Modifications of the Effect of Juvenile Idiopathic Arthritis (JIA) on Anxiety and Depression in Children and Adolescents: A Pseudo-Longitudinal Study of 192,019 Children in the United States

1.1. JIA Epidemiology

Reported prevalence estimates range from 12.8 to 45 per 100,000, and incidence rates from 7.8 to 8.3 per 100,000 person-years [6-11], although these may be underestimates [4]. JIA subtypes include oligoarticular, polyarticular (RF-positive or RF-negative), enthesitis-related, psoriatic, systemic, and undifferentiated arthritis [5, 11]. JIA can substantially affect both physical and psychological health, impacting the quality of life and increasing healthcare costs [12-14]. A systematic review observed that annual costs for JIA can vary widely from $310 up to $44,832 per patient, largely depending on factors such as the country of treatment, disease activity, JIA subtype, and use of biological therapies [12]. Although “juvenile” typically implies childhood onset, a significant proportion of children with JIA experience ongoing inflammation into adolescence and adulthood, i.e., more than a third of children with JIA continue to experience active inflammation throughout their adult years [15-17].

JIA rarely occurs in babies younger than six months [18-20]. Peak onset ages for JIA often occur between 2-5 years and again between 6-14 years [4, 19, 21, 22]. JIA has also sex differences, with girls diagnosed more frequently than boys, especially with the oligoarticular and polyarticular RF-negative subtypes [4, 13, 23, 24]. The oligoarticular subtype is the most common JIA subtype in developed countries, typically affecting girls under six years old [24]. On the other hand, the polyarticular JIA-RF negative showed a bimodal trend in girls, i.e., peaks at 2 - 4 years and 6 - 12 years, which is consistent in general for typical JIA diagnosing onset age [4].

1.2. Sex, Medical Insurance, and Psychological Comorbidities in JIA

Children with JIA often face multiple psychosocial challenges. Among these, depression and anxiety are two of the most frequently observed mental health concerns [13, 25-27]. Their prevalence ranges widely, from 7% to 36% for depression and 7% to 64% for anxiety [13]. These mental health symptoms can arise from chronic pain, reduced mobility, social isolation, and uncertainty about long-term outcomes [14, 18, 28, 29].

The sex differences in mental health are well established, with adolescent girls typically demonstrating higher rates of depression than boys [30-32]. However, studies focusing specifically on anxiety and depression among children with JIA remain limited, frequently involving small samples and a few minority participants [13, 25, 33-36].

Socioeconomic status (SES), often reflected in medical insurance status, is another important factor. Families with public insurance (frequently associated with lower SES) may face more barriers to comprehensive healthcare, potentially leading to delayed or limited treatment. By contrast, children with private insurance usually have greater access to specialized care, potentially mitigating the psychological burden [37].

1.3. Study Objectives

In this context, the primary objective of our study was to examine the association between JIA and two key mental health outcomes, depression and anxiety, in a large, nationally representative sample. Specifically, we sought to 1) assess the prevalence of JIA, depression, and anxiety in children in the United States [U.S.], 2) identify significant predictors (demographic, clinical, psychosocial) for depression and anxiety in children with and without JIA), and 3) investigate effect modification by sex in the JIA depression association and by insurance type in the JIA anxiety association. By illuminating these relationships, we aim to guide more tailored interventions, including mental health screening and support strategies, for children living with JIA.

2.1. Source of Data

The National Survey of Children’s Health (NSCH) database (January 1, 2016 – December 31, 2021) was used for the current study. The NSCH database is funded by the Health Resources and Services Administration and Child Health Bureau to collect physical and mental health, access to quality health care, and the child’s family, neighborhood, school, and social context information [37, 38] of children ages 0 to 17 surveyed in all 50 states plus the District of Columbia. Surveys are conducted via mail and web-based surveys by the U.S. Census Bureau between 2016 and 2021, including the prior version of the NSCH and a second survey that includes questions related to children with special needs: “National Survey of Children with Special Health Care Needs” (NS-CSHCN). The NSCH compared response rates across various demographic and socioeconomic subgroups to highlight disparities. The analysis examined the effectiveness of weighting adjustments to reduce nonresponse bias. The weighting process for interviewed children started with a base weight for each sampled household, followed by a nonresponse adjustment for the screener. Eligible children were then adjusted using a Child-Level Screener Factor and a within-household subsampling factor. A nonresponse adjustment for topical issues was applied, followed by a raking adjustment to demographic controls and trimming of extreme weights if necessary. Although findings indicated some differences between respondents and nonrespondents, the weighting adjustments were generally effective in minimizing the nonresponse bias and enhancing the survey's representativeness [38-41]. Additional information about the sampling and administration process, survey methodology, nonresponse bias analysis, and other pertinent information can be found on the survey’s website [38].

