Abstract
Objective:
To examine how characteristics vary between children with any mental health (MH) diagnosis who have typical spending and the highest spending; to identify independent predictors of highest spending; and to examine drivers of spending groups.
Methods:
This retrospective analysis utilized 2016 Medicaid claims from 11 states and included 775,945 children ages 3–17 years with any MH diagnosis and at least 11 months of continuous coverage. We compared demographic characteristics and Medicaid expenditures based on total healthcare spending: the top 1% (highest-spending) and remaining 99% (typical-spending). We used chi-squared tests to compare the 2 groups and adjusted logistic regression to identify independent predictors of being in the top 1% highest-spending group.
Results:
Children with MH conditions accounted for 55% of Medicaid spending among 3- to 17-year-olds. Patients in the highest-spending group were more likely to be older, have multiple MH conditions, and have complex chronic physical health conditions (p<0.001). The highest-spending group had $164,003 per-member-per-year (PMPY) in total healthcare spending, compared to $6097 PMPY in the typical-spending group. Ambulatory MH services contributed the largest proportion (40%) of expenditures ($2455 PMPY) in the typical-spending group; general health hospitalizations contributed the largest proportion (36%) of expenditures ($58,363 PMPY) in the highest-spending group.
Conclusions:
Among children with MH conditions, mental and physical health comorbidities were common and spending for general healthcare outpaced spending for MH care. Future research and quality initiatives should focus on integrating MH and physical healthcare services and investigate whether current spending on MH services supports high-quality MH care.
Keywords: Medicaid, mental health, utilization, healthcare spending
Background
One in five US children has a mental health (MH) condition,1 and MH services account for a large proportion of healthcare encounters and spending in children.2 Healthcare spending is increasing nationally,3 with a disproportionate portion of expenditures attributed to patients with the highest spending.4 Among adult high-spending patients and high-spending pediatric patients with complex chronic conditions, high spending is often driven by emergency department (ED) and inpatient visits.5–7 Among children with MH conditions, relationships between utilization and spending are not well-understood. Children with MH conditions are a priority population for understanding health services utilization and spending, given the prevalence of MH conditions, unmet MH needs, and increased use of high-cost services such as ED and inpatient visits for MH crises.8–10
Understanding needs of high-spending patients has enabled health systems to improve quality, reduce costs, and develop novel models of care coordination and delivery.10–11 For example, among adults with frequent ED visits, novel health service models that address social determinants of health have better met patients needs and reduced spending. Among children with physical health conditions, for approximately half of high-spending children with complex chronic physical health conditions, high spending is a marker of a transient health crisis, such as a hospitalization.2,7 These findings have informed quality programs designed to keep children safely out of the hospital. Less is known about what drives high spending among children with MH conditions. Understanding high spending among children with MH conditions can identify opportunities to help prevent mental and physical health crises and possibly liberate resources to ensure health services are available to more children.
We conducted this study in order to understand and compare healthcare spending for the highest-spending and typical-spending children with MH conditions. Using the Pareto principle,12 we selected the top 1% of patients with the highest healthcare spending for analysis. We hypothesized that (1) patients in each healthcare spending category (highest vs. typical) have different clinical and demographic characteristics and (2) categories of expenditures differ between the two groups. To address this knowledge gap, using a large sample of Medicaid-insured children, we examined (1) clinical and demographic characteristics associated with being in the highest spending group, and (2) the breakdown of which subcategories of healthcare utilization contributed to spending in the highest-spending and typical-spending groups. As a secondary objective, we compared general and mental health spending within and across groups in order to understand which categories of services utilization were most prevalent and the biggest drivers of expenditures among the highest-spending and typical-spending patients.
Methods
Study Design, Data Source, and Study Sample
We conducted a retrospective cross-sectional analysis of the 2016 IBM Watson Health MarketScan Medicaid claims database (Armonk, New York). This database contains paid healthcare claims for beneficiaries from 11 de-identified states in diverse geographic regions of the U.S. Our sample included children ages 3 to 17 years who were enrolled in a plan that provided both medical and MH and substance abuse coverage (i.e., did not carve out behavioral health services), had at least 11 months of continuous coverage during the study year, and had any MH diagnosis.
In order to identify children with any MH diagnoses, we examined all diagnoses from all encounters (i.e., any diagnosis in any position). Using a published definition, we classified children as having any MH diagnosis if they had at least one inpatient or ED claim or two outpatient claims associated with the same MH diagnosis.13 We categorized MH diagnoses using an algorithm based on ICD-10-CM codes that identifies pediatric MH diagnoses in administrative data sources,1,14,,16 MH diagnoses were classified into 21 mutually-exclusive categories (e.g., Depressive Disorders, Anxiety Disorders, Autism Spectrum Disorders).
Healthcare spending groups
We examined all mental and physical healthcare claims for the year and calculated total Medicaid expenditures for children with MH conditions. Per-claim spending was determined by a standardized payment per unit of service,2,17 calculated from all fee-for-service claims. We identified patients in the top one percent of expenditures as being in the highest-spending group, and the remaining 99% of patients as being in the typical-spending group.
Healthcare spending categories
For both highest-spending and typical spending patients, we examined the following specific categories of expenditures in order to identify drivers of spending: inpatient services, ED services, office visits, pharmacy, dental care, diagnostic testing (e.g., laboratory, radiology), miscellaneous therapies (e.g., physical, occupational, and speech therapy), and other services (e.g., durable medical equipment), using previously-described methods.2 For inpatient, ED, ambulatory, and pharmacy categories, we distinguished MH claims from general health claims using the methods described below.
