Abstract
BACKGROUND AND OBJECTIVES
Children with medical complexity (CMC) may be at a high risk of neurodevelopmental and mental health conditions given disease comorbidities and lived experiences. Little is known about the prevalence of these conditions at a population level. In this study, we estimated the prevalence of neurodevelopmental and mental health diagnoses in CMC relative to children without medical complexity and measured associations between these diagnoses in CMC and subsequent health care utilization and in-hospital mortality.
METHODS
We applied the Child and Adolescent Mental Health Disorders Classification System to identify neurodevelopmental and mental health diagnoses using all-payer claims data from three states (2012–2017). Poisson regression was used to compare outcomes in CMC with neurodevelopmental and mental health diagnoses to CMC without these diagnoses, adjusting for sociodemographic and clinical characteristics.
RESULTS
Among 85 581 CMC, 39 065 (45.6%) had ≥1 neurodevelopmental diagnoses, and 31 703 (37.0%) had ≥1 mental health diagnoses, reflecting adjusted relative risks of 3.46 (3.42–3.50) for neurodevelopmental diagnoses and 2.22 (2.19–2.24) for mental health diagnoses compared with children without medical complexity. CMC with both neurodevelopmental and mental health diagnoses had 3.00 (95% confidence interval [CI]: 2.98–3.01) times the number of ambulatory visits, 69% more emergency department visits (rate ratio = 1.69, 95% CI: 1.66–1.72), 58% greater risk of hospitalization (rate ratio = 1.58, 95% CI: 1.50–1.67), and 2.32 times (95% CI: 2.28–2.36) the number of hospital days than CMC without these diagnoses.
CONCLUSIONS
Neurodevelopmental and mental health diagnoses are prevalent among CMC and associated with increased health care utilization across the continuum of care. These findings illustrate the importance of recognizing and treating neurodevelopmental and mental health conditions in this population.
What’s Known on This Subject:
Children with medical complexity may have increased risks of neurodevelopmental and other mental health conditions given their complex health care needs and lived experiences; however, the prevalence of these conditions at a population level is not well understood.
What This Study Adds:
Forty-six percent of children with medical complexity had neurodevelopmental diagnoses and 37% had mental health diagnoses, more than twice the prevalence observed in children without medical complexity. These diagnoses were associated with significantly increased health care use across care settings.
Children with medical complexity (CMC) comprise a growing population of children in the United States with complex chronic medical conditions, high levels of health system use, and increased risks of care fragmentation, hospitalization, and childhood mortality.1–5 Chronic disease comorbidities, as well as disease symptoms, treatments, and other lived experiences, may bring substantial stress to CMC and their families.6–8 Correspondingly, CMC may be at significantly increased risk of neurodevelopmental and mental health conditions.
Approximately 1 in 5 children in the United States have a neurodevelopmental or mental health condition.9,10 However, few analyses have focused specifically on CMC. Most studies conducted to date have examined individual chronic disease diagnoses or have relied on survey data, which may be limited by self-report bias and small sample sizes. These studies suggest significantly increased rates of autism, attention-deficit hyperactivity disorder, depression, and anxiety in several CMC subgroups compared with children without chronic conditions.11–17 However, the authors of a 2019 systematic review noted that research on mental health outcomes in CMC was “strikingly lacking.”18
The analysis of all-payer claims data (APCD), which links health care claims across payers and settings, presents a unique opportunity to study neurodevelopmental and mental health diagnoses in a large cohort of CMC with diverse conditions. As such, in this study, we used APCD to estimate the prevalence of neurodevelopmental and mental health diagnoses in CMC compared with children without medical complexity and investigate associations between these diagnoses in CMC and subsequent health system utilization and in-hospital mortality. Understanding the prevalence of neurodevelopmental and mental health conditions and differences in health service use can inform policies and programs to support CMC and promote equitable care for this population.
Methods
Study Design and Data Sources
In this retrospective cohort study, we used APCD from Colorado, New Hampshire, and Massachusetts, which included data from Medicaid and employee-sponsored commercial plans and linked individuals within states across payers. These states were chosen to represent geographic diversity, with a particular focus on varied rural-urban composition. Each of the 3 data sets spanned a 5-year period (January 1, 2013–December 31, 2017 for New Hampshire and Massachusetts and October 1, 2012–September 30, 2017 for Colorado), with files indicating the duration of enrollment in a participating health care plan, as well as professional and facility claims. The acquisition, storage, and analysis of the APCD was guided by data use agreements with the Center for Improving Value in Health Care in Colorado, the New Hampshire Comprehensive Health Care Information System and Department of Health and Human Services, and the Massachusetts Center for Health Information and Analysis. The Dartmouth Hitchcock Institutional Review Board deemed this study exempt from further review and informed consent.
Study Population
In this study, we evaluated neurodevelopmental and mental health conditions in CMC and in children not meeting CMC diagnostic criteria as defined below (hereafter called the reference population). We used the first 3 years of each APCD data set to identify CMC and estimate the prevalence of neurodevelopmental and mental health conditions and the subsequent 2 years to examine health care utilization and in-hospital mortality. Eligible children were limited to residents of participating states and those who were 3 to 17 years of age during the 3-year diagnosis ascertainment period (n = 1 984 200 beneficiaries). We excluded children with zip codes that could not be linked to a rural-urban commuting area19 code (n = 420), children with <12 months of visibility in the diagnosis ascertainment period (n = 776 196), and those with <12 observation months in the 2-year outcome period unless death occurred (n = 257 552). From this cohort of study-eligible children, we identified CMC by applying 2 open-source algorithms that use International Classification of Diseases, Ninth or 10th revision (ICD-9/10) codes: the Complex Chronic Conditions Classification System20 and the pediatric medical complexity algorithm (PMCA).21 These algorithms were applied concurrently by using a previously published conservative approach for claims data.1
We applied the Child and Adolescent Mental Health Disorders Classification System (CAMHD-CS)22 to identify and group neurodevelopmental and mental health diagnoses in CMC and the reference population. The CAMHD-CS maps ICD-9/10 codes to 30 diagnosis groups consistent with the Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition. Children were classified as having these diagnoses if they had eligible CAMDH-CS diagnostic codes on ≥2 separate dates. We categorized eligible diagnoses as neurodevelopmental conditions or mental health conditions (Supplemental Table 4), given differences in age of onset and health services needs associated with these 2 groups of conditions. It was therefore possible for children to have diagnoses in 0, 1, or both condition groups.
