Description of primary measures
Paediatric Quality of Life Inventory
The PedsQL comprises 23 items capturing the children’s physical and emotional health and their social and school functioning.153 Each item asks the child to indicate how much of a problem a particular aspect of life has been for them over the previous month. For example, under the section ‘My health and activities’, items include ‘It is hard for me to do chores around the house’ and ‘I hurt or ache’; under the section ‘About my feelings’, items include ‘I feel afraid or scared’ and ‘I worry about what will happen to me’. The child’s answer is scored according to the frequency with which each of the 23 problems affect them, from a choice of five options: never = 0, almost never = 1, sometimes = 2, often = 3, almost always = 4.
KIDSCREEN
The 52-measure KIDSCREEN126 was completed by children aged 8–16 years in our sample as a self-report measure of their QoL, and by their parents as a proxy-report measure of their children’s QoL. The KIDSCREEN-52 is not validated for 7-year-olds, so data for children of that age in our sample are available from their parent’s proxy-report only. The KIDSCREEN questionnaire covers 10 domains: (1) physical well-being, (2) psychological well-being, (3) mood and emotions, (4) self-perception, (5) autonomy, (6) parent relation and home life, (7) financial resources, (8) social support and peers, (9) school environment and (10) social acceptance (bullying).
Children and parents rated each item using a 1- to 5-point Likert-type scale according to the frequency (1 = never, 2 = seldom, 3 = quite often, 4 = very often, 5 = always) or intensity (1 = not at all, 2 = slightly, 3 = moderately, 4 = very, 5 = extremely) that they attributed to each statement (e.g. ‘have you felt sad?’ or ‘have you been worried about the way you look?’) over the previous week. The KIDSCREEN-52 does not provide an overall measure of general QoL, but it considers each domain separately, with higher scores indicating a better QoL in each domain.
For 14 items, scores had to be reversed so that higher scores reflected a better QoL for those items. These items were:
physical well-being – (1) in general, how would you say your health is? (excellent = 5, very good = 4, good = 3, fair = 2, poor = 1)
moods and emotions – (1) have you felt that you do everything badly? (2) have you felt sad? (3) have you felt so bad that you didn’t want to do anything? (4) have you felt that everything in your life goes wrong? (5) have you felt fed up? (6) have you felt lonely? (7) have you felt under pressure? (never = 5, seldom = 4, quite often = 3, very often = 2, always = 1)
self-perception – (3) have you been worried about the way you look? (4) have you felt jealous of the way other girls and boys look? (5) would you like to change something about your body? (never = 5, seldom = 4, quite often = 3, very often = 2, always = 1)
social acceptance (bullying) – (1) have you been afraid of other girls and boys? (2) have other girls and boys made fun of you? (3) have other girls and boys bullied you? (never = 5, seldom = 4, quite often = 3, very often = 2, always = 1).
Summary scores were calculated by adding all the scores from each participant. Only summary scores from participants with complete data, where every item of the scale was answered, were summed. Rasch person parameters (PP) were assigned to each possible summary score.154 The PPs were transformed into t-scores with a mean of 50 and a SD of approximately 10.
Description of secondary measures
Child Health Utility 9D (children and young people only)
To inform a future economic evaluation, we used the CHU-9D,116 which is validated for children aged 7–17 years, to estimate incremental health gain in QALYs. This QoL measure was used in addition to PedsQL and KIDSCREEN, which do not have corresponding utility values and cannot be used to calculate QALYs in a future economic evaluation.
The questionnaire consists of nine domains, each with five statements (scored 1–5) that assess the child’s functioning ‘today’ across domains of worry, sadness, pain, tiredness, annoyance, school, sleep, daily routine and activities, for example 1 = I don’t feel sad today, 2 = I feel a little bit sad today, 3 = I feel a bit sad today, 4 = I feel quite sad today, 5 = I feel very sad today. The responses under the nine domains can be taken together as a description of the child’s QoL or ‘health state’ using a descriptive system that combines all responses across all items (e.g. 11232152). There are many different health states based on this descriptive system because of the different combinations of responses across the nine dimensions. Each of these health states has a utility value on a 0–1 scale, where 1 is perfect health and 0 is a state equivalent to being dead.
Completed CHU-9D questionnaires were scored and transformed into utility values using Statistical Product and Service Solutions (SPSS) syntax provided by the developers of the measure. Here, we describe the CHU-9D’s mean utility values and SDs for our sample as an indicator of the CHU-9D’s performance in capturing QoL for our population. There are suggestions in the literature that utility values of 0.9, 0.8, 0.7 and ≤ 0.6, correspond to excellent, very good, good, fair or poor general health, respectively.155 Differences over 0.03 units in utility-based measures of QoL in adults are considered clinically important156 but there is no established MCID for equivalent measures in children.
