PedsQL 4.0 Generic Core Scales

PedsQL 4.0 Generic Core Scales in Paediatric Trials and Economic Evaluation

In some paediatric trials the PedsQL 4.0 Generic Core Scales are the only quality of life data collected, which leaves the health economist with well-validated scores on a 0 to 100 scale and no direct route to QALYs. In NICE's 2017 appraisal of biological treatments for plaque psoriasis in children and young people (TA455), the committee noted that the trials had not collected the EQ-5D-Y and reported only the Children's Dermatology Life Quality Index and the PedsQL, so the assessment group derived EQ-5D-Y values from PedsQL scores with a mapping algorithm. That pattern of collecting the PedsQL, then mapping it to a preference-based measure, explains most of the instrument's place in paediatric cost-utility analysis. PedsQL here means the Pediatric Quality of Life Inventory developed by James Varni and colleagues, and the topic is its 23-item generic core, not the separate disease-specific modules that share the PedsQL name.

Four scales and 23 items across age-specific forms

Varni, Seid and Kurtin describe the PedsQL as a modular instrument for measuring health-related quality of life in children and adolescents aged 2 to 18. The generic core combines multidimensional child self-report and parent proxy-report scales, and its 23 items were designed to apply to healthy school and community populations as well as to children with health conditions. Disease-specific modules, such as the epilepsy module, and the PedsQL Infant Scales are listed as separate instruments by the distributor.

Mapi Research Trust, which distributes the questionnaire through its ePROVIDE platform, gives the item count of each scale.

The forms are matched to age. Child self-report covers ages 5 to 18 and parent proxy-report ages 2 to 18, and in an analysis of 8,591 children Varni, Limbers and Burwinkle concluded that children as young as 5 can reliably and validly self-report with an age-appropriate instrument. The young-child self-report form for ages 5 to 7 uses a simplified three-point scale scored 0, 2 or 4 and anchored with faces. ePROVIDE lists standard and acute versions, and its parent proxy form for toddlers aged 2 to 4 has 21 items, because the school functioning scale has 3 items in place of 5. In the UK English version studied by Upton and colleagues, the instructions ask how much of a problem each item has been during the past month, and each answer is given on a five-point scale from 0 (never a problem) to 4 (almost always a problem).

Scoring the scales on a 0 to 100 metric

Each raw answer is reversed and rescaled so that a higher score means better health-related quality of life. Kook and Varni set out the transformation as 0 becoming 100, 1 becoming 75, 2 becoming 50, 3 becoming 25 and 4 becoming 0. A scale score is then the mean of the transformed items that were answered, which lets a score be computed when some items are missing.

$$S = \frac{1}{n} \sum_{i=1}^{n} 25 \, (4-x_i)$$

where $S$ is the scale score on the 0 to 100 metric, $x_i$ is the raw answer to item $i$ coded 0 to 4 and $n$ is the number of items in the scale that were answered.

Varni, Seid and Kurtin report three summary scores alongside the four scales, the Total Scale Score, the Physical Health Summary Score and the Psychosocial Health Summary Score. A score of 100 means the respondent reported no problems on any item. It does not mean full health in the sense used by a value set, and a score of 0 has no link to being dead, so a PedsQL score cannot stand in for a utility.

Illustrative worked example of a physical functioning score

All numbers in this example are illustrative. A child answers the eight physical functioning items with the raw codes 0, 1, 0, 2, 1, 0, 0 and 1.

1. Transform each answer. The codes become 100, 75, 100, 50, 75, 100, 100 and 75.

2. Average the transformed items. The sum is 100 + 75 + 100 + 50 + 75 + 100 + 100 + 75 = 675, and 675 ÷ 8 = 84.375, so the scale score is about 84.4.

3. Handle a missing answer. If the child had skipped the fourth item, the sum of the seven answered items would be 675 − 50 = 625, and 625 ÷ 7 = 89.29 to two decimal places. The score rises because the item the child found hardest has dropped out, which shows why the number of answered items needs reporting.

Reliability and validity in the original and UK samples

The 2001 validation administered the four scales to 963 children and 1,629 parents. Internal consistency for the Total Scale Score was 0.88 for child report and 0.90 for parent report, and Varni and colleagues judged the summary scores reliable enough for group comparisons. Validity was examined through known-groups comparisons, correlations with indicators of morbidity and illness burden, and factor analysis, and the scales distinguished healthy children from children with acute or chronic health conditions.

Upton and colleagues tested the UK English version in 1,399 children and 970 of their parents. The sample included healthy children, children with asthma, diabetes or inflammatory bowel disease, and children in remission from cancer. Internal consistency exceeded 0.70 for every self-report and proxy-report scale, the measurement properties were similar to those of the original version, and the authors recommended the UK English form for assessing paediatric health-related quality of life in the UK. Evidence of this kind is what section 4.3.14 of the NICE manual asks for when it calls for a generic measure with good psychometric performance in the relevant age ranges.

