Growth in the Concurrent Use of Antipsychotics With Other Psychotropic Medications in Medicaid-Enrolled Children

NEW RESEARCH

Growth in the Concurrent Use of Antipsychotics With Other Psychotropic Medications in Medicaid-Enrolled Children Amanda R. Kreider, BS, BA, Meredith Matone, MHS, Christopher Bellonci, MD, Susan dosReis, PhD, Chris Feudtner, MD, PhD, MPH, Yuan-Shung Huang, MS, Russell Localio, PhD, David M. Rubin, MD, MSCE Objective: Second-generation antipsychotics (SGAs) have increasingly been prescribed to Medicaid-enrolled children; however, there is limited understanding of the frequency of concurrent SGA prescribing with other psychotropic medications. This study describes the epidemiology of concurrent SGA use with 4 psychotropic classes (stimulants, antidepressants, mood stabilizers, and a-agonists) among a national sample of Medicaid-enrolled children and adolescents 6 to 18 years old between 2004 and 2008. Method: Repeated cross-sectional design was used, with national Medicaid Analytic eXtract data (10.6 million children annually). Logit and Poisson regression, standardized for year, demographics, and Medicaid eligibility group, estimated the probability and duration of concurrent SGA use with each medication class over time and examined concurrent SGAs in relation to clinical and demographic characteristics. Results: While SGA use overall increased by 22%, 85% of such use occurred concurrently. By 2008, the probability of concurrent SGA use ranged from 0.22 for stimulant users to 0.52 for mood stabilizer users. Concurrent SGA use occurred for long durations (69%–89% of annual medication days). Although the highest users of concurrent SGA were participants in foster care and disability Medicaid programs or those with behavioral hospitalizations, the most significant increases over time occurred among participants who were income-eligible for Medicaid (þ13%), without comorbid ADHD (þ15%), were not hospitalized (þ13%), and did not have comorbid intellectual disability (þ45%). Conclusion: Concurrent SGA use with other psychotropic classes increased over time, and the duration of concurrent therapy was consistently long term. Concurrent SGA regimens will require further research to determine efficacy and potential drug–drug interactions, given a practice trend toward more complex regimens in less-impaired children/adolescents. J. Am. Acad. Child Adolesc. Psychiatry, 2014;53(9):960–970. Key Words: second-generation antipsychotics, pediatric psychopharmacology, polypharmacy, Medicaid, foster care

O

ver the past 2 decades, the use of secondgeneration antipsychotics (SGAs) among children and adolescents has grown significantly.1-7 Between 1993 and 2002, the rate of antipsychotics prescribing to children in an outpatient setting increased approximately

This article is discussed in an editorial by Dr. Mark Olfson, MD, MPH, on page 942. An interview with two of the authors is available by podcast at www.jaacap.org or by scanning the QR code to the right. Supplemental material cited in this article is available online.

5-fold,1,3 and, between 2002 and 2007, SGA use increased among Medicaid-enrolled children at a rate surpassing that of all other psychotropic medication classes.1,2 This prescribing increase has triggered concern because of the risk of serious metabolic side effects of SGAs in children, including weight gain, glucose intolerance, and type 2 diabetes.8-15 Although the existing literature describes prescribing trends of SGAs to children, an understanding of trends in prescriber practice around the use of SGAs in combination with other psychotropic medications is limited

JOURNAL 960

www.jaacap.org

OF THE

AMERICAN ACADEMY OF C HILD & ADOLESCENT PSYCHIATRY VOLUME 53 NUMBER 9 SEPTEMBER 2014

CONCURRENT SGA USE IN CHILDREN

but suggestive of growth. By 2007, concurrent antipsychotic treatment with another psychotropic medication class was noted in 11% of outpatient visits in which psychotropic medication was mentioned, more than double the rate from 1996.16 New combination medications, such as combination stimulant/antipsychotics, are emerging in the marketplace and are likely to increase this trend. However, to date, knowledge of the epidemiology of combination therapy, and of the relative contributions of SGA monotherapy and concurrent therapy to total use, is scant. The few studies that describe concurrent therapy, for example, have not disaggregated the sociodemographic or clinical subgroups in which this prescribing is occurring.16-18 The concern over rising concurrent SGA treatment is not trivial, as the efficacy and safety of this treatment are even more poorly understood than that of SGA monotherapy. Although there is some endorsement for adjunctive SGA treatment with an antidepressant or a mood stabilizer for treatment of bipolar disorder (BD) and major depressive disorder (MDD),19-23 much of this research has been conducted in adult populations and is limited to nonblinded designs, case reports, and retrospective chart reviews.16,19,20,22,23 There is evidence to support SGA monotherapy in children for certain conditions, such as disruptive behavior in youth with intellectual disability24; however, there is little or no evidence, except a recent report on the addition of Risperdal for children with attentiondeficit/hyperactivity disorder (ADHD) and aggression,25 showing that combinations are safe or effective.26 Therefore, if a practice norm is emerging that increasingly endorses concurrent SGA use, it is important to determine the combinations of treatment used with the greatest frequency, the population of children prescribed these combinations, and the potential downstream implications, including adherence to complex regimens, drug–drug interactions, treatment duplication, and costs.16-18,27-35 Taken together, these concerns may be less problematic if the concurrent SGA is being used for short-term behavioral control while psychosocial therapies are commenced, but population data are lacking to confirm this practice. Knowledge of the changing trends in SGA prescribing will ensure that future efficacy and safety research prioritizes the norms of emerging clinical practice around combination therapy. This is particularly important for Medicaid-enrolled children, who represent more

than one-third of the pediatric population of the United States and are prescribed SGAs at higher rates than privately insured children,36,37 and among the subgroup of children in foster care, who demonstrate higher prescription rates than the general Medicaid population.18,38-40 Our study, therefore, aimed to describe practice trends for prescribing SGAs concurrently with other psychotropic medications among Medicaid-enrolled children over time, including duration of exposure; and to capture clinical and demographic characteristics that most influenced such prescribing.

