Institute of Health & Wellbeing
Health Economics and Health Technology Assessment (HEHTA)
Use of the Child Health Utility (CHU9D) and Strengths and Difficulties
Questionnaire (SDQ) outcome measures in economic evaluations of school-based interventions:
data from a cluster-randomised controlled trial in Northern Ireland
Nicole Boyer, MSc, Sarah Miller, PhD, Prof Paul Connolly, PhD, Emma McIntosh, PhD
Background
Methods
 There is growing evidence linking early social and emotional wellbeing to later aca1-3
demic performance and various health outcomes including mental health.
 Non-randomised SDQ and CHU-9D utility values were reported with standard descriptive statistics at baseline, after intervention completion, and 12-month follow-up
(n=1277) from intervention completion.
 An economic evaluation was designed alongside the Roots of Empathy clusterrandomised trial evaluation, which is a school-based intervention for improving
pupils’ social and emotional wellbeing.
 Agreement was based on Spearman’s rank correlation coefficient and graphical
techniques.
 The Strengths and Difficulties Questionnaire (SDQ) is a behavioural screening
questionnaire for 4-17-year-old children, comprising a total difficulties score and
4
prosocial behaviour, which aims to identify positive aspects of behaviour.
 CHU9D utility values range from 0.3261(worst state) to 1(perfect health).
 The Child Health Utility 9D (CHU9D) is a generic preference-based health-related
quality of life instrument for 7-17-year-old children.5
 Deprivation was measured by Northern Ireland Multiple Deprivation Measure; range
1-890, 1 denotes most deprived.
5,6
 Two preference-based tariffs currently exist for CHU9D however, to date there is
no preference-based SDQ tariff to facilitate use of the SDQ instrument within an
economic evaluation framework.
Objective
 Normal, borderline and abnormal thresholds for SDQ total difficulties are: 0-11, 1215, and 16-40; prosocial behaviour: 6-10, 5, and 0-4.
 Regression analysis was performed to assess the effect and variation in CHU9D
utility values in relation to SDQ total difficulties and prosocial behaviour scores, age,
sex, and deprivation.
5
6
 Both standard gamble and best-worst scaling tariffs were applied to CHU9D.
 A complete case analysis was originally performed followed by multiple imputation
(MI) using chained equations, also known as full conditional specification for
Investigate the applicability of the SDQ within economic evaluation and its relationship
7
missing
data.
with the CHU9D outcome measure.
Findings
 At baseline, after intervention completion, and at 12-month follow-up, mean CHU9D utility values were 0·84 (SD 0·11), 0·85 (0·11), and 0·85 (0·10).
 Mean SDQ total difficulties scores were 12·11 (3·27), 11·78 (3·14), and 11·95 (3·07).
Table 1 OLS Regressions Predicting CHU-9D Utility Score (SG Tariff) from
SDQ, Gender, Age and Deprivation (Complete Case Analysis)
 Mean SDQ prosocial behaviour scores were 8·12 (2·18), 8·42 (2·00), and 8·62 (1·93).
 A correlation between SDQ total difficulties and CHU9D utility at baseline and 12-month follow-up was Baseline Utility
Total Difficulties
evident (r=–0·10, p=0·004 and –0·08, p=0·018). No correlation was found after intervention completion
Prosocial Behaviour
or between SDQ prosocial behaviour and CHU9D utility.
Female (Reference Male)
 Multivariate regression analysis showed that CHU9D utility was significantly associated with SDQ total P5 ≈ 9 years old (Reference P4)
difficulties score at baseline (β=–0·09, 95% CI –0·03 to –0·16; p=0·007) and after intervention comple- P6 ≈ 10 years old (Reference P4)
Deprivation
tion (–0·08, –0·01 to –0·15; p=0·016) after adjustment for deprivation (Table 1).
β Coef. Std. Err. P value
-0.09
0.03
0.007ᵻ
0.02
0.02
0.29
-0.02
0.07
0.73
-0.09
0.16
0.60
0.20
0.22
0.38
0.03
0.04
0.33
95%
-0.16
-0.02
-0.17
-0.40
-0.25
-0.03
After Intervention Utility
 CHU9D utility was significantly associated with SDQ prosocial behaviour at 12-month follow-up (0·05, Total Difficulties
-0.08
0.03
0.02ᵻ
-0.15
0·01–0·08; p=0·010) after adjustment for age (Table 1).
Prosocial Behaviour
-0.01
0.02
0.66
-0.04
Female (Reference Male)
-0.08
0.07
0.25
-0.22
 Multiple imputation and applying the alternate BWS tariff had little impact on original analysis (Table 2).
