Influenced by Georgia's settlement agreement with the United States Department of Justice relating to the enforcement of the Americans with Disabilities Act, an increasing number of individuals with intellectual and developmental disabilities (IDD) are transitioning from institutions to community living. In this study we evaluate the pattern of medication use among individuals who recently transitioned to the community (IRTC), comparing results to the IDD population already residing in the community (comparison group). Average use and prevalence rates were trended over time, between January 1, 2010, and December 31. 2012. Findings indicate a significant increase in medication use in the IRTC and comparison group, with a greater and faster increase in the IRTC population. We suggest the transition process should be examined and revised, ensuring adequate preparation time and training for each person and relevant staff, particularly on medications and challenging behaviors. Several demographic trends were also significant and are discussed.
As part of the Department of Justice (DOJ) Americans with Disabilities Act (ADA) Settlement in Georgia (Settlement Agreement, 2010), individuals with intellectual and developmental disabilities (IDD) must be provided opportunity to live in the least restrictive environment. Therefore, many individuals who had been living in an institution are transitioning to a community setting and are using supports and services, as appropriate, to help them successfully live in their communities. However, National Core Indicator (NCI) Consumer Surveys, conducted in Georgia, demonstrated an overall increase in the proportion of individuals with IDD who used psychotropic medications (unless otherwise noted, medication refers to psychotropic or anticonvulsant medication); from 36% in fiscal year 2005–2006 (Human Services Research Institute, 2007) to 51% in fiscal year 2010–2011(Human Services Research Institute, 2012). Because people transitioning into the community from institutions often present with significant physical, health and mental health challenges, it is critical to include community providers throughout the transition process and incorporate education programs to help with managing medication use.
A successful transition process must include effective management of many different types of medications and behaviors. Because individuals are moving into new environments with new roommates, new providers, new services, and new routines, feelings of frustration and fear may manifest themselves as challenging behaviors, resulting in medication intervention. Using Logistic Regression to examine behavior management problems in a population of individuals with pervasive developmental disorder and intellectual disability, Tsakanikos, Costello, Holt, Sturmey, and Bouras (2007) found that physical aggression and problems such as pestering staff independently predicted the use of antipsychotics.
Huang et al. (2007) provided psychiatric, pharmacological, and behavioral assessments and interventions to individuals with IDD who were taking psychotropic medications. As a result, half of the 40 individuals in the study received fewer medications. At the same time, providers working with this population were offered training on the assessment processes, behavior interventions, and medication management. Their knowledge of administering the assessments in the posttest results in all three areas was significantly higher than pretest results; and a high level of satisfaction with the training process was reported by most of the providers. Results indicate training and effective assessments can help build skill in providers and, at the same time, reduce problem behaviors and the use of medications to treat them.
McGillivray and McCabe (2005) examined the use of psychotropic medications to manage challenging behaviors and found individuals in institutions received a slightly greater number of medications than individuals in the community. However, the Royal College of Speech and Language Therapists (2007) has emphasized that challenging behaviors are not a diagnosis but a descriptor of behaviors in specific contexts with many possible causes. Glover, Bernard, Branford, Holland, and Strydom (2014) indicate challenging behaviors may arise from psychotic or affective illnesses, attention-deficit hyperactivity disorder, or severe autism. The rationale for individuals with IDD is to medically treat the underlying psychiatric condition to affect the frequency and severity of challenging behaviors. However, they note the concern of Paton et al. (2011) in that medication is commonly prescribed without this rationale, but rather with a focus on targeting the behavior itself—the symptom, not the cause.
Deb, Unwin, and Deb (2015) used a prospective study of 100 adults with IDD and aggressive behavior, over a 6-month timeframe in a clinic-based community setting, to examine the trajectory of use of psychotropic medications and their relationship with different demographic and other aggression related variables. Eighty-nine percent of individuals used psychotropic medications at baseline, and 90% at 6 months; and 45% and 41%, respectively, received more than one medication at each point in time. They note that in the United Kingdom, where this study took place, psychiatrists are encouraged to regularly review psychotropic medication use for individuals with IDD. However, the authors found a marginal increase in the use of these medications over the 6-month time period, which could indicate the oversight, may need to be more standardized or the process and content of review activities revised.
