The Use of EQ-5D for Assessment of Health-Related Quality of Life and Its Determinants in Children with Complex Endocrine Conditions Authors: Xanthippi Tseretopoulou, Suet Ching Chen, M. Guftar Shaikh, S. Faisal Ahmed Published: 2026 ABSTRACT AbstractIntroductionThe routine assessment of quality of life (QoL) is rarely performed in clinical paediatric endocrinology and needs further exploration.MethodsBetween October 2022 and December 2023, 216 patients (M:F, 134:82) with a median age of 11.8 years (10th, 90th: 5.6, 16) completed a total of 227 EQ-5D questionnaires, which included an assessment of 5 dimensions (mobility, self-care, participation in usual activities, pain or discomfort, and anxiety/depression) as well as a visual analogue scale (VAS) assessment of overall health. Additional data were collected to assess presenting a complaint, clinic outcome and number of encounters and specialities involved in the patient’s care within the preceding 12 months.ResultsThe median VAS was 90/100 (65, 100), while the median dimensions score was 6/15 (5, 10) for <16 years and 7/15 (5, 9) for >16 years. The VAS score was negatively associated with the dimensions score, with the strongest correlation in the >16 years group (r = −0.64, p = 0.002). The best VAS and dimensions scores were reported in the Early Puberty group at 99/100 and 5/15, respectively, while the worst scores were reported in the Excess Weight group at 77.5/100 and 8/15, respectively. In a multivariate analysis, the number of specialities involved over the previous year was associated with a worse VAS score (p = 0.001) and dimensions scores (p < 0.001).ConclusionsThe EQ-5D can be used in paediatric endocrinology with a clear association with burden of illness as reflected by the complexity of the condition. EQ-5D may represent a useful generic tool for the routine assessment of QoL in paediatric endocrinology. Introduction Standardised questionnaires of health-related quality of life (HRQoL) play an increasingly important role in healthcare research and clinical practice. The EQ-5D is a group of HRQoL instruments, initially developed in 1990 [1] as a simple, generic, standardised tool for performing comparisons across diseases as well as tracking outcomes over time, and it is now widely available in more than 150 languages [2]. Following the first version of EQ-5D (EQ-5D-3L) for adults, the increasing demand for an accurate and reliable measurement of HRQoL in children and adolescents led to the development of EQ-5D-Y in 2009 [3, 4]. A new version of the instrument (EQ-5D-Y-5L), which provides five response levels of severity, was introduced in 2024 (Kreimeier et al. [5]). Although the feasibility, validity and reliability of EQ-5D have been demonstrated in paediatric practice [6] and EQ-5D is often used in routine clinical practice [7, 8], there remains a paucity of data in the field of paediatric endocrinology. A recent systematic review of patient-reported outcome measures (PROMs) in paediatric endocrinology showed that in most cases, these tools were used in the field of diabetes [9]. Another scoping review on PROMs in paediatrics also identified some tools, mainly focused on diabetes, hypothyroidism, and one on CAH in childhood [10]. The utilisation of PROMs in paediatric endocrinology would be useful in promoting patient-centred care, and although condition-specific tools are important, generic tools also have an important role to play, as they can facilitate the comparison of QoL among different conditions and target healthcare resources accordingly. The project, described below, aimed to understand the variation in HRQoL, as assessed by the EQ-5D instrument, in patients attending a routine paediatric endocrinology clinic. Materials and Methods Design All the EQ-5D forms (EQ-5D-5L for ages above 16 years, EQ-5D-Y self-complete for ages between 8 and 15.9 years, and EQ-5D-Y proxy for ages 4–7.9 years) were obtained from the EuroQoL website, following a non-commercial registration application (https://euroqol.org/register/obtain-eq-5d/how-to-obtain-eq-5d/) [4, 11]. The printed questionnaires were distributed at the clinic waiting area, following verbal consent from patients and families prior to their clinic consultation [12, 13]. The paper copies were scanned into the electronic medical records and also anonymously saved in a secure NHS folder, in line with the General Data Protection Regulation. As the project was a health care evaluation exercise, it did not require ethics approval or informed consent. Study Population Patients were approached at the paediatric endocrine clinics in Glasgow Children’s Hospital between November 2022 and December 2023 as part of an assessment of routine health care evaluation. Eligible patients were all patients seen in the clinics of 3 supervising consultants. Exclusion criteria were (1) age below 4 years as EQ-5D is not validated in this age group and (2) difficulty to understand English. A total of 271 eligible patients were identified and 227 questionnaires were collected from 216 patients (11 patients completed 2 questionnaires at 2 separate visits). Among the 216 patients, 134 (62%) were males and 82 (38%) females. Clinical Data Collection From the EQ-5D questionnaire, the dimension data