2.2. Study Population

The 2016 NSCH was conducted from June 2016 through February 2017; the 2017 NSCH was conducted between August 2017 and February 2018; the 2018 NSCH was conducted between June 2018 and January 2019; the 2019 NSCH was conducted between June 2019 and January 2020; the 2020 NSCH was conducted between June 2020 and January 2021; and the 2021 NSCH was conducted between July 2021 and January 2022. Additional information on the sampling and administration process, survey methodology, nonresponse bias analysis, and other pertinent information can be found on the survey’s website [37]. Since 2016, NSCH data files can be combined to increase the analytic sample size and investigate the time-series trend with multiple years of non-overlapping sampling data [38]. JIA is defined as arthritis of unknown etiology that begins before age 16 [1-5]. Nevertheless, some studies and healthcare systems have expanded this definition to encompass individuals up to 18 years old to be consistent with broader pediatric categories [39]. Although symptoms can appear as early as infancy, identifying JIA in children under one year is relatively rare; nonetheless, there have been cases of diagnosis prior to their first birthday [40]. Moreover, the NSCH collected data on JIA for children under 17 years of age. Therefore, we included all children aged 0 to 17 as the study population. After approval by the institutional review board of the primary author’s university (approval #24-053), 223,195 children between 0 and 17 from the NSCH database were examined. Because the questions on anxiety problems and depression in NSCH only included children between three and seven years, 31,176 children under three years old were excluded from the current study, resulting in 192,019 remaining in the analyses. This included 550 JIA cases, and the distribution of these children’s demographic characteristics is presented in Table 1.

2.3. Predictors and Outcomes

The primary outcomes of this study are depression and anxiety. The questions that NSCH asked children between ages 3 and 17 about depression and anxiety were, “Does this child have this current or lifelong health condition?” The NSCH identified JIA by asking, “Has a doctor or other health care provider EVER told you that this child has arthritis?” and “How severe are this child’s conditions if the child has current or lifelong conditions?” Those responses for currently having the condition and the severity was mild, moderate, or severe were considered to have arthritis [11].

The predictors included the baseline information on children’s age, sex, race/ethnicity, maternal age at delivery, premature birth, low birth weight, months of breastfeeding, body mass index (BMI), sleeping less than recommended age-appropriate hours, allergic to food, drug, and/or insect, has asthma, history of a brain injury, had behavior or conduct problem, had development delay, intellectual disability, speech disability, learning disability, frequency of physical activity, had difficulties in making or keeping friends, insurance type, father’s physical and mental health status, mother’s physical and mental health status, parent or guardian divorced or separated, child’s parent or guardian died, child’s parent or guardian served time in jail or prison, child witnessed domestic violence, child was victim or witness of neighborhood violence, child lived with anyone who was mentally ill, suicidal, or severely depressed, child lived with anyone who had a problem with alcohol or drug, and child was treated or judged unfairly because of their race or ethnic group.

2.4. Statistical Analysis Methods

We summarized continuous variables with the mean (standard deviation) and presented categorical variables as frequency (percent). We evaluated differences between continuous variables with a t-test or a Mann-Whitney U test as the normality assumption was unmet. The associations between categorical variables were estimated with the Chi-square or Fisher’s Exact test. A false discovery rate (FDR) of 0.05 was applied to adjust for the inflation of significance levels due to multiple comparisons. The period prevalence was calculated using the SAS SURVEY procedure to represent the population of noninstitutionalized children nationally and in each state [38, 42].

We used an Augmented Backward Elimination (ABE) model selection method to identify predictors for depression and anxiety using the R ABE package [43]. The ABE method accommodates various model types, such as logistic, Cox, and linear regressions. We used logistic regressions for this current study. The ABE selection method uses a Change-in-Estimate criterion (τ), indicating that a variable should be included in the model if its addition significantly alters the estimate of other variables [43-45]. The ABE procedure started with a full model that included all potential predictors and evaluated the p-values of all effects. Variables with a mild significant level of 0.2 were ranked and sequentially reintroduced individually to test for confounding effects. This assessment involved examining how the inclusion of the previously excluded variable affected the estimates of other variables within the model [43, 46]. Variables with a significant Change-in-Estimate criterion (τ) greater than 0.05 are excluded from the model [43]. The full models for depression and anxiety included the significant variables in Table 1. The ABE model selection method allows the user to retain the critical variables in the model, preventing removals. In this current study, we retained JIA in the models during the selections. After we selected the parsimonious models for depression and anxiety, we further tested the interaction term.

We performed a post hoc power analysis, which showed an average achieved power of 74.1%, confirming that our sample size was adequate for detecting significant effects. This pseudo-longitudinal (repeated cross-sectional) study was conducted in accordance with the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) guidelines to ensure methodological rigor and transparency in reporting. All analyses were performed in SAS package version 9.4 (SAS Institute Inc, NC) or R package version 4.2.2 (R Core Team 2023. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna. https://www.R-project.org/).

