Association between sleep disorders and health care utilization in children with chronic medical conditions: a Medicaid claims data analysis

Abstract

Study Objectives:

To examine the risk of increased health care utilization (HU) linked to individual sleep disorders in children with chronic medical conditions.

Methods:

Medicaid claims data from a cohort of 16,325 children enrolled in the Coordinated Healthcare for Complex Kids (CHECK) project were used. Sleep disorders and chronic medical conditions were identified using International Classification of Diseases, Ninth, and 10th Revision, codes. Three HU groups were identified based on participants’ prior hospitalizations and emergency department (ED) visits in the 12 months prior to enrollment: low (no hospitalization or ED visit), medium (1–2 hospitalizations or 1–3 ED visits), and high (≥ 3 hospitalizations or ≥ 4 ED visits). The odds of being in an increased HU group associated with specific sleep disorders after controlling for confounding factors were examined.

Results:

Children with chronic medical conditions and any sleep disorder had nearly twice the odds (odds ratio = 1.83; 95% confidence interval: 1.67–2.01) of being in an increased HU group compared with those without a sleep disorder. The odds of being in the increased HU group varied among sleep disorders. Only sleep-disordered breathing (odds ratio = 1.51; 95% confidence interval : 1.17–1.95), insomnia (odds ratio = 1.46; 95% confidence interval : 1.06–2.02), and circadian rhythm sleep disorder (odds ratio = 2.45; 95% confidence interval : 1.07–5.64) increased those odds. Younger age and being White were also linked to increased HU.

Conclusions:

Sleep disorders are associated with increased risk of heightened HU (ED visits and/or hospitalizations) in children with chronic medical conditions. This risk varies by specific sleep disorders. These findings indicate the need for careful evaluation and management of sleep disorders in this high-risk cohort.

Citation:

Adavadkar PA, Brooks L, Pappalardo AA, Schwartz A, Rasinski K, Martin MA. Association between sleep disorders and health care utilization in children with chronic medical conditions: a Medicaid claims data analysis. J Clin Sleep Med. 2024;20(4):595–601.

BRIEF SUMMARY

Current Knowledge/Study Rationale: Sleep disorders are prevalent in children with chronic medical conditions, who also exhibit disproportionately high health care utilization, yet the relationship between specific sleep disorders and health care utilization in this population remains unclear. This study aimed to investigate this association.

Study Impact: This study shows that the presence of a sleep disorder is associated with a nearly 2-fold increase in the risk of increased health care utilization and provides specific odds associated with individual sleep disorders. These findings emphasize the importance of a prioritized and targeted approach in sleep management to optimize health care resources.

INTRODUCTION

The prevalence of chronic medical conditions (CMCs) is on the rise in the pediatric population and so is health care utilization (HU).^1–5 Children with CMCs have increased utilization and cost of health care due to their complex health care needs and the burden of these conditions increases with the number of chronic conditions.^4,6 Sleep disorders are known to be prevalent in this population and can exacerbate the underlying health challenges associated with CMCs, such as asthma, diabetes, obesity, and attention-deficit/hyperactivity disorder (ADHD). For instance, children with asthma often experience sleep disruptions stemming from asthma-related symptoms or the effects of asthma medications.^7,8 Additionally, sleep-disordered breathing (SDB) is more commonly observed in children with asthma and may contribute to more frequent asthma exacerbations, emergency department (ED) visits, or hospitalizations. Similarly, children with autism spectrum disorder commonly face sleep-onset and maintenance difficulties, and the resulting poor-quality sleep can further worsen daytime behavioral impairment.^7,9–11 Obstructive sleep apnea (OSA), which is more prevalent in children with ADHD, can exacerbate symptoms of inattention and hyperactivity.¹² Moreover, OSA is also more prevalent in children with obesity, which, in turn, increases the risk of weight gain due to underlying endocrine and metabolic disturbances.^13–16

These interconnected factors highlight the potential impact of sleep disorders on the exacerbation of CMCs and subsequent HU. Unfortunately, the underdiagnosis of sleep disorders in the general pediatric population and in children with CMCs obscures the role of sleep disorders in HU and may limit treatment opportunities for those who are affected.^17–20

Sleep disorders are shown to be associated with significantly increased HU in adults.⁶ Similar studies in pediatric population are limited. Previous studies that included general pediatric populations have shown increased HU and cost associated with OSA.^6,21 HU is also higher in children with CMCs compared with those without and it is vital to examine the role of sleep disorders in HU in this cohort.⁴ The aim of this study was to examine the role of sleep disorders in increased HU in chronically ill children after controlling for comorbidities and other confounding factors. We also investigated the role of specific sleep disorders in increased health care usage.

