Psychotropic Prescribing for Persons With Intellectual Disabilities and Other Psychiatric Disorders

Individuals with intellectual disabilities are often prescribed psychotropic medication to treat a comorbid psychiatric disorder or challenging behaviors, such as aggression, self-injurious behavior, property destruction, and pica. The prevalence of morbid mental illness for people with intellectual disabilities ranges from 15% to 50% (1–3). There is limited to moderate support for the efficacy of using psychotropics to treat psychopathology in this population (4–6). Whereas only 3% of this population has a psychosis-related disorder, an estimated 30%−50% of all psychotropics prescribed for this population are antipsychotics (7). A recent large-scale study of the use of psychotropics among adults with intellectual disabilities, including those with autism spectrum disorders, found a shift in practice from using psychotropics primarily to control aggressive behavior to using these medications to treat diagnosed psychiatric disorders (8). A study from 2012 examined rates of antipsychotic prescription for individuals with developmental disabilities living in the community who had experienced at least one psychiatric crisis in the past year. Close to half (49%) were prescribed antipsychotics, with 22% of that group prescribed two or more antipsychotics concurrently (9).

The prevalence of challenging behaviors among people with intellectual disabilities is reported to range from 8% to 62%; 14% to 30% are treated with psychotropic agents to manage maladaptive behaviors (5,10,11). A Dutch study of the prevalence of and rationale for antipsychotic drug use in a population of adults with intellectual disabilities found that almost a third of those studied were given antipsychotic medications. Behavioral problems were the most common reason for use (58%), whereas psychosis accounted for only 25%.

Limited empirical support for using psychotropics to treat challenging behaviors, coupled with the growing use of such agents, polypharmacy, and risk of adverse effects in a vulnerable population, are all noted concerns (4,6,8,11–15). Although a 2009 review found risperidone to be effective in reducing the frequency of challenging behaviors, controlled studies are limited (5). Sevin and colleagues (6) highlighted the need to integrate behavioral and pharmacological interventions into the treatment of persons with codiagnoses of an intellectual disability and a psychiatric disorder. The World Psychiatric Association international guide stresses the importance of identifying and addressing the underlying cause of the behavior among adults with intellectual disabilities and recommends, in the absence of a treatable general medical or psychiatric disorder, employing behavioral strategies (11).

Studies of the use of psychotropics for children with intellectual disabilities often include a broad range of developmental disorders as well as autism spectrum disorders. A 2005 study of pediatric Medicaid recipients with disabilities examined the use of stimulants, antidepressants, anticonvulsants, and antipsychotics and found variability in the use of four classes of medication across diagnostic groups as well as differences across categories of Medicaid assistance (16). One British survey found that the most frequently prescribed classes of medications for children with intellectual disabilities were stimulants and antipsychotics (17). A 2011 review of placebo-controlled randomized double-blind studies concluded that at least four studies found risperidone to be effective in reducing problem behaviors of children with intellectual disabilities and with or without autism (18).

Few studies in the past decade have compared prescribing patterns of psychotropic medication among individuals with codiagnoses with those among individuals with a sole diagnosis of a mental disorder. Some older studies revealed a practice of prescribing antipsychotic medications for those with intellectual disabilities, with little use of antidepressants among the codiagnosis population compared with those with a mental diagnosis (19,20). A 2003 retrospective study found no significant differences in polypharmacy prescribing between individuals with intellectual disabilities living in the community and patients with a sole mental disorder. Individuals with intellectual disabilities were more frequently prescribed antipsychotics and mood stabilizers and less frequently prescribed antidepressants and anxiolytics than those without intellectual disabilities (21).

Because of concerns about widespread use of psychotropic medication among individuals with an intellectual disability, a public behavioral health managed care organization designed a study at the request of a stakeholder county to examine the prescription rates of psychotropic medication for county citizens with codiagnoses of intellectual disability and mental disorder. We examined prescribing patterns for Medicaid recipients in a single county over five years by comparing prescription rates for adults and children with a codiagnosis of an intellectual disability and mental disorder with rates for those with a sole diagnosis of a mental disorder. We compared prescription patterns for four classes of medication between the two groups, as well as the pattern of prescribing three or more psychotropic medications concurrently.

