE 6. Measurement Model: Standardized Factor Loadings and SE of Observed Variables on Latent Variables

ObservedVariables

Latent Variables

ChildBehavior

SocialSupport

FamilyFunction

StressManagement

Self-Perception PsychologicalHealth

PhysicalHealth

SDICD 0.76 (0.05)SDIHYP 0.71 (0.05)EMDI 0.59 (0.05)Netsize 0.97 (0.05)Contact 0.77 (0.05)Socprov 0.78 (0.05)Famfunc 0.63 (0.05)Chips1 0.88 (0.04)Chips2 0.77 (0.04)Chips3 0.75 (0.04)Mastery 0.91 (0.06)Selfest 0.44 (0.05)Distress 0.81 (0.04)Emotional 0.72 (0.04)Mental 0.87 (0.04)Vitality 0.77 (0.04)Genhealth 0.80 (0.04)Phyfunc 0.74 (0.04)Rolephy 0.78 (0.04)Bodypain 0.80 (0.04)Chronic 0.54 (0.05)

SDICD indicates conduct disorder scale; SDIHYP, hyperactivity scale; EMDI, emotional disorder scale; Netsize, existence of possiblepeople to be contacted; Contact, number of contacts for all categories; Socprov, Social provision scale; Famfunc, family functioning;Chips1, integration, cooperation, optimism; Chips2, support, esteem, stability; Chips3, medical communication and consultation; Mastery,mastery scale; Selfest, self-esteem summary score; Distress, distress score; Emotional, role of emotional; Mental, mental health; Vitality,score of vitality; Genhealth, general health; Phyfunc, physical functioning; Rolephy, rolephysical functioning; Bodypain, bodily pain;Chronic, number of chronic conditions.

www.pediatrics.org/cgi/doi/10.1542/peds.2004-1689 e633 by guest on May 17, 2015pediatrics.aappublications.orgDownloaded from

sized model as a result of refinements to the concep-tual model. Additional refinements led to changes inthe path model as well. Ultimately, the goal is to finda model that not only fits the data well from a sta-tistical perspective but also is one in which eachparameter of the model has a substantively meaning-ful interpretation.

CONCLUSIONSThe development of interventions to reduce the

stress experienced by caregivers of children with CPis both possible and necessary. The paths thatemerged in the structural model provided evidencefor the hypothesized relationships between variablesthat influence caregiver health outcomes. In ourmodel, it seems that the family unit is the key regu-lating mechanism of health outcomes.36,46,47 Thus,rather than target the child exclusively, interventions

and preventive strategies should also target caregiv-ers, who will in turn be able to respond to the uniquecharacteristics of their child, eg, behaviors, tempera-ment, and functional limitations, in ways that shoulddecrease the impact of their childs disability onthem.The SEM analysis has made it possible to examine

potentially important interrelated factors that con-tribute to caregivers health. In future research, itwould be interesting to look at direct and indirecteffects in this model across the trajectory of the care-givers role over time and to explore how changes inindividual circumstances influence outcomes. Thesemight include, for example, increase in the size andweight of a child with significant functional limita-tions associated with their developmental disabili-ties, at a time in their lives (particularly adolescence)when the physical capabilities of their caregivers

Fig 3. Structural model of factors that influence the health of caregivers.

e634 HEALTH OF CAREGIVERS OF CHILDREN WITH CEREBRAL PALSY by guest on May 17, 2015pediatrics.aappublications.orgDownloaded from

may begin to diminish with caregivers increasingage, or the loss of a person who has shared caregiv-ing responsibilities and moderated the impact of thechilds functional limitations. This method also al-lows for comparison of our model across varied pop-ulations, such as caregivers of children, youths, andthe elderly. Additional exploration of the relation-ships among and between the factors that influencecaring for children with other developmental disabil-ities is also warranted.

ACKNOWLEDGMENTSDr Raina holds a National Health Scholar Award from the

National Health Research and Development Program and an In-vestigator Award from the Canadian Institutes of Health Re-search. Dr Rosenbaum holds a Canada Research Chair in Child-hood Disability from the Canadian Institutes of Health Research.Dr Walter holds a Senior Investigator Award from the CanadianInstitutes of Health Research. Dr Brehaut holds an Ontario Min-istry of Health and Long-Term Care Career Scientist Award.

CanChild Centre for Childhood Disability Research is a healthsystemlinked research unit funded by the Ontario Ministry ofHealth and Long-Term Care.