Sociodemographic and Clinical Characteristics
Sociodemographic characteristics included age in years at the beginning of the outcome period, sex (female, male, other/unknown), state of residence, rurality determined by linking residential zip codes to rural-urban commuting area codes,19 and primary payer. For CMC, additional characteristics derived from the PMCA and Complex Chronic Condition Classification System included the number of body systems with chronic disease diagnoses,23 a binary variable indicating technology assistance,1,20 and a binary indicator of a progressive complex disease diagnosis (eg, muscular dystrophy or cystic fibrosis).21
CMC Health Care Utilization and In-Hospital Mortality
During the 2-year outcome period, we examined differences in ambulatory clinic visits, emergency department (ED) visits, acute care hospitalizations and hospital days, and in-hospital mortality between CMC with and without neurodevelopmental and mental health conditions. Ambulatory clinic visits were defined as visits, procedures, or treatments in office or clinic settings with physicians, nurse practitioners, physician assistants, and behavioral health and social services providers, as well as at clinics with current procedural terminology codes for office or other outpatient evaluation and management services. ED visits were defined by using ED-specific revenue codes, current procedural terminology codes, and place of service codes limited to those that did not result in hospitalization on the same or subsequent day. Acute care hospitalizations included inpatient and observation stays. In-hospital mortality per 10 000 CMC was determined by using mortality discharge disposition codes and also required the termination of health care claims for in-person encounters subsequent to the hospitalization.
Statistical Analysis
We first examined differences in sociodemographic characteristics between children with and without neurodevelopmental conditions and with and without mental health conditions using descriptive statistics. We then calculated the prevalence of neurodevelopmental and mental health conditions in CMC and the reference population and estimated the relative risk of these diagnoses in CMC compared with the reference population using Poisson regression with robust error variance.24 Models were adjusted for age, sex, rurality, payer, state, and observation months during the 3-year diagnosis ascertainment period.
To investigate the differences between CMC with and without neurodevelopmental or mental health conditions in health care utilization and in-hospital mortality, we used Student’s t tests to examine unadjusted differences of means and χ2 tests for differences in proportions. Recognizing that health system use and mortality risk may differ for CMC with neurodevelopmental conditions alone, mental health conditions alone, or these conditions in combination, in our regression analyses, we estimated models using a 4-level primary predictor: (1) CMC without neurodevelopmental or other mental health conditions (reference), (2) CMC with neurodevelopmental conditions only, (3) CMC with mental health conditions only, and (4) CMC with both neurodevelopmental and mental health conditions. Models were adjusted for potential confounding on the basis of age, sex, payer, state of residence, rurality, body systems with chronic disease diagnoses, technology assistance, and progressive complex disease diagnosis. Health care utilization models were also adjusted for observation months during the outcome period. Counts of ambulatory care visits, ED visits, and acute care hospital days were modeled by using Poisson regression with a log-link function. The dichotomous outcomes, acute care hospitalization, and in-hospital mortality were modeled by using Poisson regression with robust error variance to account for model misspecification.24
Given that some CMC may have met PMCA diagnostic criteria on the basis of having a chronic mental health diagnosis, we performed a sensitivity analysis in which these individuals were excluded. All analyses described above were repeated by using the restricted cohort. Analyses were performed by using SAS software, Version 9.4; all statistical testing was 2-sided with an α of 0.05.
Results
Neurodevelopmental and Mental Health Diagnoses
Among 85 581 CMC meeting eligibility criteria, 39 065 (45.6%) had 1 or more neurodevelopmental diagnoses, and 31 703 (37.0%) had 1 or more mental health diagnoses. CMC with neurodevelopmental diagnoses were most frequently male, insured only by Medicaid, and urban residing (Table 1). They were also more likely to be assisted by technology and have ≥2 body systems affected by chronic diseases than CMC without neurodevelopmental diagnoses. Similar patterns were observed in CMC with mental health diagnoses, although CMC with mental health diagnoses were more likely to be 12 to 17 years of age and less likely to be assisted by technology. Demographic and clinical characteristics of the sample based on 4 mutually exclusive groups (CMC with neurodevelopmental conditions only, mental health conditions only, both neurodevelopmental and mental health conditions, and neither) are provided in Supplemental Table 5.
TABLE 1.