For inpatient hospital stays, we used facility, place of service, and revenue codes to determine location (e.g., dedicated MH facility) and the Diagnosis Related Group (DRG) classification system18 to identify the primary indication for hospitalization. Inpatient MH hospitalizations included all hospitalizations in specialty MH hospital units, long-term/residential MH programs, and hospitalizations in general acute-care hospital units with a MH DRG (including those with no physical health treatments, i.e., “psychiatric boarding”19,20). Inpatient general hospitalizations included hospitalizations in general hospital units for primary physical health reasons (i.e., not a MH DRG).
For ED services, we used place of service codes to determine the location and ICD-10CM diagnosis codes to determine the reason for the visit. MH emergency visits included visits to specialty MH EDs and visits to general EDs with only ICD-10-CM codes for MH diagnoses (i.e., no co-occurring non-MH code). General ED visits included visits to general EDs with any non-mental health ICD-10-CM code (i.e., visits with both MH and non-MH diagnoses were categorized as general ED visits). Thus, ED visits for suicide attempts with any physical health diagnoses were classified as general visits.
For ambulatory office visits, we used provider type codes to determine a clinician’s specialty and ICD-10-CM codes to determine the reason for the visit. MH ambulatory office visits included visits with MH specialists (i.e., psychologists, psychiatrists, psychiatric nurses, and clinicians working in ambulatory MH or substance abuse treatment facilities), and visits with clinicians from any specialty with an accompanying ICD-10-CM code for a MH diagnosis. This would capture, for example, management of psychiatric medications by a primary care provider. General ambulatory office visits included visits with non-MH-specialist clinicians that had no MH diagnoses. Clinicians who were identified in claims by credentials but not specialty (e.g., medical doctor, nurse practitioner) were classified as non-MH-specialists.
For pharmacy claims, medications were classified using the American Hospital Formulary Service Classification Compilation.21 which provides “therapeutic class” (e.g., amphetamine stimulant) and “generic drug names” (e.g., methylphenidate hydrochloride). MH medications included antidepressants, antipsychotics, stimulants, and non-stimulant medications for attention-deficit hyperactivity disorder. General medications included medications with no MH indications (e.g., analgesics, antibiotics). Since we could not determine medication indication from administrative claims, we also classified anti-epileptic drugs, benzodiazepines, and clonidine as general medications, since they are commonly used to treat seizures and hypertension, in order to avoid over-estimating MH medication utilization.
Healthcare utilization.
Using the services classification approach described above, we captured the number of patients in our MH group who utilized each type of service during the study year as well as the number of encounters among those who utilized a given service. We captured utilization data for inpatient, ED, ambulatory, and pharmacy services for both MH and general health reasons.
Covariates
We captured additional patient sociodemographic characteristics, including age, sex, race/ethnicity, Medicaid plan type (fee-for-service or capitated), basis for Medicaid eligibility (disability, other, missing). Excluding the diagnoses used to identify MH conditions, we used the remaining diagnoses to identify clinical characteristics: non-complex chronic physical health conditions22 and complex chronic physical health conditions.23
Analyses
We generated descriptive statistics to characterize the overall sample of children with MH conditions, comparing the highest- and typical-spending subgroups. We used chi-square tests to compare characteristics between groups. We measured which child characteristics were independently associated with being in the top 1% highest-spending group using a logistic regression model adjusted for all patient characteristics listed in table 1.
Table 1.
Clinical Characteristics of Pediatric Medicaid Beneficiaries with a Mental Health Diagnosis, Across Spending Group
| Typical Spending1 (99%) |
Highest Spending (Top 1%) |
|
|---|---|---|
| N | 768,185 | 7760 |
| Demographic Characteristics | ||
| Age in years, n (%) | ||
| 3–8 | 260,130 (34) | 2170 (28) |
| 9–12 | 242,800 (32) | 1835 (24) |
| 13–17 | 265,255 (35) | 3755 (48) |
| Male, N (%) | 466,797 (61) | 4481 (58) |
| Race/ethnicity, n (%) | ||
| White, non-Hispanic | 430,432 (56) | 4060 (52) |
| Black, non-Hispanic | 220,297 (29) | 2030 (26) |
| Hispanic | 38,779 (5) | 366 (5) |
| Other | 17,019 (2) | 165 (2) |
| Missing | 61,658 (8) | 1139 (15) |
| Type of Medicaid plan, n (%) | ||
| Fee-for-service | 254,276 (33) | 5197 (67) |
| Capitated | 513,909 (67) | 2563 (33) |
| Medicaid eligibility, n (%) | ||
| Disability | 81,599 (11) | 3997 (52) |
| Income | 674,247 (87) | 3702 (48) |
| Missing | 12,339 (2) | 61 (1) |
| Mental health conditions,2 n (%) | ||
| Anxiety disorders | 92,896 (12) | 1361 (18) |
| Attention deficit hyperactivity disorder | 358,121 (47) | 2338 (30) |
| Autism spectrum disorder | 53,456 (7) | 1212 (16) |
| Bipolar disorder | 22,220 (3) | 975 (13) |
| Communication disorders | 110,135 (14) | 1059 (14) |
| Depressive disorders | 107,043 (14) | 1905 (25) |
| Developmental disorder | 41,296 (5) | 1843 (24) |
| Disruptive, impulse control, and conduct disorders | 127,763 (17) | 1949 (25) |
| Intellectual disability | 24,732 (3) | 2347 (30) |
| Learning disorders | 97,286 (13) | 1388 (18) |
| Psychotic disorders | 6477 (1) | 445 (6) |
| Substance use disorders | 25,057 (3) | 566 (7) |
| Suicide attempt | 7843 (1) | 881 (11) |
| Suicide ideation | 15,227 (2) | 497 (6) |
| Tic disorders | 21,034 (3) | 715 (9) |
| Other mental health condition | 73,867 (10) | 2344 (30) |
| Number of mental health conditions, n (%) | ||
| 1 | 452,158 (59) | 2321 (30) |
| 2 | 178,323 (23) | 1779 (23) |
| 3 or more | 137,704 (18) | 3660 (47) |
| Physical health conditions,3 n (%) | ||
| None | 276,138 (36) | 418 (5) |
| Any physical health condition | 492,047 (64) | 7,342 (95) |
| Chronic, non-complex | 412,457 (54) | 1541 (20) |
| Complex chronic | 79,590 (10) | 5801 (75) |
All characteristics varied significantly (p<0.001) between spending groups;
Patients could have MH conditions in more than 1 category;
Physical health condition categories are mutually exclusive.