We thank Susanne King for the input on the conceptual frame-work of the article, Dr Steven Hanna for the input on structuralequation modeling, and the families who participated in ourstudy. We also thank Helen Massfeller, Fulvia Baldassarre, andRoxanne Cheeseman for help in the preparation of this article.

REFERENCES1. Canadian Institute of Child Health. The Health of Canadas Children: A

CICH Profile. 3rd ed. Ottawa, Canada: Canadian Institute of ChildHealth; 2003

2. Bax M. Terminology and Classification of Cerebral Palsy. Dev Med ChildNeurol. 1964;6:297

3. Stanley FJ, Blair E, Alberman E. Cerebral Palsies: Epidemiology and CausalPathways. London, England: Mac Keith Press; 2000

4. Blacher J. Sequential stages of parental adjustment to the birth of a childwith handicaps: fact or artefact? Ment Retard. 1984;22:5568

5. Breslau N, Staruch KS, Mortimer EA Jr. Psychological distress in moth-ers of disabled children. Am J Dis Child. 1982;136:682686

6. Shillitoe R, Christie M. Psychological approaches to the management ofchronic illness: the example of diabetes mellitus. In: Bennett P, Wein-man J, Spurgeon P, eds. Current Developments in Health Psychology. NewYork, NY: Harwood Academic Publishers; 1990:177208

7. King G, King S, Rosenbaum P, Goffin R. Family-centered caregivingand well-being of parents of children with disabilities: linking processwith outcome. J Pediatr Psychol. 1999;24:4153

8. Dumas D, Peron J, Peron Y. Marriage and Conjugal Life in Canada (ReportNo. 91534E). Ottawa, Ontario, Canada: Statistics Canada; 1992

9. Eicher PS, Batshaw ML. Cerebral palsy. Pediatr Clin North Am. 1993;40:537551

10. Aneshensel CS, Pearlin LI, Mullan JT, Zarit SH, Whitlatch CJ. Profiles inCaregiving; The Unexpected Career. San Diego, CA: Academic Press; 1995

11. McCubbin H, Patterson J. The Family Stress Process: The Double ABCXModel of Adjustment and Adaptation. Marriage Fam Rev. 1983;6:735

12. Sloper P, Turner S. Risk and resistance factors in the adaptation ofparents of children with severe physical disability. J Child Psychol Psy-chiatry. 1993;34:167188

13. Wallander JL, Varni JW, Babani L, DeHaan CB, Wilcox KT, Banis HT.The social environment and the adaptation of mothers of physicallyhandicapped children. J Pediatr Psychol. 1989;14:371387

14. Cadman D, Rosenbaum P, Boyle M, Offord DR. Children with chronicillness: family and parent demographic characteristics and psychosocialadjustment. Pediatrics. 1991;87:884889

15. McKinney B, Peterson RA. Predictors of stress in parents of develop-mentally disabled children. J Pediatr Psychol. 1987;12:133150

16. Dyson LL. Response to the presence of a child with disabilities: parentalstress and family functioning over time. Am J Ment Retard. 1993;98:207218

17. Raina P, ODonnell M, Schwellnus H, et al. Caregiving process and

caregiver burden: conceptual models to guide research and practice.BMC Pediatr. 2004;4:1.

18. Palisano R, Rosenbaum P, Walter S, Russell D, Wood E, Galuppi B.Development and reliability of a system to classify gross motorfunction in children with cerebral palsy. Dev Med Child Neurol.1997;39:214223

19. Haley SM, Coster WJ, Ludlow LH. Pediatric Evaluation of DisabilityInventory (PEDI). Boston, MA: Trustees of Boston University; 1992

20. Feeny D, Furlong W, Boyle M, Torrance GW. Multi-attribute healthstatus classification systems. Health Utilities Index. Pharmacoeconomics.1995;7:490502

21. Rosenbaum PL, Walter SD, Hanna SE, et al. Prognosis for gross motorfunction in cerebral palsy: creation of motor development curves.JAMA. 2002;288:13571363

22. Mittelman MS, Ferris SH, Shulman E, et al. Comprehensive supportprogram: effect on depression in spouse-caregivers of AD patients.Gerontologist. 1995;35:792802

23. Beckman PJ. Influence of selected child characteristics on stress infamilies of handicapped infants. Am J Ment Defic. 1983;88:150156

24. Friedrich WN, Wilturner LT, Cohen DS. Coping resources and parent-ing mentally retarded children. Am J Ment Defic. 1985;90:130139