METHOD Study Design and Sample Selection The study sample was drawn from a nationally representative population of children and adolescents continuously enrolled in Medicaid (defined as at least 10 of 12 months per year) between 2004 and 2008. Age was restricted to 6 to 18 years to include the youth most commonly prescribed psychotropic medications. A sampling scheme targeted 25,000 children/adolescents per state per year. Small states were oversampled in relation to larger states, and participants with foster care and Supplemental Security Income (SSI) eligibility were oversampled versus those who were incomeeligible. Each child was assigned a sampling weight equal to the inverse of the probability of being selected (Table S1, available online). Children with seizure disorder were excluded from analyses of mood stabilizer.

Data Source The data source was the national Centers for Medicare and Medicaid Services Medicaid Analytic eXtract (MAX) files for years 2004, 2006, and 2008. Child-level demographic, eligibility, encounter, and pharmacy data were extracted from the personal summary, outpatient, inpatient, and pharmacy files. Following state-level data quality reviews, Connecticut and Maine were excluded as a result of incompleteness in outpatient behavioral health encounters, and Massachusetts was excluded because children with foster care eligibility were not identifiable. Because of concerns regarding underreporting of pharmacy claims for children in Medicaid managed care arrangements,41 statelevel reviews of psychotropic medication rates within managed care and fee-for-service payer arrangements, and stratified by Medicaid eligibility group (foster care, SSI, Temporary Assistance for Needy Families [TANF]/ other), were also conducted. Five additional states (Florida, Hawaii, Nevada, Ohio, and Pennsylvania) and the District of Columbia were excluded from analysis based on the results of payment arrangement reviews (Supplement 1, available online). The final population included 42 states, comprising more than 80% of the youth Medicaid population in each study year.

JOURNAL OF THE AMERICAN ACADEMY OF C HILD & ADOLESCENT PSYCHIATRY VOLUME 53 NUMBER 9 SEPTEMBER 2014

www.jaacap.org

961

KREIDER et al.

Psychotropic Medications Psychotropic medications were identified in the pharmacy claims file using the national drug code (NDC). The file was merged with the First Data Bank to extract medication names. Analysis focused primarily on stimulants, antidepressants, SGAs, a-agonists, and mood stabilizers. Because anxiolytics and sedative/ hypnotics were used at much lower rates (<0.2% of the population), they were not reported in primary analyses of concurrent SGA use. Antidepressants included selective serotonin reuptake inhibitors (SSRI), tricyclic antidepressants (TCA), and other antidepressants. Mood-stabilizing agents included carbamazepine, valproic acid, gabapentin, lamotrigine, oxcarbazepine anticonvulsants, and lithium. Because a-agonists can be prescribed for other medical conditions (e.g., hypertension), these agents were included only if the child/ adolescent also had a claim for a medication from another psychotropic class during the year.

Psychotropic Use Patterns A “pattern” was identified when 1 or more psychotropic medication classes were used for at least 14 consecutive days. Initially, all mutually exclusive psychotropic treatment patterns were identified in each of the study years; for example, stimulant alone and stimulant with antidepressant constituted 2 separate patterns. This process resulted in the identification of 101 unique patterns. The 30 most common patterns were retained, encompassing all medication use for 99.7% of the children/adolescents in the sample. A child/adolescent was identified as a user of psychotropic medications if she or he had at least 1 psychotropic treatment pattern within a year. The primary outcome measures were derived from these 30 patterns, which encompassed the 5 major therapeutic classes, including: stimulants, antidepressants, SGA, a-agonists, and mood stabilizers. Aggregating the psychotropic use patterns allowed us to estimate, first, SGA use during a calendar year, including SGA alone only, both SGA alone and SGA concurrently with another medication class, and concurrent SGA only. Next, the overall use of the other 4 psychotropic medication classes was estimated, followed by the concurrent use of SGA with each class.

Psychotropic Use Duration Duration of psychotropic class use was estimated within each study year, based on the total days’ supply of medication. The duration of concurrent SGA use within each year was also calculated and reflected the days of overlap with another therapeutic class. Duplicate prescription fills for a medication on the same date were not counted toward duration.

Clinical Characteristics To capture differential trends in concurrent SGA prescribing across demographic and clinical characteristics, a subsample of children/adolescents enrolled in

fee-for-service Medicaid arrangements was identified. The fee-for-service sample was chosen to mitigate concerns of reliability of diagnostic encounter data in managed care payment arrangements.41,42 Fee-forservice was defined as no more than 2 months of enrollment in physical and behavioral health managed care plans for a given individual within a year, as indicated in the MAX personal summary file. Although foster care and SSI populations were overrepresented, Hispanic children/adolescents were underrepresented, and their white counterparts were somewhat overrepresented in fee-for-service Medicaid arrangements, the overall trends of concurrent SGA use in fee-forservice mirrored trends in the full sample, allowing subanalyses in this population. Clinical characteristics included diagnostic history and evidence of behavioral or psychiatric hospitalization within the year. Mental health diagnoses were classified using the DSM-IV-TR and coded using the ICD-9 classification. These included: schizophrenia, BD, depression, anxiety disorder, conduct disorder (CD), autism, ADHD, intellectual disability, developmental delay, and a composite variable of miscellaneous mental health diagnoses (Supplement 2, available online). Diagnostic codes associated with inpatient and outpatient encounter claims were used to identify psychiatric morbidity. Diagnoses were further classified as “comorbid” or “noncomorbid.” For example, if a child diagnosed with autism lacked any other mental health diagnoses within the year, she was considered to have “noncomorbid autism.” In addition, diagnostic codes associated with inpatient claims were used to identify psychiatric hospitalization.