METHODS

Study population

Study data were obtained from the Coordinated Health Care for Complex Kids (CHECK) program funded under the Centers for Disease Control and Prevention Healthcare Innovation Award (grant number 1C1CMS331342). Persons eligible for enrollment in this program were Medicaid beneficiaries 25 years of age or less residing in Cook County and having 1 or more CMCs. Participants who enrolled in the CHECK project between 2014 and 2018 were included in this analysis. Referrals came from the State of Illinois Department of Healthcare and Family Services (ie, Medicaid), Medicaid Managed Care Organizations, the University of Illinois at Chicago Hospitals, or by self-referral.^22–24 Illinois Medicaid claims data for each participant were extracted. The International Classification of Diseases, Ninth Revision (ICD-9), or Tenth Revision, Clinical Modification (ICD-10-CM), diagnostic codes were used to determine sleep disorders and CMCs. The older ICD-9 codes were converted to ICD-10-CM codes.

Participants

Only children aged 0 to 18 years with at least 1 CMC were included in this study. Participants aged 18–25 years (n = 3,052) and those who did not have a CMC but were included in the CHECK program as siblings of those with a CMC (n = 1,192) were excluded in this research.

HU categorization levels

CHECK participants were divided into 3 HU groups at the time of enrollment based on the number of ED visits and hospitalizations they had during 1 year prior to the CHECK enrollment. The High-HU group consisted of children who either had 4 or more ED visits or 2 or more hospitalizations. The Medium-HU group included children with 1 to 3 ED visits or 1 hospitalization, and the Low-HU group included those who had no ED visits or hospitalizations. These 3 groups were used to analyze the association between HU and sleep disorders. Children with sickle cell disease (n = 284) were not included in the study because they were automatically placed in the “High” HU group regardless of their number of ED visits or hospitalizations.

Sleep disorder diagnoses

Sleep disorders were identified using ICD-9 and ICD-10-CM codes. Sleep disorders related to breathing difficulties were categorized as SDB. SDB included diagnoses of apnea only, snoring only, and apnea and snoring combined. Periodic limb movement disorder and restless legs syndrome were grouped as sleep-related movement disorders. Sleepwalking, confused arousal, sleep terrors, nightmares, and other parasomnia types were grouped as parasomnia disorders.

Age groupings

In our analysis, we divided participants into 5 age subgroups: infants and toddlers (0–2 years), preschoolers (3–5 years), elementary school children (6–10 years), middle school students (11–13 years), and high school students (14–17 years).

Chronic medical conditions

Most of the participants (77%) in the study cohort had multiple CMCs. For analysis, the CMCs were grouped as follows: asthma/respiratory disorders, developmental disorders, ADHD, diabetes/metabolic disorders, overweight/obesity, prematurity, neurologic disorders, and mood disorders. Tonsillar hypertrophy was included as a clinical feature for analysis. Additionally, the association between HU and the presence of either 1 or multiple CMCs was explored.

Data analysis

A multivariable cumulative logistic regression analysis using each sleep disorder as an independent variable and age group, race, and CMCs as covariates was used to analyze the odds of being enrolled in the increased HU group.

In the analyses, statistical methods such as chi-square tests, multivariable logistic regression, and multivariable cumulative logistic regression were used. Results were considered statistically significant if the P value was less than .05. Analyses were performed using R version 3.6.0 (April 26, 2019; R Foundation for Statistical Computing, Vienna, Austria) on de-identified patient information. The University of Illinois at Chicago Institutional Review Board approval for secondary analyses of the CHECK data is covered by protocol no. 2016-0235.

RESULTS

Participant characteristics by HU groups

The study population of 16,325 children was divided into 3 levels of HU: Low (n = 7,486), Medium (n = 8,063), and High (n = 776) (see Table 1). Age varied significantly among the groups, with a mean age of 10 years in the Low-HU group vs 6.8 years in the High-HU group. The most prevalent age group in the High-HU group was 0–2 years (32.1%); the 6–10-year age group was most prevalent in both Low-HU (32.1%) and Medium-HU (26.7%) groups.

Table 1.

Participant characteristics by health care utilization groups.

