, Introduction
The global impact of the coronavirus disease 2019 (COVID-19) pandemic, stemming from the severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2), has touched the lives of millions worldwide [1]. For most people, recovery from the initial infection happens within weeks, but some experience lingering or new symptoms that persist for months—a condition known as postacute sequelae of COVID-19 (PASC) infection, or long COVID [2]. It encompasses a spectrum of effects on various organ systems and functional domains, including respiratory, cardiovascular, neurological, psychological, and cognitive aspects [3].
Children have not been spared from these effects. During the pandemic, school closures and the shift to online learning disrupted education in many countries, potentially impacting children’s health even in the short term [4]. Beyond school learning, kids faced stress and anxiety from parents or caregivers, sensing that something serious was unfolding around them. This often left them confused and eager for clear, honest information [5]. The fear and uncertainty of the outbreak, combined with the isolating effects of lockdowns, have sparked growing concerns about children’s mental and emotional well-being during and after pandemic period [6-8]. Among these concerns, PASC in children has drawn increasing attention because of its lasting impact on their quality of life (QoL). While much research has focused on adults with long COVID, the effects on children remain less understood. Symptoms like tiredness, trouble thinking, and ongoing physical discomfort can disrupt their daily lives and overall health [9]. However, standardized screening tools are lacking to identify children at high risk for PASC-related QoL impairment in clinical practice.
QoL assessment is a method of evaluating an individual's subjective perception of satisfaction with their own health. It is particularly valuable for assessing the long-term impact of diseases on pediatric populations [10]. In Taiwan, this approach has been used in child psychiatry to explore how physical health issues affect mental and social wellbeing [11,12]. This study relies on the Kinder Lebensqualität fragebogen (KINDL) questionnaire, a trusted tool that assesses different aspects of children’s QoL. It has been adapted and validated for use in Taiwan, ensuring it fits the local culture and language [13,14]. To measure physical symptoms, we used the Children’s Somatic Symptom Inventory-24 (CSSI-24), a 24-question survey that asks children to rate symptoms over the past 2 weeks on a scale from 0 (not at all) to 4 (a lot). Higher scores reflect more severe physical complaints [15]. Widely applied to children with chronic conditions, the CSSI-24 helps gauge the intensity of their symptoms [16,17].
The purpose of this study is threefold: (1) to examine how KINDL (a QoL tool) and CSSI-24 (a symptom tool) relate to each other in Taiwanese children with PASC, shedding light on how symptoms affect their well-being [13,15]; (2) to use receiver operating characteristic (ROC) analysis to find a practical KINDL score cutoff that identifies children at high risk of PASC-related QoL issues early on; and (3) to offer pediatricians clear, actionable advice on using these tools in routine checkups to better support affected kids. This research seeks to fill the gap in pediatric PASC screening, providing a starting point for better care and understanding of long COVID’s impact on children.
Methods
1. Study participants and data collection
This research was a prospective cohort study conducted at China Medical University Children's Hospital, a major medical center in central Taiwan. We enrolled children from the DISCOVER (Diagnosis and Support for COVID Children to Enhance Recovery) cohort, a program designed to study and support kids recovering from severe acute respiratory syndrome coronavirus 2 infections confirmed by reverse transcription-polymerase chain reaction or rapid antigen tests. Participants were children with PASC, defined as having symptoms lasting more than 4 weeks after their initial COVID-19 recovery [18,19]. Recruitment took place from July 1, 2022, to July 31, 2023, during Taiwan’s Omicron variant outbreak, which began in April 2022. This study was approved by the Research Ethics Committee of China Medical University Hospital (CMUH111-REC2-113 (AR-1)). Written informed consent was obtained from all participants and/or their legal guardians before enrollment. All procedures were conducted in accordance with institutional guidelines and the Declaration of Helsinki. Patients and the public were not involved in the design, conduct, reporting, or dissemination of this research.
2. Data gathering
We collected information from children and their parents during outpatient visits for post-COVID-19 concerns. This included details on COVID-19 vaccination status, how their acute infection was managed, and symptoms persisting for at least 4 weeks postrecovery, categorized based on prior studies [2,20]. Laboratory tests were also conducted, measuring markers like complete blood count, erythrocyte sedimentation rate, hsCRP, liver enzymes (aspartate aminotransferase, alanine aminotransferase), creatine phosphokinase, lactic dehydrogenase, immunoglobulin E, and other relevant biomedical indicators.