Statistical Analysis Demographic and clinical characteristics of the sample were described as percentages (weighted and unweighted), stratified by study year and age group (6–11 years and 12–18 years). Medication use was described similarly for each year, stratified by Medicaid eligibility group. Logistic regression estimated the log odds of use of each therapeutic class and of concurrent SGA, while adjusting for year, sex, age, state, and Medicaid eligibility group and including an interaction term for year and eligibility group. Model results were transformed into probabilities of medication use with robust variance estimates,43 standardized across demographic characteristics, and stratified by year and eligibility group. For duration of use, Poisson models, with an offset for the child’s Medicaid-eligible days, estimated the number of days of use of the medication(s) of interest, using a similar algorithm. Given large sample sizes, confidence intervals of all models were extremely narrow and are therefore not reported. Race was excluded from the main analysis owing to significant variation in the percentage of unknown race across eligibility groups in many states. Controlling for race in

JOURNAL 962

www.jaacap.org

OF THE

AMERICAN ACADEMY OF C HILD & ADOLESCENT PSYCHIATRY VOLUME 53 NUMBER 9 SEPTEMBER 2014

JOURNAL OF THE AMERICAN ACADEMY OF C HILD & ADOLESCENT PSYCHIATRY VOLUME 53 NUMBER 9 SEPTEMBER 2014

TABLE 1

Demographic and Clinical Characteristics of Sample Age 6e11 Yearsa % of Sample

Weighted % of Population

% Change in Population

% of Sample

% Change in Sample

Weighted % of Population

% Change in Population

2004

2008

2004e2008

2004

2008

2004e2008

2004

2008

2004e2008

2004

2008

2004e2008

53.9 46.1

54.4 45.6

þ0.5 0.5

51.3 48.7

51.4 48.6

þ0.1 0.1

53.3 46.7

54.0 46.0

þ0.7 0.7

49.5 50.5

49.9 50.1

þ0.4 0.4

48.4 23.9 12.5 7.1 8.2

46.2 22.2 14.4 7.3 10.0

2.2 1.7 þ1.9 þ0.2 þ1.8

37.1 28.5 26.0 4.9 3.5

35.8 26.0 28.5 4.9 4.8

1.3 2.5 þ2.5 0.0 þ1.3

49.3 27.1 9.7 7.1 6.8

47.0 25.9 11.2 6.9 9.0

2.3 1.2 þ1.5 0.2 þ2.2

38.5 30.6 22.0 5.6 3.3

36.5 28.6 24.9 5.5 4.5

2.0 2.0 þ2.9 0.1 þ1.2

15.2 64.2 20.6

16.0 61.9 22.1

þ0.8 2.3 þ1.5

4.1 90.2 5.8

4.2 89.5 6.4

þ0.1 0.7 þ0.6

20.8 53.4 25.9

21.6 50.9 27.5

þ0.8 2.5 þ1.6

6.0 85.9 8.0

6.3 85.0 8.7

þ0.3 0.9 þ0.7

1.6

1.7

þ0.1

0.6

0.7

þ0.1

1.7

1.8

þ0.1

0.7

0.8

þ0.1

www.jaacap.org

Note: SSI ¼ Supplemental Security Income; TANF ¼ Temporary Assistance for Needy Families. a Annual mean in sample: 487,823; weighted: 5,307,929. b Annual mean in sample: 536,328; weighted: 5,280,179.

963

CONCURRENT SGA USE IN CHILDREN

Sex Male Female Race White Black Hispanic Other Unknown Eligibility group Foster care TANF/other SSI Chronic conditions Seizure disorder

% Change in Sample

Age 12e18 Yearsb

KREIDER et al.

sensitivity analyses restricted to those children with known race did not alter results; therefore, all children/adolescents (without adjustment for race) were included in the final models. All estimates were weighted using sampling weights. A similar approach was used to capture which clinical and demographic groups experienced disproportionate increases in concurrent SGA use in a fee-for-service subsample. Separate logit models, with interaction terms between year and the demographic or clinical characteristic of interest (age group, sex, and an indicator for psychiatric hospitalization within the year) estimated the log odds of concurrent use of SGA with any other psychotropic medication class, while adjusting for all other covariates. Additional multivariate logistic regression models, stratified by diagnosis, included an interaction term between noncomorbid versus comorbid diagnosis and year, allowing evaluation of disproportionate increases in concurrent SGA use by diagnostic group. All model results were transformed into probabilities of concurrent SGA use with robust variance estimates by year,43 standardized for demographic characteristics. Analyses were conducted using SAS version 9.3 (SAS Institute, Cary, NC) and Stata version 12.1 (Stata Corp, College Station, TX). The Institutional Review Board at the Children’s Hospital of Philadelphia approved this study.

RESULTS Description of the Child and Adolescent Sample The sample included an annual average of 490,000 children 6 to 11 years old and 540,000 children and adolescents 12 to 18 years old, weighted to an annual average population of 10.6 million children/adolescents. The fee-for-service subsample included an annual average of 550,000 (weighted: 4.6 million) children/adolescents. More than one-third of participants were identified as white, more than one-fourth were black, one-fourth were Hispanic, and about 10% had race/ethnicity identified as missing or unknown. One-half of all participants were male. Five percent were Medicaid-eligible based on foster care status, 7% were in the SSI program, and the remainder (88%) were income-eligible (TANF/other) (Table 1). Trends in Psychotropic Use Overall use of psychotropic medications was relatively stable between 2004 and 2008 (Table 2). By 2008, stimulants were the most prevalent psychotropic medication class (8.3%), followed by antidepressants (3.7%), SGAs (3.3%), a-agonists (1.8%), and mood stabilizers (1.6%).

Children with foster care–based and SSI-based eligibility were 3.5 to 10 times more likely to receive each psychotropic medication class than were income-eligible children. Concurrent Psychotropic Use Patterns Despite rather flat trends in overall psychotropic use, SGA use increased over time, from 2.7% in 2004 to 3.3% in 2008 (Table 2). Increases occurred both for SGA used alone as well as SGA used concurrently with another psychotropic medication class; however, the majority of participants (85%) with any SGA use during the year used concurrent SGA (Table 2) These findings were robust to standardization (Figure 1). Among children who used each of the other psychotropic medication classes, the proportion using a concurrent SGA increased steadily. By 2008, concurrent SGA use was evident in 52% of mood stabilizer users, 37% of a-agonist users, 32% of antidepressant users, and 22% of stimulant users. Concurrent SGA rates were 30% to 60% higher among youth in foster care and those in the SSI program. The increase in concurrent SGA rates among users of each of the other medication classes was also robust to standardization (see Figure S1, available online). Standardized to clinical and demographic characteristics, concurrent SGA use increased by 9% to 14% across all indicator medication classes between 2004 and 2008, and was highest for children in foster care and SSI. Although absolute rates were lower among the 88% of youth with TANF/other eligibility, these children experienced the greatest proportional increase in concurrent SGA use: more than 20% among stimulant users (from 0.12 to 0.15), antidepressant users (from 0.18 to 0.22), and a-agonist users (from 0.20 to 0.25); and 13% among mood stabilizer users (from 0.37 to 0.42). Duration of Psychotropic Use Overall, the duration of concurrent SGA use was not short-term (Figure 2), occurring on 35% to 40% of calendar days. For those children receiving concurrent treatment, SGA use overlapped the majority of the time that an indicator medication was used. For example, concurrent SGA use was present for 146 of 197 stimulant days (74%). In addition, children/adolescents who were prescribed a concurrent SGA had a longer duration of use of the indicator medication class (e.g., stimulant) than their counterparts without concurrent SGA. This was true for all 4 of

