OBJECTIVE:

To determine if adverse family factors are associated with a higher likelihood of psychotropic polypharmacy among US youth with a mental health condition.

METHODS:

The 2009–2015 Medical Expenditure Panel Survey data were used to identify family characteristics of 5136 youth aged ≤18 years with an emotional or behavioral health condition. Family adversity was based on family size, number of parents in the household, parental education and income, and parent-reported physical and/or cognitive or mental health disability. Cluster analysis identified family adversity subgroups. Polypharmacy was defined as 3 or more psychotropic classes (eg, stimulants, antipsychotics, antidepressants, mood stabilizers, and sedatives) in at least 1 interview round in a calendar year. Weighted logistic regression evaluated associations between family adversity and psychotropic polypharmacy among youth.

RESULTS:

Nearly half (47.8%) of youth lived with parents who had a disability. Parents in the least socioeconomically disadvantaged cluster mainly had a mental illness, and 94% of parents in the most socioeconomically disadvantaged cluster had a parent-reported physical and/or cognitive disability and mental illness. Among youth, mood disorder (24.2%; 95% confidence interval [CI]: 12.6%–16.0%), antidepressant use (16.0%; 95% CI: 10.6%–21.5%), and antipsychotic use (7.5%; 95% CI: 5.4%–9.6%) were higher in the most socioeconomically disadvantaged cluster relative to the other clusters. Approximately 3% of youth received psychotropic polypharmacy. The odds of psychotropic polypharmacy were 2.7 (95% CI: 1.1–6.4) times greater among youth in the most relative to the least socioeconomically disadvantaged cluster.

CONCLUSIONS:

Higher use of psychotropic polypharmacy among youth with parents who have multiple disabilities raises concerns about oversight and monitoring of complex psychotropic treatment.

What’s Known on This Subject:

Youth exposure to adverse family experiences affects the onset, persistence, and severity of mental illness. Clustering of adverse family factors can have a greater impact on youth emotional and behavioral health than the independent effect of any single factor.

What This Study Adds:

This study contributes new knowledge about the clustering of parental disability within socioeconomically disadvantaged families and its association with childhood psychotropic use. Psychotropic medication oversight is most critical for youth in the most socioeconomically disadvantaged families with serious disability.

Fourteen percent of US children aged <18 years with a mental health diagnosis are prescribed a psychotropic medication.1,2  Increased psychotropic use in youth37  has given way to use of concomitant psychotropic medications (ie, polypharmacy), most frequently without an initial trial of evidence-based psychotherapy.810  Such use is predicated on a complex underlying mental health condition,911  but a body of literature shows that clustering of adverse family factors, such as socioeconomic status and parental disability, also impacts mental health.1216  Few researchers have investigated psychotropic polypharmacy in this context. According to the World Health Organization and the Centers for Disease Control and Prevention, disability is a physical, cognitive, or mental impairment leading to activity limitations (https://www.cdc.gov/ncbddd/disabilityandhealth/disability.html). Disability among families is rarely studied, yet such situations are not infrequent and carry important implications for child development. One-quarter of US adults has a disability,17  and youth with a disability are twice as likely as those without a disability to live with an adult who has a disability.18  Youth with a disability living with an adult with a disability is more prevalent in poverty.1719  Those who live in families of lower socioeconomic status are exposed to traumatic family circumstances, financially stressed parents with compromised parenting skills, and crime and violence in disadvantaged communities,16  which affect the onset, persistence, and severity of childhood mental illness.15 

Adverse family factors conflate a number of these socioeconomic and disability issues that can increase the likelihood of psychotropic polypharmacy among youth. Clustering of adverse family factors may directly impact marital conflict and compromise parenting practices13  and can have a greater impact on a youth’s emotional and behavioral health than the independent effect of any single factor. The greater the number of adverse experiences and the lower the parental education, the higher the risk of psychotropic use as a young adult.12  We aimed to identify clusters of parental disability and socioeconomic factors that define family adversity and determine which clusters are associated with a higher likelihood of psychotropic polypharmacy. We hypothesized that youth in clusters with more adverse family factors would have a greater likelihood of receiving psychotropic medication concomitantly from 3 or more therapeutic classes.

We used a retrospective cohort design to link youth survey data with data on the parents who live in the same household. The design allowed us to assess the association between parental socioeconomic status and disability status with the likelihood that a youth would receive 3 or more psychotropic therapeutic classes.

The data source is the 2009–2015 Medical Expenditure Panel Survey (MEPS) data, which is sponsored by the Agency for Healthcare Research and Quality and the National Center for Health Statistics. MEPS is a national household survey that comprises a constellation of large-scale surveys to gather nationally representative estimates of health care use and expenditures, health status, demographics, and socioeconomic characteristics of the civilian noninstitutionalized population in the United States (www.meps.ahrq.gov/mepsweb/). The survey has been ongoing since 1996 and is administered in an overlapping panel design with data collected through 5 rounds of interviews that span ∼2 calendar years. One designee per household is the reporting unit, and this person answers questions for all individuals living in the household. The 2015 survey is the most recent year available for which the medical condition codes were the same (ie, information required to identify the study cohort). The medical condition coding in the most recent data release (2016) differs from previous years and could not be combined with data from previous years. This research was deemed nonhuman subjects research by the University of Maryland Institutional Review Board.

The study sample included all youth ≤18 years of age who had at least one mental health condition recorded in at least one interview in survey years 2009 through 2015. The clinical classification codes in the MEPS data identified all reported mental health conditions.

The youth cohort with at least one reported mental health condition was linked to the mother and father within the same reporting unit by using the identification number unique to the reporting unit. Excluded from the analysis was any youth who did not have a parent recorded in the data (ie, these could be individuals living with a sibling or on their own). We also excluded youth in the same family to avoid overestimating the association between family adverse factors and psychotropic polypharmacy. Once we established the youth-parent linkage, we extracted information related to the youth, the parent(s), and the family household.

Demographic Characteristics

We used the MEPS Household Component to extract basic demographic characteristics for the study cohort. These included age, sex, race and/or ethnicity, and insurance status. We categorized age into 3 groups (0–4, 5–14, and 15–18 years) to correspond with preschool, elementary and middle, and high school, respectively. Insurance status was categorized as uninsured, privately insured (ie, Tricare and any other private insurance), and publicly insured (ie, Medicaid and other state-specific programs).