Variable	Health Care Utilization Groups						P
	Low (n = 7,486)		Medium (n = 8,063)		High (n = 776)
	n	%	n	%	n	%
Age, mean (SD), y	10.02 (4.88)		8.44 (5.45)		6.83 (5.72)		< .001
Age categories							< .001
0–2 y	477	6.4	1,367	17.0	249	32.1
3–5 y	1,118	14.9	1,565	19.4	154	19.8
6–10 y	2,403	32.1	2,152	26.7	165	21.3
11–13 y	1,333	17.8	1,096	13.6	63	8.1
14–18 y	2,155	28.8	1,883	23.4	145	18.7
Sex							.227
Female	3,232	43.2	3,510	43.5	360	46.4
Male	4,254	56.8	4,553	56.5	416	53.6
Race/ethnicity
White	109	1.5	184	2.3	15	1.9	< .001
Black/African-American	2,860	38.2	2,671	33.1	225	29.0
Hispanic/Latino	1,683	22.5	1,502	18.6	142	18.3
Other	44	0.6	51	0.6	5	0.60
Two or more	122	1.6	139	1.7	11	1.4
Unknown	2,668	35.6	3,516	43.6	378	48.7
Number of CMCs							< .001
One CMC	2,103	28.1	1,632	20.2	101	13.0
Two or more CMCs	5,383	71.9	6,431	79.8	675	87.0

Each P value is based on the overall association between a background category and the health care utilization categories. CMC = chronic medical condition, SD = standard deviation.

HU did not significantly vary by sex but did by race/ethnicity. Black and Hispanic children were more prevalent in the Low-HU group (38.2% and 22.5%, respectively) compared with the High-HU group (29% and 18.3%, respectively). White race was most prevalent in the Medium-HU group (2.3%) and least prevalent in the Low-HU group (1.5%).

There was significant variation in the number of children with 1 CMC compared with multiple CMCs across the HU groups. Most children in all 3 HU groups had 2 or more CMCs, with the highest prevalence in the High-HU group (87%) and the lowest in the Low-HU group (71.9%). Conversely, the prevalence of 1 CMC was highest in the Low-HU group (28.1%) and lowest in the High-HU group (13%).

Sleep disorders across the HU groups

SDB was the most prevalent sleep diagnosis across all HU groups followed distantly by insomnia (Table 2). The prevalence of SDB, insomnia, and circadian rhythm sleep disorder was highest in the High-HU group and lowest in the Low-HU group and the difference in the prevalence among those groups was statistically significant. The prevalence of nocturnal enuresis, hypersomnia, and parasomnia did not show significant variations across the 3 HU groups.

Table 2.

Distribution of sleep disorders across health care utilization groups.

ICD-10-CM Diagnostic Codes	Sleep Disorder	Health Care Utilization Groups						Comparisons, P
		Low (n = 7,486)		Medium (n = 8,063)		High (n = 776)		High vs Low	Medium vs Low
		n	%	n	%	n	%
780.51,780.53,780.57, 27.20,327.1,327.23, 327.26,786.09, 786.03,786.09, 786.03	Sleep-disordered breathing	493	6.6	1143	14.2	160	20.6	< .001	< .001
327.01, 327.02, 327.09, 780.51, 780.52	Insomnia	58	0.8	91	1.1	14	1.8	.004	.024
788.36, 788.30	Nocturnal enuresis	80	1.1	109	1.4	6	0.8	.442	.108
780.58, 327.51, 333.94	Sleep-related movement disorder	15	0.2	30	0.4	3	0.4	.298	.050
327.31, 327.30, 327.39, 307.45, 780.55	Circadian rhythm sleep disorders	6	0.1	15	0.2	4	0.5	.004	.081
347.00, 347.10	Narcolepsy	0	0.0	1	0.0	1	0.1	.994	.995
307.43, 327.10, 327.11, 327.12, 327.14, 780.54	Hypersomnia	6	0.1	11	0.1	1	0.1	.660	.294
327.40, 327.41, 327.42, 327.43, 327.44, 327.49	Parasomnias	17	0.2	28	0.3	2	0.3	.865	.166

ICD-10-CM = International Classification of Diseases, 10th Revision.

CMCs across HU groups

Asthma/respiratory disorders were observed as the most common CMCs across all HU groups (Table 3), showing no significant variations among utilization groups. However, the High-HU group exhibited significantly higher percentages of developmental disorders (28.6%), premature birth (9.9%), neurological disorders (34.1%), and mood disorders (12.8%) in comparison to the Medium-HU (23.9%, 6.2%, 27%, and 11.3%, respectively) and Low-HU (20.3%, 1.9%, 22.2%, and 9.1%, respectively) groups.

Table 3.

Distribution of chronic medical conditions across health care utilization groups.

