Variation at local government level in the support for families of severely disabled children and the factors that affect it

Authors


Dr Rob Forsyth at Institute of Neuroscience, Newcastle University, Sir James Spence Institute, Royal Victoria Infirmary, Newcastle upon Tyne NE1 4LP, UK. E-mail: r.j.forsyth@ncl.ac.uk

Abstract

Aim  The aim of this study was to examine geographical variability in the support for families caring for children with severe disabilities as well as the relationships between this variability and local government social and educational performance indicators.

Method  Data were collected from a cross-sectional, self-completed postal survey of the families of 5862 children and young people (aged 0–24y, mean 10y 7mo; 68% male) with severe disabilities resulting in a variety of impairments (21% with autism spectrum disorders, 16% with learning disabilities,* 13% with emotional and behavioural difficulties, and 13% with cerebral palsy [CP]). Data on the severity of intrinsic impairment were assessed using the Health Utilities Index, and the need for support was assessed from the results of a novel parent-completed questionnaire, the European Child Environment Questionnaire (ECEQ). These responses were related to data published by local authorities on educational and social policy.

Results  Higher levels of unmet need and lack of support, as reported by parents of children and young people with severe disabilities, are associated with greater impairment but not with socioeconomic deprivation. After controlling for impairment and diagnosis, variation at local government level is of the order of 1 to 1.5 ECEQ standard deviation scores. The best- and the worst-performing local authorities – in terms of the averages of the ‘support’ scores reported by their surveyed residents – cluster in urban areas. For children with CP, a positive correlation was found between the reported unmet educational support requirements in each local authority area and rates of mainstream school placement for children with special educational needs. This indicates that the placement of children with disabilities into mainstream schools is associated with reported unmet need (r=0.60; p=0.01). In the case of children with autism spectrum diagnoses, the provision of additional basic educational support in mainstream primary education was associated with lower average local authority scores for unmet need, suggesting that this support was appreciated by residents (r=−0.75; p=0.005).

Interpretation  Parent-reported unmet need in the care of children with disabilities shows significant geographical variation after adjustments for severity, type of impairment, and socioeconomic deprivation. Associations between some aspects of reported unmet need and local authority performance indicators suggest that support for families of children with severe disabilities may be improved by policy changes at local government level.

Abbreviations
ASD

Autism spectrum disorder

ECEQ

European Child Environment Questionnaire

HUI

Health Utilities Index

IMD

Index of Multiple Deprivation

PCA

Principal components analysis

SEN

Special educational needs

What this paper adds

  • • The unmet needs of families with children with severe disabilities can be measured.
  • • These needs show significant geographical variation at a local government level after controlling for socioeconomic deprivation and the severity and type of impairment.
  • • This variation is correlated with some performance indicators of local government provision for children with disabilities.

The International Classification of Functioning, Disability and Health (ICF) achieves a synthesis between the previously polarized medical and social models of disability. The medical model emphasized diagnosis, viewing disability as an intrinsic attribute that a person carries with him or her. The social model focused on disability as an inhibitory process imposed on individuals by discriminatory structures. The ICF regards reduced participation (involvement in life situations) as the key consequence of the disablement process.1,2 By incorporating this social perspective, the ICF model makes the testable prediction that individuals with similar levels of intrinsic impairment will achieve differing levels of participation depending on setting or context.

In a population-based study of families of children with cerebral palsy (CP) in north-east England, Hammal et al.3 confirmed that, after controlling for impairment, participation was determined in part by the family’s district of residence, although the factors that make a particular district more or less facilitatory were not identified. SPARCLE, a large cross-sectional study of the determinants of social participation of young people with CP in seven European countries, found variations in participation depending on country of residence.4 As it may not be medically possible to reduce impairment, it is important to understand what attributes make a particular context more or less facilitatory towards those with disabilities. Candidate attributes identified in qualitative work with families of children with CP included accessibility of the physical environment and societal values both at public policy level and at the personal level, in terms of the experienced attitudes of friends, family, and members of the public.5 The European Child Environment Questionnaire (ECEQ) was developed to capture parent-reported information regarding these factors.6 Because the ECEQ covers a wide range of areas, including subjective perceptions of attitudes towards and services for those with disabilities, we conceptualize the questionnaire as representing parent-reported unmet need.