Mental Health Conditions

The clinical classification codes in the MEPS data are mutually exclusive clinical categories. In this study, we used the mental and behavioral health clinical classification codes corresponding to adjustment disorder (650); anxiety disorder (651); attention-deficit/hyperactivity disorder (ADHD), conduct, and oppositional defiant disorders (652), which are referred to as disruptive behavior disorders in this study; delirium or cognitive disorders (including nonpsychotic mental disorders and mild cognitive impairment; 653); developmental disorders (including communication, learning, and motor skills disorders and developmental and intellectual disabilities; 654); mental disorders diagnosed in infancy, childhood, or adolescence (including autism, elimination disorders, and tic disorders; 655); mood disorder (including bipolar disorder and depressive disorder; 657); personality disorders (including paranoid, explosive, dependent, and antisocial disorders; 658); schizophrenia (659); alcohol-related disorders (660); substance use disorders (661); and miscellaneous mental health conditions (including dissociative, eating, sleep, and somatoform disorders; 670).

Psychotropic Medication Classification

Using the unique study identification number, we extracted all reported psychotropic medications used for youth in the study cohort. Psychotropic medications were grouped by therapeutic class and included ADHD medications (stimulants and atomoxetine), centrally acting α agonists (guanfacine and clonidine), antidepressants, antipsychotics, anxiolytics, mood stabilizers, and sedatives and/or hypnotics.

Psychotropic Polypharmacy

To identify youth receiving psychotropic polypharmacy, we identified the medications reported within each interview round in a calendar year. We defined psychotropic polypharmacy as the use of medications from 3 or more distinct therapeutic classes reported in at least 1 interview in a calendar year.

Household Characteristics

The MEPS Household Component provides information on socioeconomic status. Variables selected for analysis included the number of parents in the household, parental educational attainment, income status, and the number of individuals living in the same household. Household income status is a variable in MEPS defined as low, middle, or high income. This is based on the percentage above the poverty line and is derived from annual household income and family size. A household at <200% of the poverty line was defined as low income, between 200% and (but not including) 400% above the poverty line was defined middle income, and ≥400% above the poverty line was defined high income. Educational attainment was categorized as high school or less, more than high school up to a college education, and postgraduate education.

Parental Disability Status

Using the MEPS Household Component, we extracted information on parent-reported physical, cognitive, or mental disability associated with the parents of youth in the study cohort. Four variables denoted activity limitations: ACTLIM (ie, any limitation in work, housework, or school), WLKDIF (ie, difficulty walking 3 blocks), MILDIF (ie, difficulty walking a mile), and STPDIF (ie, difficulty standing for 20 minutes). Physical disability for any parent was defined as endorsing some or serious limitation for any of the 4 variables above. Two variables defined cognitive disability: COGLIM (ie, any cognitive disability) and DFCOG (ie, any serious cognitive limitation). Parents who reported any of the mental health conditions defined earlier were classified as having a mental disability.

Disability, assessed for both the mother and the father in each household, was based on any parent-reported physical, cognitive, or mental disability present in at least one parent in the household. We identified serious physical or cognitive disability on the basis of whether the respondent endorsed serious limitation. We also measured the overall parental impairment by examining the distribution of families in which there was no parent-reported disability or mental illness, parent-reported physical or cognitive disability only, parent-reported mental illness only, or parent-reported both physical and/or cognitive and mental disability.

Descriptive Analyses

Survey weights supplied in the MEPS data were used to obtain national population estimates. In descriptive analyses of categorical variables, we estimated weighted proportions and corresponding 95% confidence intervals (CIs). Descriptive analyses were conducted for youth and family adversity characteristics.

Cluster Analysis of Family Adversity

K-means cluster analysis was used to find clusters on the basis of centroids in the data that define clusters. The centroids of each cluster are computed repeatedly to assess for changes in clusters based on a change in the centroid. This process continues until there are no further changes in the centroids. The cluster analysis identified mutually exclusive groups, in which each cluster represents families with common characteristics based on the number of parents in the household, parental education, family income status, family size, and parental disability status measured as parent-reported physical and/or cognitive disability, mental illness, and overall impairment. We tested 2, 3, 4, and 5 clusters. The number of clusters was selected on the basis of goodness-of-fit statistics, which include the approximate r2 value, a pseudo F-statistic, and the cubic cluster criterion. The F-statistic tests the between-group and within-group variation. The cubic cluster criterion assesses the deviation in the r2 obtained from a uniform distribution and the actual r2. Larger values for all metrics indicate better clustering. These metrics and the theoretical interpretation of the clusters guided model selection.

Weighted Logistic Regression

To assess the odds that youth received psychotropic polypharmacy across family adversity clusters, we used weighted logistic regression. The independent variable was the family cluster. Because several family characteristics define the clusters, this accounts for the collinearity.

The weighted estimates for the family characteristics of the 5136 youth with at least 1 mental health condition are shown in Table 1. More than one-third (36.8%) of the sample lived in a single-parent household, and 38.4% lived in a low-income household. Overall, 16.4% of youth lived in households in which at least 1 parent had a serious disability, and 39.6% lived in households in which at least 1 parent had a mental illness. Youth who lived with a parent who reported both a physical and/or cognitive disability and a mental illness constituted 15.3% of the sample.