To assess costs overall and within spending categories, we calculated per-member-per-year (PMPY) expenditures by dividing the payments by the number of enrollees.24 We used Kruskal-Wallis tests to compare expenditures and utilization across the spending categories (i.e., inpatient, ED, ambulatory, and pharmacy).
We used SAS version 9.4 (SAS Institute, Cary, NC) for all analyses. This study of de-identified administrative claims was considered exempt from review by the Children’s Hospital of Philadelphia Institutional Review Board.
Results
Study Population
We included 775,945 3- to 17-year-olds enrolled in Medicaid who had a MH condition, which made up 19% of the overall Medicaid population in the dataset in that age range, consistent with prior data.25 Fifty-nine percent had one MH condition; 23% had two, and 18% had three or more. In addition to MH conditions, 64% had a physical health condition. Specifically, 53% had a non-complex chronic physical health condition (e.g., asthma, obesity), and 11% had a complex chronic physical health condition (e.g., muscular dystrophy, cystic fibrosis). Supplemental Table 1 shows characteristics of the children with MH conditions in our sample, compared to children in Medicaid with no MH conditions.
Total Healthcare Spending and Resource Use
Total healthcare spending for patients with a MH condition was $5.3 billion ($7676 PMPY). For comparison, total healthcare spending for all 3- to 17-year-old patients in Medicaid was $9.7 billion; thus 55% of all healthcare spending for 3- to 17-year-old patients in Medicaid was for children with a MH condition. Children with MH conditions in the top 1st percentile incurred 21% of all spending among MH patients ($1.3 billion spending for highest-spending MH patients / $5.3 billion total spending for MH patients) or 13% of spending among all 3- to 17-year-olds in Medicaid ($1.3 billion spending for highest-spending MH patients / $9.7 billion total spending for all patients).
Comparison of Medicaid Mental Health Patients with the Highest Spending vs. Typical Spending
Child sociodemographic and clinical characteristics significantly varied by spending group (Table 1). In unadjusted bivariate comparisons, children in the highest-spending group were more likely to be ages 13 to 17 years (48% vs. 35, p<.001), more often enrolled in fee-for-service Medicaid plans (67% vs. 33%, p<0.001), and more likely to be eligible for Medicaid because of disability (52% vs. 11%, p<0.001). Children in the highest-spending group were more likely to have specific MH diagnoses, with the exception of attention-deficit hyperactivity disorder. Children with the highest spending also had a higher rates of comorbidity. Highest spenders were more likely than typical spenders to have multiple MH conditions (70% vs. 41%, p<0.001), and nearly half of highest spenders had 3 or more MH conditions (47% vs. 18% of typical spenders, p<0.001). Highest spenders were also more likely to have complex chronic physical health conditions (75% vs. 10% of typical spenders, p<0.001) and chronic physical health conditions (95% vs. 64% of typical spenders, p<0.001).
Predictors of being in the highest-spending group using a multivariable regression model are summarized in Table 2. Having a complex chronic condition was the strongest independent predictor of being in the highest-spending group (AOR 19.98, 95% CI: 18.02–22.16). Other factors independently associated with being in the highest-spending group included: having a fee-for-service Medicaid plan (AOR 2.23, 95%CI: 2.11–2.36), Medicaid eligibility due to disability (AOR 2.7, 95%CI: 9.54–2.87), intellectual disability (AOR 2.51, 95%CI: 2.35–2.68), suicide attempt (AOR 5.48, 95%CI: 4.96–6.04), and having a chronic non-complex physical health condition (AOR 2.12, 95%CI: 1.9–2.36).
Table 2.