Little work has examined medication use for individuals with IDD as they transition from institutions to community programs. In one cohort study of 109 adults age 16 to 65 years, Nettestad and Linaker (2003) found neuroleptic use did not change before (before 1985) and after deinstitutionalization (after 1987). However, they concluded that challenging behaviors were the main predictor of dosage, both before and after deinstitutionalization. They posit several potential reasons for this: Difficulties determining the extent to which behaviors are from a psychiatric or behavior disorder; lack of knowledge among caretakers and the ensuing referral practice; lack of knowledge among general practitioners; and lack of access to specialized health services.
The purpose of this study is to assess the use of psychotropic and anticonvulsant medications in adults with intellectual and developmental disabilities, with a focus on individuals who had recently transitioned to the community from an institution (IRTC population). Effective transition programs should include the person, institutional and community-based staff/providers, support coordinators/case managers, and adequate training in the administration and understanding of assessments, medications and challenging behaviors, as well as appropriate interventions needed to help transition individuals successfully.
The Georgia Department of Behavior Health and Developmental Disabilities (DBHDD), in an ongoing effort to assess the appropriateness and effectiveness of the transition process and the health of individuals, requested a closer analysis of medication use for this group: before, during, and after transition from the institution. The research questions are as follows:
Does the average medication use increase after individuals transition from an institution into a community setting? Does it differ from the IDD population of individuals who had been living continuously in the community and had not transitioned from an institution?
Is there an increase in the proportion (prevalence) of individuals prescribed medication as they transition to the community? Does it differ from the IDD population of individuals who had been living continuously in the community and had not transitioned from an institution?
Does the prevalence rate vary by demographics—residential setting, gender, race/ethnicity, age or disability?
The study population consists of adults, age 18 or older on January 1, 2010, with IDD and receiving services through the Georgia Medicaid Home and Community-Based Services (HCBS) waivers or state funding. Adults in the IRTC Group (N = 325) transitioned from state hospitals into the community between July 2010 and June 2012. All other individuals receiving HCBS waiver or state funded services were included as the comparison group (N = 12,722), individuals already residing in a community-based residence. (It is not known how long individuals in the comparison group lived in their residence. Residential status is based on where the person lived when we received the data in March 2013. Though individuals may have changed residence during the study period, the most common change is from one group home to another.)
Demographic information was obtained from the Georgia Case Management System, a database hosting information for all individuals receiving waiver or state funded services in Georgia. Medication information was obtained from the Health Risk Screening Tool (HRST), a web-based application providers use to document and assess an individual's health status, including prescription drug use. The HRST is updated by providers if there is a change in a person's prescription medication, and at least annually. (Though claims data may provide complete medication information, because at least half of the study population has dual eligibility, prescription claims, paid for through the Federal Medicare Part D program, were unavailable at the time of this study.)
Medication information includes the prescription name, start and end dates, and whether the specific medication is a psychotropic or anticonvulsant. For the purpose of this study, the term “psychotropic medication” refers to any medication prescribed with the intent to affect or alter thought processes, mood, or behavior, including antipsychotic, antidepressant, mood stabilizer, antipanic, antianxiety, and anticonvulsant medications. Medication types and prescription names were reviewed and categorized by Delmarva Foundation's Registered Nurse, a Certified Developmental Disabilities Nurse with a Master's degree in Public Health, and by a Pharmacist. (Unfortunately, we do not know why medications were prescribed; and because psychotropic and anticonvulsant medications are used for various reasons in this population, we include both types in the analysis.)The transition date for each individual was provided by DBHDD, and is used as an anchor point to identify before and after transition rates when analyzing medication utilization in the IRTC population. Since the greatest number of transitions occurred in June and July 2011, July 1, 2011, was used as the anchor date to analyze comparison group results. Figure 1 shows how the time spans for the two groups overlap. The analytic periods were defined as follows and shown in Figure 1:
IRTC Group: Rates were calculated 6 and 3 months before transition, on the transition day, and 3 and 6 months after transition. The earliest pretransition date was in January 2010, and the latest posttransition date was December 2012.