TABLE 1

Description of Family Characteristics of Youth and Adolescents in the United States With a Mental Health Condition (2009–2015)

Weighted % (N = 5136)95% CI
No. parents in the household   
 1 36.8 34.3–39.3 
 Both 63.2 60.7–65.7 
Parental educational attainmenta   
 High school or less 24.9 23.0–26.8 
 Some college or college degree 32.2 30.0–34.6 
 Postgraduate degree 12.9 10.9–15.0 
Family income statusb   
 Low 38.4 36.1–40.7 
 Middle 31.5 29.0–33.9 
 High 30.1 27.4–32.9 
Family size   
 1–3 37.6 35.2–39.9 
 4–5 53.5 51.3–55.8 
 6 or more 8.9 7.6–10.2 
Parental physical or cognitive disabilityc   
 None 76.4 74.4–78.4 
 1 parent with some disability 6.9 5.8–8.0 
 Both parents with some disability 0.3 0.03–0.54 
 1 parent with serious disability 14.8 13.4–16.3 
 Both parents with serious disability 1.6 1.0–2.1 
Parental mental illness   
 No mental illness 60.4 57.9–63.1 
 Mental illness in 1 parent 32.8 30.4–35.2 
 Mental illness in both parents 6.8 5.5–8.0 
Overall parental impairment   
 No disability or mental illness 52.2 49.6–54.8 
 Physical and/or cognitive disability only 8.3 7.2–9.4 
 Mental illness only 24.2 22.1–26.4 
 Physical and/or cognitive disability and mental illness 15.3 13.6–17.1 
Weighted % (N = 5136)95% CI
No. parents in the household   
 1 36.8 34.3–39.3 
 Both 63.2 60.7–65.7 
Parental educational attainmenta   
 High school or less 24.9 23.0–26.8 
 Some college or college degree 32.2 30.0–34.6 
 Postgraduate degree 12.9 10.9–15.0 
Family income statusb   
 Low 38.4 36.1–40.7 
 Middle 31.5 29.0–33.9 
 High 30.1 27.4–32.9 
Family size   
 1–3 37.6 35.2–39.9 
 4–5 53.5 51.3–55.8 
 6 or more 8.9 7.6–10.2 
Parental physical or cognitive disabilityc   
 None 76.4 74.4–78.4 
 1 parent with some disability 6.9 5.8–8.0 
 Both parents with some disability 0.3 0.03–0.54 
 1 parent with serious disability 14.8 13.4–16.3 
 Both parents with serious disability 1.6 1.0–2.1 
Parental mental illness   
 No mental illness 60.4 57.9–63.1 
 Mental illness in 1 parent 32.8 30.4–35.2 
 Mental illness in both parents 6.8 5.5–8.0 
Overall parental impairment   
 No disability or mental illness 52.2 49.6–54.8 
 Physical and/or cognitive disability only 8.3 7.2–9.4 
 Mental illness only 24.2 22.1–26.4 
 Physical and/or cognitive disability and mental illness 15.3 13.6–17.1 
a

Missing educational attainment for 1642 observations.

b

Income status from MEPS is based on annual income and family size.

c

One observation was missing for parental disability status.

The majority of youth were between 5 and 14 years old (57.5%), boys (65.2%), non-Hispanic white (69.6%), and privately insured (53.9%). Minority race and/or ethnicity groups were 12.6% non-Hispanic African American and 10.3% Hispanic. A small proportion (3.6%) was uninsured, and 42.5% were publicly insured. Disruptive behavior disorders were the most prevalent mental health condition (68.8%), followed by mood disorders (14.3%), and developmental disorders (5.3%). Disorders present in <0.6% of the overall sample include delirium and/or cognitive disorders, personality disorders, and alcohol or substance use disorders. The sample sizes do not meet the precision standard guidelines for reporting MEPS descriptive statistics and are not reported. Less than one-third reported using an ADHD medication (ie, a stimulant or atomoxetine; 32.4%), and 5.4% used a centrally acting α agonist. Antidepressants (11.6%), antipsychotics (5.1%), and mood stabilizers (2.9%) were other psychotropic therapeutic classes reported.

A 4-cluster model had the best sample segmentation and goodness of fit. Table 2 describes each of the 4 clusters, ranging from the most socioeconomically advantaged (cluster 1) to the most socioeconomically disadvantaged (cluster 4). Cluster 1 (n = 656; 12%) represented 2-parent households (86.9%) with high income (82.7%), a postgraduate degree (59.1%), and 4 to 5 family members per household (55.5%). Cluster 2 (n = 2660; 50%) represented 2-parent households (76.1%) with low (42.9%) and middle (39.5%) income, some college (32.9%) or high school (30.6%) education, and 4 to 5 family members per household (77.7%). Cluster 3 (n = 1035; 19%) represented single-parent households (97.2%) with low income (53.5%), some college education (44.2%), and 1 to 3 family members per household (100%). Cluster 4 (n = 785; 15%) reflected 2-parent households (66.1%), with low income (55.9%), a high school education or less (38.1%), and 1 to 3 family members per household (51.9%).