Methods

Sample

Data from Medicaid recipient members from a public behavioral health managed care organization within a county in a mid-Atlantic state were examined for five consecutive fiscal years (FYs 2007–2011). Criteria for study inclusion were as follows: 80% Medicaid eligibility from the start of FY 2007 to the end of FY 2011, defined as eligibility to receive Medicaid-funded services for at least 292 days for the particular year, and having at least one behavioral health claim in each of the five years.

The members who met inclusion criteria for the study were split into two groups. Persons with mild, moderate, or severe intellectual disability who also had mental health diagnoses constituted the codiagnosis group, whereas those without a diagnosed intellectual disability made up the group solely with mental disorders.

A sequential process was used to assign members to groups. First, members with a diagnosis of mild, moderate, or severe intellectual disability were identified (N=1,350). Next, the number of days of Medicaid eligibility were computed, and the sample was reduced to include only those with 80% eligibility for all five fiscal years (N=1,095). The final codiagnosis sample was obtained by identifying members with at least one behavioral health claim in each of the five years (N=325). The same process was used to form the group solely with mental disorders; 69,005 members in the county did not have a diagnosed intellectual disability. Of these, 14,271 members were eligible for Medicaid during 80% of each fiscal year, and 1,521 had at least one behavioral health claim in each of the five years. The need for study approval was waived by the institutional review board.

Diagnoses

Paid behavioral health service claims were searched to extract diagnostic categories for both groups. The number of total primary, secondary, and tertiary diagnoses from the claims information was calculated for each member, and these diagnoses were grouped into broader categories. For example, conduct disorder included conduct disorder, oppositional defiant disorder, and impulse control disorder. Because all claims for a member were utilized, a member could be counted in more than one diagnostic category.

Psychotropic utilization and polypharmacy

Psychotropic utilization and polypharmacy were examined in paid pharmacy claims data for each group. Members with a filled prescription for agents from one of four therapeutic classes (antidepressants, antipsychotics, benzodiazepines, and mood stabilizers) were identified for each fiscal year. Binary variables were created to represent fills (coded 1) or no fills (coded 0) for the specific fiscal year; therefore, each member could have between one and four observations per year. Polypharmacy was defined as any member prescribed three or more psychotropic medications for 90 days or more per year in any of the five study years. Binary variables were created to represent three or more prescriptions (coded 1) or fewer than three prescriptions (coded 0).

Analyses

The two groups were stratified by age, with members 18 years or older in the adult category and members under age 18 assigned to the child category. Diagnostic categories and rates of psychotropic utilization and polypharmacy were examined by group and separately for adults and children. Once all diagnosis frequencies were computed, the top three in each year were identified for each group. Next, the rates of psychotropic utilization by group were compared by using a chi square test. This test was repeated for each of the four therapeutic classes over each of the five fiscal years. The polypharmacy rate by group was also examined by using a chi square test for each year.

Because repeated analyses were conducted, the level of significance was adjusted to account for potential inflation of type I error. The psychotropic utilization analyses were repeated 20 times within the adult and children groups; therefore, a p value of .0025 (.05/20) was considered significant. The polypharmacy analyses were repeated five times within the adult and children groups; therefore, a p value of .01 (.05/5) was considered significant for these findings. All analyses were performed with the PROC FREQ method with SAS software.

Results

Sample demographic characteristics

Table 1 reports the baseline demographic information for both groups, split into adult and child categories. The adult category was primarily white, with more males in the codiagnosis group (51%) and more females in the sole-disorder group (58%). The mean age was 41 for the codiagnosis group and 40 for the comparison group with mental disorders only. Most of the children were white and male (71% for both groups). There was a slight difference in mean age; members in the group solely with mental disorders were younger (8.7 years) than the codiagnosis group (10.3 years).