Clinical Characteristics Associated with Higher Spending Among Pediatric Medicaid Beneficiaries with a Mental Health Diagnosis, from an Adjusted Regression Model
| Independent predictors of being in Highest Spending Group Adjusted3 Odds Ratio (95%CI) |
|
|---|---|
| N | |
| Demographic Characteristics | |
| Age in years | |
| 3–8 | 1.01 (0.95, 1.09) |
| 9–12 | 0.91 (0.85, 0.97) |
| 13–17 | Ref |
| Male, N (%) | 1.15 (1.09, 1.21) |
| Race/ethnicity | |
| White, non-Hispanic | Ref |
| Black, non-Hispanic | 1.17 (1.1, 1.24) |
| Hispanic | 0.91 (0.82, 1.03) |
| Other | 1.07 (0.9, 1.26) |
| Missing | 0.74 (0.69, 0.8) |
| Type of Medicaid plan | |
| Fee-for-service | 2.23 (2.11, 2.36) |
| Capitated | Ref |
| Medicaid eligibility | |
| Disability | 2.7 (2.54, 2.87) |
| Income | Ref |
| Missing | 1.12 (0.86, 1.46) |
| Mental health conditions1 | |
| Anxiety disorders | 1.07 (0.99, 1.15) |
| Attention deficit hyperactivity disorder | 0.6 (0.56, 0.64) |
| Autism spectrum disorder | 0.77 (0.71, 0.82) |
| Bipolar disorder | 1.95 (1.78, 2.14) |
| Communication disorders | 0.79 (0.73, 0.85) |
| Depressive disorders | 1.28 (1.19, 1.39) |
| Developmental disorder | 1.94 (1.81, 2.07) |
| Disruptive, impulse control, and conduct disorders | 1.52 (1.42, 1.63) |
| Intellectual disability | 2.51 (2.35, 2.68) |
| Learning disorders | 0.76 (0.7, 0.81) |
| Psychotic disorders | 1.51 (1.33, 1.71) |
| Substance use disorders | 1.35 (1.21, 1.5) |
| Suicide attempt | 5.48 (4.96, 6.04) |
| Suicide ideation | 1.81 (1.61, 2.04) |
| Tic disorders | 1.31 (1.19, 1.43) |
| Other mental health condition | 1.82 (1.71, 1.93) |
| Number of mental health conditions | |
| 1 | Ref |
| 2 | 1.16 (1.08, 1.24) |
| 3 or more | 1.23 (1.11, 1.35) |
| Physical health conditions2 | |
| None | Ref |
| Chronic, non-complex | 2.12 (1.9, 2.36) |
| Complex chronic | 19.98 (18.02, 22.16) |
Patients could have MH conditions in more than 1 category;
Physical health condition categories are mutually exclusive.
The model is adjusted for all covariates represented in the table.
Distribution of Spending for the Highest Spending vs. Typical-Spending Patients
Total PMPY spending for all mental and physical healthcare was $164,003 for the highest-spending patients (i.e., top 1st percentile) and $6907 PMPY for typical-spending patients. For children in the typical-spending group, the largest proportions of expenditures were attributable to ambulatory MH care (37%, $2455 PMPY), MH pharmacy (17%, $803 PMPY), and general pharmacy (9%, $618).
The distribution of spending on general and MH services differed between spending groups (Table 3). In unadjusted comparisons, the highest-spending patients had a higher percentage of spending on physical health services (73% vs. 42%, p<0.001) and a lower percentage of spending on MH services (27% vs. 58%, p<0.00). The largest proportions of expenditures for children in the highest-spending group were attributable to inpatient general (i.e., non-MH) hospitalizations (36%; $58,363 PMPY), ambulatory MH services (14%; $22,803 PMPY), and inpatient MH hospitalization (12%; $19,691 PMPY). Compared to children with typical spending, children in the highest-spending group were more likely to use all categories of general health services, except dental care. Children in the highest-spending group were more likely to have general and MH hospitalizations as well as ED visits. Highest-spending patients were more likely to use MH ambulatory services than typical-spending patients.
Table 3.
Health Care Spending for Children with Mental Health Conditions in Medicaid, by Spending Group
| Health Service | Total Annual Spending ($) | Per-Member-Per-Year Spending1 ($) |
Within-Spending Group Cost Distribution2 (%) |
|||
|---|---|---|---|---|---|---|
| 99% Lower Spending |
Top 1% Highest Spending |
99% Lower Spending |
Top 1% Highest Spending |
99% Lower Spending |
Top 1% Highest Spending |
|
| Total | 4,683,666,570 | 1,272,666,816 | 6097 | 164,003 | 100 | 100 |
| Mental Health | 2,715,344,392 | 342,278,337 | 3535 | 44,108 | 57 | 27 |
| Inpatient | 197,361,861 | 152,802,175 | 257 | 19,691 | 2 | 12 |
| ED | 14,938,698 | 1,115,210 | 19 | 144 | 0 | 0 |
| Ambulatory | 1,885,831,728 | 176,949,889 | 2455 | 22,803 | 37 | 14 |
| Pharmacy | 617,212,104 | 11,411,064 | 803 | 1470 | 17 | 1 |
| General | 1,968,322,178 | 1,272,666,816 | 2562 | 119,895 | 43 | 73 |
| Inpatient | 192,304,830 | 930,388,479 | 250 | 58,363 | 1 | 36 |
| ED | 200,976,745 | 452,900,210 | 262 | 1234 | 6 | 1 |
| Ambulatory | 345,590,165 | 9,573,987 | 450 | 16,040 | 9 | 10 |
| Pharmacy | 474,720,667 | 124,471,351 | 618 | 26,559 | 9 | 16 |
| Dental | 183,985,509 | 1,424,673 | 240 | 184 | 6 | 0 |
| Misc. therapies | 238,610,513 | 22,790,359 | 311 | 2937 | 5 | 2 |
| Testing | 118,909,477 | 11,460,716 | 155 | 1477 | 3 | 1 |
| Other | 213,224,266 | 101,670,776 | 278 | 13,102 | 4 | 8 |
Abbreviations: ED,Emergency Department
Per-member-per-year spending varied significantly (p<0.001) across the spending groups for all health services.
