Predictors of Psychotropic Polypharmacy Among Outpatients With Psychiatric Disorders and Intellectual Disability

High rates of psychotropic medication use by adults with intellectual disability have been flagged as a national concern in recent commentaries in the British Medical Journal (1) and the British Journal of Psychiatry (2). The overuse of medications, sometimes in the absence of a psychiatric diagnosis, poses unique risks to an already vulnerable group with significant medical comorbidities, lower seizure threshold, and compromised ability to self-report side effects (3,4).

Multiple studies have documented that a substantial proportion of adults with intellectual disability who are prescribed psychotropic medications take more than one at a time, with interclass as well as intraclass polypharmacy being a concern (5–7). Although there are circumstances in which psychotropic polypharmacy may be warranted, guidelines suggest that intraclass polypharmacy be avoided, that medications for behavior management be offered in combination with psychological or other interventions, and that there be regular monitoring of medication use, with the goal of reducing multiple medications when possible (3,4). Clinical predictors of psychotropic polypharmacy include psychiatric diagnosis (5–7) and behavior problems (6,7). Residential setting has been suggested as an organizational predictor (7). In terms of demographic characteristics, there have been mixed results with regard to whether age (7,8) and gender are associated with psychotropic polypharmacy (5,7,8). It is difficult to draw conclusions about predictors of psychotropic polypharmacy because of variation in sample composition and polypharmacy definitions and because few studies have included a multivariate analysis (9).

This study investigated psychotropic medication profiles and predictors of psychotropic polypharmacy among individuals with intellectual disability at the point of entry into specialized psychiatric outpatient services for persons with intellectual disability. Studying medication profiles at intake—prior to specialized service receipt—reflects community-based prescribing practices by nonexperts in intellectual disability.

Methods

The sample consisted of 517 outpatients (females, N=199; males, N=318) who had intellectual disability and who accessed specialized psychiatric outpatient services for individuals with intellectual disability in an urban, tertiary-level mental health care facility in Ontario, Canada. Patients ranged in age from 15 to 73 (mean±SD age=31.17±12.90). During the study period, the last three institutions remaining in Ontario for people with intellectual disability were closed. No study participants were living in these institutions at the time of the study. Prior to receiving specialist care, individuals with intellectual disability received their medications from either family physicians or general psychiatrists. The intellectual disability outpatient service was one of two publicly funded specialized mental health services for adults with intellectual disability in the region.

Patient information on age, gender, place of residence, psychiatric diagnoses, level of intellectual disability, presence of autism spectrum disorder, and use of psychotropic medications was retrieved from the outpatient intake assessment, a standardized form completed by the clinical staff by interviewing patients and caregivers at the time of intake.

Level of intellectual disability was categorized as mild or severe (which included moderate, severe, and profound). Residence was categorized as living with family, living independently, or living in a supervised residential setting. Psychiatric diagnoses were categorized as anxiety, mood, psychotic, or other disorder, based on DSM-IV. Individuals with diagnoses falling into two or more diagnostic categories were categorized as having psychiatric complexity. Those with a diagnosis in a single category or no psychiatric diagnosis were categorized as having no psychiatric complexity.

Medications were categorized by drug class according to their standard use as antidepressants, anxiolytics, antipsychotics, mood stabilizers, sedatives, and stimulants. Anticonvulsants (for example, valproate and carbamazepine) were classified as mood stabilizers if the recipient did not have a seizure disorder recorded on the intake form. Psychotropic polypharmacy was operationally defined as concurrent prescriptions for three or more psychotropic medications (10). PRN psychotropic medications were not included in this count because it was not known how often PRN medications were administered.

All intake assessment forms completed were reviewed by trained research assistants and coded into categories as described above.

Demographic, clinical, and organizational characteristics of patients who were prescribed three or more psychotropic medications (polypharmacy) were compared with those of patients who were prescribed two or fewer psychotropic medications. In the bivariate analysis, unadjusted, single-predictor logistic regression models were used to identify demographic, organizational, and clinical variables associated with polypharmacy. Variables that were significant at the level of p≤.10 were subsequently included in a predictive multivariate logistic regression analysis. Associations between potentially related predictor variables were investigated to avoid collinearity, and only variables that were not significantly correlated with one another were included in the final model. All analyses were performed with SPSS, version 20.0. The Centre for Addiction and Mental Health research ethics board approved this study.

Results

Overall, 362 (70%) of 517 outpatients were prescribed at least one psychotropic medication, and 146 (40%) of those prescribed such medications had no psychiatric diagnosis recorded on their intake form. Approximately one in five individuals (N=112, 22%) who were referred to the service received three or more psychotropic medications concurrently, 30 of them in the absence of any psychiatric diagnosis. The most commonly prescribed medication class was antipsychotics (N=275, 53%), followed by antidepressants (N=151, 29%). The most common interclass polypharmacy combination was antipsychotics, mood stabilizers, and antidepressants (N=52). Only 15 of the 112 individuals who received polypharmacy were not prescribed an antipsychotic as one of the three or more medications they received. With regard to intraclass polypharmacy, 47 individuals were prescribed at least two antipsychotics concurrently.

In terms of demographic characteristics, polypharmacy was more likely among women but was not associated with age (Table 1). With regard to organizational variables, persons living with family or living independently were less likely to receive polypharmacy compared with those living in supervised residential settings. Clinically, there was no relationship between psychotropic polypharmacy and level of intellectual disability or presence of autism spectrum disorder. Psychotropic polypharmacy was more common among persons with a diagnosis of anxiety, mood, or psychotic disorder and was also more likely among persons who had psychiatric complexity. Differences between referral time periods in likelihood of polypharmacy approached significance.

