This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in the Journal of Medical Internet Research, is properly cited. The complete bibliographic information, a link to the original publication on http://www.jmir.org/, as well as this copyright and license information must be included.
Mental health problems are commonly encountered in primary care, with primary care providers (PCPs) experiencing challenges referring patients to specialty mental health care. Electronic consultation (eConsult) is one model that has been shown to improve timely access to subspecialty care in a number of medical subspecialties. eConsults generally involve a PCP-initiated referral for specialty consultation for a clinical question that is outside their expertise but may not require an in-person evaluation.
Our aim was to describe the implementation of eConsults for psychiatry in a large academic health system.
We performed a content analysis of the first 50 eConsults to psychiatry after program implementation. For each question and response, we coded consults as pertaining to diagnosis and/or management as well as categories of medication choice, drug side effects or interactions, and queries about referrals and navigating the health care system. We also performed a chart review to evaluate the timeliness of psychiatrist responses and PCP implementation of recommendations.
Depression was the most common consult template selected by PCPs (20/50, 40%), followed by the generic template (12/50, 24%) and anxiety (8/50, 16%). Most questions (49/50, 98%) pertained primarily to management, particularly for medications. Psychiatrists commented on both diagnosis (28/50, 56%) and management (50/50, 100%), responded in an average of 1.4 days, and recommended in-person consultation for 26% (13/50) of patients. PCPs implemented psychiatrist recommendations 76% (38/50) of the time.
For the majority of patients, psychiatrists provided strategies for ongoing management in primary care without an in-person evaluation, and PCPs implemented most psychiatrist recommendations. eConsults show promise as one means of supporting PCPs to deliver mental health care to patients with common psychiatric disorders.
Mental and behavioral health problems are among the most common and costly conditions in the United States [
Electronic consultations, or eConsults, are an innovative model of collaborative care developed to improve access to specialty consultation for a variety of physical health conditions. The primary goals of eConsults are to prevent unnecessary referrals for management that could occur in the primary care setting and, in so doing, to decrease wait times for those who require in-person evaluation by the specialist [
eConsults could improve access to mental health care by providing a way for PCPs and psychiatrists to collaborate and expand treatment within the primary care setting. However, acceptability by physicians or patients is not established. In addition, given the nature of behavioral health problems and the importance of the interview in diagnosis, the proportion of eConsult questions that consulting psychiatrists may feel require a face-to-face evaluation is not known. The fact that mental health is often carved out from other medical care creates additional challenges around payment and insurance coverage. Reports on the use of eConsult for mental health are limited but suggest some improvement in PCP perceptions of support for psychiatric diagnosis and treatment as well as access to mental health consultation [
Our aim was to test the feasibility of a new psychiatry eConsult program in a large academic health system. We evaluated a consecutive series of the first 50 eConsults to describe the consult questions and responses, assess consult response time and implementation, and determine the rate of conversion to in-person consultation. To our knowledge, this is one of the few studies to describe eConsults for mental health and the first to evaluate consultation content and conversion to in-person referrals.
The University of California, San Francisco (UCSF) is a multisite urban academic medical center with 8 primary care practice sites and a diverse payer mix including commercial insurance, Medicare, and Medicaid. PCPs include attending physicians, nurse practitioners, and resident physicians. All clinics use a shared EHR (Epic Systems Corp). For more details on the study site, see
The UCSF eConsult program began in 2012 with 8 internal medicine subspecialties; details of the overall program are described in
We performed a content analysis on a consecutive series of the first 50 eConsults to psychiatry, occurring between October 2014 and August 2015. We determined our initial sample size based on feasibility. We then analyzed the consults sequentially and found no new themes emerging after analysis of approximately the first 25 consults. We opted to present the entirety of the data in this paper as this adds to the richness of the findings.
Question and response texts were extracted from the EHR with personal identifiers removed. Consult type was determined by the template the PCP selected, with the 4 options being depression, anxiety, bipolar disorder, and a generic template as described above. We examined the content of the generic template and identified a subgroup of consults using this template for questions about psychotic disorders.
We developed a coding framework modeled on prior work classifying eConsult questions and responses and modified by the research team [
Within the major domains, we used a hybrid inductive-deductive process to identify categories based on a previously described taxonomy of clinical questions [
The two coders analyzed the consult questions and responses independently, and we evaluated agreement using interrater reliability to identify areas of discrepancy and refine the coding structure based on preliminary findings [
Additionally, when analyzing questions, we noted whether the PCP had expressed diagnostic uncertainty and whether this was an initial presentation for a patient or the PCP or other provider had tried one or more treatments that failed or resulted in ongoing uncontrolled symptoms. We also noted whether psychiatrist responses included multiple therapeutic options and contingencies or thresholds for starting or adjusting treatment.