25. Brehaut JC, Kohen DE, Raina P, et al. The health of primary caregiversof children with cerebral palsy: how does it compare with that of otherCanadian caregivers? Pediatrics. 2004;114(2). Available at:www.pediatrics.org/cgi/content/full/114/2/e182

26. James LR, Mulaik SA, Brett JM. Causal Analysis: Assumptions, Models, andData. 1st ed. Beverly Hills, CA: Sage Publications; 1982

27. Gowen JW, Johnson-Martin N, Goldman BD, AppelbaumM. Feelings ofdepression and parenting competence of mothers of handicapped andnonhandicapped infants: a longitudinal study. Am J Ment Retard. 1989;94:259271

28. Kazak AE, Marvin RS. Differences, difficulties and adaptation: stressand social networks in families with a handicapped child. Fam Relat.1984;33:6776

29. Dunst CJ, Trivette CM, Cross AH. Mediating influences of socialsupport: personal, family, and child outcomes. Am J Ment Defic. 1986;90:403417

30. Kitagawa EM, Hauser PM. Differential Mortality in the United States: AStudy in Socioeconomic Epidemiology. Cambridge, MA: Harvard Univer-sity Press; 1973

31. Skaff MM, Pearlin LI, Mullan JT. Transitions in the caregiving career:effects on sense of mastery. Psychol Aging. 1996;11:247257

32. Palfrey JS, Walker DK, Butler JA, Singer JD. Patterns of response infamilies of chronically disabled children: an assessment in five metro-politan school districts. Am J Orthopsychiatry. 1989;59:94104

33. Musil CM. Health, stress, coping, and social support in grandmothercaregivers. Health Care Women Int. 1998;19:441455

34. MaloneBeach EE, Zarit SH. Dimensions of social support and socialconflict as predictors of caregiver depression. Int Psychogeriatr. 1995;7:2538

35. Frey KS, Greenberg MT, Fewell RR. Stress and coping among parents ofhandicapped children: a multidimensional approach. Am J Ment Retard.1989;94:240249

36. Lambrenos K, Weindling AM, Calam R, Cox AD. The effect of a childsdisability on mothers mental health. Arch Dis Child. 1996;74:115120

37. Krause N. Perceived health problems, formal/informal support, andlife satisfaction among older adults. J Gerontol. 1990;45:S193S205

38. Kronenberger WG, Thompson RJ Jr. Psychological adaptation of moth-ers of children with spina bifida: association with dimensions of socialrelationships. J Pediatr Psychol. 1992;17:114

39. Ransom DC, Fisher L, Terry HE. The California Family Health Project:II. Family world view and adult health. Fam Proc. 1992;31:251267

40. Barakat LP, Linney JA. Children with physical handicaps and theirmothers: the interrelation of social support, maternal adjustment, andchild adjustment. J Pediatr Psychol. 1992;17:725739

41. Scharlach AE, Boyd SL. Caregiving and employment: results of anemployee survey. Gerontologist. 1989;29:382387

42. Haan MN, Kaplan GA, Syme SL. Socioeconomic status and health: oldobservations and new thoughts. In: Bunker JP, Gomby DS, Kehrer BH,eds. Pathways to Health: The Role of Social Factors. Menlo Park, CA: HenryKaiser Family Foundation; 1989:76135

43. Freeman HE. Social factors in the chronic disease. In: Levine S, ReederLG, Freeman HE, eds. Handbook of Medical Sociology. Englewood Cliffs,NJ: Prentice-Hall; 1972:63107

44. Rogot E, Sorlie PE, Johnson NJ, Glover CS, Treasure DWA. A Mortality

www.pediatrics.org/cgi/doi/10.1542/peds.2004-1689 e635 by guest on May 17, 2015pediatrics.aappublications.orgDownloaded from

Study of One Million Persons by Demographic, Social and EconomicFactors (Report No. 882896). Bethesda, MD: National Institutes ofHealth; 1981

45. Hayduk LA. Structural Equation Modeling with LISREL. Baltimore, MD:The Johns Hopkins University Press; 1987

46. Friedrich WN. Predictors of the coping behavior of mothers of handi-capped children. J Consult Clin Psychol. 1979;47:11401141

47. Erickson M, Upshur CC. Caretaking burden and social support: com-parison of mothers of infants with and without disabilities. Am J MentRetard. 1989;94:250258

48. Boyle MH, Offord DR, Hofmann HG, et al. Ontario Child Health Study.I. Methodology. Arch Gen Psychiatry. 1987;44:826831