and VAS score were collected. The answers of the dimension section of the EQ-5D-Y (proxy and self-complete) were coded as “1” if the patient reported “no problems,” “2” if “some problems,” and “3” as “a lot of problems,” and the total score was calculated based on the sum of the individual answers. For patients aged >16 years who used the EQ-5D-5L, the possible answers varied from “1” for “no problems” to “5” for “a lot of problem,” on a 5-point Likert scale. Additional clinical data were collected: the patient’s date of birth, the main endocrine presenting complaint (PC), and the outcome from the last endocrine clinic (1 = discharge; 2 = follow-up, no intervention and no test ordered; 3 = follow-up, no intervention and test ordered; and 4 = intervention requiring admission to day unit, such as dynamic function test, hormone replacement). The PCs were coded in 12 disorder groups: growth, puberty (early and late), thyroid, excess weight, DSD, pituitary, adrenal, bone mineral metabolism, oncology, gynaecology, and others. Sex and postcode were also recorded, with the latter used to obtain the Scottish index of multiple deprivation (SIMD), an integrated measure that captures area-level deprivation across multiple aspects, including unemployment, income, education, and crime rates of the data zone of an individual’s place of usual residence [14]. The number of clinical encounters (both face-to-face and virtual ones) and the number of specialities the patient met in the 12 months preceding the clinic visit were also captured. For the purposes of analysis, index patient “11111” was used, which indicates the patient who answers “No problems” (answer 1) in all dimensions, whereas the “Non-11111” group indicates patients who answered “Some problems” or “A lot of problems” to any 1 dimension. Statistical Analysis Results were reported as medians and ranges (10th, 90th centiles). Spearman’s correlation coefficient was used to evaluate the correlation between the scores (dimension, VAS) and the different clinical variables. The Mann-Whitey U test was used for comparison between 2 groups, whereas for comparisons between more than 2 groups, the Kruskal-Wallis test was used. Linear regression analysis was performed to evaluate whether any of the clinical parameters (i.e., the SIMD score) were independent predictors of the dimensions and VAS scores. Data analysis was performed using SPSS statistical software (v. 25.0, Armonk, NY: IBM Corp) and GraphPad Software (v. 10.1, Boston, MA, USA). Results Description of Participants Of the 216 participants, 21 were above 16 years and completed the adult form of the questionnaire. The median (10th, 90th) age of the whole cohort was 11.8 years (5.6, 16). The median VAS score was 90/100 (65, 100), while the median dimensions score was 6/15 (5, 10) for those below 16 years and 7/25 (5, 10) for those above 16 years. The median number of specialities involved in the care of patients during the preceding year to the clinic visit was 2 (1, 4), while the median number of clinical encounters was 2 (1, 7). The median SIMD was 4 (1, 9), with 1 indicating worse socioeconomic deprivation (Table 1). The PCs were coded in 12 different groups of condition (in declining order): growth – 55 of 227 clinic visits (24%), disorders of sex development – 49 (22%), thyroid – 42 (19%), pituitary – 16 (7%), early puberty – 15 (7%), excess weight – 14 (6%), delayed puberty – 12 (5%), adrenal – 10 (4%), bone mineral metabolism – 5 (2%), gynaecology – 5 (2%), oncology – 1 (0.4%), and other/miscellaneous – 3 (1.3%). Table 1.Patients’ characteristicsCharacteristics​ N (%)1216 (100)Median age, years11.8 (5.6, 16)Median SIMD4 (1, 9)Median N specialities22 (1, 4)Median N encounters22 (1, 7)Median EQ-5D VAS (out of 100)90 (65, 100)Median EQ-5D Dimension score <16 years (out of 15)6 (5, 10)Median EQ-5D Dimension score 16–18 years (out of 25)7 (5, 10)All parameters are reported as medians and 10th, 90th centiles.SIMD, Scottish Index of Multiple Deprivation; VAS, Visual Analogue Score. 1 N indicates the number of patients. The number of questionnaires was 227 as 11 patients completed the questionnaire twice. 2The number of clinical encounters (both face-to-face and virtual ones) and the number of specialities the patient met in the preceding to the clinic visit 12 months. Patients’ characteristics All parameters are reported as medians and 10th, 90th centiles. SIMD, Scottish Index of Multiple Deprivation; VAS, Visual Analogue Score. 1 N indicates the number of patients. The number of questionnaires was 227 as 11 patients completed the questionnaire twice. 