Comparison Group: The time period between January 2010 and December 2012 represents the same span of time as for the IRTC group, 6 months before the first transition in July 2010 and 6 month after the last transition in July 2012. Rates were calculated at 3-month intervals, consistent with the IRTC Group.
Medication use at each point in time was determined by the prescriptions documented in HRST (all analyses were completed with the SAS 9.3). Prescriptions without an end date were assumed to be active, however, it is not known which medications are actually being taken, only those prescribed. The average number of medications prescribed per person was calculated to trend medication use for each group. Prevalence rates were defined as the percentage of people with one or more active prescription. A 95% confidence level with a 5% error rate (confidence interval (CI) of +/– 5%) was the standard used to determine statistical significance for average use rates—if intervals do not overlap we report a statistically significant difference. The difference in proportions Z-test was used to determine statistical significance for prevalence rates (Bohrnstedt & Knoke, 1988).
Prevalence rates were stratified by race, gender, disability, age group, and residential setting. If the N size of a subgroup was smaller than 30, results were suppressed. Disability is identified by the “primary disability” noted in the data and is not just individuals with an intellectual disability. Therefore, someone may have a primary disability of Autism and may or may not also have ID. An independent home or apartment is the person's “own place” and individuals in a parent's or relative's home are in a “family home.” Group homes (2 to 5 residents) and host homes (1 or 2 residents) are licensed by the state.
The demographic distribution of both study groups is shown in Table 1 (a small proportion, 4% or less, of “other/don't know” in each category is not shown). Compared to the comparison group, the IRTC group had a higher percentage of males (64.6% vs. 57.5%), was more likely to be white (60.6% vs. 52.1%), and much more likely to be living in group homes with two to five other individuals (80.0% vs. 20.5%). (The IRTC residential setting is where the individuals lived post transition from the institution.) The majority of transitioning individuals had ID at more severe levels (72.0%) (formerly called “profound”) whereas most of the comparison group had ID at less severe levels (73.9%) (formerly designated mild or moderate ID). Individuals who transitioned from an institution were more likely to be 50 or older (50.2% vs. 25.6%).
At some point during the study time period, 151 (40.3%) of the 325 individuals who transitioned to the community and 4,371 (34.4%) of the 12,722 individuals in the comparison group were prescribed at least one psychotropic medication. The list of most frequently prescribed medications, by generic name, is provided in Table 2, based on at least one prescription per person, including the most frequent use for each. However, medications, particularly in this population, may be used for more than one type of diagnosis. Findings indicate the following:
Risperdone, an antipsychotic, was the drug most likely to be prescribed in the comparison group and was more likely to be prescribed for the comparison group than the IRTC group, 25.9% and 13.9%, respectively.
Valproic Acid, a drug used as an antipsychotic or antiseizure medication (can also be used as a mood stabilizer) was the drug most likely to be prescribed for the IRTC population. However, the percent of individuals prescribed this medication was similar to the comparison group, 24.5% and 22.8%, respectively.
On average, in the broader categories the two groups had a similar proportion of individuals prescribed medications, with the exception of antipsychotics. Individuals in the IRTC group were much less likely to have a prescription for these, 35.8% vs. 61.3%.
Average Medication Use
The average number of medications was calculated for the IRTC group at the specified times and for the comparison group at 3-month intervals. The number of medications per individual ranged from one to six. The “shorter” line in Figure 2 shows results for individuals who transitioned, anchored around the transition date and positioned midpoint on the comparison group “longer” line. Even though the line is centered, the timeframe is from the first pretransition date (January 2010) to the last posttransition date (December 2012). There is not a one-to-one correspondence with the dates representing the comparison group rates. However, the timeframes are the same.