49. Ware JE Jr, Sherbourne CD. The MOS 36-Item Short-Form Health Sur-vey (SF-36). I. Conceptual framework and item selection. Med Care.1992;30:473483

50. King GA, King SM, Rosenbaum PL. How mothers and fathers viewprofessional caregiving for children with disabilities. Dev Med ChildNeurol. 1996;38:397407

51. Statistics Canada. NLSCY Data Dictionary Cycle 2, Release 3. Ottawa,Canada: Statistics Canada; 2003

52. Statistics Canada. NPHS 1996/97 Household Component Questionnaire.Ottawa, Canada: Statistics Canada; 1998

53. Statistics Canada. NPHS 1994-95 Household Component Questionnaire.Ottawa, Canada: Statistics Canada; 1996

54. Pearlin LI, Mullan JT, Semple SJ, Skaff MM. Caregiving and the stressprocess: an overview of concepts and their measures. Gerontologist.1990;30:583594

55. Cutrona CE, Russel DW. The provision of social relationships andadaptation in stress. In: Jones WH, Perlmann D, eds. Advances in Per-sonal Relationships. Greenwich, CT: JAI Press; 1987:3767

e636 HEALTH OF CAREGIVERS OF CHILDREN WITH CEREBRAL PALSY by guest on May 17, 2015pediatrics.aappublications.orgDownloaded from

DOI: 10.1542/peds.2004-1689 2005;115;e626Pediatrics

Walter, Dianne Russell, Marilyn Swinton, Bin Zhu and Ellen WoodParminder Raina, Maureen O'Donnell, Peter Rosenbaum, Jamie Brehaut, Stephen D.

The Health and Well-Being of Caregivers of Children With Cerebral Palsy

ServicesUpdated Information &

tmlhttp://pediatrics.aappublications.org/content/115/6/e626.full.hincluding high resolution figures, can be found at:

References

tml#ref-list-1http://pediatrics.aappublications.org/content/115/6/e626.full.hat:This article cites 36 articles, 10 of which can be accessed free

Citations

tml#related-urlshttp://pediatrics.aappublications.org/content/115/6/e626.full.hThis article has been cited by 27 HighWire-hosted articles:

Subspecialty Collections

_psychology_subhttp://pediatrics.aappublications.org/cgi/collection/psychiatryPsychiatry/Psychology

_subhttp://pediatrics.aappublications.org/cgi/collection/neurologyNeurology

_subhttp://pediatrics.aappublications.org/cgi/collection/disabilitiesDisabilitiesthe following collection(s):This article, along with others on similar topics, appears in

Permissions & Licensing

mlhttp://pediatrics.aappublications.org/site/misc/Permissions.xhttables) or in its entirety can be found online at: Information about reproducing this article in parts (figures,

Reprints http://pediatrics.aappublications.org/site/misc/reprints.xhtml

Information about ordering reprints can be found online:

rights reserved. Print ISSN: 0031-4005. Online ISSN: 1098-4275.Grove Village, Illinois, 60007. Copyright 2005 by the American Academy of Pediatrics. All and trademarked by the American Academy of Pediatrics, 141 Northwest Point Boulevard, Elkpublication, it has been published continuously since 1948. PEDIATRICS is owned, published, PEDIATRICS is the official journal of the American Academy of Pediatrics. A monthly

by guest on May 17, 2015pediatrics.aappublications.orgDownloaded from

DOI: 10.1542/peds.2004-1689 2005;115;e626Pediatrics

Walter, Dianne Russell, Marilyn Swinton, Bin Zhu and Ellen WoodParminder Raina, Maureen O'Donnell, Peter Rosenbaum, Jamie Brehaut, Stephen D.

The Health and Well-Being of Caregivers of Children With Cerebral Palsy

http://pediatrics.aappublications.org/content/115/6/e626.full.htmllocated on the World Wide Web at:

The online version of this article, along with updated information and services, is

of Pediatrics. All rights reserved. Print ISSN: 0031-4005. Online ISSN: 1098-4275.Boulevard, Elk Grove Village, Illinois, 60007. Copyright 2005 by the American Academy published, and trademarked by the American Academy of Pediatrics, 141 Northwest Pointpublication, it has been published continuously since 1948. PEDIATRICS is owned, PEDIATRICS is the official journal of the American Academy of Pediatrics. A monthly

by guest on May 17, 2015pediatrics.aappublications.orgDownloaded from