JOURNAL 964

www.jaacap.org

OF THE

AMERICAN ACADEMY OF C HILD & ADOLESCENT PSYCHIATRY VOLUME 53 NUMBER 9 SEPTEMBER 2014

CONCURRENT SGA USE IN CHILDREN

TABLE 2

Psychotropic Medication Use and Concurrent Second-Generation Antipsychotic (SGA) Use % With Psychotropic Class (Weighted)

All children 6e18 years old Stimulant SGA SGA alone only SGA alone and concurrent Concurrent SGA Only Antidepressant Mood stabilizer a-agonist Any psychotropic SSI subgroup Stimulant SGA Antidepressant Mood stabilizer a-agonist Any psychotropic Foster care subgroup Stimulant SGA Antidepressant Mood stabilizer a-agonist Any psychotropic

% With Concurrent SGA (Weighted)

2004

2008

Change (2004e2008)

2004

2008

Change (2004e2008)

7.7 2.7 0.4 0.6 1.7 4.5 1.6 1.3 11.6

8.3 3.3 0.5 0.9 1.9 3.7 1.6 1.8 12.0

þ0.6 þ0.6 þ0.1 þ0.3 þ0.2 0.8 0.0 þ0.5 þ0.4

19.5 n/a n/a n/a n/a 27.6 47.8 32.8 n/a

22.4 n/a n/a n/a n/a 31.7 52.1 36.7 n/a

þ2.9 n/a n/a n/a n/a þ4.1 þ4.3 þ3.9 n/a

23.5 14.1 13.3 9.3 5.9 38.6

24.2 15.5 10.9 8.2 7.6 38.4

þ0.7 þ1.4 2.4 1.1 þ1.7 0.2

32.4 n/a 43.9 50.3 44.2 n/a

35.6 n/a 47.1 54.7 48.7 n/a

þ3.2 n/a þ3.2 þ4.4 þ4.5 n/a

21.1 12.9 15.8 6.7 4.8 32.5

22.8 14.5 12.7 6.2 6.1 33.0

þ1.7 þ1.6 3.1 0.5 þ1.3 þ0.5

33.3 n/a 42.4 64.6 46.2 n/a

36.5 n/a 46.9 68.3 50.5 n/a

þ3.2 n/a þ4.5 þ3.7 þ4.3 n/a

Note: “Concurrent SGA” refers to the concomitant use of SGA with the listed psychotropic medication class. SSI ¼ Supplemental Security Income. n/a ¼ not applicable.

the indicator medication classes. As a result, the duration of concurrent SGA use among users of concurrent SGA nearly approximated, and sometimes exceeded, the duration of use of the indicator medication for children who were not using SGAs. Duration of use was highest for FIGURE 1 Standardized trends over time in secondgeneration antipsychotic (SGA) use.

children in foster care, which was on average 10 to 20 days longer than in the overall Medicaid population. Clinical and Demographic Antecedents of Concurrent SGA Use Analysis in a fee-for-service subsample of children revealed disproportionate increases in concurrent SGA use among children/adolescents without comorbid ADHD, those without psychiatric hospitalization, and those without comorbid intellectual disability. Individuals with ADHD, who represented the largest clinical diagnostic category (>16% of the sample), experienced large proportional increases in concurrent SGA use; specifically, those without comorbid ADHD (40% of youth with ADHD overall) experienced a 15% increase (Table 3). Youth without comorbid intellectual disability, albeit only a small proportion of the sample (1.3%), experienced the largest proportional increase in concurrent SGA use (45%). The probability of concurrent SGA use among hospitalized children was as high as 0.242

JOURNAL OF THE AMERICAN ACADEMY OF C HILD & ADOLESCENT PSYCHIATRY VOLUME 53 NUMBER 9 SEPTEMBER 2014

www.jaacap.org

965

KREIDER et al.

FIGURE 2 Annual duration of psychotropic medication use and concurrent second-generation antipsychotic (SGA) use in days (2008). Rx ¼ treatment.

JOURNAL 966

www.jaacap.org

OF THE

AMERICAN ACADEMY OF C HILD & ADOLESCENT PSYCHIATRY VOLUME 53 NUMBER 9 SEPTEMBER 2014

CONCURRENT SGA USE IN CHILDREN

TABLE 3 Concurrent Second-Generation Antipsychotic (SGA) Use by Demographic and Clinical Characteristics in a Fee-for-Service Subsample Standardized Probability of Using SGA Concurrently With Another Psychotropic Class

% of Subsample Characteristic Demographic Sex Male Female Age group 6e11 12e18 Hospitalizations Psychiatric hospitalization No psychiatric hospitalization Diagnosis ADHD Noncomorbid Comorbid CD Noncomorbid Comorbid Developmental delay Noncomorbid Comorbid Depression Noncomorbid Comorbid Anxiety Noncomorbid Comorbid BD Noncomorbid Comorbid Intellectual disability Noncomorbid Comorbid Autism Noncomorbid Comorbid Schizophrenia Noncomorbid Comorbid All children in fee-for-service

2004 (n ¼ 586,951)

2008 (n ¼ 527,442)

