Addressing medical needs of adolescents and adults with autism spectrum disorders in a primary care setting

A chart review of all patients who were seen in the CAST clinic at The Ohio State University between April 2014 and April 2015 was conducted. All patients who telephoned to enroll in the CAST clinic during this time period were considered for inclusion. Diagnosis of ASD was confirmed at the time of their visit by review of prior medical and psychological evaluations. Patients who did not attend their initial scheduled visit or who were younger than 15 years of age were excluded from the data collection.

A retrospective cross-sectional study was conducted. Information collected from the electronic medical records included demographic data (age, gender, race/ethnicity), vital signs (height, weight, body mass index (BMI)), diagnoses and medical comorbidities, and medications at the time of the initial visit. Also from the medical records we extracted data from our structured pre-assessment visit documentation. Prior to the patient’s first office visit at CAST, a standardized pre-visit telephone assessment was completed by either the CAST nurse or transition coordinator. The same set of questions was asked of each patient (see Figure 1). The assessment was designed to identify any barriers to achieving a successful medical visit that may exist. Individuals providing the information included the patient or their primary caregiver. If concerns were raised during the telephone assessment, individualized plans were created by the staff member performing the assessment to overcome these barriers within the existing medical office setting. Data collected from the individualized plans and the initial visit included patient concerns identified during the pre-visit assessment, if the patient needed a unique clinical intervention(s), description of the unique clinical intervention, any outcomes associated with the intervention (i.e. were patients able to tolerate the necessary parts of the clinical visit), and any adverse events that may have occurred at the initial visit.

The pre-visit assessment was completed via the telephone and included information about diagnoses, communication skill level, intelligence quotient (IQ) range (e.g. above, below, normal), previous experiences in medical settings (specifically if any event triggered the patient to struggle in a medical setting), and if the patient had any physical limitation. The IQ range or presence of intellectual disability was based on the patient or caregiver’s assessment and was not verified objectively by testing. Any person reporting below normal IQ or providing a reported IQ lower than 70 was considered to have ID for the purposes of this study analysis. Patients and caregivers were also questioned about a history of aggressive behavior. This included both physical aggression and verbal or emotional outbursts. Any mention of physical, verbal, or emotional aggression in either the pre-visit assessment or the documentation from the initial clinical visit was recorded as being positive for aggressive behavior, regardless of whether this was current or prior behavior.

Data were collected for medication use by drug therapy category and described as total number of different categories of medications taken daily. In order to more completely examine trends in medication use in the ASD population, a Medical Regimen Complexity Index (MRCI) score was calculated for each patient (George et al., 2004). The MRCI uses number of medications, medication form (e.g. tablet and cream), frequency (e.g. once daily prn and twice daily), and additional directions (e.g. crush tablet and as directed) to calculate an overall score per patient.

As an example, a medication dosed twice daily and administered as a liquid formulation would score a 2 for the liquid form and a 2 for the frequency. The MRCI is a standardized instrument with predefined weighted values that is applied for each drug in the patient’s medication regimen. The MRCI score is then determined by summing the individual values for each drug. The score has no upper limit, and as the score increases, the regimen complexity increases. Utilizing this tool allows for a more complete understanding of medication administration and related adherence barriers that might otherwise be overlooked by simply counting the number or drug classes a patient is receiving.

This method has been validated for chronic diseases (Libby et al., 2013) and has been suggested as a method to identify patients who may benefit from medication therapy management (Hirsch et al., 2014).

Patient characteristics were summarized as mean and standard deviation for continuous variables and number and percentage for categorical variables. For each comorbidity (ID, seizure, anxiety, depression, aggressive behavior, obesity, ADHD, and hypertension), MRCI scores were compared between patients with and without the comorbidity by Wilcoxon rank-sum tests. Associations between comorbidities and medication types were assessed by chi-square tests. Use of specialized intervention was compared for each comorbidity by chi-square test. All statistical tests were evaluated at the α = 0.05 significance level.

A total of 10 adults (9 male and 1 female), aged 18–30 years, with ASD participated in the focus group. The participants were recruited through members of the Aspirations Ohio program; a social and vocational skills support group provided by the Nisonger Center, a University Center for Excellence in Developmental Disabilities located in Columbus, Ohio. Participants were invited to meet in a private room at a familiar location to the participants for the focus group session and were situated around a large table for collaborative discussion. At the beginning of the session, all participants signed consent forms allowing the session to be recorded for the purpose of research and analysis. None of these individuals had intellectual disability and all were their own guardians and able to communicate verbally.

