Prevalence of Psychotherapy Surrounding Initiation of Psychotropic Polypharmacy in the Medicaid-Insured Population, 1999–2010

The 2014 National Survey on Drug Use and Health (NSDUH) estimated that 18.1% of the United States adult population had a mental illness in the past year (approximately 43.6 million people) (1). The prevalence of mental illness in the past year was estimated at 15.4% among adults with private insurance and at almost twice that amount, 29.7%, among persons ages 18 and older who are insured by Medicaid. Among the U.S. pediatric population, an estimated 13% of children between the ages of eight and 15 had any mental disorder in the previous year, and 46.3% of 13- to 18-year-olds had lifetime prevalence of any mental disorder (2).

Published guidelines for the management of the most common psychiatric disorders consistently suggest providing initial behavioral or psychotherapy followed by concomitant pharmacotherapy (3–6). This guidance is especially appropriate for pediatric patients, for whom behavioral therapy is recommended as the first line of treatment for conditions such as attention-deficit hyperactivity disorder (ADHD), depression, and conduct disorder (7–9). Combining medication use with psychotherapy in the treatment of psychiatric disorders is strongly recommended, based on a diverse body of literature that shows that combined treatment has higher effectiveness on symptom control compared with only pharmacotherapy. For treatment of ADHD among adults and children, for example, multiple studies consistently report on the substantial contribution of behavioral therapy strategies, such as cognitive-behavioral therapy and family or group therapy, in achievement of symptom resolution, improvement of social behavior, and in some cases decreased need for pharmacotherapy augmentation (10–12). Notwithstanding these recommendations, multiple reports on the treatment of psychiatric disorders continue to show overutilization of psychotropic drugs, including increasing trends in use of psychotropic medication polypharmacy (PMP) (13–20).

PMP is defined as the use of two or more psychotropic medications from the same or different drug classes, and it is one of the most common patterns of prescribing in the United States (21). To illustrate, an analysis of a national sample of ambulatory care visits identified that the percentage of outpatient psychiatry visits for which two or more psychotropic drugs were prescribed increased from 42.6% in 1997 to 59.8% in 2006 (22). In contrast, it has been estimated that the use of psychotherapy is well below that of PMP. In a recent study of Medicaid-insured children with ADHD, 7% of patients received psychotherapy treatment alone and 32% received both medication and psychotherapy (23).

Although evidence of the prevalence of PMP and trends toward increased use of PMP among various patient populations is abundant, to date no study has evaluated the prevalence of psychotherapy relative to the initiation of PMP. Given that pharmacotherapy alone might not be sufficient to achieve desired outcomes among many patients with psychiatric disorders, efforts should be made to evaluate the extent to which ancillary treatment strategies are adopted and provide information about patterns of utilization of these strategies. The initiation of PMP among patients receiving psychotropic monotherapy might indicate that patients require additional treatment strategies. Therefore, this period may be particularly important for optimizing the care of Medicaid patients with mental illness. The aim of this study was to quantify the prevalence of psychotherapy among Medicaid pediatric and adult patients within six months of initiating PMP.

Methods

Study Design and Data Sources

We conducted a retrospective, cross-sectional analysis of a cohort of Medicaid-insured pediatric and adult patients ages 64 and younger to determine the prevalence of psychotherapy in the six months before and after the initiation of PMP. Prevalence was estimated in two-year periods and stratified by age group and state. The data set for this study was established from the Medicaid Analytic eXtract (MAX) files from 1999 to 2010 for 29 states. This large database contains individual-level information on monthly Medicaid enrollment, demographic variables, and adjudicated inpatient and outpatient medical encounter claims as well as detailed prescription dispensing data. The 29 states selected for the study are those with the largest fee-for-service (FFS) population, representing over 80% of the entire Medicaid FFS population, per 2002 estimates. The study was approved by the institutional review and privacy boards of the University of Florida and the privacy office of the Centers for Medicare and Medicaid Services.