In a previous study of young children with disabilities we hypothesized that participation would be partly determined by attributes of the experienced environment that could be reported for each child using the ECEQ and then compared across the cohort and we confirmed that attributes of the environment, as reflected by ECEQ scores, correlated with reduced social participation.7 We examined the relative contributions of intrinsic impairments to reduced participation, as emphasized by the medical model, and of environmental and contextual factors, as emphasized by the social model and reflected in the ECEQ, and found them to have similar effect sizes.7

In this study, we extend the use of the ECEQ as a parent-reported indicator of attributes of the local environment to a large UK data set of families of children with severe disabilities.1 The aim of the study was to identify geographical variability in unmet need – as reflected by the ECEQ – at local government level as the variation in participation by children with CP found in the study by Hammal et al.3 was demonstrated on this geographic scale. We also sought correlation of any identified variability in unmet need with local government policies as reflected in published performance indicator data.

Method

We use the term ‘local authority’ to denote the UK local government bodies responsible for delivering social and educational services and, within national policy constraints, setting local social and educational policy at the time of data collection (2007).

Sample recruitment

Those who participated were families of children with disabilities known to the Family Fund, a UK charity that distributes national government funds to the families of children with severe disabilities arising from a range of medical conditions and with annual incomes below a threshold (approximately £25 000 per annum in 2007). Families were approached as they contacted the Family Fund for any reason and were invited, without prejudice, to participate in a postal questionnaire study. Ethical approval for the study was granted by the trustees of the Family Fund. All data collection took place during 2007.

Data collection

Data collected comprised the family’s postcode of residence, the child’s medical diagnoses, the parent-completed Health Utilities Index 3 (HUI), and ECEQ instruments. Family Fund staff coded medical diagnosis information provided by the child’s general practitioner and/or paediatrician according to a thesaurus of standard terms. In order to facilitate subgroup analysis, diagnoses were collapsed into 10 broad categories: autism spectrum disorders (ASD), emotional and behavioural disorders, CP, learning disabilities, sensory impairments, and central nervous system, non-neurological, neuromuscular, orthopaedic, and other disorders.

The HUI was chosen as the measure of disease severity because of its applicability across a range of diagnoses, the availability of a one-dimensional summary score, and its strictly ‘within the body’ focus. It is a multidimensional ordinal classification system of impairments of vision, hearing, speech, ambulation, dexterity, emotion, and cognition that can be aggregated into a multiattribute utility function score. This is achieved using weightings derived from public consultation exercises in which adults are asked to rate the desirability of arbitrary combinations of the previously mentioned impairments on a scale where immediate death has a utility of zero and perfect health a utility of 1.8 It is used as a relative, not an absolute, scale of impairment in this study, and any inference that the health status of those children in this study with negative HUI scores should be regarded as literally ‘worse than immediate death’ is inappropriate.

The ECEQ is a novel instrument developed to capture the environment relevant to children with CP as part of SPARCLE.4,9 It offers a subjective parent-reported metric of the environment according to personal experience.6 The questions in ECEQ probe the issues that are identified as important to families with children with disabilities in qualitative work,5 for example physical attributes of home, school, and public spaces; the provision, appropriateness, and adequacy of services; and the experienced attitudes of extended family, friends, professionals, and members of the public.6 The standard form of the items establishes whether a given factor is present or absent and its importance to the family (Table I). ECEQ scoring is discussed below.