The distribution of spending varied significantly (p<0.001) across the spending groups for all health services.
Inpatient and Emergency Department Utilization and Spending
Use of healthcare services and number of encounters varied significantly by spending group (Table 4). Of all MH patients, 3.4% were hospitalized for a MH condition and 2.6% for a general health indication. In unadjusted comparisons, hospital use was more common in the highest-spending group compared with typical-spending patients: 21% vs. 3.2% for MH hospitalization; 46% vs. 2.1% for general hospitalization (p<0.001 for both). The median length of stay for a MH hospitalization was 6 days (IQR 4, 8), and the median length of stay for a general health hospitalization was 3 days (IQR 2, 5). Among 29,766 general health hospitalizations, 46% also had the patient’s MH diagnosis associated with the hospitalization. Among all patients in the sample, 39% had an ED visit. ED visits were more common among the highest-spending patients compared with typical-spending patients: 14% vs. 4% for MH ED visits; 59% vs. 38% for general ED visits (p<0.001 for all).
Table 4.
Use of Healthcare Services and Number of Encounters Among Children with Mental Health Conditions in Medicaid, by Spending Group
| 99% Lower Spending | Top 1% Highest Spending | |||
|---|---|---|---|---|
| Number of patients using service |
Encounters among utilizers |
Number of patients using service |
Encounters among utilizers |
|
| n (%)1 | Median, (IQR) | n (%)1 | Median, (IQR) | |
| Mental Health | ||||
| Inpatient | 24,838 (3) | 1 (1, 2) | 1418 (21) | 2 (1, 4) |
| ED | 33,217 (4) | 1 (1, 1) | 892 (13) | 2 (1, 4) |
| Ambulatory | 746,479 (97) | 10 (4, 28) | 6350 (92) | 15 (4, 75) |
| Pharmacy | 386,575 (50) | - | 3235 (47) | - |
| General | ||||
| Inpatient | 16,399 (2) | 1 (1, 1) | 3535 (46) | 2 (1, 4) |
| ED | 295,696 (38) | 1 (1, 2) | 4566 (59) | 2 (1, 3) |
| Ambulatory | 640,116 (83) | 3 (2, 6) | 6800 (88) | 38 (10, 120) |
| Pharmacy | 619,160 (81) | - | 7546 (97) | - |
Health care utilization varied significantly (p<0.001) between spending groups for each category of health service.
- Column is not relevant for the service category.
Ambulatory and Pharmacy Utilization and Spending
Nearly all patients with any MH diagnosis used ambulatory care; 97% had an ambulatory MH visit, and 83% had an ambulatory general health visit. 42% of healthcare spending for patients in the sample was attributable to ambulatory visits ($2.2 billion). Most patients (81%) had at least one prescription drug claim for a general health medication; 50% had a claim for a MH-specific medication (i.e., antidepressant, antipsychotic, stimulant, or non-stimulant medication for attention-deficit hyperactivity disorder).
Discussion
The study shows that children with MH conditions accounted for 19% of the Medicaid population and 55% of Medicaid spending among 3- to 17-year-olds. Among children with any MH condition, half of healthcare spending was for general health services, and half was for mental health services. We identified characteristics and utilization patterns for patients with typical spending and patients with the top 1% highest spending. The highest-spending patients were more likely to have complex and chronic physical health problems, >2 MH diagnoses, and fee-for-service payment plans. For the highest-spending patients, the largest proportion of spending was for general health hospitalizations, and the most commonly-used services were pharmacy services (i.e., medications). For patients with typical spending, ambulatory MH visits were the most commonly used services and accounted for the largest proportion of spending. Among all MH patients, comorbid conditions were common; 41% of typical-spending patients and 70% of high-spending patients had more than 1 MH condition. Nearly two-thirds of typical-spending MH patients had a chronic physical health condition, and among the highest-spending patients, 95% of had a chronic physical health conditions.
Among children with MH conditions, aggregate spending for general health care was higher than spending for MH care, and most high spending was driven by general health care, specifically inpatient general health hospitalizations. One possible explanation that might contribute to this finding is that lower spending for MH services reflects cost containment efforts in MH care, such as narrow mental healthcare provider networks26 and utilization management programs.27 Studies have shown that some cost containment programs for MH services in Medicaid have resulted in poor quality MH care. For example, youth enrolled in behavioral health utilization management programs are authorized to receive fewer services than recommended by evidence-based guidelines.28 Studies of children with ADHD using Medicaid have shown that the proportion receiving no treatment at all ranges from 10 to 44%, medication adherence was low, and a larger than recommended proportion received medication only without concomitant recommended psychotherapy, and 10% of children with ADHD receive no treatment.29–30 This prior research raises the question whether Medicaid expenditures for MH services are adequate to support recommended MH treatments. Future research to further investigate our findings could focus on understanding to what extent children using Medicaid have unmet need for MH services and whether reimbursement rates are adequate to support clinicians in delivering high-quality evidence-based services.