Revised Child Anxiety and Depression Scale
The brief version of the RCADS is a 25-item questionnaire that derives from the full 47-item RCADS157 and assesses children’s depression and anxiety. It is validated as a self-completed measure for 8- to 18-year-olds and it can also be completed by their parents as a proxy-report. We had two children younger than 8 years in our sample, for whom we only obtained parent proxy-report data.
Items are rated on a four-point Likert scale from 0 to 3, where 0 = never, 1 = sometimes, 2 = often and 3 = always. Examples of items are ‘I worry when I think I have done poorly at something’ (anxiety) and ‘nothing is much fun anymore’ (depression). Raw scores are obtained by adding the scores of all the items. Missing data for raw scores for up to two missing items can be prorated using the remaining items within a scale (sum of the completed items divided by the number of the completed items, then multiplied by the total number of items and rounded to the nearest integer).
Raw scores are transformed into t-scores by matching the raw score to its corresponding age and sex normed t-scores available in the questionnaire’s user guide.111 Clinical cut-off points for the t-scores are 0–64 pre-clinical range, 65–69 borderline clinical range and ≥ 70 clinical range. Higher scores denote greater clinical need. We analysed the t-scores as a continuous variable because the small sample size did not allow grouping the children into clinical categories.
Strengths and Difficulties Questionnaire
The 25-item SDQ110,158 is validated as a self-report measure for 11- to 16-year-olds and as a parent proxy-report measure for all ages. For children aged < 11 years in our sample, we obtained parent proxy-report data only. Over two-thirds of our sample were aged < 11 years, so we have fewer child self-report data than parent proxy-report data on the SDQ.
The SDQ assesses positive and negative attributes of a child’s behaviours and experiences over the previous 6 months under the following five domains:
emotional symptoms (e.g. ‘I worry a lot’); items 3, 8, 13, 16, 24
conduct problems (e.g. ‘I fight a lot. I can make other people do what I want’); items 5, 7, 12, 18, 22
hyperactivity/inattention (e.g. ‘I am constantly fidgeting or squirming’); items 2, 10, 15, 21, 25
peer relationship problems (e.g. ‘I am usually on my own. I generally play alone or keep to myself’); items 6, 11, 14, 19, 23
prosocial behaviour [e.g. ‘I often volunteer to help others (parents, teachers, children)’]; items 1, 4, 9, 17, 20.
Children or parents mark a box for ‘not true = 0, somewhat true = 1, certainly true = 2’. Items 7, 11, 14, 21 and 25 need to be reverse coded. The resultant score for each subscale is 0–10 and the total score ranges from 0 to 40. A total ‘difficulties’ score was generated by summing scores from four scales, excluding the prosocial behaviours one. If one of the four component scores is missing, then the total score is not counted. Higher mean scores for SDQ total and for the four subscales (other than prosocial behaviours scale) indicate a higher level of difficulties. For prosocial behaviours, higher mean scores indicate less difficulties. Each 1-point increase in the total difficulties score corresponds with an increase in the risk of mental health problems.
A range of scores can be used to categorise difficulties into four groups: close to average, slightly raised, high and very high. For child self-report, the corresponding range of scores are:
total difficulties – 0–14, 15–17, 18–19, 20–40
emotional problems – 0–4, 5, 6, 7–10
conduct problems – 0–3, 4, 5, 6–1
hyperactivity – 0–5, 6, 7, 8–10
peer problems – 0–2, 3, 4, 5–10
prosocial – 7–10, 6, 5, 0–4.
For parent proxy-report the range of scores are:
total difficulties – 0–13, 14–16, 17–19, 20–40
emotional problems – 0–3, 4, 5–6, 7–10
conduct problems – 0–2, 3, 4–5, 6–10
hyperactivity – 0–5, 6–7, 8, 9–10
peer problems – 0–2, 3, 4, 5–10
prosocial – 8–10, 7, 6, 0–5.
Mental Health Literacy questionnaire (children and young people only)
We computed the total MHLq score as the sum of all 33 items – the minimum possible score is 33 and the maximum possible score is 165. If any of the items on the questionnaire were missing, then we did not obtain a total score. We also computed subscores for the three MHLq domains by summing up the relevant items, that is:
Help seeking and first-aid skills (10 items – 1, 5, 6, 8, 10, 13, 19, 20, 24, 29). Example items – ‘If a friend of mine developed a mental disorder, I would talk to the form teacher or other teacher’; ‘If I had a mental disorder I would seek my friends’ help’. The minimum possible score is 11 and the maximum possible score is 55.