Results in Figure 2 show an increase for the IRTC group, from 0.68 (CI 0.54–0.83) to 1.84 (CI 1.65–2.03), whereas average use for the comparison group increased from 1.01 (CI 0.97–1.05) to 1.98 (CI 1.94–2.02). The 95% confidence intervals for both groups from the first to the last time period do not overlap and indicate the increases for both groups were statistically significant. However, the increase as individuals transitioned into the community was greater and faster. Before the transition date, the average IRTC use rate (0.68) was significantly lower than in the comparison group on January 1, 2010 (0.97). After transitioning to the community, there was no significant difference between the two groups (1.84 and 1.98, respectively).
Prevalence rates show the percent of adults prescribed one or more psychotropic medication and were calculated for the same time periods as indicated in Figure 2. The IRTC rate increased over 20 percentage points from 23% 6 months before transition to 44% 6 months post transition (Figure 3). There was a slower increase in the comparison group, 13 percentage points over the time period (from 19% to 32%). Although the initial prevalence rates were similar for the two groups (January 1, 2010, for the comparison group and 6 months pretransition for the IRTC group), by the end of the time period (December 31, 2012, and 6 months posttransition), individuals who transitioned to the community were significantly more likely to be taking one or more medication (difference between 44% and 32%: Z-score = 4.22, p < .000).
Prevalence Rates by Demographic Characteristics
The demographic composition of the two different groups, IRTC and comparison, has been shown to be different. Therefore, we examine the demographic differences in medication use only within the IRTC group. Prevalence rates for individuals using one or more psychotropic medication are shown in Table 3, by race, gender, disability, residence, and age group. Rates are shown for three data points: 6 months before transition, on the day of transition, and 6 months after transition.
Within the IRTC group, prevalence increased significantly over the time period across all demographic categories and some differences between demographic characteristics were also statistically significant. After transition, individuals in the IRTC group with ID at more severe levels were significantly less likely to take medication than were individuals with ID at less severe levels, 40.2 percent versus 53.9%, respectively (Z-score = 2.2 (p < .028)). The increase in prevalence from pretransition to posttransition for individuals with ID at less severe levels was greater than for individuals with ID at more severe levels, 27 percentage points and 22 percentage points, respectively. Therefore, as individuals with ID at less severe levels transitioned from institutions, it was more likely for them to begin taking psychotropic medications.
Results vary by age group, comparing prevalence between age groups as well as the percentage point increases shown from pretransition to posttransition. Younger adults (age 18–29 years) were significantly more likely to have been prescribed medications than were individuals in all other age groups combined—before, during, and after transition (Z-score = 3.2, p < 0.001 posttransition). Six months posttransition, close to 70% of young adults had at least one prescription for a psychotropic or anticonvulsant medication. Individuals age 60 years and over were least likely to have been prescribed medications. The 26 percentage point increase in prevalence rates for adults age 18–29 and 30–39 years, from pretransition to posttransition, was much greater than for individuals age 60 years and over, who showed only a 19 percentage point increase.
Other findings by demographic characteristics point to what may be some interesting trends, but did not reflect statistically significant results. This could be due, in part, to the small sample size in the IRTC group. The difference between African American and white individuals consistently showed a higher prevalence rate for African Americans. Though not significant at p < .05, the differences at each point in time indicate a less than 10% chance of error (p < .10).
Rate Increases Pretransition
Results for the IRTC group also indicate an increase in both the average use and prevalence rates from 6 months before transitioning up to the transition day. Overall, the average use (Figure 2) increased from 0.68 (CI 0.54–0.83) to 1.17 (CI 1.00–1.35) over this time period. The confidence intervals do not overlap, indicating this may be a significant factor for individuals preparing to move into the community. Prevalence rates also increased as individuals prepared to transition from the institution, from approximately 21% to 32%. This increase is also shown to be statistically significant, with a Z-score of 3.2 (p < .001).
Due in part to the effect of the DOJ ADA Settlements, individuals are moving into community settings from institutional life at an increased pace, not only in Georgia but in states across the nation. Often these individuals had lived in an institution for many years, perhaps most of their lives. Although there is a limited amount of information in the literature regarding individuals as they transition to the community, some findings in this study are similar to other published results.