2004

2008

Relative Change, %

54.11 45.89

54.83 45.17

0.041 0.023

0.046 0.025

12.8 10.2

46.48 53.52

47.13 52.87

0.028 0.036

0.031 0.041

10.3 12.9

1.54 98.46

1.61 98.39

0.242 0.030

0.242 0.034

0.0 13.4

6.17 8.36

6.51 9.97

0.106 0.280

0.122 0.304

14.6 8.8

2.17 6.88

2.16 7.77

0.106 0.314

0.103 0.321

2.3 2.2

2.76 3.30

3.04 4.13

0.011 0.153

0.010 0.160

9.6 4.7

1.27 4.38

1.06 3.97

0.126 0.299

0.106 0.302

15.9 0.9

0.47 1.52

0.63 2.27

0.033 0.205

0.032 0.216

3.5 5.4

0.27 2.10

0.36 2.58

0.519 0.658

0.540 0.670

4.1 1.8

1.30 1.98

1.28 2.23

0.042 0.268

0.061 0.283

44.9 5.4

0.64 1.41

1.00 2.40

0.200 0.372

0.180 0.383

9.5 3.1

0.05 0.33 100.00

0.04 0.27 100.00

0.439 0.679 0.033

0.556 0.651 0.036

26.5 4.1 11.8

Note: ADHD ¼ attention-deficit/hyperactivity disorder; BD ¼ bipolar disorder; CD ¼ conduct disorder.

but was stable over time; conversely, such use, although less prevalent among the 98% of youth who were not hospitalized, rose by 13% (to 0.034). There were also modest differences in the growth of concurrent SGA use between boys and girls (13% versus 10%) and older versus younger children (13% versus 10%).

DISCUSSION This national study found that between 2004 and 2008, Medicaid-enrolled children and adolescents increasingly used SGAs concurrently with other psychotropic medications—a trend supported by recent research among another population of publicly insured youth44—and that

JOURNAL OF THE AMERICAN ACADEMY OF C HILD & ADOLESCENT PSYCHIATRY VOLUME 53 NUMBER 9 SEPTEMBER 2014

www.jaacap.org

967

KREIDER et al.

this concurrent therapy was not short term. On average, 85% of children using SGAs used the SGA concurrently with other medications during the year. The findings indicate that the previously reported increase in SGA use is largely occurring in the setting of psychotropic polypharmacy. Determining the subgroups of children increasingly exposed to combination therapy can help to inform priorities around specific populations for whom efficacy and safety data for combination therapy are most needed. Although previous research demonstrated that children with ADHD account for the majority of SGA use among children in Medicaid,2 these results suggest that among this group, those children without comorbid ADHD were driving the increase in concurrent use of SGA through 2008. This polypharmacy practice also grew disproportionately among income-eligible children and occurred entirely among children in ambulatory settings, rather than among the small numbers of youth in hospital settings. The implication is that “medication add-on” is becoming the therapeutic option for clinically challenging youth in the public sector and is reaching beyond youth with serious mental illness in hospital settings. The aforementioned trends indicate a growing and pervasive use of polypharmacy and raise a number of practice implications. One implication is that the growth in concurrent SGA prescribing among all children, but particularly incomeeligible children and children without comorbid ADHD, suggests a changing trend in prescribing practice that increasingly favors concurrent SGA use in less-impaired youth. If this trend is replicated in other studies, it raises a concern about the growing reliance on polypharmacy to manage the emotional and behavioral disorders of children. The finding that concurrent SGA use is rising among children with ADHD is consistent with prior studies within the Medicaid program, which have characterized the high rates of SGA use among youth with ADHD and oppositional defiant disorder (ODD) or CD and suggested that polypharmacy trends are targeting off-label indications, such as aggression.2,45 Unfortunately, these data cannot identify the prescribers of concurrent SGA regimens; however, previous research has demonstrated that concurrent psychotropic therapies are prescribed to children by psychiatrists more often than by primary care physicians.16,33 Another implication relates to the significance of the long-term concurrent SGA use estimated in this study, in the absence of rigorous scientific

supporting evidence or data on the potential implications in terms of drug interactions. Although this study was not able to assess whether psychosocial therapies were being used concomitantly with SGA polypharmacy, the limited existing evidence suggests that evidencebased psychosocial therapies are used infrequently.46 The duration of use noted in this study raises concern that SGA polypharmacy is not being used as a short-term intervention for crisis response to challenging behaviors or in tandem with nonpharmacologic therapy. These practice trends are seemingly at odds with the American Academy of Child and Adolescent Psychiatry (AACAP) Practice Parameter guidelines that call for caution and a clear rationale before using combinations of psychotropic medications.47 Demonstrating that extended duration of concurrent SGA use is not uncommon has mixed implications for clinical practice. On one hand, the evidence of longer durations and consistent refills could be interpreted as evidence that families perceive the effectiveness of the medication when used in longer rather than shorter durations. On the other hand, such evidence needs to be weighed against concerns about metabolic effects that may be higher in children than in adults,14,48 and the potential that adverse effects might be potentiated in the presence of some medications (e.g., antidepressants) versus reduced in the setting of others (e.g., stimulants). At the very least, the long durations of concurrent SGA use illustrate that metabolic monitoring guidelines, long established for adults,49 are equally if not more important in children; however, routine monitoring for adverse events has not been common practice to date.50 There is also no infrastructure for systematic collection of such information. Finally, if the standard of care is moving in the direction of longerterm use, identifying the safest doses for such use and adequate routine monitoring protocols become urgent yet unresolved issues. This study is not without limitations. Although we report subanalyses of children in fee-forservice Medicaid arrangements for whom diagnostic data are more reliable, misclassification of clinical rationale and lack of specificity of clinical severity nevertheless remain concerns.41,42 Given large proportional increases among incomeeligible children, those without comorbid ADHD, and those without evidence of psychiatric hospitalization, we expect that much of this rationale relates to treatment of aggressive behaviors, but we are unable to confirm this with

JOURNAL 968

www.jaacap.org

OF THE

AMERICAN ACADEMY OF C HILD & ADOLESCENT PSYCHIATRY VOLUME 53 NUMBER 9 SEPTEMBER 2014

CONCURRENT SGA USE IN CHILDREN

chart review. Second, the data describe filled prescriptions and not actual consumption. Third, these data are aggregated across the country and do not describe regional or provider variation in concurrent SGA prescribing. Finally, these data could not allow us to report the entire duration of therapy because of left and right censoring of cross-sectional data, but the long annual durations that were observed are noteworthy and suggest sustained treatment when an SGA was prescribed. The present study illustrates that the trend of increasing SGA use, which is occurring in the context of stable or declining use of other medication, is due in large part to sustained concurrent use of SGAs with other medications, an exposure that is known to have serious side effects and unknown long-term effects and drug– drug interactions. Even more problematic is that the exposure to concurrent SGA is increasing disproportionately among youth with less perceived comorbidity and impairment. Such trends indicate a growth in off-label prescribing among children for whom evidence of benefit is lacking. In this context, the next decade of pediatric psychiatric prescribing will increasingly face the specter of federal and state policies that limit access to SGAs and impose greater oversight and monitoring requirements. As this occurs, it would be ideal if the development and implementation of state monitoring efforts were situated within a comprehensive behavioral health model that encourages multidisciplinary and collaborative approaches to treating youth with the most severely impairing conditions. It will also be important to