Table I.   Structure of a typical European Child Environment Questionnaire item
Does your child receive practical physical help from teachers, therapists, and helpers at school?
A Not needed or not relevant to my child
B Mostly yes, and this helps my child a lot in everyday life
C Mostly yes, and this helps my child a little in everyday life
D Mostly no, but this only restricts my child a little in everyday life
E Mostly no, and this restricts my child a lot in everyday life

The Index of Multiple Deprivation (IMD) 200710 is a deprivation score for very small areas known as lower super output areas (of which there are 32 482 in England). It combines 37 census-derived indicators of income, employment, health, and disability. The IMD score of the child’s locality and the identity of the relevant local authority were derived from the home postcode for the families in this sample. This was also used to link children to published data for their local authority in relation to education and social policy. These data included the local authority rates of issuance of formal statements of special educational needs (SEN).11 Under UK SEN regulations, such statements constitute a formal recognition of the additional needs for educational support for a child that a school cannot meet without additional resources, a commitment to their provision, and regular reviews of the situation. Additionally, rates of placement in specialist segregated versus mainstream schooling were obtained.11 Social care data included local authority expenditure on vulnerable children as a proportion of expenditure on all children’s services and the use of direct payments (the direct allocation of government money to families in order to allow them to purchase their own services).12–14

Data analysis

This study used the same approach to the quantification of ECEQ data as our previous work.7 The ECEQ item A to E responses (Table I) were treated as unordered nominal categorical data. A post hoc factor structure was allowed to emerge from the ECEQ data using a form of principal components analysis (PCA), known as categorical PCA (CATPCA), adapted to handle categorical data.15,16 For continuous numerical data, PCA is a method of calculating the directions along which the variance of a data set is maximized. The goal is to convert a set of variables into a new set of fewer components that is an exact mathematical transformation of the original data. This is achieved by finding a particular linear combination of the original variables for each component that maximizes the variance in the data set accounted for. Such an approach is problematic for categorical data such as the ECEQ responses. CATPCA incorporates an optimal scaling step that amounts to further maximization of this variance accounted for not only over component variable loadings but also over admissible non-linear functions of the original data. These user-selected functions are typically spline or other non-linear transformations that assume the data to be either nominal or ordinal. For this study, the ECEQ data were treated as nominal (with no ordering assumed). Dimensions were added sequentially to the PCA solution, whilst component loadings remained less than 0.3 and items showed a clear, interpretable theme.

The data set comprises children with a range of medical diagnoses. It was hypothesized that ECEQ responses may depend on the diagnosis: for example, accessibility issues might be particularly relevant to the carers of a child with severe CP, whereas respite services could be a priority for the carers of a child with severe ASD. The approach used permitted the creation of diagnosis-specific versions of the ECEQ within with the subgroups with ASD or CP with tailored structures and summary scores. This allowed analyses to examine separately the experiences of children with ASD and CP living in the same local authority.

Categorical Principle Components Analysis was performed using SPSS version 16 for Macintosh OS X (SPSS Inc, Chicago, IL, USA). Relationships between ECEQ dimension scores (as dependent factors) and potential predictor variables were explored using linear regression and mixed-effects modelling to accommodate correlations of ECEQ patient scores within a local authority.17,18

Results

Approximately 12 000 questionnaires were distributed. Data were available for 5862 children, of whom 68% were male and 32% were female. Ages at 1 January 2008 ranged from 0.05 months to 23.7 years (mean 10y 7mo; SD 3y 11mo). HUI scores ranged from −0.36 to 1.0 (mean 0.33; SD 0.34). Diagnosis categories were identified for 99% of children. The largest categories were ASDs (21%), learning disabilities (16%), emotional and behavioural difficulties (13%), and CP (13%). A total of 1603 children lived in Scotland, Wales, or Northern Ireland: local authority and IMD data were available for the 4259 English residents only. IMD scores showed a bias towards greater deprivation (higher IMD scores), with the proportions below the 20th, 40th, 60th, and 80th centiles of the national distribution being 4%, 11%, 23%, and 41% respectively. The English children lived in 149 local authorities, with the number of children per local authority varying widely (median 22; range 3–146).