Self-report, parent proxy-report and who answers

A PedsQL score always has a respondent attached to it. Upton and colleagues found that the correlation between proxy reports and self-reports was higher for children with health problems than for healthy children, so parent and child scores are not interchangeable and the gap between them varies by population.

The respondent matters for economic evaluation because QALYs inherit it, so an analysis needs to state who completed each questionnaire and at what ages. A trial that uses self-report at older ages and parent proxy-report at younger ages produces a mixed series whose differences partly reflect who answered.

Mapping PedsQL scores to utilities for QALYs

Because the scores carry no preference weights, a cost-utility analysis that has only PedsQL data predicts utilities with a mapping function. Two published algorithms show the approach, one predicting EQ-5D utilities and one predicting CHU9D utilities.

Khan and colleagues used a survey of children aged 11 to 15 in four secondary schools in England. Their models were estimated on 559 respondents with complete PedsQL and EQ-5D scores and validated on separate data for 337 respondents. Ordinary least squares models using the four scale scores, their squared terms and interactions predicted best, with a mean squared error of 0.036 and a mean absolute error of 0.115 with age and gender included, or 0.114 without them. Prediction errors were larger for children whose EQ-5D utility was below 0.6, so mapped values are least reliable for the sickest children. The assessment report for NICE TA455 tabulated Khan's best-fitting algorithms for mapping the PedsQL to the EQ-5D-Y.

Lambe and colleagues developed algorithms predicting CHU9D index scores from the PedsQL using the PREDNOS randomised trial of prednisolone for childhood corticosteroid-sensitive nephrotic syndrome, and a generalised linear model performed best. The NIHR report of that trial states that the proxy-reported versions of both questionnaires were used, so the algorithm describes parent-reported health in a single clinical population.

A third route builds a preference-based measure out of PedsQL items themselves, so that health state utility values could be scored from the questionnaire without mapping. In 2024 Kim DeLuca and colleagues described the PedsUtil health state classification system, selected with Rasch analysis from Australian longitudinal data, with one item in each of seven dimensions (physical functioning, pain, fatigue, emotional functioning, social functioning, school functioning and school absence). They reported that preference elicitation to build its scoring system was still in progress.

Illustrative worked example of mapping in a cost-utility analysis

The numbers in this example are illustrative and do not come from any published algorithm. A trial compares a new treatment with usual care in children aged 8 to 12, and the only quality of life measure is the child-report PedsQL. At 12 months the mean Total Scale Score is 75 in the treatment arm and 65 in the usual care arm, the arms were equal at baseline, and the difference is assumed to hold for the whole year. The treatment adds GBP 1,500 per child, and nothing is discounted.

$$u = \alpha + \beta T$$

where $u$ is the predicted utility, $T$ is the PedsQL Total Scale Score, $\alpha$ is the intercept and $\beta$ is the change in predicted utility for each one-point change in $T$.

1. Utilities under the first algorithm. With $\alpha$ = 0.35 and $\beta$ = 0.006, 0.006 × 75 = 0.45 and 0.35 + 0.45 = 0.80 in the treatment arm, while 0.006 × 65 = 0.39 and 0.35 + 0.39 = 0.74 in the usual care arm.

2. QALYs and the ratio. The utility gain of 0.80 − 0.74 = 0.06 held for one year gives 0.06 QALYs, and 1,500 ÷ 0.06 = 25,000, so the ICER is GBP 25,000 per QALY.

3. A second algorithm with a flatter slope. With $\beta$ = 0.004, the same 10-point difference gives 10 × 0.004 = 0.04 QALYs, and 1,500 ÷ 0.04 = 37,500, so GBP 37,500 per QALY.

The children's answers are identical in steps 2 and 3, yet the ratio rises by half because the slope of the mapping changed. An analysis built on mapped PedsQL data therefore needs to name its algorithm, the population in which it was estimated and its prediction error, and to test the result against a credible alternative.

How the Generic Core Scales differ from paediatric utility measures

The Generic Core Scales are a generic patient-reported outcome profile without preference weights, which places them beside preference-based instruments rather than among them. The CHU9D was built for valuation, with nine dimensions and a value set that turns answers into utilities on the scale where full health is 1 and dead is 0, and its questions ask about today, not the past month. The EQ-5D-Y is the youth version of the EQ-5D, originally with three severity levels in each dimension, and Kreimeier and colleagues later developed a five-level version. NICE does not recommend any specific paediatric measure, according to section 4.3.14 of the NICE technology appraisal and highly specialised technologies manual (PMG36, last updated 13 October 2026).