Our finding that nearly two-thirds of children with any MH condition and 95% of highest-spending MH patients had a co-occurring chronic physical health condition supports the importance of ongoing efforts to integrate MH services into health care across service settings.31–32 Such efforts are particularly important for children with chronic and complex chronic physical health conditions. Children with chronic and complex chronic physical health conditions were disproportionately represented among the highest-spending population. For some comorbid mental and physical health conditions (e.g., depression and inflammatory bowel disease), prior evidence shows that MH services can help prevent disease exacerbations that are costly to children, families, and payers.33–34 Incorporating MH services into routine care for children with chronic physical illnesses may improve quality of life and generate cost offsets due to fewer disease exacerbations requiring hospitalization. Investing in preventive MH services for children with medical complexity may be of particular interest for accountable care organizations and systems working with value-based or risk-sharing payment models.
General health hospitalizations were the largest contributor to spending in the top 1% highest-spending group. As in other populations of high-spending patients, these hospitalizations could be a marker of a transient health crisis.4 Previous research has shown that general health hospitalizations are more common among patients with MH conditions than their similarly physically ill peers with no MH conditions,35–36 and when these hospitalizations occur, they are more likely to be longer,15 have complications,37 and result in readmissions.38 Therefore, our findings provide background health services data for future research or quality improvement initiatives investigating whether providing ambulatory MH services integrated with physical healthcare can prevent general health hospitalizations. Future research to investigate whether the costs of additional ambulatory services are offset by reductions in hospitalizations and ED visits could inform patient-centered care models designed to keep children safely at home and out of the ED and hospital.
In order to achieve integration of MH and medical care across service settings, interdisciplinary workforce development is required. First, creative solutions to address the longstanding MH workforce shortage39 are needed. Pipeline programs to incentivize training in child and adolescent MH disciplines will help build the necessary workforce to support better MH services integration. Nurse practicioner, physician assistant, and social work training programs may be well-positioned for rapid workforce development, since these programs require a shorter training period and less financial investment from students than most medical and psychology doctorate training programs. Second, medical clinicians require better training and professional development in MH care skills. These skills are not adequately taught in current pediatric medical training,40 and ongoing efforts to improve MH training will help better prepare pediatric clinicians to collaborate in MH care. Interdisciplinary professional development opportunities to exchange knowledge and skills can help foster mutual understanding of each discipline’s clinical approach and can foster collaboration.
Our findings should be considered in the context of this study’s strengths and limitations. First, our ability to examine data from 11 state Medicaid program offers a more comprehensive look at Medicaid costs than is typically available. In order to to ensure complete capture of claims, we restricted our analyses to Medicaid plans that provided both medical and behavioral health services; utilization and costs for children receiving MH and substance abuse services through carve-out arrangements or private insurance plans could differ. Second, we were only able to identify spending covered by the Medicaid plan; services paid for out-of-pocket or by other funding mechanisms, such as local governments or federal block grants are not included. Third, we categorized services delivered by clinicians whose credentials were not specified as being delivered by generalists. If any of these clinicians were, in fact, MH specialists, we may have miscategorized some MH services as being for general health and thus underestimated spending on MH services. We mitigated this limitation by also using diagnosis codes to determine which encounters included MH services. Substance abuse and suicide ideation diagnosis codes are thought to underestimate the prevalence of these disorders41; thus we likely underestimate the prevalence of substance abuse and suicide ideation in young people using Medicaid. Certain medications with multiple indications (anti-epileptic drugs, benzodiazepines, and clonidine) were classified as general medications; therefore, we underestimate utilization of these medications for MH indications. In spite of these limitations, the present analyses provide an overview of health services use for children with MH conditions enrolled in Medicaid.
Conclusions
In this sample of children from 11 states using Medicaid in 2016, the highest-spending patients with MH conditions were more likely to have multiple MH conditions and chronic and complex physical health conditions. Drivers of high spending among children with MH conditions included inpatient hospitalizations for physical or MH reasons. Reducing hospitalizations is a target area to reduce overall costs. Future research and quality initiatives to determine whether increased ambulatory services can lead to fewer hospitalizations could inform the design of patient-centered care models and inform strategies to reduce costs.
Supplementary Material
What’s New:
Children with mental health conditions had more spending for general and physical healthcare than for mental healthcare. Characteristics associated with higher spending included older age, having multiple MH conditions, and having a complex chronic physical health condition (p<0.001).
Acknowledgments:
The authors acknowledge Amy R. Pettit, PhD, consultant and Adjunct Fellow, University of Pennsylvania Center for Public Health Initiatives, for her feedback and assistance with editing.
Funding Source: Dr. Doupnik was supported by grant K23MH115162 from the National Institute of Mental Health. Dr. Feinstein was supported by the Eunice Kennedy Shriver National Institute of Child Health & Human Development under Award Number K23HD091295. Dr. Zima received funding from the Behavioral Health Centers of Excellence for California (SB852). Drs. Berry and Hall and Mr. Rodean were supported by grant HRSA-17–060-147599 (Children with Special Healthcare Needs Research Network) from the Maternal and Child Health Bureau.
The sponsors had no role in the study’s design, data collection, data analysis or interpretation, writing of the report, or the decision to submit the article for publication.