TABLE 2

Description of Family Characteristics Across Clusters of Family Adversity

Family CharacteristicsCluster 1 (N = 656)Cluster 2 (N = 2660)Cluster 3 (N = 1035)Cluster 4 (N = 785)
Weighted %95% CIWeighted %95% CIWeighted %95% CIWeighted %95% CI
No. parents in the householda         
 1 13.1 9.4–16.9 23.9 21.4–26.5 97.2 95.9–98.5 33.9 28.5–39.4 
 Both 86.9 83.1–90.6 76.1 73.5–78.4 2.8 1.5–4.1 66.1 60.6–71.5 
Parental educational attainmenta,b         
 High school or less 4.8 2.6–7.1 30.6 27.7–33.5 23.2 19.4–27.0 38.1 32.1–44.0 
 Some college or college degree 21.4 16.7–26.1 32.9 29.3–36.4 44.2 39.0–49.4 31.2 25.5–36.9 
 Postgraduate degree 59.1 53.1–65.0 0.00 0.00–0.00 0.5 0.0–0.9 2.7 0.45–5.0 
Family income statusa,c         
 Low 2.9 1.5–4.2 42.9 39.5–46.2 53.5 48.8–58.2 55.9 50.7–61.1 
 Middle 14.5 10.8–18.1 39.5 36.1–42.8 32.1 27.4–36.7 30.4 25.9–34.9 
 High 82.7 78.8–86.6 17.7 14.6–20.7 14.4 10.8–18.0 13.7 8.7–18.6 
Family sized         
 1–3 40.2 35.1–45.2 5.8 4.5–7.2 100.0 100.0–100.0 51.9 45.9–57.8 
 4–5 55.5 50.6–60.5 77.7 75.3–80.1 0.00 0.00–0.00 44.4 38.2–50.5 
 6 or more 4.3 2.4–6.2 16.4 14.3–18.6 0.00 0.00–0.00 3.8 2.4–5.1 
Parental physical or cognitive disabilitya         
 None 88.3 84.8–91.8 91.4 89.7–93.0 87.9 84.9–90.7 1.0 0.0–1.9 
 1 parent with some disability 5.7 3.3–8.2 2.5 1.8–3.3 5.8 3.5–8.2 22.9 18.7–27.2 
 Both parents with some disability 0.3 0.0–0.7 0.2 0.0–0.5 0.3 0.0–0.8 0.6 0.0–1.1 
 1 parent with serious disability 5.7 3.6–7.9 5.3 3.9–6.6 6.0 4.3–7.8 67.0 62.9–71.0 
 Both parents with serious disability 0.04 0.00–0.12 0.6 0.1–1.1 0.0 0.0–0.0 8.5 5.5–11.6 
Parental mental illnessa         
 No mental illness 51.1 45.0–57.1 79.3 76.7–81.8 70.3 65.6–75.1 4.9 2.9–7.0 
 Mental illness in 1 parent 37.6 32.2–43.0 17.8 15.6–20.0 29.7 24.9–34.4 74.8 69.7–79.8 
 Mental illness in both parents 11.3 7.6–15.0 2.9 1.9–3.9 0.00 0.00–0.00 20.3 15.6–24.9 
Overall parental impairmentd         
 No disability 44.7 38.7–50.8 70.6 67.7–73.5 58.2 53.6–62.8 0.0 0.0–0.0 
 Physical and/or cognitive disability only 6.3 4.0–8.7 8.7 7.0–10.3 12.1 9.3–15.0 4.9 2.9–7.0 
 Mental illness only 43.6 37.7–49.3 20.7 18.2–23.3 29.7 24.9–34.4 1.0 0.0–1.9 
 Physical and/or cognitive disability and mental illness 5.4 2.6–8.2 0.0 0.0–0.0 0.0 0.0–0.0 94.1 91.7–96.4 
Family CharacteristicsCluster 1 (N = 656)Cluster 2 (N = 2660)Cluster 3 (N = 1035)Cluster 4 (N = 785)
Weighted %95% CIWeighted %95% CIWeighted %95% CIWeighted %95% CI
No. parents in the householda         
 1 13.1 9.4–16.9 23.9 21.4–26.5 97.2 95.9–98.5 33.9 28.5–39.4 
 Both 86.9 83.1–90.6 76.1 73.5–78.4 2.8 1.5–4.1 66.1 60.6–71.5 
Parental educational attainmenta,b         
 High school or less 4.8 2.6–7.1 30.6 27.7–33.5 23.2 19.4–27.0 38.1 32.1–44.0 
 Some college or college degree 21.4 16.7–26.1 32.9 29.3–36.4 44.2 39.0–49.4 31.2 25.5–36.9 
 Postgraduate degree 59.1 53.1–65.0 0.00 0.00–0.00 0.5 0.0–0.9 2.7 0.45–5.0 
Family income statusa,c         
 Low 2.9 1.5–4.2 42.9 39.5–46.2 53.5 48.8–58.2 55.9 50.7–61.1 
 Middle 14.5 10.8–18.1 39.5 36.1–42.8 32.1 27.4–36.7 30.4 25.9–34.9 
 High 82.7 78.8–86.6 17.7 14.6–20.7 14.4 10.8–18.0 13.7 8.7–18.6 
Family sized         
 1–3 40.2 35.1–45.2 5.8 4.5–7.2 100.0 100.0–100.0 51.9 45.9–57.8 
 4–5 55.5 50.6–60.5 77.7 75.3–80.1 0.00 0.00–0.00 44.4 38.2–50.5 
 6 or more 4.3 2.4–6.2 16.4 14.3–18.6 0.00 0.00–0.00 3.8 2.4–5.1 
Parental physical or cognitive disabilitya         
 None 88.3 84.8–91.8 91.4 89.7–93.0 87.9 84.9–90.7 1.0 0.0–1.9 
 1 parent with some disability 5.7 3.3–8.2 2.5 1.8–3.3 5.8 3.5–8.2 22.9 18.7–27.2 
 Both parents with some disability 0.3 0.0–0.7 0.2 0.0–0.5 0.3 0.0–0.8 0.6 0.0–1.1 
 1 parent with serious disability 5.7 3.6–7.9 5.3 3.9–6.6 6.0 4.3–7.8 67.0 62.9–71.0 
 Both parents with serious disability 0.04 0.00–0.12 0.6 0.1–1.1 0.0 0.0–0.0 8.5 5.5–11.6 
Parental mental illnessa         
 No mental illness 51.1 45.0–57.1 79.3 76.7–81.8 70.3 65.6–75.1 4.9 2.9–7.0 
 Mental illness in 1 parent 37.6 32.2–43.0 17.8 15.6–20.0 29.7 24.9–34.4 74.8 69.7–79.8 
 Mental illness in both parents 11.3 7.6–15.0 2.9 1.9–3.9 0.00 0.00–0.00 20.3 15.6–24.9 
Overall parental impairmentd         
 No disability 44.7 38.7–50.8 70.6 67.7–73.5 58.2 53.6–62.8 0.0 0.0–0.0 
 Physical and/or cognitive disability only 6.3 4.0–8.7 8.7 7.0–10.3 12.1 9.3–15.0 4.9 2.9–7.0 
 Mental illness only 43.6 37.7–49.3 20.7 18.2–23.3 29.7 24.9–34.4 1.0 0.0–1.9 
 Physical and/or cognitive disability and mental illness 5.4 2.6–8.2 0.0 0.0–0.0 0.0 0.0–0.0 94.1 91.7–96.4 
a

P < .0001.

b

Missing educational attainment for 1642 observations in total.

c

Income status from MEPS is based on annual income and family size.

d

Not estimable because of 0 cell sizes.

Parental disability coincided with socioeconomic disadvantage clusters (Table 2). Parents in the most socioeconomically advantaged cluster (cluster 1) either had no disability (44.7%) or had a mental illness only (43.6%). Clusters 2 and 3 had the majority of parents without any overall impairment. The most socioeconomically disadvantaged group (cluster 4) had at least 1 parent with a reported serious disability (67.0%) or a mental illness (74.8%), and nearly all had some reported physical and/or cognitive and a mental illness (94.1%).