Characteristics of CMC With and Without Diagnoses of Neurodevelopmental Conditions and Other Mental Health Conditions
| CMC Characteristics | Full Cohort | Neurodevelopmental Conditions | Other Mental Health Conditions | |||||||
|---|---|---|---|---|---|---|---|---|---|---|
| + Diagnosis | No Diagnosis | + Diagnosis | No Diagnosis | |||||||
| (n = 85 581) | (n = 39 065, 45.6%) | (n = 46 516, 54.4%) | (n = 31 703, 37.0%) | (n = 53 878, 63.0%) | ||||||
| n % | n % | n % | n % | n % | ||||||
| Age | ||||||||||
| 3–7 y | 21 215 | 24.8% | 11 024 | 28.2% | 10 191 | 21.9% | 5259 | 16.6% | 15 956 | 29.6% |
| 8–11 y | 24 008 | 28.1% | 12 808 | 32.8% | 11 200 | 24.1% | 9420 | 29.7% | 14 588 | 27.1% |
| 12–17 y | 40 358 | 47.2% | 15 233 | 39.0% | 25 125 | 54.0% | 17 024 | 53.7% | 23 334 | 43.3% |
| Sexa | ||||||||||
| Female | 38 664 | 45.2% | 12 907 | 33.0% | 25 757 | 55.4% | 13 819 | 43.6% | 24 845 | 46.1% |
| Male | 46 829 | 54.7% | 26 134 | 66.9% | 20 695 | 44.5% | 17 864 | 56.3% | 28 965 | 53.8% |
| No. body systems affected by chronic diseases | ||||||||||
| 1b | 23 828 | 27.8% | 1892 | 4.8% | 21 936 | 47.2% | 3745 | 11.8% | 20 083 | 37.3% |
| 2 | 43 666 | 51.0% | 24 855 | 63.6% | 18 811 | 40.4% | 19 573 | 61.7% | 24 093 | 44.7% |
| 3 | 11 068 | 12.9% | 7073 | 18.1% | 3995 | 8.6% | 5167 | 16.3% | 5901 | 11.0% |
| ≥4 | 7019 | 8.2% | 5245 | 13.4% | 1774 | 3.8% | 3218 | 10.2% | 3801 | 7.1% |
| Technology assistancec | 4196 | 4.9% | 2544 | 6.5% | 1652 | 3.6% | 1407 | 4.4% | 2789 | 5.2% |
| Progressive complex conditiond | 24 901 | 29.1% | 10 125 | 25.9% | 14 776 | 31.8% | 8245 | 26.0% | 16 656 | 30.9% |
| Primary payer | ||||||||||
| Medicaid only | 48 391 | 56.5% | 25 620 | 65.6% | 22 771 | 49.0% | 20 053 | 63.3% | 28 338 | 52.6% |
| Commercial only | 27 888 | 32.6% | 8966 | 23.0% | 18 922 | 40.7% | 8136 | 25.7% | 19 752 | 36.7% |
| Both Medicaid and Commercial | 9302 | 10.9% | 4479 | 11.5% | 4823 | 10.4% | 3514 | 11.1% | 5788 | 10.7% |
| Rurality | ||||||||||
| Urban-residing | 80 181 | 93.7% | 36 809 | 94.2% | 43 372 | 93.2% | 29 789 | 94.0% | 50 392 | 93.5% |
| Rural-residing | 5400 | 6.3% | 2256 | 5.8% | 3144 | 6.8% | 1914 | 6.0% | 3486 | 6.5% |
| State | ||||||||||
| Colorado | 17 524 | 20.5% | 7839 | 20.1% | 9685 | 20.8% | 5464 | 17.2% | 12 060 | 22.4% |
| Massachusetts | 61 737 | 72.1% | 28 837 | 73.8% | 32 900 | 70.7% | 23 884 | 75.3% | 37 853 | 70.3% |
| New Hampshire | 6320 | 7.4% | 2389 | 6.1% | 3931 | 8.4% | 2355 | 7.4% | 3965 | 7.4% |
| Observation mo, cohort creation period, mean (SD) | 33.5 (5.1) | 34.0 (4.4) | 33.0 (5.5) | 33.9 (4.5) | 33.2 (5.3) | |||||
| Observation mo, outcome period, mean (SD) | 21.9 (3.5) | 22.4 (3.1) | 21.5 (3.7) | 22.1 (3.6) | 21.8 (3.5) | |||||
Sex missing or other for n = 88 (<1%) CMC.
Includes 1008 CMC with only technology assistance codes who do not have another chronic disease diagnosis.
Determined using the Complex Chronic Conditions Classification System.
Determined using the PMCA.
Among children without medical complexity, 100 849 (11.7%) had 1 or more neurodevelopmental conditions, and 126 046 (14.6%) had 1 or more mental health conditions. In this reference population, 71 162 (70.6%) of children with neurodevelopmental conditions were male (compared with 370 154 [48.4%] of children without these conditions), and 60 187 (59.7%) were Medicaid-insured (compared with 390 760 [51.1%] of children without these conditions, Supplemental Table 6). Children with mental health conditions were most often 12 to 17 years of age (n = 63 338, 50.3%), male (n = 67 076, 53.2%), and insured exclusively by Medicaid (n = 75 482, 59.9%).
In Table 2, we summarize the frequency and prevalence of neurodevelopmental and mental health conditions in CMC and the reference population. Overall, 21.0% (n = 181 709) of children in the reference population had 1 or more neurodevelopmental or mental health conditions, compared with 59.4% (n = 50 848) of CMC. Adjusting for baseline sociodemographic characteristics, CMC had 2.51 times the risk of these conditions overall (rate ratio [RR] = 2.51, 95% confidence interval [CI]: 2.48–2.53). The adjusted risk of neurodevelopmental conditions specifically was 3.46 times greater in CMC than children in the reference population (95% CI: 3.42–3.50), and attention-deficit hyperactivity disorder and developmental delay or unspecified neurodevelopmental disorder were the most common diagnoses. The most substantive differences in risk between CMC and the reference population were observed for developmental delay or unspecified neurodevelopmental disorders (RR = 8.42, 95% CI: 8.22–8.62) and intellectual disability (RR = 7.25, 95% CI: 7.01–7.50).
TABLE 2.