Performance indicators published by local authorities about educational and social provision for children with additional needs are summarized in Table II. The data show considerable variation between local authorities in the use of available policy instruments, including statements of SEN and direct payments.

Table II.   Local authority performance indicators
IndicatorRange (mean, SD)
  1. ‘Maintained’ denotes a local authority-funded public sector school. For further details see Method.

Gross expenditure on vulnerable children not in local authority care as a percentage of gross expenditure on all children’s services (year ended 31 March 2007)9–61 (40, 7)
Average gross weekly expenditure on direct payments per child receiving direct payments at 31 March 2007£0–1349 (£88, £96)
% Children with statements placed in mainstream schools19–71 (48, 11)
% Children with statements placed in maintained special schools10–63 (36, 10)
% Children with statements placed in non-maintained independent special schools1–21 (5, 3)
% Primary school-age children with a statement of special educational needs0–3 (1.5, 0.5)
% Primary school-age children with special educational needs but without a statement of special educational needs10–27 (18, 3)

The ECEQ has 67 items, and ECEQ returns were complete for 3377 of 5862 (58%) questionnaires. For one item (‘Has your child got the school placement you think he or she needs?’), 19% of data were missing, but the average missing data rate of the remaining items was 5.4% (Appendix I). Cases with missing data were omitted from the PCA analysis. A five-factor structure for the ECEQ data was adopted, the dimensions accounting for 16%, 15%, 6%, 4%, and 3% respectively, of the total variance (totalling 45%). Inspection of factor loadings (Table III; Appendix II) suggested dimension names: support, physical accessibility, respite, transport, and family and friends. These derived ECEQ dimension scores are standardized to a mean of zero and SD of 1, with positive scores implying high unmet need.

Table III.   ECEQ component loadings from PCA analysis
 Dimension 1Dimension 2Dimension 3Dimension 4Dimension 5
  1. Component loadings for a five-dimension categorical principal components analysis (PCA) model of the European Child Environment Questionnaire (ECEQ) data. The top 10 items for each component are shown where loading more than 0.3.

Suggested nameFew hour breaks0.57Ramps in public0.74Overnight care0.43Adequate buses0.75Encouragement family
Home assistant0.57Ramps school0.73Active care0.42Accessible buses0.72Positive-attitude family0.43
Active care0.57Doorways0.73Few hour breaks0.39Accessible trains0.62Encouragement circle0.41
Social services coordinate0.54Lifts in public0.71Home assistant0.36Accessible taxis0.61Emotional support family0.40
Practical help circle0.54Got modified wheelchair/buggy0.71Social services coordinate0.34Adequate vehicle0.52Positive-attitude circle0.40
Overnight care0.54Equipment grants0.68Emotional support circle0.34  Emotional support circle0.35
Counselling0.52Smooth pavements in public0.65Positive-attitude circle0.33    
Emotional support teachers0.52Room in public0.63Encouragement circle0.33    
Encouragement teachers0.52Toilets in public0.61Practical help circle0.31    
Practical help public0.51Toilets school0.61Parent support groups0.31    
Support Physical accessibility Respite Transport Family and friends 

Figure 1 shows a scatterplot of the first two ECEQ dimensions, support and physical accessibility, for children with ASD and CP. This shows that as unmet need for support increases, there is greater variation in the scores for physical accessibility. Children with CP cluster towards the right side of the graph, implying higher unmet needs in relation to physical accessibility, which reflects its greater importance for these children.

Figure 1.

 Relation between first and second European Child Environment Questionnaire (ECEQ) dimension scores and diagnosis. Scatterplot of the ‘support’ and ‘physical accessibility’ ECEQ scores for children with autism spectrum diagnoses (open circles, n=900) and CP (filled circles, n=536). For clarity, data for English children only are shown.