Chronic Medical Condition	Health Care Utilization Groups						Comparisons, P
	Low (n = 7,486)		Medium (n = 8,063)		High (n = 776)		High vs Low	Medium vs Low
	n	%	n	%	n	%
Asthma/respiratory disorders	5,950	79.5	6,573	81.5	612	78.9	.686	.001
Developmental disorders	1,523	20.3	1,925	23.9	222	28.6	< .001	< .001
ADHD	574	7.7	679	8.4	68	8.8	.278	.085
Diabetes/metabolic disorders	1,725	23.0	1,672	20.7	167	21.5	.337	< .001
Overweight/obesity	1,584	21.2	1,576	19.5	155	20.0	.441	.013
Prematurity	144	1.9	499	6.2	77	9.9	< .001	< .001
Neurologic disorders	1,664	22.2	2,180	27.0	265	34.1	< .001	< .001
Mood disorders	678	9.1	912	11.3	99	12.8	< .001	< .001
Tonsillar hypertrophy	253	3.4	400	5.0	65	8.4	< .001	< .001

ADHD = attention-deficit/hyperactivity disorder.

Sleep disorders and risk of increased HU

Figure 1 indicates that children diagnosed with a sleep disorder had an 80% increased risk (odds ratio [OR] = 1.83; 95% confidence interval [CI]: 1.67–2.01) of being in an increased HU group, compared with those without a sleep disorder, even after controlling for confounding factors such as age, race/ethnicity, and CMCs. The presence of circadian rhythm sleep disorder was linked most strongly to the risk of being in an increased HU group (OR = 2.45; 95% CI: 1.07–5.64) compared with those without it. SDB (OR = 1.51; 95% CI: 1.17–1.95) and insomnia (OR = 1.46; 95% CI: 1.06–2.02) were associated with an approximately 50% increase in increased HU risk. Other sleep disorders were not associated with the risk of increased HU. The small sample size for narcolepsy made it difficult to interpret the results of that analysis.

Figure 1. Odds ratios: risk of higher health care utilization by sleep disorders.

Any sleep disorder = cumulative logistic regression of utilization risk with any sleep disorder and age, race, and chronic medical conditions as covariates. Specific sleep disorders = cumulative logistic regression OR with each sleep disorder and age, race, and chronic medical conditions as covariates. ORs and 95% LCLs and UCL for age, race, and chronic medical conditions not shown. LCL = lower confidence limit, OR = odds ratio, UCL = upper confidence limit.

Demographic factors and risk of increased HU

We also examined the risks of increased HU for each CMC and demographic factors after controlling for confounding factors (Figure 2). Children with Black race and Hispanic ethnicity were associated with an almost 50% lower risk of increased HU compared with White children (Black: OR = 0.58; 95% CI: 0.46–0.73; Hispanic/Latinx: OR = 0.57; 95% CI: 0.45–0.72). The risk of increased HU decreased as age increased (OR = 0.94; 95% CI: 0.94–0.95).

Figure 2. Risk of higher health care utilization by CMCs.

Cumulative logistic regression of utilization risk with each CMC and age, race, and any sleep disorder as covariates. ORs and 95% LCLs and UCL for age, race, and any sleep disorder not shown. ADHD = attention-deficit/hyperactivity disorder, CMC = chronic medical condition, LCL = lower confidence limit, OR = odds ratio, UCL = upper confidence limit.

CMCs and risk of increased HU

After controlling for other CMCs, sleep disorders, and demographic factors, children with asthma/respiratory disorders had an over 50% increased risk of being in the higher HU group (OR = 1.53; 95% CI: 1.40–1.67). The HU risks were doubled for prematurity (OR = 2.31; 95% CI: 1.94–2.76) and neurologic disorders (OR = 2.28; 95% CI: 1.87–2.78) and increased by 75% for mood disorders (OR = 1.75; 95% CI: 1.57–1.95). Interestingly, developmental disorders (OR = 0.49; 95% CI: 0.40–0.60) were associated with a decreased risk of HU. Furthermore, obesity and diabetes/metabolic disorders were not associated with increased HU risk.

DISCUSSION

This study highlights the role of sleep disorders in increased HU measured as the number of ED visits and/or hospitalizations in children with CMCs. The HU in the cohort of this study, mostly composed of children with multiple CMCs, is likely markedly higher than that of the general pediatric population, emphasizing the relevance of these findings. This study shows that the presence of sleep disorders nearly doubles the odds of increased HU in this population even after controlling for comorbid medical conditions and demographics. Children with sleep disorders were more likely to have a greater number of ED visits and/or hospitalizations during a 1-year pre-enrollment period compared with those without sleep disorders. This study also unveils variations in the risk of increased HU across specific sleep disorders and identifies the specific sleep disorders associated with heightened HU.