As of December 31, 2012, 32% of Georgia adults with IDD were prescribed at least one psychotropic medication. This is similar to the prevalence reported by the cross-sectional study from Australia (Doan, Lennox, Taylor-Gomez, & Ware, 2013), which was conducted between 2000 and 2002 among adults with IDD living in the community of Brisbane, Australia. Of the 117 participants, 35% was prescribed psychotropic medications. In addition, similar to trends noted in the NCI data (Human Services Research Institute, 2007, 2012), findings in this study indicate that psychotropic medication use has increased significantly among all individuals with IDD who are receiving services through the Georgia Home and Community-Based Services Waivers or through state funding.
Key findings indicate psychotropic medication use and prevalence have significantly increased in this population in Georgia, and for individuals who have recently moved from an institutional setting to the community the increase has been significant and dramatic. Findings (Figure 2) suggest that individuals are not only prescribed more psychotropic medications (increased average number from .68 to 1.84) as they move through the transition process, but individuals are significantly more likely to be prescribed at least one of these types of medications as well (increased prevalence rate from 21% to 44%). These results appear to be independent of the overall trend in the population, indicating something about the transition process may be affecting medication use and the management of new and additional behavioral challenges.
In institutions, services may be provided by the same providers and psychologist for many years, providing consistency for the person and a reduced need for medication therapies. Average use rates for individuals living in the institutions (0.68 at 6 months pretransition) were actually lower than for individuals living in the community (1.01 on January 1, 2010). However, when moving into the community, individuals may be medicated to compensate for the less restrictive environment with inconsistent provider supports to address complex issues and new types of behavioral challenges. This highlights the importance of training for direct care providers and staff who invariably present at a clinic to request intervention, often medical prescriptions, for challenging behaviors. Huang et al. (2007) indicated training and effective assessments can help build skill in providers and, at the same time, reduce problem behaviors and the use of medications to treat them.
Data in this study also indicate medication use in the IRTC group began to increase before the actual transition day. The reason for this is not clear. We suggest that as individuals begin discussing and learning about transitioning, they may become frustrated or fearful of the change; and this may manifest in challenging behaviors and increased medication use. Whatever the reason, the increased use of psychotropic medications before leaving the institution reinforces the importance of using a thorough and effective transition process where the person has adequate time to work with a support coordinator/case manager and learn about new providers and the new environment.
Increases in the prevalence and average medication use during the transition process were significant across all the different demographic categories over time; and appear to indicate that most demographic groups were equally affected through the transition process. However, though the sample sizes were relatively small, some relationships within demographic categories were robust enough to be statistically significant at this level of analysis (without controlling for other factors). Older individuals who transitioned were much less likely to be taking a psychotropic medication than their younger counterparts. Furthermore, individuals with ID at more severe levels who transitioned to the community were much less likely to take a psychotropic medication than individuals with ID at less severe levels. Perhaps more complex disabilities or advanced age limit the degree to which individuals are able to exhibit challenging behaviors that would increase the likelihood of taking psychotropic medications. Clearly age and type of disability are factors to take into consideration when designing and implementing a transition process for the person.
Findings in this study may possibly indicate other interesting variances by race and gender for individuals being discharged from an institution. Though not found to be statistically significant in this study within a 5% error rate, African American individuals and men were consistently more likely to be prescribed psychotropic medications than were white individuals or women. We have not established that a pattern exists, but these findings may be worth a deeper analysis if more data are available.
The differences between individuals living in group homes and host homes were not statistically significant at any point during the transition process. However, it is of note that only five individuals who had transitioned from an institution were living in a family home, all others had moved into licensed small (N = 49) or large (N = 260) group home settings. Trends identified in the comparison group suggest that individuals living in a large group home were much more likely to be prescribed these medications than individuals in a family home, and the gap between them has grown. (Comparison group data for each demographic was not shown but is available upon request. Comparison group results indicated the prevalence rate for individuals who had been living in a group home (32.6%) was 20 points higher than for individuals in a family home (11.8%) at the beginning of the study period (January 1, 2010). That difference had grown to almost 39 points by the end of the study period, when 58.2% of individuals had been prescribed at least one of these medications, as compared to 19% of individuals living with parents or relatives. This finding is consistent with other studies and reiterates the benefit of living in a more family-oriented and familiar setting (Tsiouris, Kim, Brown, Pettinger, & Cohen, 2013).