A linear regression of children’s support scores (the first ECEQ dimension) as a function of HUI, IMD, and diagnosis (treated as a factor with 10 levels) with two- and three–way interactions showed a significant association with HUI (p<0.001) but not with IMD (p=0.36). The importance of a child’s local authority in determining support can be studied by examining the model’s residuals, for example the differences between individual children’s actual support scores and their predicted scores from the regression model given their diagnosis, HUI, and IMD. If support is independent of local authority then there should be as many children with positive residuals as negative residuals in each local authority, and box plots of the spread of these residuals by local authority should be centred on zero. However, this is not what was found: Figure 2 shows a map of the average of the residuals for the children in each local authority, which can be thought of as an average local authority score for the support of children with disabilities. Each circle represents one local authority and its area is proportional to this score: positive scores (meaning that, on average, families in that local authority report greater unmet need than would be expected) are shown as black circles. Local authorities with negative scores are shown in grey. The plot demonstrates spatial correlation: local authorities with positive scores (black circles) show clustering. The local authorities with large residuals, whether positive or negative, tend to be in more densely populated areas including Greater London and the Manchester/Liverpool conurbation.

Figure 2.

 Plot of average local authority (LA) hindering/supporting scores. Map of the average local authority score for the first European Child Environment Questionnaire (ECEQ) dimension (support). Circle areas are proportional to the score, with the scale shown in ECEQ standard deviation units. Grey circles represent local authorities with negative residuals; black circles represent local authorities with positive residuals. These scores are averages, for all children in a given local authority, of the residual scores from a regression of ECEQ1 as a function of the Health Utilities Index, the Index of Multiple Deprivation, and diagnosis category (see text).

The presence of spatial correlation prevents the modelling of local authority effects on support of children with disabilities as a random effect. A subgroup analysis was therefore performed in the 12 largest local authorities (each contributing more than 65 children) with a fixed effect model, that is, these individual local authorities were studied in particular, rather than as representative of all local authorities. The presence of children with both ASD and CP in each of these local authorities was capitalized on by using each group to report independently on the characteristics of their shared local authority. Separate PCA reductions of the ECEQ data for the children with ASD and CP generated diagnosis-specific dimension scores (denoted ASDECEQ 1–5 and CPECEQ 1–5 respectively). The first 10 items loading on each dimension together with provisional names for each are shown in Table IV.

Table IV.   Factor loadings for autism spectrum disorder (ASD) and CP-specific ECEQ dimension structures
CPECEQ1: Physical accessibilityCPECEQ2: School supportCPECEQ3: School attitudesCPECEQ4: Public transportCPECEQ5: Family and friends
  1. Component loadings for five-dimension diagnosis-specific categorical principal components analysis (PCA) models of the European Child Environment Questionnaire (ECEQ) data for the subgroups with autism and with CP.