Frequency of Neurodevelopmental and Other Mental Health Diagnoses in CMC and in the Reference Population of Children Without Medical Complexity
| Non-CMC, n = 865 451 | CMC, n = 85 581 | Risk Ratioa (95% CI) | |
|---|---|---|---|
| At least 1 neurodevelopmental or other mental health condition | 181 709 (21.0%) | 50 848 (59.4%) | 2.51 (2.48–2.53) |
| ≥1 Neurodevelopmental conditions | 100 849 (11.7%) | 39 065 (45.6%) | 3.46 (3.42–3.50) |
| Attention-deficit hyperactivity disorder | 66 973 (7.0%) | 20 171 (23.6%) | 2.63 (2.58–2.67) |
| Developmental delay or unspecified neurodevelopmental disorder | 14 132 (1.6%) | 13 534 (15.8%) | 8.42 (8.22–8.62) |
| Specific learning disorders | 12 697 (1.5%) | 8793 (10.3%) | 6.61 (6.43–6.80) |
| Autism spectrum disorder | 13 389 (1.5%) | 8045 (9.4%) | 5.46 (5.31–5.61) |
| Communication disorders | 9325 (1.1%) | 4957 (5.8%) | 5.36 (5.17–5.55) |
| Intellectual disability | 7061 (0.8%) | 6281 (7.3%) | 7.25 (7.01–7.50) |
| Motor disorders | 4062 (0.5%) | 2624 (3.1%) | 6.38 (6.07–6.71) |
| Neurocognitive disorders | 2929 (0.3%) | 2206 (2.6%) | 6.46 (6.11–6.83) |
| ≥1 Mental health conditions | 126 046 (14.6%) | 31 703 (37.0%) | 2.22 (2.19–2.24) |
| Trauma and stressor-related disorders | 59 915 (6.9%) | 12 297 (14.4%) | 1.80 (1.77–1.84) |
| Anxiety disorders | 41 238 (4.8%) | 11 763 (13.7%) | 2.52 (2.47–2.57) |
| Disruptive, impulse control, and conduct disorders | 27 051 (3.1%) | 9334 (10.9%) | 2.95 (2.88–3.02) |
| Depressive disorders | 23 808 (2.8%) | 7825 (9.1%) | 2.54 (2.47–2.60) |
| Miscellaneousb | 18 744 (2.2%) | 7417 (8.7%) | 3.19 (3.11–3.28) |
| Mental health symptomsc | 9557 (1.1%) | 4396 (5.1%) | 4.04 (3.90–4.19) |
| Elimination disorders | 4240 (0.5%) | 2161 (2.5%) | 4.94 (4.69–5.20) |
| Bipolar and related disorders | 2856 (0.3%) | 1445 (1.7%) | 3.85 (3.61–4.10) |
| Suicide or self-injury | 2376 (0.3%) | 1153 (1.3%) | 3.63 (3.39–3.90) |
| Obsessive–compulsive and related disorders | 2277 (0.3%) | 969 (1.1%) | 3.78 (3.51–4.08) |
| Schizophrenia spectrum and other psychotic disorders | 2031 (0.2%) | 1356 (1.6%) | 5.06 (4.72–5.42) |
| Feeding and eating disorders | 1023 (0.1%) | 1280 (1.5%) | 11.12 (10.23–12.09) |
| Somatic symptom and related disorders | 660 (0.1%) | 819 (1.0%) | 10.83 (9.76–12.01) |
| Otherd | 2665 (0.3%) | 1351 (1.6%) | 4.44 (4.15–4.75) |
| No. unique neurodevelopmental disorders and/or mental health conditions | |||
| 0 | 683 726 (79.0%) | 34 701 (40.6%) | 0.50 (0.49–0.50) |
| 1 | 104 097 (12.0%) | 18 075 (21.1%) | 1.59 (1.57–1.62) |
| 2 | 41 398 (4.8%) | 12 803 (15.0%) | 2.76 (2.70–2.81) |
| 3 | 18 846 (2.2%) | 8310 (9.7%) | 3.84 (3.74–3.94) |
| ≥4 | 17 384 (2.0%) | 11 692 (13.7%) | 5.64 (5.50–5.77) |
All groups of neurodevelopmental and mental health diagnoses are based on the CAMHD-CS. A child can have zero or multiple diagnoses.
Adjusted for age, sex, payer, rurality, state, and months of health care enrollment or claims in the diagnosis ascertainment period.
Most common ICD-9/10 codes included within the miscellaneous group are unspecified episodic mood disorder (5224 [46.4%] CMC; 12 750 [68.1%] in the reference population), unspecified nonpsychotic mental disorder (1889 [25.5%] CMC; 3523, 18.8% in the reference population), and other emotional disturbances of childhood and adolescence (970 [8.6%] CMC; 2601 [13.9%] in the reference population), accounting for 6069 (81.8%) of CMC and 15 187, 81.0% of children in the reference population included in this category.
Most common ICD-9/10 codes included within the mental health symptoms are unspecified mental or behavioral problem (1635 [37.2%] CMC; 4559 [47.7%] in the reference population), other and unspecified special symptoms or syndromes, not elsewhere classified (1497 [34.1%] CMC; 2719 [28.5%] in the reference population), and attention or concentration deficit (549 [12.5%] CMC; 1444 [15.1%] in the reference population), accounting for 3289 (74.8%) of CMC and 7939 (83.1%) of children in the reference population.
Other includes the following groups, combined because of small cell sizes: fetal or newborn damage related to maternal substance abuse, maternal mental illness or substance abuse during pregnancy, delivery, or postpartum, personality disorders, sexuality and gender identity disorders, sleep–wake disorders, substance abuse-related medical illness, substance-related and addictive disorders, and accidental or undetermined poisoning.
In adjusted regression analysis, CMC were 2.22 (95% CI: 2.19–2.24) times more likely to be diagnosed with a mental health condition than children in the reference population. The most common mental health conditions for both CMC and children without medical complexity were trauma and stressor-related disorders (n = 59 915, 6.9% in the reference population and n = 12 297, 14.4% of CMC) and anxiety disorders (n = 41 238, 4.8% in the reference population and n = 11 763, 13.7% of CMC). The largest risk ratios observed in CMC compared with the reference population were for feeding and eating disorders (RR = 11.12, 95% CI: 10.23–12.09) and somatic symptoms and related disorders (RR = 10.83, 95% CI: 9.76–12.01). In total, 13.7% (n = 11 692) of CMC had 4 or more neurodevelopmental or mental health conditions, compared with 2.0% (n = 17 384) of children in the reference population, reflecting a 5.64 times (95% CI = 5.50–5.77) greater adjusted risk.
Figure 1 reveals prevalence estimates for neurodevelopmental and mental health conditions in CMC, alone and in combination, based on demographic and clinical characteristics. The lowest prevalence of neurodevelopmental and mental health conditions was observed in CMC with a chronic disease affecting just 1 body system; 79.3% (18 897) of CMC in this group did not have any neurodevelopmental or mental health diagnoses. In contrast, only 15.2% (n = 1067) of CMC with chronic diseases affecting 4 or more body systems did not have any neurodevelopmental or mental health diagnoses. In this group, 35.8% (n = 2511) had both neurodevelopmental and mental health conditions, 10.1% (n = 707) had mental health diagnoses alone, and 39.0% (n = 2734) had neurodevelopmental diagnoses only.
FIGURE 1.