2The number of clinical encounters (both face-to-face and virtual ones) and the number of specialities the patient met in the preceding to the clinic visit 12 months. Generic Associations The VAS score showed an inverse correlation with the dimensions score in the whole cohort, with Spearman’s correlation coefficient, r = −0.35, p < 0.001. At the analysis of the groups separately, according to age, both in the group below 16 years and in the group above 16 years, the association was inverse (Spearman’s correlation coefficient, r = −0.32, p < 0.001 and Spearman’s correlation coefficient, r = −0.64, p = 0.002, respectively). All correlations were inverse, as a higher VAS score indicates better HRQoL, as demonstrated by a lower dimensions score too. Figure 1 shows the correspondence between the 2 scores as box plots, where the VAS score is divided into 3 different groups: VAS 30–53, 54–77 and 78–100 (Fig. 1). The median dimensions score of females above 16 years was worse (8/25) compared to that of males (6/25) of the same age group (p < 0.01). The VAS score of the index patient “11111” was higher (95/100) compared to that (85/100) of the “Non-11111” group, p < 0.0001 (Fig. 2). Box plots showing the correlation between EQ-5D VAS and dimensions score VAS scores are grouped in 3 categories: 30–53, 54–77, and 78–100 with 100 indicating the best HRQoL. The number of cases in each VAS category were 16, 34, and 177, respectively. Box plots comparing the EQ-5D VAS score of those cases that reported no problems in any dimension and those cases that reported a problem in any of the dimension. *p < 0.0001. Association of VAS Scores with Clinical Variables The number of specialities was negatively associated with the VAS score (Spearman’s correlation coefficient, r = −0.20, p = 0.003) (Fig. 3). The best VAS scores were reported in the Early Puberty group with a median score of 99/100, while the worst ones were reported in the Excess Weight group with a median score of 77.5/100. The median scores were the same across all clinic outcome groups (90/100), with both discharge and follow-up groups, having the same 10th and 90th centiles, 65 and 100, respectively. No statistically significant associations were detected between the VAS score and age (Spearman’s correlation coefficient, r = −0.11, p = 0.1), SIMD (coefficient, r = −0.03, p = 0.65), and the number of clinical encounters (coefficient, r = −0.07, p = 0. 35). The association between number of specialities and the EQ-5D VAS score. Association of Dimensions Scores with Clinical Variables The number of specialities was positively associated with the dimensions score in the below-16-years group (Spearman’s correlation coefficient, r = 0.25, p = 0.0003). The PC with the best median dimensions score was early puberty with 5/15, whereas the worst median was excess weight and pituitary with 8/15. The worse dimension scores were observed in the intervention group, median 5/15 (5, 11), while the median dimensions scores were the same across the groups’ discharge, follow-up, and test: 6/15 (10th, 90th: 5, 8.7) for discharge (5, 10) for follow-up and (5, 10) for test. No statistically significant associations were detected between the dimensions score and the number of clinical encounters in neither the below nor above 16 years (coefficient, r = 0.11, p = 0.1 and r = 0.36, p = 0.11, respectively) or between the SIMD score (coefficient, r = −0.08, p = 0.25 and r = 0.14, p = 0.53, respectively). No association was detected between the dimensions score and age either (coefficient, r = −0.02, p = 0.78 and r = 0.007, p = 0.97, respectively). Multivariate Analysis Analysis of the continuous and categorical clinical parameters (complaint group, sex, SIMD, age, outcome, number of encounters, and number of specialities involved in the care of the patient) showed that the number of specialities involved over the previous year was the only statistically significant independent determinant of the VAS score in multiple linear regression analysis (β = −3.7, p = 0.001) with a negative correlation. Discussion This is the first report of the EQ-5D being used in paediatric endocrinology, and the answers of the 216 patients enable us to draw some conclusions regarding their perceived HRQoL which could be utilised in improving the healthcare system delivery in this speciality. A weak correlation was noted between the two measures of the questionnaire, the VAS and dimensions score, with the exception of the older group, above 16 years old, who completed the adult version of EQ-5D; in this subgroup the inverse association between the two was stronger. This may reflect the higher level of understanding in the adolescent group or a more focused approach in completing the questionnaire. Previous studies have shown that among the adult population, there is a variable level of correlation between VAS scores and EQ-5D-3L value sets, which represent the dimensions scores in the current analysis. The correlation becomes stronger 1 year after follow-up [15]. In primary care, the correlation between EQ-VAS score and EQ-5D-5L dimensions score was also reported to be variable with correlation coefficients ranging between 0.211 and 0.623, which is in agreement with the findings of this work [16]. In children, weaker and variable correlations have been reported between the two scores, according to the different dimensions [17]. The higher (better) VAS score of the index patient “11111” was higher in our cohort compared to that of the “Non-11111” patients, as expected and shown also in previous studies [18]. In the current cohort, the worst VAS and dimensions score were observed in the excess weight group, which is in agreement with previous reports showing worse EQ-5D scores in morbid obesity compared to the normal weight population [19]. The indirect marker of case complexity, the number of specialities involved in the patient’s care during the