3. Measuring physical symptoms
To assess physical symptom burden in PASC children, we used the CSSI-24, developed by Walker et al. [15] This 24-item questionnaire asks participants to rate symptoms over the past 2 weeks on a 5-point scale (0=not at all, 4=a lot), with scores ranging from 0 to 96. Higher scores indicate greater physical distress. We treated CSSI-24 scores as a continuous variable and used ROC analysis to pinpoint a threshold for significant symptom burden.
4. Assessing QoL
QoL was measured using the revised KINDL questionnaire (KINDL-R), which is a traditional Chinese version culturally adapted for use in Taiwan, and consists of 24 items rated on a 5-point scale (1=never, 5=all the time).21) Scores are converted to a 0–100 scale across 6 areas—physical, psychological, self-esteem, family, friends, and school—plus an overall QoL score, where higher values reflect better QoL. We used age-specific versions: the 7–11-year-old version for younger children (completed with parental help) and the 12–18-year-old version for adolescents (self-completed). Previous research confirms KINDL’s reliability and validity across these age groups and reporting methods [22-26].
5. Statistical analysis
We summarized participant characteristics (e.g., age, sex, clinical data) with descriptive statistics. Spearman correlation tested the relationship between KINDL scores (total and domain-specific) and CSSI-24 scores or lab results. ROC analysis, guided by Youden’s Index, identified an optimal KINDL cutoff for detecting high-risk children, prioritizing sensitivity due to the lack of a control group [27].
To explore PASC symptoms, we developed a structured questionnaire based on existing literature and clinical observations [2,15]. Although not validated, it captured common post-COVID-19 complaints in children. Principal component analysis (PCA) with parallel analysis (PA) condensed these symptoms into key categories, excluding those reported by fewer than 10 participants [28]. PA, using 1000 Monte Carlo simulations, determined the number of components, followed by PCA with Oblimin rotation to define symptom clusters. These clusters were then correlated with KINDL and CSSI-24 scores using Pearson correlations.
We built 7 multiple regression models to identify factors affecting QoL, with each model using a different QoL score (6 domains plus total) as the outcome. Predictors included age, sex (reference: male), vaccination status (reference: unvaccinated), symptom clusters, and CSSI-24 scores. Finally, we compared our participants’ QoL to prepandemic norms using independent t-tests: children (7–11 years) against a 2010 cohort [25], and adolescents (12–18 years) against a 2008 cohort [13].
Results
Our study included 84 participants: 46 children (average age, 8.74±1.77 years; 41.3% girls) and 38 adolescents (average age, 14.50±1.56 years; 44.7% girls). Most participants (n=65, 77.4%) had received COVID-19 vaccination prior to infection, while 19 (22.6%) were unvaccinated. The average CSSI-24 score across all participants was 16.45±11.94. Table 1 provides detailed breakdowns by age group.
KINDL and CSSI-24 scores showed a significant negative correlation (r=-0.44, P<0.001), meaning higher symptom severity was tied to lower QoL (Fig. 1A). ROC analysis set a KINDL cutoff at 74.75 (area under the curve [AUC], 0.82), separating 72 participants scoring ≤74 (average CSSI-24: 18.43±11.95) from 12 scoring >75 (average CSSI-24: 6.53±4.55) (Fig. 1B). A heatmap of lab results revealed KINDL scores positively linked to eosinophil and lymphocyte percentages, but negatively tied to hsCRP and neutrophil percentages (Fig. 1C).
PCA and PA grouped PASC symptoms into 3 clusters (Fig. 1D): cluster I, psychological distress and cognitive impairment (17.18% variance, eigenvalue=5.84), cluster 2, physical fitness and discomfort (13.61% variance, eigenvalue=4.63), and cluster 3, somatic symptoms (9.97% variance, eigenvalue=3.39). Factor loadings are in Supplementary Table 1. All 3 clusters correlated significantly with CSSI-24 scores (r=0.30 to 0.53, P=0.005 to <0.001). Cluster 1 affected physical (r=-0.25, P=0.02), psychological (r=-0.25, P=0.03), self-esteem (r=-0.25, P=0.02), friends (r=-0.31, P=0.004), school (r=-0.28, P=0.01), and total QoL (r=-0.37, P=0.001). Cluster 2 impacted physical (r=-0.51, P<0.001), psychological (r=-0.28, P=0.01), school (r=-0.28, P=0.01), and total QoL (r=-0.41, P<0.001). Cluster 3 only affected physical QoL (r=-0.32, P=0.003) (Table 2).