Percentage of CMC with cooccurring diagnoses of neurodevelopmental conditions and other mental health conditions by demographic and clinical characteristics. a Includes 1008 CMC with only technology assistance codes; b n = 706; c n = 1186.
CMC Health Care Utilization and In-Hospital Mortality
On average, CMC with neurodevelopmental diagnoses had more ambulatory clinic visits, more ED visits, more hospitalizations, and more days in hospital during the 2-year outcome period than CMC without these diagnoses (Table 3). One of the largest differences was in the average number of clinic visits; CMC with neurodevelopmental diagnoses had a mean of 28.7 (95% CI: 28.2–29.2) clinic visits over 2 years, compared with 15.4 (95% CI: 15.2–15.6) in CMC without these diagnoses. The in-hospital mortality rate in CMC with neurodevelopmental conditions was also significantly higher than that observed in CMC without these diagnoses: 69 deaths per 10 000 CMC versus 41 deaths per 10 000 CMC (P < .001). Similar patterns were observed when comparing health care utilization in CMC with and without mental health diagnoses, although there were no observable differences in mortality risk. CMC with mental health diagnoses had an average of 34.3 (95% CI: 33.7–34.9) clinic visits over 2 years compared with 13.9 (95% CI: 13.7–14.9) visits in those without these diagnoses.
TABLE 3.
Health System Utilization and In-Hospital Mortality During the 2-Year Outcome Period in CMC With and Without Cooccurring Diagnoses of Neurodevelopmental Condition and Other Mental Health Conditions
| Full Cohort | Neurodevelopmental Conditions | Other Mental Health Conditions | |||||
|---|---|---|---|---|---|---|---|
| (n = 85 581) | + Diagnosis (n = 39 065) | No Diagnosis (n = 46 516) | P a | + Diagnosis (n = 31 703) | No Diagnosis (n = 53 878) | P a | |
| Ambulatory clinic visits, mean (95% CI) | |||||||
| Total no. clinic visits per CMC | 21.4 (21.1–21.7) | 28.7 (28.2–29.2) | 15.4 (15.2–15.6) | <.001 | 34.3 (33.7–34.9) | 13.9 (13.7–14.1) | <.001 |
| No. clinic visits with a principal diagnosis of a neurodevelopmental condition | 4.3 (4.1–4.5) | 8.8 (8.5–9.1) | 0.6 (0.5–0.7) | <.001 | 8.0 (7.7–8.3) | 2.1 (2.0–2.2) | <.001 |
| No. clinic visits with a principal diagnosis of a mental health condition | 6.7 (6.5–6.9) | 8.9 (8.6–9.2) | 4.9 (4.7–5.1) | <.001 | 15.0 (14.5–15.5) | 1.9 (1.8–2.0) | <.001 |
| No. clinic visits without a principal diagnosis of a neurodevelopmental or mental health condition | 10.4 (10.3–10.5) | 11.0 (10.9–11.1) | 9.9 (9.8–10) | <.001 | 11.3 (11.2–11.4) | 9.9 (9.8–10.0) | <.001 |
| ED visits | |||||||
| CMC with ≥1 ED visit, n (%) | 36 883 (43.1%) | 18 718 (47.9%) | 18 165 (39.1%) | <.001 | 16 219 (51.2%) | 20 664 (38.4%) | <.001 |
| No. ED visits per CMC, mean (95% CI) | 2.4 (2.4–2.4) | 2.5 (2.5–2.5) | 2.2 (2.2–2.3) | <.001 | 2.7 (2.7–2.8) | 2.1 (2.1–2.1) | <.001 |
| ED visits with a principal diagnosis of a neurodevelopmental condition, n (%) | 1638 (1.9%) | 1517 (3.2%) | 121 (0.3%) | <.001 | 1081 (2.4%) | 557 (1.3%) | <.001 |
| ED visits with a principal diagnosis of a mental health condition, n (%) | 6588 (7.5%) | 4353 (9.3%) | 2235 (5.5%) | 5521 (12.4%) | 1067 (2.5%) | ||
| ED visits without a principal diagnosis of a neurodevelopmental or mental health condition, n (%) | 79 417 (90.6%) | 40 959 (87.5%) | 38 458 (94.2%) | 37 826 (85.1%) | 41 591 (96.2%) | ||
| Acute care hospitalizations | |||||||
| CMC with ≥1 acute care hospitalization, n (%) | 9416 (11.0%) | 5012 (12.8%) | 4404 (9.5%) | <.001 | 4500 (14.2%) | 4916 (9.1%) | <.001 |
| No. d in hospital per CMC, mean (95% CI) | 14.5 (13.6–15.4) | 18.1 (16.5–19.7) | 10.5 (9.6–11.4) | <.001 | 19.2 (17.6–20.8) | 10.2 (9.2–11.2) | <.001 |
| Hospitalizations with a principal diagnosis of a neurodevelopmental condition, n (%) | 1130 (2.7%) | 1078 (4.6%) | 52 (0.3%) | <.001 | 702 (3.5%) | 428 (1.9%) | <.001 |
| Hospitalizations with a principal diagnosis of a mental health condition, n (%) | 4064 (9.5%) | 2700 (11.6%) | 1364 (7.0%) | 3724 (18.3%) | 340 (1.5%) | ||
| Hospitalizations without a principal diagnosis of a neurodevelopmental or mental health condition, n (%) | 37 503 (87.8%) | 19 434 (83.7%) | 18 069 (92.7%) | 15 849 (78.2%) | 21 609 (96.6%) | ||
| In-hospital mortality, n (deaths/10 000 CMC) | 110 (12.9) | 69 (17.7) | 41 (8.8) | <.001 | 42 (13.2) | 68 (12.6) | .80 |
Derived from t tests to examine unadjusted differences of means and χ2 tests for differences in proportions.