Additional information about consults was obtained via chart abstraction, including completion time by the specialist and implementation of consultant recommendations by the PCP. We considered the recommendation to be implemented if the PCP placed an EHR order for one of the recommended medications, tests, or referrals (including in-person psychiatric evaluation). We also considered a recommendation to have been implemented if the PCP documented taking any of the actions recommended by the consultant in the EHR (including documentation of telephone communications, electronic messages to patients, and clinic notes). Simply documenting the recommendation without any action did not count as implementation. We did not categorize whether PCPs implemented all of the recommendations as many responses included multiple therapeutic options for the PCP to choose between. The time frame for documentation was limited to the 6 months following the eConsult, and assessment of PCP implementation of specialist recommendations was limited to eConsults that recommended a change to existing care plan to be carried out within the study window. Both reviewers extracted implementation data for all 50 consults independently with moderate levels of agreement (kappa was 0.55). We used this initial data to clarify our classification system and then reviewed the data again, resolving any remaining disagreements via discussion with the research team.
We used descriptive statistics to report patient characteristics, question and response characteristics, and conversion to in-person consultations. Interrater reliability for categorization of consultative focus and implementation of recommendations was determined using Cohen’s kappa statistic.
The UCSF Committee on Human Research approved this study.
The consult patients ranged from 22 to 80 years of age; 54% (27/50) were female, 62% (31/50) white, and 14% (7/50) were African American (
Patient demographics for the first 50 psychiatry eConsults.
Patient characteristics | Total | ||
Age, years, median (range) | 48.5 (22-80) | ||
Female | 27 (54) | ||
Male | 23 (46) | ||
Commercial | 19 (38) | ||
Medicare | 5 (10) | ||
Medicaid | 16 (32) | ||
Medicare-Medicaid | 10 (20) | ||
White | 31 (62) | ||
African American | 7 (14) | ||
Asian-Pacific Islander | 6 (12) | ||
Hispanic | 1 (2) | ||
Other/unknown | 5 (10) | ||
English | 49 (95) | ||
Non-English | 1 (2) |
Depression was the most common consult template selected by PCPs with 40% (20/50) of consults, followed by the generic template with 24% (12/50) of consults, anxiety with 16% (8/50), psychosis with 14% (7/50), and bipolar with 6% (3/50).
For consults that used the generic template, 37% (7/19) of the questions pertained to psychotic disorders. The remaining 12 of the 19 consults placed using the generic template included less common psychiatric disorders, questions about medical comorbidities, access to care, and others. PCPs expressed diagnostic uncertainty in 22% (11/50) of cases, and for 58% (29/50) of patients, PCPs reported that the patient had persistent or worsening symptoms despite having initiated one or more treatments. PCPs noted prior treatment failures for 80% (16/20) of patients with depression, 88% (7/8) of patients with anxiety, 33% (1/3) of patients with bipolar disorder, 43% (3/7) of patients with psychosis, and 20% (2/12) of patients with other diagnoses.
She is a 64-year-old woman with a history of chronic pain and depression, who has failed multiple selective serotonin reuptake inhibitors and serotonin norepinephrine reuptake inhibitors in past and was recently started on bupropion, but bupropion has not been effective and patient has also had an episode concerning for possible seizure activity. In addition to depressed mood, significant concerns include fatigue, decreased appetite, and significant weight loss.
If the patient is more concerned about pain and its treatment, I would initiate duloxetine. Start at 30 mg PO daily, and pending tolerance, increase to 60 mg PO daily. If improving sleep, increasing appetite, and weight gain is more desirable, mirtazapine would be a good idea. Start at 7.5 mg PO daily, and pending tolerance, increase to anywhere from 15 mg to 45 mg. Sedative effects tend to occur at 15 mg or below, less sedation above 15 mg.
Most (49/50, 98%) consult questions pertained to management; only in 4% (2/50) was the PCP seeking consultation on diagnosis. Within the management domain, the majority were queries about medication management, including 74% (37/50) of questions about medication choice and 32% (16/50) of questions about drug side effects or interactions. PCPs asked about navigating the health care system in 14% (7/50) of the questions, specifically for assistance helping patients access specialty mental health treatment. For example, PCPs asked for recommendations for services based on insurance status (particularly for patients with public insurance) or for specific types of treatment such as counseling for posttraumatic stress disorder or diagnostic evaluation for eating disorders. Other questions pertained to recommendations about psychotherapy or in-person evaluation.
Psychiatrist responses focused on both diagnosis and management, with 60% (30/50) of responses addressing the diagnosis and all 50 (100%) making recommendations about management (
Management responses also focused on medications, with 72% (36/50) offering advice on choice of medication. Responses also included assistance with specific medication titration instructions in 42% (21/50) of cases, information about side effects in 48% (24/50) of cases, and conditions for starting or adjusting treatment in 34% (17/50) of cases. In 50% (25/50) of responses, consultants offered multiple therapeutic options to be guided by PCP choice or clinical conditions. For example, the consultant might recommend several different medication options depending on specific clinical situations, as seen in
Psychiatry consultants recommended in-person evaluation or treatment in 26% (13/50) of cases and agreed with ongoing management in primary care for the remaining 74% (37/50) of cases (
Breakdown of primary care provider consult question and psychiatrist response text. Question and response texts were coded for 2 domains (diagnosis and management) and for presence and absence of specific categories within these. There could be multiple domains or categories per question or response.