Cerebral palsy
 Doorways0.67Encouragement teachers0.69Teachers understand condition0.41Adequate buses0.69Emotional support family0.50
 Room in public0.66Emotional support teachers0.67Emotional support teachers0.41Accessible buses0.66Encouragement family0.43
 Lifts in public0.65Encouragement classmates0.66Encouragement teachers0.40Accessible taxis0.53Adequate buses0.38
 Ramps in public0.64Positive teacher attitude0.65Encourage independence at school0.34Accessible trains0.50Positive attitude family0.37
 Enlarged rooms at home0.64Practical help teachers0.64Right school0.34  Positive attitude circle0.36
 Active care0.64Positive friends attitude0.64Practical help teachers0.33  Accessible buses0.34
 Home modification grants0.63Emotional support classmates0.62Encouragement classmates0.31  Encouragement circle0.34
 Home assistant0.62Teachers understand condition0.59Lifts school0.30  Accessible taxis0.33
 Got modified wheelchair/buggy0.59Encourage independence at school0.59Child attends school0.30  Emotional support circle0.32
 Toilets in public0.58Included at school0.58      
ASDECEQ1: School attitudesASDECEQ2: School supportASDECEQ3: RespiteASDECEQ4: Public transportASDECEQ5: Family and friends
Autism
 Encouragement teachers0.65Emotional support teachers0.53Overnight care0.49Adequate buses0.70Positive attitude family0.45
 Positive teacher attitude0.64Teachers understand condition0.53Active care0.48Accessible buses0.66Encouragement family0.44
 Emotional support teachers0.64Encourage independence at school0.51Positive attitude circle0.45Accessible trains0.64Positive attitude circle0.41
 Encourage independence at school0.61Encouragement teachers0.50Encouragement circle0.45Adequate vehicle0.56Emotional support family0.40
 Teachers understand condition0.59Special staff school0.48Few hour breaks0.44Accessible taxis0.54Encouragement circle0.40
 Professionals listen0.59Positive teacher attitude0.46Social services coordinate0.43  Emotional support circle0.36
 Included at school0.58Lunchtime support0.44Emotional support circle0.42    
 Special staff school0.56Communication aids school0.43Home assistant0.40    
 Positive friends attitude0.54Practical help teachers0.42Practical help circle0.39    
 Encouragement classmates0.54Right school0.38Encouragement family0.33    

These diagnosis-specific scores were regressed as linear functions of local authority (as a factor with 12 levels), IMD, and HUI. The local authority intercepts from these regressions represent average diagnosis-specific local authority scores for each ECEQ dimension after adjustment for HUI and IMD. In these single-diagnosis subgroups, these local authority scores are more varied than in Figure 2: they range from −2.3 ECEQ SD (facilitating) for one local authority in relation to education for children with CP (CPECEQ3) to +1.0 (hindering) for one local authority in relation to physical accessibility for children with CP (CPECEQ1).

The first ASD and CP dimensions (ASDECEQ1 and CPECEQ1) reflect disparate environmental attributes (school attitudes and physical accessibility respectively; Table IV) and local authority scores for these are, as would be expected, not correlated. Similar considerations apply to the corresponding third dimensions. However, the corresponding second dimensions are nearly identical in terms of their factor loadings, which strongly reflect educational factors (Table IV). The local authority scores for these second dimensions are, however, correlated (r=0.53; p<0.05), supporting the hypothesis that these scores reflect a single underlying educational support trait of the local authority, being experienced and reported on independently by families of children with ASD and CP.

Finally, the relationships between these local authority scores and local authority performance indicators were examined. The local authority scores for the educational ECEQ dimensions for ASD and CP (ASDECEQ2 and CPECEQ2) showed positive correlations with local authority-level data on the percentage of children with statements in mainstream schools (ASDECEQ2: r=0.48; p<0.07; CPECEQ2: r=0.69; p=0.01). ASDECEQ2 was strongly negatively correlated with the proportion of primary school children with SEN but without statements (r=−0.75; p=0.005). There were no significant correlations between relevant ECEQ dimensions and local authority spending on children in need or direct payments.

Discussion

The key finding of this paper is the demonstration of variability in support for families with children with severe disabilities living in different local government areas. These residence effects can be as large as SD 1 to 1.5 (ECEQ; Fig. 2). This study adds to a growing literature offering empirical support for an integrated socio-medical model of disability.3,4,7,19,20 Effects on participation have been demonstrated at international and local government scales,3,4 begging the question of what determines the facilitatory/hindering properties of a given location. We have only been able to give some answers to this question. Our finding of limited correlations between the average facilitating/hindering scores for each of the larger local authorities and their published performance indicator data suggests that support for families with children with disabilities might be somewhat improved by policy changes at local government level, although much of the variability in local authority scores remains unexplained. We report a further novel finding of the clustering of local authorities with large average support scores (whether positive or negative) in some more densely populated areas of England, although the reasons for this have not been established. Although the specifics of these relationships may apply only to the group of families we studied (relatively low-income families of children with severe disabilities living in the UK), the ICF predicts that geographical variability in the factors supporting children with disabilities and their families will be a general finding, and this paper demonstrates the feasibility of examining such effects in detail.