The KIDSCREEN-10 is the nearest non-preference-based neighbour. It condenses ten items into one Rasch-scaled index reported as T-values against European norms, while the PedsQL keeps four scale scores and three summary scores on a 0 to 100 metric. The PedsQL forms also start at age 2 through parent proxy-report, whereas the KIDSCREEN-10 was validated for ages 8 to 18. Within the PedsQL family, the disease-specific modules are separate questionnaires listed under their own entries, so a module score measures something different from a generic core score.

Limitations and common misreadings

Most errors with PedsQL data in economic evaluation come from treating the score as more than it is. The scale is well suited to comparing groups of children, and less suited to valuing health.

• Reading the score as a utility. A Total Scale Score of 80 is not a utility of 0.80, because 0 and 100 are the ends of a response scale and not anchors for dead and full health.
• Mixing respondents. Self-reports and parent proxy-reports agree to different degrees in different populations, so pooling them, or switching respondent during follow-up, builds the change of respondent into the effect estimate.
• Moving a mapping algorithm out of its population. Khan's algorithms came from secondary school pupils in England and Lambe's from parent reports in a nephrotic syndrome trial, so prediction error is likely to grow in younger children, other conditions or another respondent type.
• Ignoring the recall period. The standard form asks about the past month, while the CHU9D asks about today, so a mapping between the two links answers about different periods.
• Overlooking licensing. The ePROVIDE page states that commercial and funded academic users complete a licence agreement and pay a user fee, which belongs in trial budgets and in plans to reuse the questionnaire.

Sources

• Khan KA, Petrou S, Rivero-Arias O, Walters SJ, Boyle SE. Mapping EQ-5D utility scores from the PedsQL generic core scales. PharmacoEconomics. 2014;32(7):693-706. doi:10.1007/s40273-014-0153-y
• Kim DeLuca E, Dalziel K, Wittenberg E, Henderson NC, Prosser LA. Selecting PedsQL items to derive the PedsUtil health state classification system to measure health utilities in children. Health and Quality of Life Outcomes. 2024;22:53. doi:10.1186/s12955-024-02268-5
• Kook SH, Varni JW. Validation of the Korean version of the pediatric quality of life inventory 4.0 (PedsQL) generic core scales in school children and adolescents using the Rasch model. Health and Quality of Life Outcomes. 2008;6:41. doi:10.1186/1477-7525-6-41
• Kreimeier S, et al. EQ-5D-Y-5L: developing a revised EQ-5D-Y with increased response categories. Quality of Life Research. 2019;28(7):1951-1961. doi:10.1007/s11136-019-02115-x
• Lambe T, et al. Mapping the Paediatric Quality of Life Inventory (PedsQL) Generic Core Scales onto the Child Health Utility Index-9 Dimension (CHU-9D) score for economic evaluation in children. PharmacoEconomics. 2018;36(4):451-465. doi:10.1007/s40273-017-0600-7
• Mapi Research Trust. Official PedsQL Generic Core Scales, ePROVIDE instrument page, accessed 17 October 2026. https://eprovide.mapi-trust.org/instruments/pediatric-quality-of-life-inventory-generic-core-scales
• National Institute for Health and Care Excellence. Adalimumab, etanercept and ustekinumab for treating plaque psoriasis in children and young people (TA455). Published 2017. https://www.nice.org.uk/guidance/ta455
• National Institute for Health and Care Excellence. NICE technology appraisal and highly specialised technologies guidance: the manual (PMG36), section 4.3.14. Published 31 January 2022, last updated 13 October 2026. https://www.nice.org.uk/process/pmg36/chapter/economic-evaluation-2
• NIHR Journals Library. Economic analysis: the mapping exercise. In: Sixteen-week versus standard eight-week prednisolone therapy for childhood nephrotic syndrome: the PREDNOS RCT. Health Technology Assessment. 2019;23(26). doi:10.3310/hta23260. NCBI Bookshelf NBK541714. https://www.ncbi.nlm.nih.gov/books/NBK541714/
• Upton P, Eiser C, Cheung I, Hutchings HA, et al. Measurement properties of the UK-English version of the Pediatric Quality of Life Inventory 4.0 (PedsQL) generic core scales. Health and Quality of Life Outcomes. 2005;3:22. doi:10.1186/1477-7525-3-22
• Varni JW, Limbers CA, Burwinkle TM. How young can children reliably and validly self-report their health-related quality of life? An analysis of 8,591 children across age subgroups with the PedsQL 4.0 Generic Core Scales. Health and Quality of Life Outcomes. 2007;5:1. doi:10.1186/1477-7525-5-1
• Varni JW, Seid M, Kurtin PS. PedsQL 4.0: reliability and validity of the Pediatric Quality of Life Inventory version 4.0 generic core scales in healthy and patient populations. Medical Care. 2001;39(8):800-812. doi:10.1097/00005650-200108000-00006