Regression models (Table 3) showed that, after adjusting for age and sex, vaccination (β=0.194, P=0.037), cluster 1 (β=-0.251, P=0.012), cluster 2 (β=-0.237, P=0.018), cluster 3 (β=-0.209, P=0.049), and CSSI-24 scores (β=-0.487, P<0.001) significantly predicted total QoL (R2=0.46, adjusted R2=0.41). Cluster 2 and CSSI-24 drove physical QoL; all clusters plus CSSI-24 influenced psychological QoL; CSSI-24 alone affected family and school QoL; and cluster 1 impacted friends QoL. In brief, Component 2 and CSSI-24 score were significant explaining factors for physical QoL; all 3 components and CSSI-24 score were significant explaining factors for psychological QoL; CSSI-24 score was the only significant explaining factor for both family and school QoL; Component 1 was the only significant explaining factor for friend QoL.
In the quality-of-life section (Table 4), the average KINDL score of the children in this study was 67.60±11.41. There was no significant difference (P=0.962) compared to the self-assessment scores of children from the 2010 generation (67.71±15.21). However, the KINDL scores of the children in this study were significantly lower than the parental evaluation scores of the 2010 child generation (75.42±11.27, P<0.001). The average KINDL score of adolescents in this study was 61.43±15.48. Compared with the self-assessment scores of children from the adolescent generation in 2008 (57.68±11.25), the scores were significantly higher (P=0.044). Table 4 also reports a detailed comparison of each KINDL domain. The important findings are summarized as follows: for the children group, there was no significant difference (all P>0.05) between the KINDL scores reported by this sample and the children in the 2010 cohort. However, compared to the scores reported by parents in the 2010 child cohort, current children have significantly lower scores in terms of physical QoL (P<0.001), psychological QoL (P=0.018), self-esteem QoL (P<0.001), and friend QoL (P=0.004). For the adolescent population, compared with the scores reported by adolescents in the 2008 adolescent cohort, the quality of family life (P<0.001) and school life (P=0.049) in this sample were significantly higher.
Discussion
This study provides valuable insights into the interplay between QoL and somatic symptom burden in children with PASC, or pediatric long COVID. By leveraging the KINDL questionnaire and the CSSI-24, the research reveals a significant negative correlation (r=-0.44, P<0.001) between these 2 measures. This finding indicates that children experiencing more severe physical symptoms, such as fatigue or pain, tend to report lower QoL, affecting their daily functioning and well-being. Additionally, PCA categorized PASC symptoms into 3 distinct clusters—psychological distress and cognitive impairment, physical fitness and discomfort, and somatic symptoms—each linked to specific QoL domains. A KINDL cutoff score of 74.75, derived from ROC analysis (AUC, 0.82), was proposed as a potential screening threshold to identify high-risk children in clinical settings. These results carry important implications for understanding PASC’s impact and improving pediatric care.
1. Comparison of QoL scores with historical cohorts
Our results showed that the overall QoL of PASC children was not significantly different from the 2010 child-reported KINDL scores (P=0.962), suggesting that the perceived QoL of PASC children was comparable to that of children before the pandemic. However, compared with the 2010 parent-reported KINDL scores, our sample showed significantly lower QoL (P<0.001). This also shows that the disruptions in education, social interactions, and physical health of PASC children due to the long-term impact of the COVID-19 pandemic have indeed affected their QoL.
Furthermore, adolescents in our study reported significantly higher overall QoL than the 2008 adolescent cohort (P=0.044), a finding that warrants further exploration. Due to lifestyle changes during the epidemic, such as distance education that reduced the pressure of commuting to study for Taiwanese teenagers, and the increased interaction and involvement of parents with teenagers in their lives, some of the negative psychological effects of long-term infection with COVID-19 among teenagers may have been alleviated. Future studies should investigate whether these trends persist across different pediatric populations.