Study Population

Patients were selected for the study if they were enrolled in Medicaid under FFS or Primary Care Case Management (PCCM). We excluded patients enrolled in comprehensive managed care because of concerns about the completeness and usability of claims for research purposes (24). We established ten cohorts of patients between the ages of zero and 64 who were enrolled in Medicaid between 1999 and 2010 and who had two years of continuous enrollment after receipt of a prescription for a psychotropic medication, resulting in ten two-year blocks. For each two-year block, we identified the first dispensing of a psychotropic medication (cohort entry date) and required each qualifying patient to have continuous enrollment for at least two years thereafter. We further restricted our study population to patients who started PMP between six months after cohort entry and six months before cohort exit (the end of the two-year follow-up period). This strategy ensured that prevalence estimates for psychotherapy reflected the enrollment period that was included in the assigned two-year block (Figure 1). Patients could be included in multiple blocks if they met all inclusion criteria. For additional stratified analyses by age and state, we assigned patients to a group by using information from the earliest two-year block. For example, a child in Florida who entered Medicaid at age 5 during 2001 and remained continuously enrolled until 2006 would contribute data only for the Florida and five-and-younger groups.

Data Analysis

For each patient from the two-year cohorts, we extracted all prescription pharmacy claims for all psychotropic medications with FDA approval for treatment of psychiatric disorders. Drug classes included alpha agonists, antidepressants, antipsychotics, lithium, anticonvulsants, and sedatives/hypnotics/anxiolytics. Medications were identified by National Drug Code, and the duration of active pharmacotherapy was determined from the dispensing date plus the reported days of supply listed in the billing claim and a ten-day grace period to account for late refills. For drugs with missing or erroneous information on days’ supply, we imputed this value by using the calculated median of days’ supply for all claims for the drug from our data. PMP was defined by an overlap of greater than 45 days in the active periods of two or more psychotropic medications with different active ingredients. This overlap method is consistent with previous studies published on polypharmacy, and it is a generally accepted threshold (25). The first fill date of the second medication contributing to the overlap definition was considered to be the date of PMP initiation.

Prevalence of psychotherapy was estimated as the proportion of PMP patients with a claim for psychotherapy within six months before and within six months after the start of PMP, respectively. We determined psychotherapy from selected outpatient and inpatient claims by using Current Procedure Terminology (CPT) and Healthcare Common Procedure Coding System (H) codes (26). We included CPT codes for individual psychotherapy (90804–90829) and psychotherapy with family members and group settings (90857–90876) and selected H codes for services used in billing for provision of behavioral, psychosocial, or psychiatric support services. [A list of codes used is available as an online supplement to this article.] All data management and statistical analyses were conducted with SAS, version 9.4, and ArcGIS 10.2.2.

Results

Pediatric Population

Most of the included pediatric patients were male and nonwhite (Table 1). The decline in the number of patients in the sample size for each cohort in the last years of the study is consistent with the shift from FFS and PCCM to managed care programs, a shift that led to increasing exclusion of Medicaid beneficiaries from our analysis. Approximately 10% of children included in each of the cohorts were in foster care, and more than 30% required cash assistance. More than 20% of children across the cohorts received Medicaid because of disability, and there was an increasing trend in the number of pediatric patients who qualified on the basis of poverty. There was an increasing trend in the prevalence of PMP among children who received at least one psychotropic medication, with estimates ranging from 21.2% to 27.7% between 1999–2001 and 2008–2010 (results presented elsewhere) (27).

Among children who initiated PMP (N=397,728), we found increasing prevalence in the use of prior psychotherapy across our study period, starting with 21% of patients in the 1999–2000 cohort to a maximum of 40% of patients in the 2005–2007 cohort (Figure 2). From 1999–2001 to 2008–2010, there was an 11% increase in the prevalence of psychotherapy before PMP initiation and a 12% increase in the prevalence of psychotherapy after the start of PMP. Pediatric cohorts consistently showed higher prevalence of psychotherapy after PMP initiation than before PMP initiation, with estimates of prevalence of psychotherapy after PMP initiation ranging between 25.4% in the 1999–2000 cohort to 44.05% in the 2005–2007 cohort (Figure 2) [see online supplement].