In adjusted regression analyses, CMC with neurodevelopmental conditions or mental health conditions alone or in combination experienced significantly higher health care utilization across all measures (Fig 2). Compared with CMC without neurodevelopmental or mental health conditions, CMC with both condition types had 3.00 (95% CI: 2.98–3.01) times the number of ambulatory visits, 69% more ED visits (RR = 1.69, 95% CI: 1.66–1.72), 1.58 times the risk of being hospitalized (95% CI: 1.50–1.67), and 2.32 times the number of days in hospital (95% CI: 2.28–2.36). In adjusted analyses, the risk of in-hospital mortality in CMC with both neurodevelopmental and mental health conditions was not significantly different from the CMC without these conditions (RR = 1.09, 95% CI: 0.60–1.98).
FIGURE 2.
Regression analyses demonstrating the likelihood of health care utilization and in-hospital mortality in CMC with cooccurring diagnoses of neurodevelopmental and other mental health conditions, adjusted for demographic and clinical characteristics. Counts of ambulatory care visits, ED visits, and acute care hospital days were modeled by using Poisson regression with a log-link function. The dichotomous outcomes (acute care hospitalization and in-hospital mortality) were modeled by using Poisson regression with robust error variance to account for model misspecification. Models were adjusted for age, sex, payer, state of residence, rurality, body systems with chronic disease diagnoses, technology assistance, and progressive complex disease diagnosis. Health care utilization models were also adjusted for observation months during the outcome period. a In the model for in-hospital mortality, diagnoses pertaining to the following body systems were combined into a single group to support stability in modeling a rare outcome: dermatologic, craniofacial, ophthalmologic, and ear, nose, and throat.
Sensitivity Analysis
In the sensitivity analysis restricted to children who qualified as CMC when chronic mental health diagnoses were excluded from the PMCA, 22 038 (25.8%) CMC were excluded (Supplemental Table 7). In this restricted cohort of 63 543 CMC, 46.6% (n = 29 582) had 1 or more neurodevelopmental or mental health diagnoses and the most common diagnoses were similar to those of the original cohort (Supplemental Table 7). The distribution of sociodemographic characteristics was also similar to the original cohort; however, there were differences in the number of body systems impacted by complexity, technology dependence, and the presence of a progressive complex condition (Supplemental Table 8). Likewise, the patterns of health care utilization remained consistent with the original cohort, with the mean volume of ambulatory clinic visits by CMC with neurodevelopmental (27.8, 95% CI: 27.2–28.4) or other mental health conditions (32.3, 95% CI: 31.6–33.0) being significantly higher compared with CMC without these conditions (Supplemental Table 9). Supplemental Fig 3 reveals adjusted regression analyses that are consistent with the main findings in the original cohort.
Discussion
In this cohort study representing children from 3 states, 45.6% of CMC had 1 or more neurodevelopmental conditions, and 37.0% had 1 or more mental health conditions, reflecting 3 times greater risk of neurodevelopmental diagnoses and 2 times greater risk of mental health diagnoses than observed in children without medical complexity. CMC with neurodevelopmental and mental health conditions used significantly more health care services across the continuum of care from ambulatory to inpatient care than CMC without these diagnoses, and CMC with both neurodevelopmental and mental health conditions experienced the highest utilization. These findings speak to the tremendous importance of the screening, diagnosis, and treatment of these conditions in CMC and can be used to advocate for resources to support their unique health care needs.
Limited research has been conducted to examine the prevalence of neurodevelopmental and mental health conditions in CMC to date. However, studies have reported an increased prevalence of these diagnoses in children with specific complex conditions, such as congenital heart disease, dystonia, and cancer,11–17 as well as in children with special health care needs or chronic physical conditions.25,26 In this study, we build on this evidence base, demonstrating a substantially increased risk of these conditions overall and also illustrating specific diagnoses disproportionately experienced by this population. The significantly increased risks of developmental delay (experienced with 8 times greater frequency in CMC than the reference population) and intellectual disability (experienced with 7 times greater frequency) may result from genetic, metabolic, or other conditions that include both physical and neurodevelopmental manifestations or may be secondary to medical treatments and disrupted routines and learning opportunities due to complex health care needs.27,28 The results of this study justify focused efforts to support consistent and timely neurodevelopmental screening for CMC, as well as efforts to integrate neurodevelopmental disorder prevention and treatment in their care plans.29
Among both CMC and children without medical complexity, trauma and stressor-related disorders and anxiety disorders were the most prevalent mental health conditions, with CMC experiencing 1.8 and 2.5 times the risk of these conditions compared with their peers without medical complexity. These findings are consistent with previous work by Zima et al, who found that trauma and stressor-related disorders were the most common mental health conditions experienced by a Medicaid-insured national sample.22 The twofold greater risk of these conditions in CMC is likely multifactorial; systematic reviews reveal that the diagnosis of a severe chronic condition may be traumatizing in and of itself, and pediatric medical traumatic stress may result from pain, injury, disease symptoms, and their treatments.30,31 The substantially increased risk of all mental health conditions observed in this analysis is particularly notable given established challenges in screening, diagnosing, and accessing mental health treatment for CMC.32–35
Several previous studies reveal patterns of increased health care use in children with neurodevelopmental or mental health conditions. For example, comorbid mental health conditions have been associated with longer lengths of stay for pediatric medical and surgical hospitalizations,36 a higher likelihood of recurrent hospitalizations for diabetic ketoacidosis,37 and higher overall health care costs for children with chronic physical health conditions.38 Building on this literature, we observed that CMC with neurodevelopmental or mental health conditions, either alone or in combination, had substantially increased health care utilization across care settings. Taken together with past research revealing that the most common reasons for outpatient visits among CMC are mental health-related,39 our findings speak to the importance of supporting research and public policy to develop and implement effective neurodevelopmental and mental health treatments for CMC.