REFERENCES 1. Cooper WO, Arbogast PG, Ding H, Hickson GB, Fuchs DC, Ray WA. Trends in prescribing of antipsychotic medications for US children. Ambul Pediatr. 2006;6:79-83. 2. Matone M, Localio R, Huang YS, dosReis S, Feudtner C, Rubin D. The relationship between mental health diagnosis and treatment with second-generation antipsychotics over time: a national study of U.S. Medicaid-enrolled children. Health Serv Res. 2012;47:1836-1860. 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;63:679-685. 4. Patel NC, Crismon ML, Hoagwood K, et al. Trends in the use of typical and atypical antipsychotics in children and adolescents. J Am Acad Child Adolesc Psychiatry. 2005;44:548-556. 5. Zito JM, Burcu M, Ibe A, Safer DJ, Magder LS. Antipsychotic use by Medicaid-insured youths: impact of eligibility and psychiatric diagnosis across a decade. Psychiatr Serv. 2013;64:223-229. 6. Zito JM, Safer DJ, DosReis S, et al. Psychotropic practice patterns for youth: a 10-year perspective. Arch Pediatr Adolesc Med. 2003; 157:17-25. 7. 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;69:1247-1256.

pursue safety and efficacy trials that focus most specifically on youth in ambulatory settings who are receiving combination therapies for off-label treatment of disruptive behaviors. &

Accepted June 16, 2014. Mss. Kreider and Matone are with PolicyLab at the Children’s Hospital of Philadelphia. Dr. Bellonci is with Tufts University School of Medicine. Dr. dosReis is with the Department of Pharmaceutical Health Services Research at the University of Maryland School of Pharmacy. Drs. Feudtner and Rubin are with PolicyLab at the Children’s Hospital of Philadelphia, the Department of Pediatrics at the Perelman School of Medicine at the University of Pennsylvania, and the Division of General Pediatrics at the Children’s Hospital of Philadelphia. Ms. Huang is with the Division of General Pediatrics at the Children’s Hospital of Philadelphia. Dr. Localio is with the Department of Biostatistics and Epidemiology at the Perelman School of Medicine at the University of Pennsylvania. This study was funded through AHRQ R01 HS01855001A1. Data for this study were made available through Centers for Medicare and Medicaid Services agreements 20927 and 23593. Dr. Localio served as the statistical expert for this research. Disclosure: Dr. Bellonci is president of the American Association of Children’s Residential Centers and a member of the Corporation of Walker, a multiservice agency in Needham, MA (volunteer, not a paid role). He has also been or continues to be a consultant for the Annie E. Casey Foundation, Casey Family Services, Center for Healthcare Strategies, Children’s Rights, and the US Department of Justice. Dr. Rubin, Ms. Kreider, and Ms. Matone are consultants for the Commonwealth of Pennsylvania on psychotropic medication use in the state’s Medicaid-enrolled children, funded by a grant from Casey Family Programs. Drs. dosReis, Feudtner, Localio, and Ms. Huang report no biomedical financial interests or potential conflicts of interest. Correspondence to David M. Rubin, MD, MSCE, 34th St. and Civic Center Boulevard, CHOP North, Room 1533, Philadelphia, PA 19104; e-mail: [email protected] 0890-8567/$36.00/ª2014 American Academy of Child and Adolescent Psychiatry http://dx.doi.org/10.1016/j.jaac.2014.05.010

8. Andrade SE, Lo JC, Roblin D, et al. Antipsychotic medication use among children and risk of diabetes mellitus. Pediatrics. 2011;128: 1135-1141. 9. Correll CU. Monitoring and management of antipsychotic-related metabolic and endocrine adverse events in pediatric patients. Int Rev Psychiatry. 2008;20:195-201. 10. Correll CU, Carlson HE. Endocrine and metabolic adverse effects of psychotropic medications in children and adolescents. J Am Acad Child Adolesc Psychiatry. 2006;45:771-791. 11. Correll CU, Manu P, Olshanskiy V, Napolitano B, Kane JM, Malhotra AK. Cardiometabolic risk of second-generation antipsychotic medications during first-time use in children and adolescents. JAMA. 2009;302:1765-1773. 12. De Hert M, Dobbelaere M, Sheridan EM, Cohen D, Correll CU. Metabolic and endocrine adverse effects of second-generation antipsychotics in children and adolescents: a systematic review of randomized, placebo controlled trials and guidelines for clinical practice. Eur Psychiatry. 2011;26:144-158. 13. Panagiotopoulos C, Ronsley R, Davidson J. Increased prevalence of obesity and glucose intolerance in youth treated with secondgeneration antipsychotic medications. Canadian J Psychiatry. 2009;54:743-749.

JOURNAL OF THE AMERICAN ACADEMY OF C HILD & ADOLESCENT PSYCHIATRY VOLUME 53 NUMBER 9 SEPTEMBER 2014

www.jaacap.org

969

KREIDER et al.