Financial Disclosure: The authors have no financial relationships relevant to this article to disclose.
Abbreviations:
- ED
Emergency Department
- DRG
Diagnosis Related Group
- ICD-10-CM
International Classification of Diseases, 10th revision, Clinical Modification
- IQR
Interquartile Range
- MH
Mental Health
- PMPY
Per-member-per-year
- US
United States
Footnotes
Declarations of interest: None
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
References
- 1.Perou R, Bitsko RH, Blumberg SJ, et al. Mental health surveillance among children—United States, 2005–2011. MMWR Surveillance Summaries. 2013;62(Suppl 2):1–35. [PubMed] [Google Scholar]
- 2.Kuo DZ, Hall M, Agrawal R, et al. Comparison of health care spending and utilization among children with Medicaid insurance. Pediatrics. 2015;136(6):e1521–9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Hartman M, Martin AB, Benson J, and Catlin A. National Health Care Spending in 2018: Growth Driven by Accelerations in Medicare and Private Insurance Spending. Health Affairs. 2020;39(1):8–17. [DOI] [PubMed] [Google Scholar]
- 4.Johnson TL, Rinehart DJ, Durfee J, et al. For Many Patients Who Use Large Amounts of Health Care Services, The Need Is Intense Yet Temporary. Health Affairs. 2015;34(8): 1312–1319. [DOI] [PubMed] [Google Scholar]
- 5.Jiang HJ, Weiss AH, and Barrett ML. Characteristics of Emergency Department Visits for Super-Utilizers by Payer, 2014. Healthcare Cost and Utilization Project, Statistical Brief #221. Agency for Healthcare Research and Quality; February 2017. Available: https://www.hcup-us.ahrq.gov/reports/statbriefs/sb221-Super-Utilizer-ED-Visits-Payer2014.jsp [PubMed] [Google Scholar]
- 6.Jiang HJ, Weiss AH, Barrett ML, and Sheng M. Characteristics of Hospital Stays for Super-Utilizers by Payer, 2012. Healthcare Cost and Utilization Project, Statistical Brief #190. Agency for Healthcare Research and Quality; May 2015. Available: https://www.hcup-us.ahrq.gov/reports/statbriefs/sb190-Hospital-Stays-Super-UtilizersPayer-2012.jsp [PubMed] [Google Scholar]
- 7.Berry JG, Hall M, Neff J, et al. Children with medical complexity and Medicaid: spending and cost savings. Health Affairs. 2014; 33(12):2199–2206. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Zima BT, Rodean J, Hall M, et al. Psychiatric disorders and trends in resource use in pediatric hospitals. Pediatrics. 2016;138(5):e20160909–e20160909. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Plemmons G, Hall M, Doupnik S, et al. Hospitalization for suicide ideation or attempt: 2008–2015. Pediatrics. 2018;141(6):e20172426. [DOI] [PubMed] [Google Scholar]
- 10.Ilkowitz JT, Choi S, Rinke ML, et al. Pediatric type 1 diabetes. Quality Management in Health Care. 2016;25(4):231–237. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Snyder A, Marton J, McLaren S, et al. Do high fidelity wraparound services for youth with serious emotional disturbances save money in the long-term? Journal of Mental Health Policy and Economics. 2017;20(4):167–175. http://www.ncbi.nlm.nih.gov/pubmed/29300703. Accessed September 7, 2018. [PubMed] [Google Scholar]
- 12.Pareto V Manual of Political Economy (A Critical And Variorum Edition) 2014. Oxford, United Kingdom: Oxford University Press. [Google Scholar]
- 13.Witt WP, Fullerton CA, Chow C, et al. Effect of having a usual source of care on health care outcomes among children with serious emotional disturbance. Academic Pediatrics. 2017;17(1):45–52. [DOI] [PubMed] [Google Scholar]
- 14.Bardach NS, Coker TR, Zima BT, et al. Common and costly hospitalizations for pediatric mental health disorders. Pediatrics. 2014;133(4):602–609. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Doupnik SK, Lawlor J, Zima BT, et al. Mental health conditions and medical and surgical hospital utilization. Pediatrics. 2016;138(6):e20162416–e20162416. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Callahan ST, Fuchs DC, Shelton RC, et al. Identifying suicidal behavior among adolescents using administrative claims data. Pharmacoepidemiology and Drug Safety. 2013;22(7):769–775. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Hoefgen ER, Andrews AL, Richardson T, et al. Health care expenditures and utilization for children with noncomplex chronic disease. Pediatrics. 2017;140(3):e20170492. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Centers for Medicare and Medicaid Services. Design and Development of the Diagnosis Related Group (DRG). Baltimore, MD; 2016. https://www.cms.gov/ICD10Manual/version34-fullcode-cms/fullcode_cms/Design_and_development_of_the_Diagnosis_Related_Group_(DRGs)_PBL-038.pdf. Accessed March 29, 2019. [Google Scholar]
- 19.Claudius I, Donofrio JJ, Lam CN, Santillanes G. Impact of boarding pediatric psychiatric patients on a medical ward. Hospital Pediatrics. 2014;4(3):125–132. [DOI] [PubMed] [Google Scholar]
- 20.Bender D, Pande N, Ludwig M. A Literature Review: Psychiatric Boarding. 2008. http://aspe.hhs.gov/daltcp/reports/2008/PsyBdLR.pdf. Accessed March 29, 2019.