Diagnoses

The top three diagnoses for both groups in the adult category are shown in Table 2. Depression, schizophrenia, and “other” disorders were the most frequently diagnosed disorders in the codiagnosis group, whereas depression, schizophrenia, and substance abuse were the three top diagnoses for the comparison group. The rankings of the diagnoses did not change for the group with mental disorders only across any of the five study years.

Children in the codiagnosis group were most frequently diagnosed as having autism, conduct disorder, and attention-deficit hyperactivity disorder (ADHD) (Table 3), respectively, in FY 2007 and FY 2008; however, conduct disorder moved to third ranked in FY 2009 and was replaced by “other” disorder in FY 2010. The conduct disorder diagnosis reappeared as third most frequent in FY 2011. Children in the group with mental disorders were most frequently diagnosed as having ADHD, autism, or conduct disorder, in that order, for all five fiscal years.

Psychotropic utilization

Of the 325 members with codiagnoses, 24 adults and ten children did not have psychotropic medication claims. Of the 1,521 members with a sole diagnosis, 160 adults and 131 children did not have psychotropic medication claims. Table 4 displays the number of adult members with prescriptions for the four therapeutic classes of psychotropic medication for each fiscal year, along with the results of the chi square test to compare rates between groups. Adults with diagnoses of mental disorders only were prescribed antidepressants at a significantly higher frequency than those in the codiagnosis group for all five study years. The adults in the codiagnosis group also were prescribed antipsychotics less frequently than those in the comparison group, with significant differences in FY 2007 through FY 2009. Because the frequency of antipsychotic prescriptions increased in the codiagnosis group and decreased in the comparison group over the five-year period, the difference between the groups became nonsignificant by FY 2010. No significant differences were noted for the group comparisons of benzodiazepines and mood stabilizers.

There was a significant difference in the frequency of mood stabilizer prescriptions for members under age 18, with a greater proportion of the codiagnosis group receiving prescriptions versus the comparison group (Table 5). Although the codiagnosis group was also prescribed benzodiazepines at a higher rate than the other group, the only significant difference in rates occurred in FY 2008. There were no statistically significant differences between diagnostic groups for prescriptions of antipsychotics and antidepressants. In FY 2008, the number of members with codiagnoses and prescriptions for antipsychotics (N=54) and benzodiazepines (N=18) was higher than in any other year, and the number of members with codiagnoses and prescriptions for mood stabilizers was the highest (N=35), although this number also occurred in two other fiscal years (FY 2007 and FY 2010).

Polypharmacy

For adults and children, there were no significant differences with regard to polypharmacy between the two groups across the five study years. Polypharmacy was more frequent for adults and children in the codiagnosis group, and there were variations across the years, with large differences between the two groups in FY 2010 and FY 2011. For children, the largest (nonsignificant) difference between the two groups occurred in FY 2008, which was also the same year that there was a spike in the prescriptions for benzodiazepines and antipsychotics in the codiagnosis group.

Discussion

The demographic characteristics of the two groups of Medicaid recipients—those with codiagnoses of an intellectual disability and a mental disorder and those with mental disorders only—were fairly similar, although there was a higher proportion of men in the adult codiagnosis group, reflective of the higher rate of intellectual disability among males (22). Among children under age 18, 71% of both diagnosis groups were male, perhaps because male children were also more likely to have one of the top three mental disorders (autism, conduct disorder, and ADHD).

In our sample, all individuals with an intellectual disability had a comorbid psychiatric disorder. This known comorbidity made it difficult to compare our findings with those of studies where the presence of other psychiatric disorders is unknown. With regard to psychiatric diagnoses, some adults in the group with mental disorders only received a diagnosis of substance abuse, whereas adults in the codiagnosis population did not. For children, the top three mental disorders were the same in both diagnosis groups. However, the prevalence of autism was consistently higher in the codiagnosis group, range 65%−77%, compared with 35%−41% in the other group. The higher prevalence of autism in the codiagnosis group was not unexpected. The prevalence of autism diagnoses in the comparison group was higher than expected, even taking into account the rising prevalence of autism spectrum disorders in the general population (23).