Youth characteristics within each cluster are shown in Table 3. Youth in cluster 4 were more likely to have public insurance and a mood disorder. Psychotropic polypharmacy was 4.2% among youth in more socioeconomically disadvantaged families with parental disability (cluster 4) compared with 1.7% for youth in the least socioeconomically disadvantaged families with less parental disability (cluster 1). Youth in cluster 4 were more likely to be taking antidepressant, antipsychotic, and anxiolytic medications relative to those in the other clusters.

TABLE 3

Description of Characteristics of Youth Within Clusters of Family Adversity

Child CharacteristicsCluster 1 (N = 656)Cluster 2 (N = 2660)Cluster 3 (N = 1035)Cluster 4 (N = 785)
Weighted %CIWeighted %CIWeighted %CIWeighted %CI
Psychotropic polypharmacya: ≥3 therapeutic classes 1.7 0.6–2.8 2.2 1.3–3.0 2.7 1.0–4.3 4.2 2.1–6.4 
Age group, yb         
 0–4 4.8 2.5–7.1 5.9 4.6–7.3 3.7 1.9–5.4 4.0 2.4–5.7 
 5–14 54.7 48.9–60.4 60.9 58.0–63.8 58.3 53.4–63.2 49.9 44.6–55.4 
 15–18 40.6 34.9–46.2 33.2 30.3–36.0 38.1 33.3–42.8 45.9 40.4–51.6 
Sex         
 Male 65.8 59.9–71.8 67.4 64.5–70.4 64.8 59.9–69.5 58.6 53.0–64.1 
 Female 34.2 28.2–40.1 32.6 29.6–35.5 35.2 30.5–40.0 41.4 35.9–46.9 
Race and/or ethnicityc         
 Non-Hispanic white 77.9 73.8–82.1 67.5 64.4–70.6 63.1 58.5–67.7 72.3 67.1–77.5 
 Non-Hispanic African American 7.5 4.9–10.2 12.5 10.4–14.6 19.5 15.6–23.4 11.2 8.3–14.1 
 Hispanic 4.9 3.2–6.7 12.7 10.7–14.7 11.9 8.9–14.9 8.3 5.7–10.9 
 Other 9.6 6.3–12.8 7.3 5.1–9.5 5.5 3.5–7.5 8.1 4.8–11.5 
Insurance statusc         
 Uninsured 1.6 0.3–2.8 3.8 2.7–5.0 4.2 2.2–6.2 4.7 2.0–7.4 
 Private 87.1 83.6–90.6 51.6 48.2–54.9 37.9 32.9–42.9 34.3 29.0–39.7 
 Public 11.3 8.1–14.6 44.6 41.5–47.7 57.9 52.9–62.8 60.9 55.5–66.4 
Mental health conditionsd         
 Disruptive behavior disorderb 60.5 54.6–66.3 72.3 69.4–75.1 69.3 64.8–73.8 69.5 63.7–75.3 
 Anxiety disordera 28.3 23.2–33.3 13.9 11.7–15.9 17.5 13.7–21.3 24.1 18.5–29.7 
 Adjustment disordere 1.1 0.0–2.2 0.33 0.0–0.7 1.8 0.3–3.4 2.2 0.8–3.6 
 Developmental disorderse 7.1 4.4–9.9 5.9 4.6–7.2 3.1 1.7–4.6 3.5 2.0–5.1 
 Mood disorderc 13.7 9.5–17.9 10.6 8.8–12.4 16.1 12.3–19.8 24.2 20.0–28.4 
 Mental disorders diagnosed in infancy, childhood, or adolescence 9.1 6.3–12.0 7.7 6.2–9.1 10.1 7.0–13.2 9.6 6.1–13.0 
 Schizophrenia and/or psychotic disorders 0.1 0.0–0.3 0.2 0.0–0.6 0.4 0.0–0.9 0.1 0.0–0.3 
 Other mental health conditions 4.2 2.0–6.4 3.8 2.4–5.1 3.0 1.4–4.6 5.9 2.9–8.9 
Psychotropic medication class         
 Stimulants and/or atomoxetine 36.0 31.2–40.8 31.5 28.4–34.6 31.1 26.4–35.7 31.9 26.7–37.2 
 Antidepressantsb 13.9 10.4–17.4 9.1 7.4–10.8 11.4 8.2–14.6 16.0 10.6–21.5 
 Antipsychoticsb 5.4 3.7–7.2 3.4 2.0–4.7 4.2 1.8–6.6 7.5 5.4–9.6 
 Anxiolyticse 1.1 0.1–2.0 0.6 0.26–0.91 1.3 0.38–2.1 2.2 0.77–3.6 
 Mood stabilizersb 4.7 2.5–6.9 1.7 0.9–2.5 3.3 1.8–4.8 3.2 1.7–4.8 
 Sedatives 0.3 0.0–0.8 0.4 0.0–0.7 0.5 0.0–1.1 0.9 0.1–1.7 
 Central-acting antiadrenergice 3.7 1.9–5.4 4.7 3.5–6.0 8.3 5.4–11.2 6.0 4.2–7.9 
Child CharacteristicsCluster 1 (N = 656)Cluster 2 (N = 2660)Cluster 3 (N = 1035)Cluster 4 (N = 785)
Weighted %CIWeighted %CIWeighted %CIWeighted %CI
Psychotropic polypharmacya: ≥3 therapeutic classes 1.7 0.6–2.8 2.2 1.3–3.0 2.7 1.0–4.3 4.2 2.1–6.4 
Age group, yb         
 0–4 4.8 2.5–7.1 5.9 4.6–7.3 3.7 1.9–5.4 4.0 2.4–5.7 
 5–14 54.7 48.9–60.4 60.9 58.0–63.8 58.3 53.4–63.2 49.9 44.6–55.4 
 15–18 40.6 34.9–46.2 33.2 30.3–36.0 38.1 33.3–42.8 45.9 40.4–51.6 
Sex         
 Male 65.8 59.9–71.8 67.4 64.5–70.4 64.8 59.9–69.5 58.6 53.0–64.1 
 Female 34.2 28.2–40.1 32.6 29.6–35.5 35.2 30.5–40.0 41.4 35.9–46.9 
Race and/or ethnicityc         
 Non-Hispanic white 77.9 73.8–82.1 67.5 64.4–70.6 63.1 58.5–67.7 72.3 67.1–77.5 
 Non-Hispanic African American 7.5 4.9–10.2 12.5 10.4–14.6 19.5 15.6–23.4 11.2 8.3–14.1 
 Hispanic 4.9 3.2–6.7 12.7 10.7–14.7 11.9 8.9–14.9 8.3 5.7–10.9 
 Other 9.6 6.3–12.8 7.3 5.1–9.5 5.5 3.5–7.5 8.1 4.8–11.5 
Insurance statusc         
 Uninsured 1.6 0.3–2.8 3.8 2.7–5.0 4.2 2.2–6.2 4.7 2.0–7.4 
 Private 87.1 83.6–90.6 51.6 48.2–54.9 37.9 32.9–42.9 34.3 29.0–39.7 
 Public 11.3 8.1–14.6 44.6 41.5–47.7 57.9 52.9–62.8 60.9 55.5–66.4 
Mental health conditionsd         
 Disruptive behavior disorderb 60.5 54.6–66.3 72.3 69.4–75.1 69.3 64.8–73.8 69.5 63.7–75.3 
 Anxiety disordera 28.3 23.2–33.3 13.9 11.7–15.9 17.5 13.7–21.3 24.1 18.5–29.7 
 Adjustment disordere 1.1 0.0–2.2 0.33 0.0–0.7 1.8 0.3–3.4 2.2 0.8–3.6 
 Developmental disorderse 7.1 4.4–9.9 5.9 4.6–7.2 3.1 1.7–4.6 3.5 2.0–5.1 
 Mood disorderc 13.7 9.5–17.9 10.6 8.8–12.4 16.1 12.3–19.8 24.2 20.0–28.4 
 Mental disorders diagnosed in infancy, childhood, or adolescence 9.1 6.3–12.0 7.7 6.2–9.1 10.1 7.0–13.2 9.6 6.1–13.0 
 Schizophrenia and/or psychotic disorders 0.1 0.0–0.3 0.2 0.0–0.6 0.4 0.0–0.9 0.1 0.0–0.3 
 Other mental health conditions 4.2 2.0–6.4 3.8 2.4–5.1 3.0 1.4–4.6 5.9 2.9–8.9 
Psychotropic medication class         
 Stimulants and/or atomoxetine 36.0 31.2–40.8 31.5 28.4–34.6 31.1 26.4–35.7 31.9 26.7–37.2 
 Antidepressantsb 13.9 10.4–17.4 9.1 7.4–10.8 11.4 8.2–14.6 16.0 10.6–21.5 
 Antipsychoticsb 5.4 3.7–7.2 3.4 2.0–4.7 4.2 1.8–6.6 7.5 5.4–9.6 
 Anxiolyticse 1.1 0.1–2.0 0.6 0.26–0.91 1.3 0.38–2.1 2.2 0.77–3.6 
 Mood stabilizersb 4.7 2.5–6.9 1.7 0.9–2.5 3.3 1.8–4.8 3.2 1.7–4.8 
 Sedatives 0.3 0.0–0.8 0.4 0.0–0.7 0.5 0.0–1.1 0.9 0.1–1.7 
 Central-acting antiadrenergice 3.7 1.9–5.4 4.7 3.5–6.0 8.3 5.4–11.2 6.0 4.2–7.9 
a