Pre-prepared, open-ended questions were used to facilitate discussion and were designed to prompt participants to consider and share with the group how their current health care clinic makes them feel, how they think it should make them feel, and to discuss any barriers before, during, or after the clinic visit. One question in particular focused on which step(s) of the clinic experience were the most problematic. The approach used to analyze and synthesize the information collected at the focus group followed a typical process used by design researchers (Kolko, 2010): individual statements were extracted from audio recordings, arranged and clustered by similarity into categories to identify patterns, summarized into higher-level insights, and then configured into visual representations for communication purposes. Two of these visual representations are included (see Figures 2 and 3). They are the result of synthesized insights derived from the statements made by the individuals when prompted to describe the most stressful parts of the clinic visit. In Figure 2, the x-axis represents steps in a typical office visit. Each of the steps is represented here to show the full clinic journey and where high stress areas were identified along that journey in their discussion. The heights (stress) associated with each step of the journey are relative and not tied to specific numbers. The line is meant to visually indicate where higher areas of stress peak at specific stages of the clinic visit in relation to other stages. The “loop” in Figure 3 was created after the focus group as a visual representation of the verbal discussion had by three of the participants when describing how social interaction creates a positive feedback loop of increasing anxiety and stress that were carried into the clinical setting. A summary of the findings of this focus group (along with ideas generated by the participants with autism, shown in Figure 2) is also described.

From the total charts reviewed (n = 126), 112 completed a pre-visit telephone assessment. Of these, only 74 charts were reviewed to evaluate for interventions completed based on pre-visit assessment findings. The other 38 charts were not included due to missing information regarding whether individualized interventions were planned based on the pre-visit assessment findings. These data were unavailable for chart review because it was recorded in a location that was no longer visible 1 year after the initial visit date. For this subset of patients, the age ranged from 15 to 45 years with a mean age of 20.8 years. There was no significant difference in medical diagnoses from the 74 charts compared to the total reviewed charts. In all, 23% (17 of 74) of patients had documented changes to the standard medical visit format. Comorbidities that were significantly associated with need for modification of the routine office visit included ID (40%, p < 0.001), seizures (45%, p = 0.01), and history of aggressive behavior (57%, p < 0.001). Of the 17 patients who had changes to the standard medical visit, 15 of 17 (88%) had at least one of these three comorbidities, with 6 patients having 2 of them and 8 patients having all 3. Specific needs and barriers included difficulty with the waiting room or waiting in general (n = 12) not liking noises including loud noises or children crying (n = 11), aversion to needles (n = 6), difficulty being touched (n = 6), history of aggression in a medical facility (n = 3), not liking bright lights (n = 2), and being unable to tolerate vital signs (n = 1).

Based on these concerns, our team created individualized plans to assist patients in successfully completing their initial medical appointment. Our most common interventions included rooming the patient immediately upon arrival and completing registration in the exam room (n = 16), not performing vital signs at the time of intake (n = 2), notifying the patient prior to touching them (n = 2), waiting in the car until the physician was ready to see the patient (n = 2), turning the lights out in the room prior to rooming (n = 1), scheduling them in the first appointment of the session (n = 1), and having a security guard present but outside the room and not visible to the patient (n = 1).

All 74 patients were successfully roomed and evaluated by the physician. Of those who underwent a pre-visit assessment and implemented personalized modifications, only seven patients had “adverse events” which consisted of being unable to get complete vital signs (n = 5) and being unable to complete portions of the physical exam (n = 2). Of the 14 patients who did not have a pre-visit assessment completed, one patient had an “adverse event” consisting of being unable to get complete vital signs.

To characterize the use of medications in adolescent and adult patients with ASD, medication use for the 126 patient charts were reviewed. Medication use at the time of the initial visit included 51 (41%) on selective serotonin reuptake inhibitors (SSRIs)/serotonin–norepinephrine reuptake inhibitors (SNRIs), 48 (38%) on atypical (second generation) antipsychotics, 39 (31%) on antiepileptics, 32 (25.4%) on non-stimulant ADHD treatments, 32 (25%) on sleep aids, 27 (21%) on stimulants, 29 (23%) on benzodiazepines, 19 (15%) on stool softeners, 12 (10%) on non-SSRI/SNRI antidepressants, 11 (9%) on first-generation antipsychotics, and 9 (7.1%) on antihypertensives. Additionally, 39 patients (31%) reported current use of complementary and alternative medicines with an additional 13 patients (10%) reporting previous use of complementary and alternative medicines. The average number of different medication classes taken was 3.1 for patients with ID and 1.8 for patients without ID. Further statistical analysis was performed to better understand the relationships of different comorbidities with specific medication class usage with a particular focus on ID, seizures, and history of aggressive behavior. These data are provided in Table 2. (Of note the analyses were not adjusted for multiple measures.) Comparisons for individuals with anxiety (n = 65) and without anxiety (n = 65) for SSRI/SNRI showed more frequent use in patients with an anxiety diagnosis, 35 (53.8%) versus no anxiety 16 (26.2%), p < 0.001. For the obese (n = 52) versus non-obese (n = 74) patients, use of atypical antipsychotics was not statistically significantly higher (21 vs 27, p = 0.66), respectively. Comparatively, typical antipsychotic use was higher for obese (9) versus non-obese (2), p ⩽ 0.001). The average MRCI score for the total population was 14.7 (range, 0–79.5). MRCI scores were significantly higher for patients with ID (mean MRCI score of 20.4, p < 0.001 vs non-ID), patients with seizures (mean MRCI score of 21.7, p < 0.001), and patients with a history of aggressive behavior (mean MRCI score of 25.8, p < 0.001) (refer to Table 3).