Antidepressants were prescribed at a lower rate for the adult codiagnosis group than for the comparison group, replicating previous findings. However, the prescription rates in the codiagnosis group (range 24%−31%) were slightly higher than previously reported (8). Given the concerns about the increasing use of second-generation antipsychotics and associated adverse effect profiles, it is of interest that, for three of the five years in the study, the prescription rates for antipsychotic medication were lower in the adult codiagnosis group than in the comparison group. The average utilization rate of 32% for antipsychotics in the codiagnosis population is consistent with the rate reported in community samples (21,22,24) and lower than reported rates of 45%−49% in two large-scale studies (8,9).

Among adults, the difference in antipsychotic prescribing between the groups diminished by FY 2010 and remained nonsignificant in FY 2011. This could reflect the heightened awareness of adverse effects of second-generation antipsychotics, particularly the increased risk of metabolic syndrome. Future studies might explore this trend. Among children, there were no significant differences between the two groups in prescribing patterns for antipsychotic medications. Rising rates of prescriptions for second-generation antipsychotics for children with a range of disorders and challenging behaviors have been reported (3,25,26). For the adult population, there were no differences in prescriptions for mood stabilizers between the two groups. For youths under 18, there was a sustained and significant difference in prescriptions for mood stabilizers between the two groups. Seizure disorder is a known comorbidity among individuals with intellectual disability, and some mood stabilizers are also anticonvulsants, which may account for the higher rate of these prescriptions.

There were no differences between the two groups with respect to polypharmacy. This result held for both children and adults. Our study defined polypharmacy as three or more simultaneous medications, so it is difficult to compare our results with findings from studies that defined polypharmacy differently (based on classes of medications, for example) (8).

The use of a 90-day time frame for concurrent prescriptions is a strength of our investigation. The extended period decreased the chance of temporary medication overlap during a taper or change of medications. Another strength is the use of a longitudinal perspective, which permitted us to examine trends in the data over a five-year period. This study adds to the literature on prescribing practices for individuals with both intellectual disabilities and mental disorders.

As with other studies using claims data, our study was limited by the lack of mental health diagnostic data generated from clinical or standardized interviews. The sample captured only individuals with intellectual disabilities living in the community and excluded those from larger congregate care settings, so it did not address individuals who may have more severe cognitive limitations and may show more challenging behaviors. Because this was a single-site study it may be difficult to generalize the results to a wider population. The pharmacy data were limited to claims data processed by the general medical health plan and does not include prescription data from inpatient mental health services, which may show a different pattern. Pharmacy data provide information on prescribed medications but do not include the prescribers’ clinical rationale for why the medications were prescribed.

Conclusions

Between 2007 and 2011, prescribing patterns for psychotropic medications for individuals with both an intellectual disability and a mental disorder differed, according to class of medication and age, from the prescribing patterns for those with a sole mental disorder. Although the county stakeholder expressed concerns about antipsychotic prescribing for individuals with intellectual disabilities and anticipated higher rates of antipsychotic prescribing in the codiagnosis group, we found that this was not the case. More antipsychotics have recently received U.S. Food and Drug Administration approval for additional indications, so future studies may yield different findings in regard to prescribing patterns for individuals with mental illness and with or without intellectual disabilities. Although our study did not find differences in antipsychotic prescribing between the two groups, the rates (31%−46%) raise questions about the judicious use of antipsychotics and the role of psychosocial interventions in both groups.

As states prepare to comply with federal expectations regarding oversight of the use of antipsychotics by children in foster care, our findings may prompt further research into the behavioral health needs and treatment of children in foster care who also have intellectual limitations. The higher rates of prescriptions for mood stabilizers that we found in the children’s codiagnosis group raises the question: are the medications being used as anticonvulsants or to treat behavioral health conditions?

The literature provides some evidence for the use of psychotropics by individuals with intellectual disabilities for both psychopathology and challenging behaviors. However, best-practice guidelines underscore the need to identify underlying causes of behavior, to consider nonmedication strategies first in the absence of a treatable physical or psychiatric cause for the behavior, and to carefully monitor effectiveness and adverse effects (11,27).