The results of this study should be interpreted considering its strengths and limitations. Although the study cohort is large, it may not be nationally representative. For example, neurodevelopmental and mental health condition prevalence rates were highest in Massachusetts, which is a state with above-average mental health resources, as well as an established behavioral health screening policy.40,41 Only neurodevelopmental and mental health conditions that had been coded in health care claims were identifiable, perhaps underestimating disease prevalence. Relatedly, APCD only includes encounters processed by participating insurance plans (excluding encounters paid for out-of-pocket), and as a result, both disease prevalence and the number of health care encounters may have been underestimated. However, we are reassured that the overall prevalence of neurodevelopmental and mental health conditions in the reference population is consistent with previous national analyses.9,10 APCD lacks details that may be associated with the diagnosis of neurodevelopmental and mental health conditions, such as parental education and socioeconomic status, limiting our ability to identify factors contributing to these diagnoses. Finally, we acknowledge that the study years preceded the coronavirus disease 2019 pandemic and therefore do not reflect more recent potential changes in the prevalence of mental health conditions.42–45
Conclusions
More than 40% of CMC in our cohort had neurodevelopmental conditions, and more than one-third had other mental health conditions. CMC with these diagnoses experienced higher health care utilization across the continuum of care. These findings illustrate the importance of recognizing and treating neurodevelopmental and mental health conditions in this population. Investment in evidence-based strategies to prevent, diagnose, and treat mental health conditions in CMC is needed.
Supplementary Material
Glossary
- APCD
all-payer claims data
- CAMHD-CS
Child and Adolescent Mental Health Disorders Classification System
- CI
confidence interval
- CMC
children with medical complexity
- ED
emergency department
- ICD-9/10
International Classification of Diseases, Ninth or 10th revision
- PMCA
pediatric medical complexity algorithm
- RR
rate ratio
Footnotes
Ms Arakelyan and Dr Leyenaar conceptualized and designed the study and drafted the manuscript; Mr Freyleue and Dr Schaefer conceptualized and designed the study, conducted the analyses, and reviewed and revised the manuscript for critical content; Drs Ahuja, Austin, Chien, Goodman, Moen, O’Malley, Simon, and Van Cleave conceptualized and designed the study and reviewed and revised the manuscript for critical content; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
COMPANION PAPER: A companion to this article can be found online at www.pediatrics.org/cgi/doi/10.1542/peds.2024-067259.
FUNDING: Funded by the National Institutes of Health (NIH). Research reported in this publication was supported by the National Institute on Minority Health and Health Disparities of the NIH under award R01MD014735. The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIH. The funder did not play a role in the design and conduct of the study, the collection, management, analysis, and interpretation of the data, the preparation, review, or approval of the manuscript, or the decision to submit the manuscript for publication.
CONFLICT OF INTEREST DISCLOSURES: The authors have indicated they have no potential conflicts of interest to disclose.
References
- 1.Leyenaar JK, Schaefer AP, Freyleue SD, et al. Prevalence of children with medical complexity and associations with health care utilization and in-hospital mortality. JAMA Pediatr. 2022;176(6):e220687. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Shumskiy I, Richardson T, Brar S, et al. Well-child visits of Medicaid-insured children with medical complexity. J Pediatr. 2018;199:223–230.e2 [DOI] [PubMed] [Google Scholar]
- 3.Kuo DZ, Goudie A, Cohen E, et al. Inequities in health care needs for children with medical complexity. Health Aff (Millwood). 2014;33(12):2190–2198 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Kuo DZ, Cohen E, Agrawal R, et al. A national profile of caregiver challenges among more medically complex children with special health care needs. Arch Pediatr Adolesc Med. 2011;165(11):1020–1026 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Cohen E, Kuo DZ, Agrawal R, et al. Children with medical complexity: an emerging population for clinical and research initiatives. Pediatrics. 2011;127(3):529–538 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Aryee E, Perrin JM, Clancy S, et al. Mental health of caregivers of children with medical complexity during COVID-19. J Dev Behav Pediatr. 2023;44(3):e212–e217 [DOI] [PubMed] [Google Scholar]
- 7.Yu JA, Bayer ND, Beach SR, et al. A national profile of families and caregivers of children with disabilities and/or medical complexity. Acad Pediatr. 2022;22(8):1489–1498 [DOI] [PubMed] [Google Scholar]
- 8.Thomson J, Shah SS, Simmons JM, et al. Financial and social hardships in families of children with medical complexity. J Pediatr. 2016;172:187–193.e1 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Whitney DG, Peterson MD. US national and state-level prevalence of mental health disorders and disparities of mental health care use in children. JAMA Pediatr. 2019;173(4):389–391 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Agency for Healthcare Research and Quality. Child and Adolescent Mental Health. In: 2022 National Healthcare Quality and Disparities Report. AHRQ; 2022:73–81 [PubMed] [Google Scholar]
- 11.Gonzalez VJ, Kimbro RT, Cutitta KE, et al. Mental health disorders in children with congenital heart disease. Pediatrics. 2021;147(2):e20201693. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Friend AJ, Feltbower RG, Hughes EJ, et al. Mental health of long-term survivors of childhood and young adult cancer: a systematic review. Int J Cancer. 2018;143(6):1279–1286 [DOI] [PubMed] [Google Scholar]
- 13.Bates L, Taylor M, Lin JP, et al. Mental health and behaviour in children with dystonia: anxiety, challenging behaviour and the relationship to pain and self-esteem. Eur J Paediatr Neurol. 2021;35:40–48 [DOI] [PubMed] [Google Scholar]
- 14.Brinkman TM, Li C, Vannatta K, et al. Behavioral, social, and emotional symptom comorbidities and profiles in adolescent survivors of childhood cancer: a report from the childhood cancer survivor study. J Clin Oncol. 2016;34(28):3417–3425 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Baumer NT, Hojlo MA, Pawlowski KG, et al. Co-occurring conditions in Down syndrome: findings from a clinical database. Am J Med Genet C Semin Med Genet. 2023;193(4):e32072. [DOI] [PubMed] [Google Scholar]