≥3 therapeutic classes reported in the same interview round in a calendar year.

b

P < .01.

c

P < .0001.

d

Disruptive behavior disorder includes ADHD, conduct disorder, and oppositional defiant disorder. Other mental health conditions include dissociative disorders, eating disorders, factitious disorders, psychogenic disorders, sexual and gender identity disorders, sleep disorders, somatoform disorders, and other miscellaneous mental conditions. The sample sizes for delirium and/or cognitive disorders, personality disorders, and alcohol or substance use disorders did not meet the precision standard guidelines for reporting MEPS descriptive statistics and thus are not reported.

e

P < .05.

To assess the association between each cluster and psychotropic polypharmacy, we conducted a weighted logistic regression with specific cluster comparisons (Table 4). Youth in cluster 4 had significantly higher odds of receiving psychotropic polypharmacy than youth in cluster 1 (odds ratio = 2.7; 95% CI: 1.1–6.4) and cluster 2 (odds ratio = 2.0; 95% CI: 1.1–3.9).

TABLE 4

Weighted Logistic Regression Analysis of the Likelihood of Polypharmacy as a Function of Family Cluster

Cluster ComparisonOdds Ratio95% CI
2 vs 1 1.33 0.54–3.29 
3 vs 1 1.64 0.69–3.92 
4 vs 1 2.69 1.13–6.42 
3 vs 2 1.23 0.55–2.77 
4 vs 2 2.02 1.05–3.89 
4 vs 3 1.64 0.74–3.64 
Cluster ComparisonOdds Ratio95% CI
2 vs 1 1.33 0.54–3.29 
3 vs 1 1.64 0.69–3.92 
4 vs 1 2.69 1.13–6.42 
3 vs 2 1.23 0.55–2.77 
4 vs 2 2.02 1.05–3.89 
4 vs 3 1.64 0.74–3.64 

In this nationally representative sample of youth with a mental health condition, 47.8% were living with parents who reported a physical and/or cognitive or mental health disability, but the extent of impairment differed by socioeconomic factors. Disability was present in less than half of families with less socioeconomic disadvantage, and when present, it tended to be exclusively a mental illness. Overall, psychotropic polypharmacy was rare (2.7%) but highest (4.2%) among youth from the most socioeconomically disadvantaged families and for whom most parents (94%) reported a serious physical and/or cognitive and mental health disability. One in 5 youth in the most socioeconomically disadvantaged families were reported to have mental illness in both parents.

Our study contributes to the body of literature on family adversity and youth mental health outcomes by accounting for the family effect of parental disability on the socioeconomic environment. This enriches what is already known about the effect of socioeconomic status on youth risk for developing a range of disabling chronic cognitive or mental health conditions across childhood development and into adolescence and adulthood.2025  Poverty has differential effects for psychotropic class use as well.1,26,27  Stimulants are more prevalent in affluent families and antidepressants are more prevalent in low-income families.1  The current study highlights the confluence of parental disability across ranges of socioeconomic status and the association with youth mental health needs. The discussion focuses on the implications of this finding for the clinical management of youth with a mental health condition in pediatric practices.