Adult Population

Baseline characteristics for adult Medicaid-insured patients differed from those of the pediatric population (Table 2). For example, the majority of adult patients who used at least one psychotropic drug claim were female rather than male, the proportion of patients who qualified for Medicaid benefits because of disability was larger, and the vast majority received cash assistance versus 47% or less among children. The prevalence of PMP among patients who received at least one psychotropic medication was consistently higher among adults than among pediatric cohorts, ranging from 48.5% in the 1999–2001 cohort to 58% in the 2007–2009 group (results presented elsewhere) (28).

When evaluating the prevalence of psychotherapy among adult PMP initiators (N=773,205), we saw that at best 16.7% received this treatment strategy in the six months prior to the start of PMP (Figure 2). Prevalence of psychotherapy after the start of PMP was only slightly higher, ranging from 12.6% in the 1999–2001 cohort to 19.3% in the 2003–2005 block. Among adults, the increase in prevalence of psychotherapy from 1999–2001 to 2008–2010, both before (2%) and after (3%) PMP start, was smaller than among pediatric patients [see online supplement].

Age Group and State-Level Analyses

Results for the age-stratified analysis in Table 3 show that pediatric patients between the ages of six and 14 had the highest prevalence of psychotherapy both before (33%) and after (40%) initiation of PMP. The prevalence of psychotherapy among adults remained lower than among pediatric patients and showed a decreasing trend with increasing age. The prevalence of psychotherapy before PMP initiation declined from 19.9% among patients ages 18 to 29 to 10.3% in the 50- to 64-year-old group. Similarly, prevalence of psychotherapy after start of PMP decreased from 26.0% to 13.0% in the same age groups.

We observed significant variation in the prevalence of psychotherapy surrounding PMP initiation at the state level [see online supplement]. We suppressed results for six of the 29 states in our data set because of small sample size and unreliable prevalence estimates. Among children, prevalence of psychotherapy prior to PMP initiation was highest in Indiana, Kansas, and Missouri. Among adults, the prevalence was highest in Indiana, Kansas, and New Jersey [see online supplement].

Discussion

We aimed to quantify the prevalence of psychotherapy use among Medicaid-insured patients within six months of initiating PMP. Our analysis revealed several important findings. First, utilization of psychotherapy surrounding initiation of polypharmacy was low for both pediatric and adult Medicaid-insured patients. This finding suggests a potential overreliance on the use of medications, including the practice of polypharmacy, and underutilization of nonpharmacological modalities indicated for the treatment of mental illness. Our results suggest that among pediatric patients who started PMP, fewer than half received any kind of psychotherapy in the six months surrounding the addition of one or more psychotropic drugs to their pre-existing pharmacotherapy regimen. The six-month period after initiation of the first psychotropic medication is a time in which patients’ monotherapy treatment should be optimized. As previously pointed out, psychotherapy can play an important role in achieving desired clinical end points and its use might prevent the start of polypharmacy in some cases.

Second, the proportion of pediatric patients who received psychotherapy was lower than recommended in clinical guidelines both before and after the start of PMP, although it was higher than the proportion among adults. For example, the American Academy of Family Physicians recommends behavioral therapy as a component of treatment for depression among children and adolescents, especially among those with moderate to severe depression (8). Likewise, the guidelines for treatment of ADHD from the American Academy of Pediatrics strongly recommend that behavioral therapy ought to be the first line of treatment for patients younger than five years of age and should also be given in combination with pharmacotherapy for elementary school–aged children. Besides patients with depression and ADHD, psychotherapy is also recommended for pediatric patients presenting with other conditions, such as conduct disorders, anxiety, and posttraumatic stress disorder, among others (9,29,30). Among all Medicaid-insured children, the prevalence of psychotherapy before and after PMP initiation was lowest for patients five years of age and younger. This finding is concerning, given that psychotherapy is strongly and consistently recommended for pediatric patients with psychiatric disorders and, more important, for the youngest of these patients. Although we found an increasing trend in prevalence in the use of psychotherapy surrounding PMP initiation, the increase was not great, and the gap between recommended mental health care and utilization remains a concern.

