How to Transition Patients Back From Psychiatric Care to Primary Care: An Access Initiative Protocol Development

We conducted a descriptive analysis of EHR data to summarize the demographic and clinical characteristics of patients seen in outpatient psychiatric clinics from April 1, 2018, to March 30, 2019. We limited our analysis to patients ages 19–49 years because pediatric and geriatric patients have different clinical needs. We used ICD-10 codes to describe mental and general medical health conditions. We obtained EHR data for 2,837 patients; 988 patients (35%) were ages 19–29, 1,013 patients (36%) were ages 30–39, and the remaining 836 patients (30%) were ages 40–49. Furthermore, 1,540 patients (54%) had private insurance, and 1,918 (68%) were overweight or obese. Anxiety disorders (N=1,775, 63%), depressive disorders (N=1,477, 52%), attention-deficit hyperactivity disorder (ADHD) (N=696, 25%), bipolar disorders (N=671, 24%), trauma- and stress-related disorders (N=564, 20%), and substance-related and addictive disorders (N=453, 16%) were among the most common psychiatric conditions. Diseases of the nervous system (N=470, 17%) and endocrine, nutritional, and metabolic diseases (N=434, 15%) were common diagnostic groups for general medical health. Additionally, of the patients with depressive disorder, 60% (N=886) had been prescribed an antidepressant.

To enhance the bidirectional flow of patients, the workflow for new and existing patients consisted of two efforts: identifying patients currently being treated in the psychiatric clinic who could transition back to primary care and creating a series of tracks for patients being referred from primary care to psychiatry. The work group determined criteria for identifying patients most appropriate for transition back to PCPs, taking into account their number and severity of psychiatric diagnoses, number of psychotropic medications, and comorbid medical conditions. (See the online supplement to this column for a figure showing the referral process.) Patients potentially suitable for transfer back to primary care included patients with ADHD who had been stable for 6 months while taking a stimulant, patients with anxiety disorders (including obsessive compulsive disorders and trauma-related disorders) taking ≤2 scheduled psychotropic medications, and patients with mood disorders taking ≤2 scheduled medications. The initial evaluation provided information to direct patients either to the “treatment optimization” track for referral back to PCPs or to the “specialized treatment optimization” track for referral to non-PCP settings.

After initial assessment, psychiatric providers could see a patient in the treatment optimization track for between one and four visits to establish a diagnosis and initiate treatment. Once stabilized, patients were referred back to a PCP with documentation to support ongoing treatment within a primary care environment and additional psychiatric assistance, should that need arise. The specialized treatment optimization track included the psychiatry continuity track, psychotherapy track, community referral track, and subspeciality clinic track.

Anticipating that this shift in pathway development and management of patients would include challenges, we established processes for ongoing feedback with notable stakeholders to ensure built-in opportunities for quality improvement as the project was conceptualized and implemented. Quarterly meetings were set up with primary care leadership. Additionally, a multidisciplinary team was created to conduct several PCP group meetings to discuss referral to primary care, criteria for referral back to primary care, number of consultation visits, and mobilization of clinic support for the access initiative. A semistructured interview guide was developed by a work group consisting of a nurse practitioner with integrated care expertise, a psychiatrist, a registered nurse, a clinical psychologist, a quality improvement team lead, four PCPs, and a primary care clinic administrator. The interviews were conducted to ensure that the clinics’ needs were being identified, prioritized, and met.

In addition, the adult division director from the outpatient psychiatric clinics met with each adult psychiatric provider to set a personalized goal for the number of new intakes conducted per week. The number of new intakes was defined as the number of patients scheduled for new psychiatric evaluation. The number of no-shows was defined as the number of patients with a new psychiatric evaluation who did not show up for the appointment and did not cancel or notify of their inability to attend. Personalized goals for new intakes were dependent on total outpatient full-time equivalents, current caseload, and provider preference. All outpatient providers committed to an increase in the number of new patient intakes per week compared with the previous year's schedule. Criteria were determined for suitable candidates for transition by using the access initiative work group and primary care input. The criteria were shared with psychiatric providers, along with direction to evaluate their current patient populations for patients who met these parameters and were appropriate for transition to primary care.

The second measure taken to develop a bidirectional workflow was to create a new system for the management of incoming referrals. Patients in the treatment optimization track were slated to be sent back to PCPs. For patients in this track, a standard evaluation and treatment recommendations were used to guide the management of patients’ conditions in the primary care environment. The work group gave a presentation to PCPs about the new referral process and met quarterly with the primary care administration group, who communicated with PCPs about progress.