The socioeconomic factors in clusters 3 and 4 were similar, but the extent of parental disability was different, with nearly all parents in cluster 4 reporting both physical and/or cognitive disability and mental illness. This corresponded with a higher proportion of youth in cluster 4 reporting psychotropic polypharmacy and a significantly higher use of antidepressant, antipsychotic, and anxiolytic medications relative to those in the other clusters. Parents are central to the monitoring plan with the prescriber.28  The prescribing clinician often relies on parental report of outcomes and side effects to gauge any necessary treatment changes, and in families in which one or both parents reported a physical and/or cognitive disability along with mental illness, this can be challenging. Given the potential for suicidal thoughts and behaviors with antidepressants and the metabolic side effects with antipsychotics, monitoring these regimens requires considerable parental input. The study findings call attention to the importance of ensuring that youth receive adequate monitoring in families in which parental disability and socioeconomic hardship pose barriers to safe and effective use of psychotropic medications.

The clinical management of mental health conditions for youth is difficult when they are living in families faced with high needs and adverse experiences, such as cluster 4 in this study. Such families are often in crisis and at risk for poor adherence and lack of access to evidence-based psychosocial interventions (for example, access to cognitive behavioral therapy).29  Despite the lack of evidence, psychotropic polypharmacy may be prescribed in primary practice to alleviate burden and further disruption in children and their families. The concern is that psychotropic medications are added too quickly and often before a trial of evidence-based psychosocial services.8  Youth living in socioeconomically disadvantaged environments (ie, Medicaid enrolled) are more likely than youth in the general population (30% vs 5%)30,31  to receive psychotropic medications that pose significant health risks due to side effects,31  such as excessive weight gain and incident diabetes with antipsychotics.

The study has several strengths as well as limitations. A strength is that data are nationally representative of US households. In addition, we used the most recent national-level data. The study is limited in that we were unable to assess continuity of care and longitudinal outcomes. The diagnoses are self-reported and not confirmed by a clinical evaluation. Duration of treatment and dosing were not available in the survey data nor was timing of treatment initiation, and so it was not possible to assess the quality or appropriateness of care.

Our finding that parental physical, cognitive, and mental health disability in the most socioeconomically disadvantaged families are associated with a higher likelihood of psychotropic polypharmacy among US youth with a mental health condition raises concerns about parental management and oversight of psychotropic medications. Screening tools for parental disability in routine primary care are needed.32  It will be important to further investigate the reasons for such patterns and identify family situations that can be targets for family-based interventions to reduce the use of psychotropic polypharmacy regimens in some of the most vulnerable youth.

Dr dosReis conceptualized the study, drafted the initial manuscript, and was primarily responsible for all revisions of earlier drafts of the manuscript; Ms Zhang downloaded the public use files, worked with the Agency for Healthcare Research and Quality to obtain access to the detailed files at the data coordinating center in Rockville, Maryland, conducted the analyses, prepared the data output and tables for the manuscript, drafted the initial manuscript, and reviewed and revised earlier drafts of the manuscript; Drs Qato and Camelo Castillo provided methodologic input on the variable construction, reviewed the data output, interpreted the results, and reviewed and edited multiple drafts of the manuscript; Dr Reeves provided clinical input on the characterization of the polypharmacy measure, interpreted the findings from a clinical perspective, and reviewed and edited multiple drafts of the manuscript; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

FUNDING: No external funding.