Originator | n (%) | ||
Diagnostic questions | 2 (4) | ||
49 (98) | |||
Medication choice | 37 (74) | ||
Drug side effects or interactions | 16 (32) | ||
Navigating the health care system | 7 (14) | ||
Psychotherapy recommendations | 3 (6) | ||
Need for in-person referral | 3 (6) | ||
30 (60) | |||
Interpretation of overall diagnosis | 28 (56) | ||
Additional diagnostic testing | 3 (6) | ||
Obtain additional history | 7 (14) | ||
50 (100) | |||
Medication choice | 36 (72) | ||
Conditions for adjusting treatment | 17 (34) | ||
Specific medication titration instructions | 21 (42) | ||
Information about side effects | 24 (48) | ||
Provided multiple therapeutic options | 25 (50) | ||
Recommended psychotherapy | 12 (24) | ||
Navigating health care system | 11 (22) |
aPCP: primary care provider.
Referral for in-person consultation by consult category.
On chart review, PCPs implemented psychiatrist recommendations in 76% (38/50) of consults. The implementation reflected the consultant responses, with the majority of cases consisting of PCPs ordering, modifying, or discontinuing medications. In cases where the recommendations were not implemented, situations varied. For some the patient either improved without intervention or did not follow up. In other cases, PCPs either documented reasons for choosing not to implement recommendations or did not acknowledge the consultant recommendations in the chart. All consults received responses, and the average consultant response time was 1.4 business days, with a median response time of 1 day and range of 0 to 30 days. In comparison, mean and median wait times for an in-person consult in the psychiatry practice during the final 3 months of the study period were 46 days and 38 days, respectively.
This study describes the early experience of a large health practice with an innovative program using eConsults from primary care to psychiatry for management of mental health problems. We found that psychiatry eConsults were a feasible way to support PCPs in managing complex psychiatric issues in the primary care setting, with PCPs using the service and frequently implementing at least one of the recommendations. For the majority of patients, particularly those with depression and anxiety, the consulting psychiatrist supported ongoing management within primary care without requesting an in-person psychiatric evaluation and provided a range of strategies that facilitated ongoing primary care–based treatment. In addition, eConsults provided consultation that circumvented common barriers to mental health treatment, including access to treatment providers and insurance coverage since eConsults were available to patient with all payer types without prior authorization from their insurer.
PCPs mainly used eConsults for guidance about treatment, particularly the management of psychotropic medications. PCPs generally made their own diagnostic assessments and used eConsults when confronted with treatment resistance or treatment failure. They also used eConsults to query about systems of care, either navigating the system more broadly or specific recommendations about psychotherapy or psychiatry. Although they were rarely directly asked, the psychiatric consultants often offered suggestions regarding diagnosis, which generally involved summarizing or clarifying the diagnosis. This may reflect a role for eConsults as both management and educational tools.
Psychiatry eConsults share several key features with eConsults in other specialties. eConsults to psychiatry provided timely recommendations for care, with average response time similar to reported values of less than 3 days [
In our study, some PCPs had queries regarding helping their patients navigate the health care system (14%). In particular, this was common for patients who had public insurance that excluded them from obtaining mental health services in our institution due to a carve-out arrangement. System navigation was not a common theme reported for consults for medical subspecialties [
Given the challenges of meeting growing mental health care demands, there is a need for novel approaches to integrated behavioral health. There is evidence that collaborative management between PCPs and behavioral health providers can provide high-quality care, improve access, reduce stigma, and be cost effective [
Our analysis has several limitations. First, as a largely qualitative analysis, our sample may not be completely representative of all possible eConsults to psychiatry. Our study was also conducted in a single institution with a robust EHR, and the results may not be generalizable to other institutions with different capabilities. However, we feel our study provides an in-depth look at the nature and scope of PCP questions regarding management of mental health conditions that may be more broadly applicable. Further, our initial interrater reliability was moderate for some portions of our content analysis, although we used this information to identify discrepancies in the coding process, refine our categories, and ultimately create a rigorous coding scheme derived through consensus within the research team. Another important limitation of our analysis is the lack of availability of data for mental health care utilization. Because of separate payment structures for mental health as well as a high percentage of psychiatry referrals outside of our health system, we were unable to track the percentage of eConsults compared to total psychiatry referrals or the impact of eConsult on utilization. We were also unable to evaluate treatment outcomes or compare these with outcomes for traditional care.
Overall, our data suggest that eConsults show promise as one means of supporting primary care providers to deliver mental health care to patients with common psychiatric disorders, particularly those with depression and anxiety. By extending and enhancing traditional psychiatric consultation within the primary care setting, the eConsult model adds to the burgeoning literature on innovative models for integrated mental health. Future efforts should be directed at evaluating the effectiveness of this intervention in terms of clinical and systems outcomes as well as PCP and patient satisfaction to improve overall understanding of best practices for behavioral health integration.
Study setting.
Development and implementation of eConsult.
eConsult templates.
electronic health record
primary care provider
University of California, San Francisco
work relative value unit
The authors would like to acknowledge Jeffrey Tice, MD, James Bourgeois, MD, and Lyudmila Popova, PhD, for comments on the manuscript.
None declared.