- 16.Quittner AL, Goldbeck L, Abbott J, et al. Prevalence of depression and anxiety in patients with cystic fibrosis and parent caregivers: results of the International Depression Epidemiological Study across nine countries. Thorax. 2014;69(12):1090–1097 [DOI] [PubMed] [Google Scholar]
- 17.Kogon AJ, Hooper SR. Mental health and social-emotional functioning in children and adolescents with chronic kidney disease [published online ahead of print October 25, 2023]. Pediatr Nephrol. doi:10.1007/s00467-023-06151-1 [DOI] [PubMed] [Google Scholar]
- 18.Barnert ES, Coller RJ, Nelson BB, et al. Key population health outcomes for children with medical complexity: a systematic review. Matern Child Health J. 2019;23(9):1167–1176 [DOI] [PubMed] [Google Scholar]
- 19.University of Washington Rural Health Research Center. Rural-urban commuting area codes. Available at: https://depts.washington.edu/uwruca/. Accessed May 28, 2017
- 20.Feudtner C, Feinstein JA, Zhong W, et al. Pediatric complex chronic conditions classification system version 2: updated for ICD-10 and complex medical technology dependence and transplantation. BMC Pediatr. 2014;14(1):199. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 21.Simon TD, Haaland W, Hawley K, et al. Development and validation of the pediatric medical complexity algorithm (PMCA) Version 3.0. Acad Pediatr. 2018;18(5):577–580 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Zima BT, Gay JC, Rodean J, et al. Classification system for International Classification of Diseases, Ninth Revision, Clinical Modification and Tenth Revision pediatric mental health disorders. JAMA Pediatr. 2020;174(6):620–622 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Arakelyan M, Freyleue SD, Schaefer AP, et al. Rural-urban disparities in health care delivery for children with medical complexity and moderating effects of payer, disability, and community poverty. J Rural Health. 2024;40(2):326–337 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Zou G. A modified poisson regression approach to prospective studies with binary data. Am J Epidemiol. 2004;159(7):702–706 [DOI] [PubMed] [Google Scholar]
- 25.Geweniger A, Haddad A, Barth M, et al. Mental health of children with and without special healthcare needs and of their caregivers during COVID-19: a cross-sectional study. BMJ Paediatr Open. 2022;6(1):e001509. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Adams JS, Chien AT, Wisk LE. Mental illness among youth with chronic physical conditions. Pediatrics. 2019;144(1):1–9 [DOI] [PubMed] [Google Scholar]
- 27.Chua WD, Pipan M, Rosen AR. Neurologic and developmental considerations in children with medical complexity. Curr Probl Pediatr Adolesc Health Care. 2021;51(9):101073. [DOI] [PubMed] [Google Scholar]
- 28.Benavente-Fernández I, Siddiqi A, Miller SP. Socioeconomic status and brain injury in children born preterm: modifying neurodevelopmental outcome. Pediatr Res. 2020;87(2):391–398 [DOI] [PubMed] [Google Scholar]
- 29.Gall VN, Buchhalter J, Antonelli RC, et al. Improving care for families and children with neurodevelopmental disorders and co-occurring chronic health conditions using a care coordination intervention. J Dev Behav Pediatr. 2022;43(8):444–453 [DOI] [PubMed] [Google Scholar]
- 30.Pinquart M. Posttraumatic stress symptoms and disorders in children and adolescents with chronic physical illnesses: a meta-analysis. J Child Adolesc Trauma. 2018;13(1):1–10 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Cuneo AA, Sifflet C, Bardach N, et al. Pediatric medical traumatic stress and trauma-informed care in pediatric chronic illness: a healthcare provider survey. J Pediatr. 2023;261(113580):113580. [DOI] [PubMed] [Google Scholar]
- 32.Inkelas M, Raghavan R, Larson K, et al. Unmet mental health need and access to services for children with special health care needs and their families. Ambul Pediatr. 2007;7(6):431–438 [DOI] [PubMed] [Google Scholar]
- 33.An R. Unmet mental health care needs in U.S. children with medical complexity, 2005-2010. J Psychosom Res. 2016;82:1–3 [DOI] [PubMed] [Google Scholar]
- 34.Ramgopal S, Goodman DM, Kan K, et al. Children with medical complexity and mental and behavioral disorders in the emergency department. Hosp Pediatr. 2023;13(1):9–16 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Penwill NY, Wong CT, Taylor DB, et al. Hospitalists’ perceptions of pediatric mental health boarding: quality of care and moral distress. Hosp Pediatr. 2023;13(3):233–245 [DOI] [PubMed] [Google Scholar]
- 36.Doupnik SK, Lawlor J, Zima BT, et al. Mental health conditions and medical and surgical hospital utilization. Pediatrics. 2016;138(6):e20162416. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 37.Malik FS, Hall M, Mangione-Smith R, et al. Patient characteristics associated with differences in admission frequency for diabetic ketoacidosis in United States children’s hospitals. J Pediatr. 2016;171:104–110 [DOI] [PubMed] [Google Scholar]
- 38.Suryavanshi MS, Yang Y. Clinical and economic burden of mental disorders among children with chronic physical conditions, United States, 2008–2013. Prev Chronic Dis. 2016;13(150535):E71. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 39.Kuo DZ, Melguizo-Castro M, Goudie A, et al. Variation in child health care utilization by medical complexity. Matern Child Health J. 2015;19(1):40–48 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 40.Hacker K, Penfold R, Arsenault LN, et al. The impact of the Massachusetts behavioral health child screening policy on service utilization. Psychiatr Serv. 2017;68(1):25–32 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.McBain RK, Kofner A, Stein BD, et al. Growth and distribution of child psychiatrists in the United States: 2007–2016. Pediatrics. 2019;144(6):2007–2016 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 42.Straub L, Bateman BT, Vine S, Huybrechts KF. Prevalence of mental health diagnoses in commercially insured children and adolescents in the US before and during the COVID-19 pandemic. JAMA Netw Open. 2023;6(5):e2314415. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 43.Samji H, Wu J, Ladak A, et al. Review: mental health impacts of the COVID-19 pandemic on children and youth - a systematic review. Child Adolesc Ment Health. 2022;27(2):173–189 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 44.Racine N, McArthur BA, Cooke JE, et al. Global prevalence of depressive and anxiety symptoms in children and adolescents during COVID-19: a meta-analysis. JAMA Pediatr. 2021;175(11):1142–1150 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 45.Radhakrishnan L, Leeb RT, Bitsko RH, et al. Pediatric emergency department visits associated with mental health conditions before and during the COVID-19 pandemic - United States, January 2019–January 2022. MMWR Morb Mortal Wkly Rep. 2022;71(8):319–324 [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.