14. Safer DJ. A comparison of risperidone-induced weight gain across the age span. J Clin Psychopharmacol. 2004;24:429-436. 15. Sikich L, Hamer RM, Bashford RA, Sheitman BB, Lieberman JA. A pilot study of risperidone, olanzapine, and haloperidol in psychotic youth: a double-blind, randomized, 8-week trial. Neuropsychopharmacology. 2004;29:133-145. 16. Comer JS, Olfson M, Mojtabai R. National trends in child and adolescent psychotropic polypharmacy in office-based practice, 1996-2007. J Am Acad Child Adolesc Psychiatry. 2010;49:1001-1010. 17. dosReis S, Zito JM, Safer DJ, Gardner JF, Puccia KB, Owens PL. Multiple psychotropic medication use for youths: a two-state comparison. J Child Adolesc Psychopharmacol. 2005;15:68-77. 18. Zito JM, Safer DJ, Sai D, et al. Psychotropic medication patterns among youth in foster care. Pediatrics. 2008;121:e157-e163. 19. Pathak S, Johns ES, Kowatch RA. Adjunctive quetiapine for treatment-resistant adolescent major depressive disorder: a case series. J Child Adolesc Psychopharmacol. 2005;15:696-702. 20. Emiroglu FN, Gencer O, Ozbek A. Adjunctive olanzapine treatment in bipolar adolescents responding insufficiently to mood stabilizers. Four case reports. Eur Child Adolesc Psychiatry. 2006; 15:500-503. 21. Delbello MP, Schwiers ML, Rosenberg HL, Strakowski SM. A double-blind, randomized, placebo-controlled study of quetiapine as adjunctive treatment for adolescent mania. J Am Acad Child Adolesc Psychiatry. 2002;41:1216-1223. 22. Kafantaris V, Coletti DJ, Dicker R, Padula G, Kane JM. Adjunctive antipsychotic treatment of adolescents with bipolar psychosis. J Am Acad Child Adolesc Psychiatry. 2001;40:1448-1456. 23. Pavuluri MN, Henry DB, Carbray JA, Sampson G, Naylor MW, Janicak PG. Open-label prospective trial of risperidone in combination with lithium or divalproex sodium in pediatric mania. J Affect Disord. 2004;82(Suppl 1):S103-S111. 24. Unwin GL, Deb S. Efficacy of atypical antipsychotic medication in the management of behaviour problems in children with intellectual disabilities and borderline intelligence: a systematic review. Res Dev Disabil. 2011;32:2121-2133. 25. Aman MG, Bukstein OG, Gadow KD, et al. What does risperidone add to parent training and stimulant for severe aggression in child attention-deficit/hyperactivity disorder? J Am Acad Child Adolesc Psychiatry. 2014;53:47-60. 26. Linton D, Barr AM, Honer WG, Procyshyn RM. Antipsychotic and psychostimulant drug combination therapy in attention deficit/ hyperactivity and disruptive behavior disorders: a systematic review of efficacy and tolerability. Curr Psychiatry Rep. 2013;15:355. 27. Colley CA, Lucas LM. Polypharmacy: the cure becomes the disease. J Gen Intern Med. 1993;8:278-283. 28. Duffy FF, Narrow WE, Rae DS, et al. Concomitant pharmacotherapy among youths treated in routine psychiatric practice. J Child Adolesc Psychopharmacol. 2005;15:12-25. 29. Dufresne RL. Issues in polypharmacotherapy: focus on major depression. Psychopharmacol Bull. 1996;32:547-553. 30. Gorard DA. Escalating polypharmacy. Q J Med. 2006;99:797-800. 31. Kingsbury SJ, Yi D, Simpson GM. Psychopharmacology: rational and irrational polypharmacy. Psychiatr Serv. 2001;52:1033-1036. 32. Martınez-Mir I, Garcıa-L opez M, Palop V, Ferrer JM, Rubio E, Morales-Olivas FJ. A prospective study of adverse drug reactions in hospitalized children. Br J Clin Pharmacol. 1999;47:681-688.

33. Safer DJ, Zito JM, DosReis S. Concomitant psychotropic medication for youths. Am J Psychiatry. 2003;160:438-449. 34. Stahl SM. Focus on antipsychotic polypharmacy: evidence-based prescribing or prescribing-based evidence? Int J Neuropsychopharmacol. 2004;7:113-116. 35. Turner S, Nunn AJ, Fielding K, Choonara I. Adverse drug reactions to unlicensed and off-label drugs on paediatric wards: a prospective study. Acta Paediatr. 1999;88:965-968. 36. Crystal S, Olfson M, Huang C, Pincus H, Gerhard T. Broadened use of atypical antipsychotics: safety, effectiveness, and policy challenges. Health Aff. 2009;28:w770-w781. 37. Coyer C, Kenney GM. The composition of children enrolled in Medicaid and CHIP: variation over time and by race and ethnicity. Washington, DC: Urban Institute; 29 March 2013. 38. Raghavan R, Zima BT, Andersen RM, Leibowitz AA, Schuster MA, Landsverk J. Psychotropic medication use in a national probability sample of children in the child welfare system. J Child Adolesc Psychopharmacol. 2005;15:97-106. 39. Zito JM, Safer DJ, Zuckerman IH, Gardner JF, Soeken K. Effect of Medicaid eligibility category on racial disparities in the use of psychotropic medications among youths. Psychiatr Serv. 2005;56: 157-163. 40. dosReis S, Zito JM, Safer DJ, Soeken KL. Mental health services for youths in foster care and disabled youths. Am J Public Health. 2001;91:1094-1099. 41. Byrd VLH, Dodd AH. Assessing the usability of encounter data for enrollees in comprehensive managed care across MAX 2007–2009. Washington, DC: Mathematica Policy Research; December 2012. 42. Nysenbaum J, Bouchery E, Malsberger R. The availability and usability of behavioral health organization encounter data in MAX 2009. Washington, DC: Mathematica Policy Research; July 2013. 43. Graubard BI, Korn EL. Predictive margins with survey data. Biometrics. 1999;55:652-659. 44. Murphy AL, Gardner DM, Cooke C, Kisely S, Hughes J, Kutcher SP. Prescribing trends of antipsychotics in youth receiving income assistance: results from a retrospective population database study. BMC Psychiatry. 2013;13:198. 45. dosReis S, Yoon Y, Rubin DM, Riddle MA, Noll E, Rothbard A. Antipsychotic treatment among youth in foster care. Pediatrics. 2011;128:e1459-e1466. 46. Garland AF, Brookman-Frazee L, Hurlburt MS, et al. Mental health care for children with disruptive behavior problems: a view inside therapists’ offices. Psychiatr Serv. 2010;61:788-795. 47. Findling RL, Drury SS, Jensen PS, et al. Practice parameter for the use of atypical antipsychotic medications in children and adolescents. J Am Acad Child Adolesc Psychiatry. In press. 48. Bobo WV, Cooper WO, Stein CM, et al. Antipsychotics and the risk of type 2 diabetes mellitus in children and youth. JAMA Psychiatry. 2013;70:1067-1075. 49. American Diabetes Association, American Psychiatric Association, American Association of Clinical Endocrinologists, North American Association for the Study of Obesity. Consensus development conference on antipsychotic drugs and obesity and diabetes. J Clin Psychiatry. 2004;65:267-272. 50. Morrato EH, Nicol GE, Maahs D, et al. Metabolic screening in children receiving antipsychotic drug treatment. Arch Pediatr Adolesc Med. 2010;164:344-351.