However, it is not clear to what degree the increased prevalence rate for the IRTC group may have been due to living in a group or host home or the degree to which the transition process itself may have affected medication use. The type of residence should not affect the increase in medication use before transition, as found in this study. However, living in a more family-oriented and familiar setting may provide more stability for the person, as well as more concern and aversion to medication use. This type of familial environment is likely related to a reduced incidence of challenging behaviors, as well. It is also true that individuals living in the family home may have fewer complex disabilities, which alone reduces the level of support needed and psychotropic medication use. We do not know the complexity of disabilities individuals in this study had; but given the findings, it may be worth further study to determine how and why most individuals transitioning from an institution do not seem to have a family home option available to them, where medication is less likely to be used.
There are important findings in this study, and there were also several limitations. Medication use is based on the number of prescriptions and not what is actually taken by the individual. Although we diligently eliminated duplicates, a better analysis could be completed, knowing what is actually being used by the person. We did not have the ability to determine the reason the person was prescribed the medication, and there could be several in this population. Anticonvulsant medication could have been prescribed for a specific behavior or for epilepsy, and we do not know if that reason changed from pretransition to posttransition.
The actual dosage, or change in dosage, was part of a comment field and not always entered into the system. Therefore, we do not know if the dose was increased or decreased for the same prescription over time and this could be a critical piece of information in future analysis if it is available. Finally, the HRST is updated annually unless there are changes to the medications or doses. However, it is completed by providers who are generally already overwhelmed with paperwork and may not always complete updates as needed. This may affect results because we may not capture short-term fluctuations in prescribing across the 3-month intervals used in the analysis.
Findings in this study highlight a steady and significant increase in the use of psychotropic medications in Georgia from 2010 to 2012, as well as a more pronounced and rapid increase among individuals who transitioned from the State's institutions to the community. It is important to monitor and control the use of psychotropic medication among people with IDD who are more vulnerable than individuals in the general population. They may be subject to overmedication and the use of physical and chemical restraints to address challenging behaviors (off-label use). Because people all over the United States are transitioning from institutions to less restrictive environments, the findings in this study have broad implications.
Each state should consider conducting an analysis of the current transition process to ensure specific, competency-based training is occurring with direct service providers to help them adequately assess and manage the use of psychotropic medication and challenging behaviors, particularly for providers who may not be familiar with the more complex behavioral and medical issues of individuals coming to them from institutions. Demographic characteristics should be included as mitigating factors, particularly age and disability. The transition process should begin early and include all relevant staff, providers, and family members, as well as, at a minimum, a pharmacy review and medication reduction plan for people prescribed psychotropic or similar types of medications.
The medical community could benefit from educational programs on challenging behaviors typically associated with IDD psychiatric or behavioral disorders that may actually be medical issues, such as a urinary tract infection or dental issue that could exhibit as unusual or challenging behaviors not resultant of a mental illness. Results in this study suggest some of the challenges of transitioning individuals to the community are being addressed through chemical controls, challenges that may be reduced through an effective transition process that helps ensure the continued health and safety of individuals striving to maintain an everyday life in the community.
This article was presented at the 2013 National Home and Community Based Services Conference (HCBS) in Washington, DC, as well as the American Network of Community Options and Resources (ANCOR) webinar on February 18, 2014.
This article is funded by the Georgia Department of Behavioral Health and Developmental Disabilities, as part of the Georgia Quality Management System.
We thank Marion Olivier, director Georgia Quality Management System, and Eddie Towson, Georgia Quality Management Director, Georgia Department of Behavioral Health and Developmental Disabilities, and Lori Reid PhD, Lead Analyst for Delmarva Foundation, for their valuable insight and input into earlier versions of this study; the Georgia Department of Behavioral Health and Developmental Disabilities for use of the data; and Larry Polnicky, Health Risk Screening, Inc., for his tireless efforts to obtain pharmacy data from the Georgia Health Risk Screening Tool database.
Susan Kelly and Yani Su, Delmarva Foundation for Medical Care, Tallahassee, Florida.