preceding 12 months, supports this finding, with obesity being related to a higher number of specialities involved. In the multivariate analysis, the number of specialities was an independent predictor of VAS score with a negative correlation, showing that in cases where a higher number of specialities is involved, which could be attributed to the complexity of the condition, the perceived HRQoL is worse. It is worth noting that there were some presenting complaining groups which traditionally require a multi-disciplinary team, such as DSD, where the median VAS score was higher compared to that of the excess weight group. Further work is required to delineate whether a different clinic setting, i.e., professionals from different disciplines reviewing the patient in one clinic, could be a factor that improves the HRQoL. In a systematic review evaluating the HRQoL of adults, with the use of EQ-5D-5L, the VAS and dimensions score varied between people with the same condition but different severity/complexity. For example, patients with asymptomatic prostate cancer showed better scores compared to those receiving chemotherapy, and patients with HIV and comorbidities had a worse score compared to those with HIV without comorbidities [20]. The age and the SIMD, the latter representing the socioeconomic status of the families, were not associated with the VAS and dimensions score. The validation of EQ-5D in adult UK population has provided tables with the expected health state according to age, with younger age being associated with a better HRQoL which declines as age progresses [21], but norms for paediatric population in UK have not been developed yet. UK norms for self-rated health status based on educational qualifications and social class exist [22], but again this refers to adult population and does not take into consideration all the parameters that SIMD does, including health, access to services, income, employment, education, housing, and crime [14]. There were a limited number of children and adolescents who completed the questionnaire twice. Of the 11 patients who completed the questionnaire twice, there was one who showed a high level of variation, but it is difficult to clarify whether this was related to the condition for which they were seen in the clinic or their general health. The dimensions score showed a smaller percentage change compared to the VAS score between the first and second questionnaire. Further studies will be required to evaluate temporal changes in a larger cohort and aim to identify factors that can improve HRQoL. The strengths of this work include the large cohort and the fact that this is the first report of EQ-5D questionnaires in paediatric endocrinology. This is a generic, validated tool with easy access which could be distributed in clinic areas without the need for additional resources. Useful information can be drawn about the HRQoL which can be utilised in targeting resources effectively and promoting further research and quality improvement work in paediatric endocrinology. Resources, including access to psychology services or the setup of multidisciplinary clinics for patients with specific conditions, could improve HRQoL. Multiple clinical variables have been analysed, and a clear association was shown between the scores and condition burden. EQ-5D is a generic tool, which adds complexity when condition-specific variables need to be analysed. In cases of additional clinical conditions, not related to the endocrine condition, or an acute condition (such as an injury or an additional learning difficulty), the patient’s response may be altered. The questionnaires were distributed in a single tertiary paediatric endocrinology centre, and therefore, further work will be required to collect data from multiple UK centres and internationally. The use of the EQ-5D in the routine clinical setting shows its feasibility and clear association with burden of illness as reflected by the complexity of the condition. EQ-5D may represent a useful generic tool for the routine assessment of QoL in paediatric endocrinology. Distribution of the questionnaire to a wider cohort and different centres could facilitate data collection which could then contribute to the improvement of clinical settings and further research in this field. Statement of Ethics As per local guidance, ethical approval and informed consent was not required for this work because all data were collected as part of an ongoing health care evaluation exercise and a clinical service review using clinical tools that are widely and routinely used in the clinical service. Conflict of Interest Statement The authors have no conflicts of interest to declare. Funding Sources The study was not supported by any sponsor or funder. Author Contributions X.T. and S.F.A. conceptualised and designed the project, coordinated and supervised data collection, performed the initial analyses, drafted the initial manuscript, and approved the final manuscript as submitted. X.T. undertook the statistical analyses. All authors, X.T., S.C.C., M.G.S., and S.F.A., contributed to data acquisition and revision of the manuscript, have read and approved the final report, and take public responsibility and accountability for the results.