Circadian rhythm sleep disorders were associated with a more than 2-fold increased odds of a higher number of ED visits or hospitalizations compared with the rest of the cohort, even after controlling for other medical conditions, demographics, and other sleep disorders. This could be attributed to disrupted sleep patterns, impaired daytime functioning, and metabolic dysfunction associated with circadian rhythm misalignment, and the presence of comorbidities such as ADHD and other neuropsychiatric disorders in these children.^25,26

The observed heightened risk of increased HU associated with SDB aligns with prior research conducted in the general pediatric population.²⁷ OSA can lead to sleep fragmentation, increased inflammatory cytokines, sympathetic activation, and hypoxic burden, all of which can exacerbate symptoms of chronic illnesses and increase the need for health care services.^15,28 Similarly, insomnia can lead to sleep deprivation, daytime sleepiness, sympathetic activation, and neurocognitive impairment, all of which can impact the management of CMCs and increase HU.²⁸

The low prevalence of sleep disorders such as circadian rhythm sleep disorders, insomnia, sleep-related movement disorders, or hypersomnia in this cohort, aligning with widely acknowledged underdiagnosis of sleep disorders,^17,18 should be considered when assessing the clinical implications of these findings.^27,28

The notable finding of the High-HU group being significantly younger compared with the lower HU groups suggests that the presence of more specialized health care needs in younger children, such as those with congenital anomalies and prematurity, likely contributes to this discrepancy.⁴ Additionally, the lower parental threshold to seek medical care for their young children likely plays a role in the heightened utilization of health care services within this group.

In addition to the primary aim of investigating the relationship between sleep disorders and HU, this study also examined the role of demographics and CMCs. The finding that ethnic minority children in our cohort, where asthma was the predominant CMC, exhibit a reduced risk of heightened HU, as measured by the number of ED visits and hospitalizations, compared with White children is not consistent with earlier data indicating that Black children with asthma are more prone to ED visits than White children.^29–31 The role of racial disparities in sleep-related health care–seeking behaviors among caregivers needs to be examined as data suggest that, despite worse sleep quality and higher rates of sleep disorders, parents and caregivers in ethnic minority groups may be less likely to seek medical attention for their children.^32–34 Also, it is important to note that this discrepancy may be influenced by the minimal representation of White children in this cohort.

Developmental disorders were associated with lower odds of increased HU in this study. We speculate that the involvement of early intervention visits likely contributes to early detection and timely interventions, mitigating the development or progression of other medical conditions in children with developmental disorders. Also, outpatient care, which is likely a bigger component of HU in children with developmental disorders was not measured as an HU metric in our study.

Unlike the findings of the other studies, obesity was not associated with increased HU risk in this study.^35,36 This may suggest that obesity itself is not associated with ED visits or hospitalizations. Outpatient care is likely a major component of obesity-related HU; however, it was not measured as a metric of HU in this study.

Strengths and limitations

The study’s strengths include its large cohort size of over 16,000 children with CMCs and distinct HU groups, allowing for a comprehensive evaluation of the impact of sleep disorders. The study has limitations worth noting. First, only Medicaid claims data for ICD-9 and ICD-10-CM codes for sleep disorders and CMCs were used in the analysis and no attempt to validate this information with chart reviews was made. Also, as mentioned previously, HU was limited to ED visits and hospitalizations. Although those metrics contribute significantly to health care use, they do not fully capture the health care burden in children with CMCs. In addition, the suspected underdiagnosis of sleep disorders in the general pediatric population, as well as in children with CMCs, may have attenuated the measured association between sleep disorders and HU.^17,18 Finally, this cohort is predominantly urban, with most participants from ethnic minority groups with low socioeconomic status. Thus, it is difficult to generalize the findings to the broader population.

CONCLUSIONS

Based on an analysis of Medicaid claims data from a substantial cohort of children with CMCs, this study provides evidence of an association between the presence of sleep disorders and increased HU, even after controlling for comorbidities and demographics. Additionally, the study demonstrates that this association varies among different sleep disorders. While further examination of this association is needed, these findings offer yet another reason to screen children with CMCs for sleep disorders.

DISCLOSURE STATEMENT

All authors have seen and approved the manuscript. Work for this study was performed at the University of Illinois at Chicago. The authors report no conflicts of interest.

ABBREVIATIONS

ADHD

attention-deficit/hyperactivity disorder

CHECK

Coordinated Health Care for Complex Kids

CI

confidence interval

CMC

chronic medical condition

ED

emergency department

HU

health care utilization

ICD

International Classification of Diseases

OR

odds ratio

OSA

obstructive sleep apnea

SDB

sleep-disordered breathing