Knowledge/stereotypes on mental health problems (15 items – 3, 4, 7, 11, 12, 15–18, 22, 23, 25, 26, 28, 31). Example items – ‘People with schizophrenia usually have delusions (i.e. they may believe they are constantly being followed and observed)’, ‘People with mental disorders come from families with little money.’ The minimum possible score is 15 and the maximum possible score is 75.
Self-help strategies (8 items – 2, 9, 14, 21, 27, 30, 32, 33). ‘Good sleep helps to improve mental health’, ‘Doing something enjoyable helps to improve mental health’. The minimum possible score is 8 and the maximum possible score is 40.
Parenting scale (parents only)
Parents who completed the questionnaire described their parenting style over the previous 2 months by using a 7-point visual continuum between two polarised statements (e.g. ‘When my child misbehaves . . . I do something right away – I do something later’). Each point in the continuum receives a score of 1–7, where 7 is the ineffective/maladaptive/unhelpful end of the item. The following items have 7 on the left side: 2, 3, 6, 9, 10, 13, 14, 17, 19, 20, 23, 26, 27 and 30. The rest are reversed scored and have the 7 on the right side.
We computed the total parenting scale score by adding the scores for all the items and then dividing by the number of scored items to arrive at an average total response for parenting. We obtained subscores by calculating the average of responses on the items specific to each domain, as follows:
laxness – items 7, 8, 12, 15, 16, 19, 20, 21, 24, 26, 30 (11 items) (e.g. ‘I threaten to do things that . . . I’m sure I can carry out – I know I won’t actually do’)
over-reactivity – items 3, 6, 9, 10, 14, 17, 18, 22, 25, 28 (10 items) (e.g. When I’m upset or under stress . . . I am picky and on my child’s back – I am not more picky than usual’)
verbosity – items 2, 4, 7, 9, 11, 23, 29 (seven items) (e.g. ‘If my child talks back or complains when I handle a problem . . . I ignore the complaining and stick to what I said – I give my child a talk about not complaining’)
items not on a domain but included in the total score – items 1, 5, 13, 27 (four items) (e.g. ‘When my child pesters me . . . I can ignore the pestering – I can’t ignore the pestering’).
The Parent Stress Index – Short Form (parents only)
Out of the 36 items, 33 items use a Likert scale response option of 1 (strongly disagree) to 5 (strongly agree); for example, ‘I feel trapped by my responsibilities as a parent’. Out of the three remaining items, item 22 asks whether the parent feels that they are 1 = a very good parent, 2 = a better than average parent, 3 = an average parent, 4 = a person who has some trouble being a parent, or 5 = not very good at being a parent. Item 32 asks whether the parent found that getting their child to do something or to stop doing something was 1 = harder, 2 = somewhat harder, 3 = about as hard as, 4 = somewhat easier, or 5 = much easier than they expected. Finally, item 33 asks how many things that the child does annoy their parent and the response options were 1 = 1–3, 2 = 4–5, 3 = 6–7, 4 = 8–9, 5 = 10 +. All items in the scale are reverse-coded so that ascending PSI-SF values indicate ‘worse’ scores, except for items 22 and 33 that are already phrased in this direction and, therefore, are not reversed scored for the analysis.
Scores were calculated separately for the three subscales by summing scores of the 12 items on each subscale, with possible scores in each subscale ranging from 12 to 60. Parental distress was calculated as the sum of items 1–12, parent–child dysfunctional interaction is calculated as the sum of items 13–24 and difficult child is calculated as the sum of items 25–36. The three subscales are computed if all items are answered or if a maximum of one item is missing. A total score for the PSI-SF is calculated by summing the three subscale scores, with possible scores ranging from 36 to 180. Higher scores on the PSI-SF total and its subscales indicate greater parental stress.
Measures for resource use
Child and Adolescent Service Use Schedule
The collection of child resource utilisation data was piloted using the CA-SUS.130,131 We used the tool to identify the most important aspects of resource utilisation and to assess the feasibility of collecting information on the receipt of care and services in relation to children’s needs and services from the NHS, social care, education, and voluntary and third-sector organisations. We did not calculate costs for this resource use as part of this feasibility study.
The CA-SUS was adapted for our study in consultation with Professor Sarah Byford, the designer of the CA-SUS questionnaire. This adaptation involved removing the sections on out-of-pocket expenses and employment; removing the question on education type; removing the follow-on questions asking name of hospital for the hospital service use questions; removing complementary therapists (e.g. homeopath) from the list of community services, adding NHS walk-in services and NHS Direct to the list of community services; and simplifying the questions in the criminal justice services section. A copy of the CA-SUS version used in our study can be found on the project web page [www.journalslibrary.nihr.ac.uk/programmes/hta/142901/#/ (accessed 1 March 2020)].