JOURNAL 970

www.jaacap.org

OF THE

AMERICAN ACADEMY OF C HILD & ADOLESCENT PSYCHIATRY VOLUME 53 NUMBER 9 SEPTEMBER 2014

CONCURRENT SGA USE IN CHILDREN

SUPPLEMENT 1: TECHNICAL APPENDIX Identifying Underreporting in Managed Care To identify children in managed care, we used the following variables in the MAX data: EL_PHP_TYPE_1_1 - EL_PHP_TYPE_1_12; EL_ PHP_TYPE_2_1 - EL_PHP_TYPE_2_12; EL_PHP_ TYPE_3_1 - EL_PHP_TYPE_3_12; and EL_PHP_ TYPE_4_1 - EL_PHP_TYPE_4_12. These identify, for each month of a child’s enrollment in Medicaid, the types (up to 4) of prepaid plans for which the child was eligible. If, for at least 1 month in the year, 1 of these 4 variables for a child had the value 01, or “eligible is enrolled in a medical or comprehensive managed care plan this month (e.g., HMO),” we considered that child to be enrolled in a managed care plan. A sensitivity analysis revealed that the majority of children who had at least 1 month of enrollment in a managed care plan were enrolled in managed care for 10 or more months of the year. We then attempted to identify major discrepancies between medication rates in fee-for-service and managed care. Ultimately, we identified 6 states in which the medication rates within at least 1 eligibility group were at least 5 times higher in feefor-service than in managed care. These were the District of Columbia (DC), Florida, Hawaii, Nevada, Ohio, and Pennsylvania. The largest discrepancies occurred within the “Temporary Assistance for Needy Families (TANF)/other” eligibility group, which is of particular concern because it makes up the majority of Medicaid-eligible children. These states additionally had some of the lowest overall medication rates in managed care (20th percentile). Although we accepted that there may be a degree of underreporting in other states, we hoped that by excluding states with what appeared to be severe underreporting from the primary analysis, we would reduce bias in our models.

disorders because of conditions classified elsewhere (293, 294); delusional disorders (297); other nonorganic psychoses (298); dissociative and somatoform disorders (300.10–300.19, 300.30– 300.99); personality disorders (301.10–301.30, 301.50–301.99); special symptoms or syndromes, not elsewhere classified (307); acute reaction to stress (308); adjustment reaction (309); and disturbance of emotions specific to childhood and adolescence (313.90–313.99). A separate covariate was identified for children who received a diagnosis of seizure disorder (345) to control for the overlapping use of anticonvulsant agents for mood stabilization in this population. TABLE S1

SUPPLEMENT 2: DIAGNOSES Ten mental health diagnostic categories were identified in the DSM-IV-TR and coded using the ICD-9 classification: schizophrenia (295), bipolar disorder (BD; 296.00–296.10, 296.36–296.89), depression (296.20–296.35, and 311), anxiety disorder (300.00–300.29 and 301.4), conduct disorder (CD; 312.00–313.89), autism (299), attentiondeficit/hyperactivity disorder (ADHD; 314), intellectual disability (formerly known as mental retardation; 317–319), developmental delay (315), and a composite variable of miscellaneous mental health diagnoses inclusive of the following: mental JOURNAL OF THE AMERICAN ACADEMY OF C HILD & ADOLESCENT PSYCHIATRY VOLUME 53 NUMBER 9 SEPTEMBER 2014

Sampling Frequencies by State and Eligibility

State Alaska Alabama Arkansas Arizona Colorado District of Columbia Delaware Hawaii Iowa Idaho Indiana Kansas Kentucky Louisiana Minnesota Missouri Mississippi Montana North Dakota Nebraska New Hampshire New Jersey New Mexico Nevada Oklahoma Oregon Rhode Island South Carolina South Dakota Utah Vermont Washington Wisconsin West Virginia Wyoming

K_FC

K_SSI

K_OTH

1 1 1 2.25024 1.54088 1 1 1 1.28536 1 1.97504 1.36688 2.79624 2.83728 1.5888 2.583511 1 1 1 1 1 2.99 1 1 1.49864 1.800529 1 2.26184 1 1 1 2.19648 2.5716 1 1

1 3.565956 1.89301 2.25024 1.54088 1 1 1 1.28536 1 1.97504 1.36688 2.79624 2.83728 1.5888 1 2.369263 1 1 1 1 2.99 1 1 1.49864 1 1 2.26184 1 1 1 2.19648 2.5716 1.229634 1

1.392684 13.52672 16.17504 17.9008 6.53368 2.131896 1.462024 2.663924 6.20808 3.301055 19.64312 4.39216 13.54216 29.42416 11.23096 21.25288 10.95888 1 1 3.856272 1.53831 18.02464 8.78742 1.999467 14.40168 5.58016 2.852407 16.94256 1.438564 1.844971 1.394804 21.11768 13.31336 6.01224 1

Note: K_ refers to sampling weight for children with Medicaid eligibility based on foster care (FC), Supplemental Security Income (SSI), and Temporary Assistance for Needy Families (TANF)/other (OTH).

www.jaacap.org 970.e1

KREIDER et al.

FIGURE S1 Trends in psychotropic medication use by class (Panel A) and with concurrent antipsychotics (Panel B). Note: SSI ¼ Supplemental Security Income; TANF ¼ Temporary Assistance for Needy Families.

JOURNAL 970.e2 www.jaacap.org

OF THE

AMERICAN ACADEMY OF C HILD & ADOLESCENT PSYCHIATRY VOLUME 53 NUMBER 9 SEPTEMBER 2014