- 21.American Society of Health-System Pharmacists. AHFS pharmacologic-therapeutic classification. Available at: http://www.ahfsdruginformation.com/ahfs-pharmacologic-therapeutic-classification/. Accessed August 30, 2019
- 22.Berry JG, Ash AS, Cohen E, et al. Contributions of children with multiple chronic conditions to pediatric hospitalizations in the united states: a retrospective cohort analysis. Hospital Pediatrics. 2017;7(7). [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Feinstein JA, Russell S, DeWitt PE, et al. R Package for pediatric complex chronic condition classification. JAMA Pediatrics. 2018; 172(6):596–598. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Romley JA, Axeen S, Lakdawalla DN, et al. The relationship between commercial health care prices and medicare spending and utilization. Health Services Research. 2015;50(3):883–896. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.MACPAC Report to Congress on Medicaid and CHIP. Chapter 4: Behavioral Health in the Medicaid Program—People, Use, and Expenditures. June 2015. Available: https://www.macpac.gov/wp-content/uploads/2015/06/Behavioral-Health-in-theMedicaid-Program%E2%80%94People-Use-and-Expenditures.pdf
- 26.Zhu JM, Zhang Y, Polsky D. Networks In ACA Marketplaces Are Narrower For Mental Health Care Than For Primary Care. 2017:1624–1631. doi: 10.1377/hlthaff.2017.0325 [DOI] [PubMed] [Google Scholar]
- 27.Wickizer TM, Lessler D. Utilization Management: Issues, Effects, and Future Prospects. Annu Rev Public Health. 2002;23(1):233–254. doi: 10.1146/annurev.publhealth.23.100901.140529 [DOI] [PubMed] [Google Scholar]
- 28.Mora Ringle VA, Scott Hickey J, Jensen-Doss A. Patterns and predictors of compliance with utilization management guidelines supporting a state policy to improve the quality of youth mental health services. Child Youth Serv Rev. 2019;96:194–203. doi: 10.1016/j.childyouth.2018.11.035 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Hoagwood KE, Kelleher K, Zima BT, et al. Ten-Year Trends in Treatment Services for Children with Attention Deficit Hyperactivity Disorder Enrolled in Medicaid. Health Affairs. 2016;35(7):1266–1270. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30.Zima BT, Bussing R, Tang L, et al. Quality of Care for Childhood Attention Deficit Hyperactivity Disorder in a Managed Care Medicaid Program. J Am Acad Chil Adolesc Psychiatry. 2010;49(12)1225–1237. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31.Doupnik SK, Walter JK. Collaboration is key to improving hospital care for patients with medical and psychiatric comorbidity. Hospital Pediatrics. 2016;6(12):760–762. doi: 10.1542/hpeds.2016-0165. [DOI] [PubMed] [Google Scholar]
- 32.Richardson LP, Ludman E, McCauley E, et al. Collaborative care for adolescents with depression in primary care: a randomized clinical trial. JAMA. 2014;312(8):809. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 33.Deter H-C, von Wietersheim J, Jantschek G, et al. High-utilizing Crohn’s disease patients under psychosomatic therapy. Biopsychosocial Medicine. 2008;2:18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 34.Ellis DA, Naar-King S, Templin T, et al. Multisystemic Therapy for Adolescents with Poorly Controlled Type 1 Diabetes. Diabetes Care. 2008;31(9)1746–1747. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 35.Myrvik MP, Burks LM, Hoffman RG, Dasgupta M, Panepinto JA. Mental health disorders influence admission rates for pain in children with sickle cell disease. Pediatr Blood Cancer. 2013;60(7):1211–1214. doi: 10.1002/pbc.24394 [DOI] [PubMed] [Google Scholar]
- 36.Garrison MM, Katon WJ, Richardson LP. The impact of psychiatric comorbidities on readmissions for diabetes in youth. Diabetes Care. 2005;28(9):2150–2154. [DOI] [PubMed] [Google Scholar]
- 37.Doupnik SK, Mitra N, Feudtner C, Marcus SC. The Influence of Comorbid Mood and Anxiety Disorders on Outcomes of Pediatric Patients Hospitalized for Pneumonia. Hosp Pediatr. 2016;6(3):135–142. doi: 10.1542/hpeds.2015-0177 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 38.Doupnik SK, Lawlor J, Zima BT, et al. Mental Health Conditions and Unplanned Hospital Readmissions in Children. An Off Publ Soc Hosp Med J Hosp Med. 2018;13(7):445–452. doi: 10.12788/jhm.2910 [DOI] [PubMed] [Google Scholar]
- 39.Hyde PS Report to Congress on the Nation’s Substance Abuse and Mental Health Workforce Issues. US Dept. for Health and Human Services, Substance Abuse and Mental Health Services, 2013. http://www.cimh.org/sites/main/files/fileattachments/samhsa_bhwork_0.pdf. Accessed September 17, 2019. [Google Scholar]
- 40.Raval GR, & Doupnik SK. Closing the Gap: Improving Access to Mental Health Care Through Enhanced Training in Residency. Pediatrics. 2018;139(1), e20163181. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 41.Nguyen TQ, Simpson PM, Braaf SC, et al. Level of agreement between medical record and ICD-10-AM coding of mental health, alcohol and drug conditions in trauma patients. Health Information Management Journal, 2018; 48(3), 127–134. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.