     
  • ADHD

    attention-deficit/hyperactivity disorder

  •  
  • CI

    confidence interval

  •  
  • MEPS

    Medical Expenditure Panel Survey

1
Merikangas
KR
,
He
JP
,
Rapoport
J
,
Vitiello
B
,
Olfson
M
.
Medication use in US youth with mental disorders
.
JAMA Pediatr
.
2013
;
167
(
2
):
141
148
2
Olfson
M
,
Druss
BG
,
Marcus
SC
.
Trends in mental health care among children and adolescents
.
N Engl J Med
.
2015
;
372
(
21
):
2029
2038
3
Olfson
M
,
Blanco
C
,
Liu
L
,
Moreno
C
,
Laje
G
.
National trends in the outpatient treatment of children and adolescents with antipsychotic drugs
.
Arch Gen Psychiatry
.
2006
;
63
(
6
):
679
685
4
Olfson
M
,
Blanco
C
,
Liu
SM
,
Wang
S
,
Correll
CU
.
National trends in the office-based treatment of children, adolescents, and adults with antipsychotics
.
Arch Gen Psychiatry
.
2012
;
69
(
12
):
1247
1256
5
Olfson
M
,
Gameroff
MJ
,
Marcus
SC
,
Jensen
PS
.
National trends in the treatment of attention deficit hyperactivity disorder
.
Am J Psychiatry
.
2003
;
160
(
6
):
1071
1077
6
Olfson
M
,
Marcus
SC
,
Weissman
MM
,
Jensen
PS
.
National trends in the use of psychotropic medications by children
.
J Am Acad Child Adolesc Psychiatry
.
2002
;
41
(
5
):
514
521
7
Olfson
M
,
Marcus
SC
.
National patterns in antidepressant medication treatment
.
Arch Gen Psychiatry
.
2009
;
66
(
8
):
848
856
8
Finnerty
M
,
Neese-Todd
S
,
Pritam
R
, et al
.
Access to psychosocial services prior to starting antipsychotic treatment among Medicaid-insured youth
.
J Am Acad Child Adolesc Psychiatry
.
2016
;
55
(
1
):
69
76.e3
9
Lohr
WD
,
Creel
L
,
Feygin
Y
, et al
.
Psychotropic polypharmacy among children and youth receiving Medicaid, 2012-2015
.
J Manag Care Spec Pharm
.
2018
;
24
(
8
):
736
744
10
Wu
B
,
Bruns
EJ
,
Tai
MH
,
Lee
BR
,
Raghavan
R
,
dosReis
S
.
Psychotropic polypharmacy among youths with serious emotional and behavioral disorders receiving coordinated care services
.
Psychiatr Serv
.
2018
;
69
(
6
):
716
722
11
Winterstein
AG
,
Soria-Saucedo
R
,
Gerhard
T
,
Correll
CU
,
Olfson
M
.
Differential risk of increasing psychotropic polypharmacy use in children diagnosed with ADHD as preschoolers
.
J Clin Psychiatry
.
2017
;
78
(
7
):
e744
e781
12
Björkenstam
E
,
Hjern
A
,
Mittendorfer-Rutz
E
,
Vinnerljung
B
,
Hallqvist
J
,
Ljung
R
.
Multi-exposure and clustering of adverse childhood experiences, socioeconomic differences and psychotropic medication in young adults
.
PLoS One
.
2013
;
8
(
1
):
e53551
13
Hosokawa
R
,
Katsura
T
.
A longitudinal study of socioeconomic status, family processes, and child adjustment from preschool until early elementary school: the role of social competence
.
Child Adolesc Psychiatry Ment Health
.
2017
;
11
:
62
14
Lund
IO
,
Skurtveit
S
,
Handal
M
, et al
.
Association of constellations of parental risk with children’s subsequent anxiety and depression: findings from a HUNT survey and health registry study
.
JAMA Pediatr
.
2019
;
173
(
3
):
251
259
15
McLaughlin
KA
,
Breslau
J
,
Green
JG
, et al
.
Childhood socio-economic status and the onset, persistence, and severity of DSM-IV mental disorders in a US national sample
.
Soc Sci Med
.
2011
;
73
(
7
):
1088
1096
16
Tracy
M
,
Zimmerman
FJ
,
Galea
S
,
McCauley
E
,
Stoep
AV
.
What explains the relation between family poverty and childhood depressive symptoms?
J Psychiatr Res
.
2008
;
42
(
14
):
1163
1175
17
Okoro
CA
,
Hollis
ND
,
Cyrus
AC
,
Griffin-Blake
S
.
Prevalence of disabilities and health care access by disability status and type among adults - United States, 2016
.
MMWR Morb Mortal Wkly Rep
.
2018
;
67
(
32
):
882
887
18
Shandra
CL
,
Avery
RC
,
Hogan
DP
,
Msall
ME
.
Child and adult disability in the 2000 Census: disability is a household affair
.
Disabil Health J
.
2012
;
5
(
4
):
241
248
19
Centers for Disease Control and Prevention
.
CDC health disparities and inequalities report —United States, 2013
.
MMWR Morb Mortal Wkly Rep
.
2013
;
62
(
suppl 3
):
189
20
Spencer
NJ
,
Blackburn
CM
,
Read
JM
.
Disabling chronic conditions in childhood and socioeconomic disadvantage: a systematic review and meta-analyses of observational studies
.
BMJ Open
.
2015
;
5
(
9
):
e007062
21
Jones
A
.
Race, socioeconomic status, and health during childhood: a longitudinal examination of racial/ethnic differences in parental socioeconomic timing and child obesity risk
.
Int J Environ Res Public Health
.
2018
;
15
(
4
):
E728
22
Easterlin
MC
,
Chung
PJ
,
Leng
M
,
Dudovitz
R
.
Association of team sports participation with long-term mental health outcomes among individuals exposed to adverse childhood experiences
.
JAMA Pediatr
.
2019
;
173
(
7
):
681
688
23
Halonen
JI
,
Kivimäki
M
,
Vahtera
J
, et al
.
Childhood adversity, adult socioeconomic status and risk of work disability: a prospective cohort study
.
Occup Environ Med
.
2017
;
74
(
9
):
659
666
24
Collishaw
S
,
Goodman
R
,
Pickles
A
,
Maughan
B
.
Modelling the contribution of changes in family life to time trends in adolescent conduct problems
.
Soc Sci Med
.
2007
;
65
(
12
):
2576
2587
25
Ford
T
,
Goodman
R
,
Meltzer
H
.
The relative importance of child, family, school and neighbourhood correlates of childhood psychiatric disorder
.
Soc Psychiatry Psychiatr Epidemiol
.
2004
;
39
(
6
):
487
496
26
Hjern
A
,
Weitoft
GR
,
Lindblad
F
.
Social adversity predicts ADHD-medication in school children--a national cohort study
.
Acta Paediatr
.
2010
;
99
(
6
):
920
924
27
Amone-P’Olak
K
,
Ormel
J
,
Oldehinkel
AJ
,
Reijneveld
SA
,
Verhulst
FC
,
Burger
H
.
Socioeconomic position predicts specialty mental health service use independent of clinical severity: the TRAILS study
.
J Am Acad Child Adolesc Psychiatry
.
2010
;
49
(
7
):
647
655
28
Walkup
J
;
Work Group on Quality Issues
.
Practice parameter on the use of psychotropic medication in children and adolescents
.
J Am Acad Child Adolesc Psychiatry
.
2009
;
48
(
9
):
961
973
29
Alegria
M
,
Vallas
M
,
Pumariega
AJ
.
Racial and ethnic disparities in pediatric mental health
.
Child Adolesc Psychiatr Clin N Am
.
2010
;
19
(
4
):
759
774
30
Rubin
D
,
Matone
M
,
Huang
YS
,
dosReis
S
,
Feudtner
C
,
Localio
R
.
Interstate variation in trends of psychotropic medication use among Medicaid-enrolled children in foster care
.
Child Youth Serv Rev
.
2012
;
34
:
1492
1499
31
Jonas
BS
,
Gu
Q
,
Albertorio-Diaz
JR
.
Psychotropic Medication Use Among Adolescents: United States, 2005-2010
.
Hyattsville, MD
:
National Center for Health Statistics
;
2013
32
Chung
EK
,
Siegel
BS
,
Garg
A
, et al
.
Screening for social determinants of health among children and families living in poverty: a guide for clinicians
.
Curr Probl Pediatr Adolesc Health Care
.
2016
;
46
(
5
):
135
153

Competing Interests

POTENTIAL CONFLICT OF INTEREST: Dr dosReis has received grant funding from the National Institute of Mental Health, the Patient-Centered Outcomes Research Institute, the Food and Drug Administration, the Pharmaceutical Research and Manufacturers of America Foundation, and GlaxoSmithKline. Dr Reeves has received grant funding from the National Institute of Mental Health and Patient-Centered Outcomes Research Institute. Dr Camelo Castillo has received funding from the Food and Drug Administration and Patient-Centered Outcomes Research Institute. Dr Qato has received funding from the National Institutes of Health; Ms Zhang has indicated she has no potential conflicts of interest to disclose.

FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.