Limitations

The absence of relationships between ECEQ scores and IMD is consistent with our previous work7 and suggests that the ECEQ is measuring something independent of general socioeconomic deprivation; however, it must be remembered that our sample was biased towards relatively low-income families. Another important limitation is the return rate: 5862 replies from 12 000 invitations. Replies were facilitated by provision of a stamped addressed envelope and, although a ‘first-wave’ response of approximately 50% is typical of postal questionnaires,21 funding restrictions precluded follow-up mailings to pursue non-responders. Thus, it is not possible to model factors that may differentially affect the response rate.22

The PCA approach to ECEQ data reduction used in this study and in our previous paper7 is post hoc and therefore exploratory. As a result, the factor structure identified in this study (Table III) is not identical to that of our previous study, although it is very similar.7 This approach also allows the derivation of the separate diagnosis-specific ECEQ scores for the groups with ASD and CP (as exemplars of children with predominantly behavioural/social and motor impairments respectively).

Caution should be exercised in the interpretation of the relationships between performance indicators and unmet need. The positive correlation between average local authority scores for the ASD and CP-specific educational ECEQ dimensions and school placement data implies that mainstream schooling is associated with greater parent-reported unmet need. The strong inverse correlation between local authority scores for the ASD-specific educational ECEQ dimension and proportions of primary school children with SEN but without statements may suggest that willingness by a local authority to acknowledge SEN early, even if thresholds for formal statements of SEN are not reached, is appreciated by families. Messages in relation to social policy are less clear. Neither local authority data on spending on children in need nor implementation of direct payments show relationships to relevant local authority ECEQ scores; however, direct payment data were not available for four of the large local authorities studied here.

Measuring the environment

The ICF model has highlighted the importance of measuring both participation1,7,20 and relevant properties of the environment.23 Both are complex, multidimensional constructs.2,23 The challenges of measuring environmental factors have been emphasized by Whiteneck and Dijkers,23 who contrast the use of objective data reported at the macro-level with personal reports at the micro-level. Our use of performance indicators and the ECEQ correspond to these two approaches. The limited correlation between them found in this study suggests that, until the relationships between them are better understood, both have value and should continue to be used. The demonstration of relationships between the ECEQ and HUI scores, and the interaction between ECEQ dimension scores and diagnosis (Fig. 1), confirm that the parent-reporter’s perspective is influenced by the severity and nature of their child’s impairment. However, this is to be expected and can be seen as providing a degree of construct (convergent) validity for the ECEQ.

There is considerable current interest in the added value of the personal reports of service users in service evaluation.24 Our study confirms the feasibility of collecting such data systematically for families living with a severely disabled child and their value in the evaluation of relevant policy initiatives. Studies are now needed with more representative samples in terms of residence, socioeconomic factors, and diagnostic groups.

Contributors

RF designed and coordinated the study, performed the statistical analysis, wrote the paper, and took overall responsibility for the delivery of the work. RF had full access to all the data in the study, final responsibility for the decision to submit for publication, and is guarantor. AC gave general advice and contributed to the paper. RMcN and PJ contributed to statistical analyses. KC obtained and collated public-domain performance indicator data. MW coordinated data collection at Family Fund and oversaw data quality.

Funding

This study was financed by the Family Fund. The funders had no role in the study design, analysis or interpretation of data, writing of the report, or decision to submit for publication.

Footnotes

  • *

    North American usage: mental retardation.

Acknowledgements

We are grateful to the staff of the Family Fund for support in data collection.

Ancillary