3.1. Study Sample Characteristics

There were 192,019 children collected in the NSCH database from 2016 to 2021. We found that the period prevalence of JIA was 2.723 per 1,000. Table 1 demonstrates the demographics and health status of the two groups (children with vs. without JIA). In total, 550 children diagnosed with JIA [213 (38.7%) boys; 337 (61.3%) girls] were reported between 2016 and 2021. The mean age of children with JIA was 13.40 (SD=3.66) years. The median (IQR) age of children with JIA was 14 (11, 16). Mothers’ age at delivery was significantly younger in children with JIA compared to children without JIA (29.1 vs. 30.1, p=0.0004). Children with JIA had a higher percentage of reported depression (21.5% vs. 4.5%, p<0.0001) and anxiety (36.6% vs. 10.5%, p<0.0001) than children without JIA. The period prevalence of depression among children with JIA compared to those without JIA was 231.12 vs. 35.92 per 1,000, while the prevalence of anxiety was 384.32 vs. 81.88 per 1,000. Children with JIA were also more likely to have allergies to food, drugs, or insects (48.8% vs. 23.7%, p<0.0001), to have asthma (25.7% vs. 8.4%, p<0.0001), developmental delay (16.8% vs. 5.5%), learning disability (22.9% vs. 7.3%, p<0.0001), being more challenging to make or keep friends (15.5% vs. 4.0%, p<0.0001). The frequency of no physical activity (0 days per week) was found to be higher in children with JIA compared to non-JIA ones (18.6% vs. 7.5%, p<0.0001). The insurance distribution was different between JIA and non-JIA children, i.e., 71.3% of non-JIA children had private insurance, while 53.3% of children with JIA had private insurance. On the contrary, public insurance had more children with JIA than without JIA ones (33.5% vs. 20.3%).

3.2. Risk Factors for Depression and Effect Modification

Table 2 shows the risk factors for depression. In the depression model, the ABE selected risk factors included children’s age, sex, race, maternal age at delivery, premature birth, months of breastfeeding, overweight or obese, sleeping less than recommended age-appropriate hours, allergic to food, drug, or insect, asthma, behavior or conduct problem, developmental delay, intellectual disability, speech disability, learning disability, physical activity, difficulties in making or keeping friends, insurance type, parents divorced or separated, victim or witness of neighborhood violence, lived with anyone who was mentally ill, suicidal, or severely depressed. Brain injury and JIA were not predictors for depression (Odds Ratio [OR] = 1.26, p = 0.2356; OR = 1.22, p = 0.4310).

4.1. Interventions for Children with JIA

Given these findings, interventions should address both the clinical management of JIA and the psychological challenges these children face. Collaborative care models, where pediatric rheumatologists, mental health professionals, and social workers coordinate, may help detect and treat anxiety or depression early [ 49 ]. One promising strategy is Cognitive Behavioral Therapy (CBT), which supports children with chronic pain by helping them develop coping skills and manage negative thought patterns and stress [ 50 ]. In addition, telemedicine and app-based interventions offer accessible mental health support and symptom-tracking tools, particularly for children with public insurance who face healthcare barriers [ 51 ]. Another viable approach is Mindfulness-Based Stress Reduction (MBSR), which can enhance coping and emotional regulation; this technique has shown success in pediatric populations managing chronic pain [ 52, 53]. Meanwhile, integrated psychoeducation programs guide parents and children in recognizing symptom changes, adhering to medication plans, and employing stress-reduction strategies, all of which can positively influence health outcomes [54]. Addressing low SES barriers entails forming partnerships with community organizations and broadening insurance coverage for mental health services, thus mitigating financial and logistical obstacles to care [55]. By incorporating these strategies, healthcare systems can significantly reduce the mental health burden among children with JIA and minimize downstream complications related to both disease progression and untreated psychiatric conditions.

4.2. Strengths and Limitations

This study’s major strength lies in its large, nationally representative sample, enhancing statistical power and generalizability. By utilizing the NSCH dataset from 2016 to 2021, we were able to obtain a representative study sample size that significantly increased the statistical power of our research. This allowed us to delve deeper into the multifaceted impact of JIA on affected children and their families, providing a comprehensive analysis of how JIA influences both physical and mental health outcomes. The large sample size also enabled us to explore various demographic, clinical, and psychosocial factors in greater detail, shedding light on the complex interplay between these variables and mental health issues in children with JIA. Our findings underscore the importance of addressing JIA patients' physical and psychological needs to improve their overall well-being and quality of life. Future research should continue to build on these insights, focusing on longitudinal studies to better understand JIA's causal relationships and long-term effects on mental health.

Despite the strengths of this study, there are several limitations: (1) The cross-sectional design focuses on associations rather than causal pathways. (2) The dataset lacked genetic and some perinatal information, which could have provided more profound insights into the predictors of depression and anxiety in Children with JIA. (3) The study did not account for different JIA subtypes, which could have varying impacts on mental health outcomes [10]. (4) Potential under-diagnosis of JIA in the comparison group might have led to underestimating the findings. (5) The results are based on children in the U.S., so caution is needed when generalizing these findings to other populations. Factors such as healthcare access [56], treatment, genetic predispositions [57], and sociocultural factors [58] may influence outcomes in ways that differ from those observed in U.S. children. Therefore, caution is warranted when applying these results to populations outside the U.S.