Three individuals from the cohort of 517 outpatients were not included in the multivariate logistic regression analysis because of missing data. Anxiety, mood, and psychotic disorder diagnoses were significantly associated with psychiatric complexity, so they were excluded from the analysis. When variables that reached the p≤.10 threshold were entered simultaneously, three of the four variables (being female, living in a supervised residential setting, and psychiatric complexity) remained significant predictors of polypharmacy, but referral time period did not.

Discussion

This was the first study to investigate psychotropic prescribing trends among adults with intellectual disability upon referral to specialized intellectual disability psychiatric outpatient services. One in five individuals was prescribed three or more psychotropic medications concurrently, with an adjusted analysis showing unique associations between demographic, organizational, and clinical variables and polypharmacy. Interventions aimed at reducing or preventing polypharmacy should consider the diverse range of potential contributors rather than focus only on clinical issues.

This study adds to a growing body of literature on frequent psychotropic medication use in the intellectual disability population. By the time individuals with intellectual disability are referred to specialized psychiatric services, the majority are already receiving psychopharmacological treatment, and specialists are often called upon to reduce unnecessary polypharmacy (11). This speaks to the importance of educating nonspecialist providers about prescribing guidelines. That said, it may be difficult for nonspecialist providers to adhere to guidelines because they do not have the time, skill, or resources to comprehensively assess and treat difficulties that arise in the treatment of persons with intellectual disability and psychiatric disorders. Future research should explore the factors that lead clinicians to prescribe as they do and identify strategies to help them provide guideline-recommended care.

The fact that women were more likely than men in this study to be prescribed multiple psychotropic medications is worthy of discussion. Recent research demonstrates that prior to psychotropic medication initiation, women with intellectual disability may have a more medically complex set of health challenges compared with men, with higher rates of diabetes and overall morbidity (12,13) and greater risk of serious antipsychotic side effects (the most commonly prescribed medication in the current study), such as blood clots (12). This may be one reason why women with intellectual disability have more difficulty than men with psychotropic medication adherence (13). Dosage and monitoring of psychotropic medications need to take gender into account, particularly in the presence of polypharmacy.

Several studies have reported that adults with intellectual disability are more likely to be prescribed polypharmacy if they have a psychiatric diagnosis (12,13). In this study, we report that individuals with a diagnosis in more than one psychiatric category (psychiatric complexity) were more likely to receive multiple medications than individuals with a diagnosis in a single category or no diagnosis. It is possible that providing access to specialized mental health services for persons with intellectual disability sooner in the illness trajectory is helpful in minimizing polypharmacy among persons with intellectual disability and psychiatric complexity. That said, it is important to remember that making a psychiatric diagnosis in the intellectual disability population can be challenging, so the absence of a psychiatric diagnosis does not always mean that the physician has no concerns about psychiatric symptoms. This may be the case for at least some of the individuals in this study who were prescribed one or more medications and who did not report having a psychiatric diagnosis. It is also possible that some individuals in the study had a psychiatric diagnosis but it was not recorded by a clinician at the time of intake. Consultation with a specialist would be important for these individuals as well.

An individual's residential setting was a predictor of polypharmacy in this study. Other studies have also reported increased polypharmacy for persons living in supervised residential settings, compared with those living independently or with family (7). However, the connection between polypharmacy and living situation remained in this study even when psychiatric complexity was taken into account. This association cannot be accounted for by the fact that persons in residential care had a more severe intellectual disability. In both our study and a prior study (14), polypharmacy rates were not associated with intellectual disability severity. One explanation is that the threshold for prescribing multiple medications may be lower for individuals with full-time paid support. An alternative explanation is that persons living in supervised residential settings have more severe psychiatric symptoms (as opposed to more psychiatric diagnoses), that do not respond as well to monotherapy. Further education about the risks and benefits of multiple psychotropic medications, how to monitor medication use, and the role medications should play as part of a range of treatment services (15) could benefit residential service providers.

This study was limited in that it does not include drug dosage or indication, it cannot distinguish between persons who were prescribed medications through consultations with community psychiatrists and those whose care was managed exclusively by family physicians, and it does not consider the impact of prior mental health contacts or PRN use on prescribing practices. It also cannot identify whether referring physicians used medications in concert with nonpharmacological interventions, as per guidelines. This study, like other chart reviews, was limited by variability in the quality of information by chart. It is possible that the psychiatric diagnoses reported by patients and caregivers at time of intake, which were based on DSM-IV criteria, were inaccurate. It may also be the case that the subgroup we described as not having a psychiatric diagnosis may in fact have had psychiatric diagnoses that were not reported at intake. With new diagnostic criteria for both intellectual disability and autism spectrum disorder in the DSM-5, as well as continued refinement of prescribing guidelines for this population, moving forward there may be shifts in how medications are prescribed for those with intellectual disability, which is important to study. Finally, it would be important in future research to evaluate whether specialized consultations, such as those provided by the clinic described in this report, were associated with a reduction in polypharmacy or whether polypharmacy continued even after consultation.

Conclusions

This study found high rates of polypharmacy among adults with intellectual disability seeking specialized psychiatric care. As physicians, we must address medication prescription problems among some of our most vulnerable psychiatric patients and develop strategies to improve prescribing practices by nonspecialist providers. Collaborative policy and practice efforts, such as StompLD in the United Kingdom (11), can be instructive with regard to developing system-level responses. Understanding the demographic, organizational, and clinical characteristics that place individuals at greater risk of polypharmacy may help in the design of future intervention efforts.