2. Clinical impact of KINDL and CSSI-24
The negative correlation between KINDL and CSSI-24 highlights the substantial impact of somatic symptoms on pediatric QoL. This finding supports the importance of using the assessment tools in pediatric clinic. It’s beneficial and meaningful for patients who need early intervention for symptom burden. The integration of QoL assessments in routine practice could help physicians recognize children who may require additional medical evaluation, psychological support, or rehabilitative care.
In the view of the influence of long COVID symptom on children with PASC, 3 components of the symptom cluster identified by using PCA analysis. And understand how PASC affect QoL form multiple aspect. Psychological distress and cognitive impairment were associated with both lower psychological and social functioning scores, whereas physical discomfort had the strongest impact on physical QoL. These patterns suggest that comprehensive assessments of both physical and psychological health are necessary for effective clinical management.
3. The clinical utility of the KINDL cutoff score
Our ROC curve analysis revealed a score of 74.75 might be a useful threshold for identifying children at risk of lower QoL due to persistent somatic symptoms. In a clinical setting, this cutoff score could help pediatrician quickly screen and identify children who need more attention or early intervention. However, it is essential to interpret this threshold with caution. While sensitivity and specificity values indicate its discriminative potential, additional validation in larger and more diverse populations is necessary to refine its applicability.
A key point in applying this cutoff is balancing sensitivity and specificity. If the threshold is too high, a large proportion of children may be classified as high-risk cluster, potentially leading to unnecessary interventions. On the contrary, if the threshold is too low, some children with significant impairment or needs may be missed. Future studies should examine how this cutoff performs in real-world clinical settings to optimize its practical use.
Interestingly, we observed that higher KINDL scores were associated with increased lymphocyte and eosinophil counts, while lower QoL correlated with elevated neutrophil counts and systemic inflammation markers such as hsCRP. These results suggest that immune system dysregulation may contribute to symptom persistence in PASC children. Elevated hsCRP has been implicated in chronic inflammatory conditions and may serve as a potential biomarker for QoL impairments in long COVID.
Although KINDL and CSSI-24 provide valuable patient-reported assessments, further studies should explore whether integrating hematologic markers into clinical screening could improve early detection of high-risk children.
4. Study limitations
This study has several limitations need to consider. First, the relatively small sample quantity limited the the generalizability of the findings.
Second, use of a nonstandardized questionnaire for PASC symptom collection (clinical chief complain, and capture commonly reported post-COVID-19 symptoms), self-reported measures rather than objective clinical assessments, which may introduce variability. However, this approach was necessary due to the evolving nature of long COVID symptoms, and the lack of a universally accepted pediatric-specific assessment tool at the time of data collection.
Third, this study utilized a cross-sectional design, did not established causal relationship between somatic symptoms and QoL outcomes. Longitudinal studies are needed to determine how these relationships evolve over time.
5. Future directions
Further research should focus on validating the KINDL cutoff in larger cohorts, ensuring its robustness in different pediatric populations. Additionally, future studies should explore whether interventions targeting specific symptom clusters can improve QoL outcomes in children with PASC. By improving screening criteria and enhancing symptom management strategies, healthcare providers can better support the long-term health of pediatric patients affected by PASC.
Future research should also investigate the potential role of immune markers in pediatric long COVID screening. The positive correlation between QoL and lymphocyte/eosinophil counts may indicate an adaptive immune response in children with better health status, and higher neutrophil and hsCRP levels in those with lower QoL suggest an ongoing inflammatory state. Longitudinal studies are needed to determine whether these hematologic markers can complement existing patient-reported outcomes and enhance clinical decision-making.
In conclusion, this study demonstrated evidence that KINDL and CSSI-24 are valuable tools for assessing the impact of PASC on QoL. The proposed KINDL cutoff score of 74.75 offers a potential benchmark for identifying high-risk children, but further validation is necessary. Integrating QoL assessments into routine clinical practice could enhance early detection and intervention, improving health outcomes for children experiencing persistent post-COVID symptoms.
PDF Links
PubReader
ePub Link
XML Download
PubMed
Download Citation