P5 ≈ 9 years old (Reference P4)
-0.01
0.15
0.93
-0.31
 Figure 1 reports the distributions of responses of each dimension of the CHU9D.
P6 ≈ 10 years old (Reference P4)
0.28
0.22
0.20
-0.15
Deprivation
0.07
0.04
0.04ᵻ
0.00
Table 2 OLS Regressions Predicting CHU9D After MI for SG Tariff and BWS Tariff Utilities
12-Month Follow-Up Utility
SG Tariff Utility
BWS Tariff Utility
Total Difficulties
-0.05
0.04
0.19
-0.12
β Coef. Std. Err. P value 95%
CI
β Coef. Std. Err. P value 95%
CI
Prosocial Behaviour
0.05
0.02
0.01ᵻ
0.01
Total Difficulties
-0.08
0.02
0.00 -0.12 -0.05 -0.06
0.02
0.00
-0.10 -0.03
Female (Reference Male)
0.00
0.07
0.98
-0.14
Prosocial Behaviour
0.03
0.02
0.14 -0.01 0.06
0.00
0.02
0.89
-0.03 0.04
P5 ≈ 9 years old (Reference P4)
0.13
0.19
0.49
-0.24
Female (Reference
-0.01
0.03
0.85 -0.07 0.06 -0.04
0.03
0.22
-0.12 0.02
P6 ≈ 10 years old
0.58
0.26
0.03ᵻ
0.07
Male)
Deprivation
0.06
0.04
0.10
-0.01
P5 ≈ 9 years old
0.09
0.07
0.20 -0.04 0.22
0.03
0.07
0.62
-0.10 0.16
ᵻSignificant variables reported were rerun in model dropping all non(Reference P4)
significant variables
P6 ≈ 10 years old
0.36
0.10
0.00 0.02 0.09
0.36
0.10
0.00
0.17 0.56
(Reference P4)
Deprivation
0.05
0.02
0.00 0.02 0.09
0.08
0.02
0.00
0.05 0.11
Interpretation
The SDQ and CHU9D are able to measure outcomes in children
aged 8–10 years within an educational setting and there is initial
evidence that they are related in their measurement properties. A
majority of children report no problems in each dimension of
CHU9D, except for tired where nearly 80% of children report
some level of tiredness. To our knowledge, the SDQ and CHU9D
have not yet been used to predict longer-term outcomes within an
economic evaluation context. This is an important avenue for further research because issues remain as to whether these childhood measures could be extrapolated into adulthood. A decision
analytic model for long-term analysis is now being developed.
Funding
CI
-0.03
0.05
0.12
0.23
0.64
0.10
-0.02
0.03
0.06
0.29
0.71
0.14
0.02
0.08
0.14
0.50
1.08
0.14
Figure 1 Frequency of Responses for Each Level of Each Dimension of CHU9D
800
F
r
e
q
u
e
n
c
y
700
600
500
400
300
200
100
0
12345
Worried
12345
Sad
12345
Pain
12345
Tired
12345
Annoyed
12345
School
Work
12345
Sleeping
12345
Daily
Routine
12345
Activities
Dimensions and Levels of CHU-9D
Baseline
After Intervention
12-Month Follow-Up
References [1] Petrides KV, Frederickson N, Furnham A. The role of trait emotional intelligence in academic performance and deviant behaviour at school. Personality & Indiv. Diffs. 2004;36(2):277-93. [2] Ciarrochi J, Deane FP, Anderson S. Emotional intelligence
This project was funded by the National Institute for Health
moderates the relationship between stress and mental health. Personality and Individual Differences. 2002; 32(3):197-209. [3] NaResearch (NIHR) Public Health Research programme.
tional Institute for Health and Care Excellence, 2008. Promoting children’s social and emotional wellbeing in primary education. Public Health Guidance 12. [4]Goodman R. The Strengths and Difficulties Questionnaire: a research note. Journal of child psychology
and psychiatry, and allied disciplines. Jul 1997;38(5):581-586. [5] Stevens K. Valuation of the Child Health Utility 9D Index. PharmacoEconomics. Aug 1 2012;30(8):729-747. [6] Ratcliffe J, Flynn T, Terlich F, Stevens
K, Brazier J, Sawyer M: Developing Adolescent-Specific Health State Values for Economic Evaluation: An Application of Profile Case Best-Worst Scaling to the Child Health Utility 9D. PharmacoEconomics 2012,
30:713-727. [7] Lee KJ, Carlin JB: Multiple imputation for missing data: fully conditional specification versus multivariate normal imputation. Am J Epidemiol 2010, 171:624-632.