Polypharmacy safety issues are not negligible and should be considered in the benefit-risk assessment of a pharmacotherapy regimen prior to the start of treatment. PMP carries inherent risks, given that patients are exposed to multiple drugs with distinct side effect profiles and potential for interactions. Common drug-drug interactions among psychotropic medications have been found to lead to respiratory depression, risk of cardiac arrhythmias through QT prolongation, depression of the central nervous system, and manifestation of serotonin syndrome (31,32). In light of the potential for these safety concerns, the disconnect between the underutilization of effective evidence-based psychotherapy approaches and overutilization of highly risky polypharmacy practices needs to be explored.

In order to achieve the appropriate equilibrium between pharmacological treatment and psychotherapy for mental disorders, both types of treatment must be accessible to patients (33). Recent reports found that the number of psychiatrists practicing in the United States did not increase between 2000 and 2013, and, if adjusted for population size, the mean number of psychiatrists actually decreased by almost 10% in this period (34). Adding to these shortages, clinics may refuse to take Medicaid-insured patients because of unsatisfactory reimbursement agreements. Psychiatrists in fact have been found to be the least likely medical specialty to accept insurance plans from Medicaid (35). Primary care providers are left with the responsibility to address complex mental health care needs but may not have the training or the resources required for proper psychotherapeutic management of these conditions (36). Future studies should aim at evaluating the specific barriers and challenges faced by Medicaid-insured patients in obtaining psychotherapy and determine what groups benefit the most from psychotherapy in order to optimize resource utilization. Similarly, psychotherapy might not be initiated for certain patients for multiple reasons, and future research should also evaluate the conditions that lead to underutilization of this treatment strategy.

Our study is the first to evaluate the prevalence of psychotherapy in relationship to initiation of polypharmacy. Previous studies have reported prevalence of these two treatment modalities separately, but no information to date has been provided for a stage in patients’ mental health treatment when simple psychotropic monotherapy is not adequate and a more comprehensive approach appears necessary. We used a large database that includes data for all medical encounters for pediatric and adult patients enrolled in Medicaid programs in 29 states, which allowed age-specific and spatial analyses of treatment disparities.

Aside from these strengths, results should be carefully interpreted in light of the data used. First, we used only Medicaid data, and results should be generalized only to patients insured under this program. Use of psychotherapy and other metrics of adequate care have shown pronounced differences between Medicaid and privately insured populations. It is important to note that besides factors related to physicians and patients in choosing psychotherapy, Medicaid enrollees may face additional barriers in receiving adequate mental health care. A study by Melfi and colleagues (37) evaluating treatment modalities for depression in an adult population showed that whereas 45.0% of patients with private insurance received any kind of psychotherapy, the same was true for only 20.0% of patients insured by Medicaid.

Second, psychotherapy use was ascertained from billing codes; thus encounters that were paid out of pocket might have been missed. In addition to CPT codes, however, our study also included H codes, which have not been considered in previous studies and which should aid in comprehensive capture of all related services that were reimbursed by Medicaid (23). Pediatric patients might also receive psychotherapy through participation in school-based programs, for example, and these too would not be captured by the claims data.

Third, our analyses included results for only Medicaid patients in the FFS and PCCM programs and therefore cannot provide assessments of patients enrolled in managed care programs or in plans with mental health carve-outs, both increasingly common in Medicaid benefit schemes (38). Finally, our state analysis showed some states with very low prevalence of psychotherapy surrounding PMP [see online supplement]. This finding might reaffirm the issue raised earlier on the move toward managed care and carving out of mental health services in some states that might reflect in low capture of psychotherapy from our data. Notwithstanding these limitations, this study suggests underutilization of psychotherapies among Medicaid-insured patients receiving psychiatric polypharmacy.

Conclusions

Although prevalence of psychotherapy immediately before and after initiation of psychotropic polypharmacy among Medicaid-insured patients has slightly increased in the past decade, it remains low, particularly among adults. Future research should determine barriers that are limiting the use of psychotherapy and potential treatment strategies to counter increasing psychotropic polypharmacy trends.