On a monthly basis, the following outcomes were tracked: number of new intakes and number of no-shows for first-time intake appointment. The outpatient psychiatric clinics had approximately 62.5 new patient intakes per month from January through June 2019, with a no-show rate of 17.5/month. However, during the access initiative period between July 2019 and October 2020, an average of 116.0 new patient intakes were completed per month, with a no-show rate of 17.8/month. Between July 2019 and June 2020, an average of 21.0 patients per month were sent back to primary care, a rate that decreased to an average of 10.0 patients per month for July 2020 to October 2020.

Time constraints and reliance on clinical interviewing to assess psychiatric disorders are frequently noted barriers to PCPs managing psychiatric disorders in primary care and result in delay of treatment or inappropriate referrals (6). Furthermore, the expertise of psychiatric providers can be leveraged to evaluate patients, increasing accuracy in diagnosis of mood disorders, such as bipolar disorder, that are often underrecognized in the primary care environment (7). In line with our commitment to treatment optimization, patients with severe, chronic mental illness need a psychiatric specialist to manage their cases. With the bidirectional model, these patients are referred to the continuity track and are seen by the same providers on a long-term basis. Research has shown that patients with chronic mental illness who require continuity care prefer and value this type of delivery of care (8).

Toward the end of the study period, fewer patients were referred back to PCPs. A few psychiatrist-related factors associated with this outcome have been identified. Some psychiatrists felt that it was more time-consuming to go through the steps to refer patients back to PCPs than to see the patients themselves. Some psychiatrists voiced dissatisfaction with seeing mostly patients with serious psychiatric illness, indicating that when they referred all stable patients back to PCPs, their resulting day-to-day schedules were overloaded with patients with acute illness, leading to dissatisfaction and increasing the chance of burnout. Many patients preferred frequent interaction with a provider—they missed seeing a psychiatrist on a monthly basis—which prompted some patients to make psychiatric appointments even if they were referred back to PCPs. Also, despite initial communication with patients about the initiative to transfer some patients back to PCPs, some patients who wanted to receive psychotherapy treatment continued returning to psychiatric clinics for such services. These outcomes showcase the importance of considering patient preference for treatment options.

Planning for the access initiative began in March 2019, and implementation began in July 2019 and ended in June 2021. When the COVID-19 pandemic started, the psychiatry department began transitioning to virtual appointments for outpatient visits, and in June 2020, the department converted all appointments to virtual visits. The access initiative was halted as the pandemic got worse. The department is evaluating the access initiative and modifying its protocol to restart the project.

There were some limitations to this study. First, the findings from this study may not be generalized to settings other than academic psychiatric clinics. Second, the study was observational; without a control (comparison) group, we cannot draw a conclusion about the effectiveness of the access initiative. Third, the study did not observe patient outcomes to measure potential improvement in psychiatric and general health outcomes. Last, it would have been helpful to conduct an EHR data analysis for the PCP clinics and to determine whether capacity did in fact increase. Before the access initiative, a majority of psychiatric providers shared in the stakeholder meetings that they did not spend much time or directed effort in transitioning patients back to primary care; establishing a baseline data point for this measure would have been beneficial to further evaluate our results.

The feedback provided by the PCPs highlighted the need for education in particular areas. For instance, decision making that may seem straightforward from the perspective of psychiatric providers (e.g., about medication [i.e., for ADHD] or which stable psychiatric disorders are suitable for transition back to primary care) may not match PCP preferences. Evaluating early how to support PCPs, adapt to specific needs, and promote communication was important. A previous study that explored communication between psychiatric providers and PCPs in part by reviewing medical records for these encounters found deficits in continuity in communication of pertinent clinical information, despite use of a mutual-access EHR (9). Furthermore, the findings of this review suggest that PCPs accessed mental health medications regularly but did not access mental health records regularly, indicating the importance of understanding how different provider specialties communicate and use the EHR (9).

Limited access to behavioral health services is a problem nationwide. A delay in access to treatment can have significant consequences for patients and their families. Our academic outpatient psychiatric clinic developed a new access initiative in partnership with PCPs to address this problem. After implementation of this access initiative, we were able to enhance timely access to psychiatric diagnostic evaluations by increasing the number of new intakes performed, and we subsequently reduced the no-show rate for these appointments.