Appropriateness of Telemedicine Versus In-Person Care: A Qualitative Exploration of Psychiatrists’ Decision Making

The pace of adoption of telemedicine by U.S. psychiatrists has been staggering. Most psychiatrists reported no experience with telemedicine before the COVID-19 pandemic, but in the spring of 2020, 85% reported seeing more than three-fourths of their patients via telemedicine (1). Although overall telemedicine use has declined from its peak early in the COVID-19 pandemic (2), it has remained a dominant model among specialty behavioral health providers. As of December 2020, more than half of all behavioral health visits were estimated to be via telemedicine (3).

It is clear that telemedicine will become a permanent feature of practice and that psychiatrists will increasingly offer hybrid care models consisting of both telemedicine and in-person visits (4). Although the future of telemedicine policy in other clinical areas is uncertain, payers appear to be committed to covering telemedicine visits for behavioral health care in a hybrid model. In the Consolidated Appropriations Act of 2021, Congress permanently expanded telemedicine coverage for behavioral health but required that clinicians have some in-person visits (5).

The potential advantage of a hybrid approach is that it allows clinicians and their patients to leverage the relative strengths of each modality. For example, telemedicine can increase access and convenience, and in-person visits can better support general medical examinations and onsite laboratory testing. However, hybrid models add another layer of complexity, requiring clinicians to use their judgment to make individualized decisions about the appropriateness of different modalities. Given that many patients and clinicians are becoming comfortable with in-person care as the pandemic progresses (6), psychiatrists are increasingly in a position to determine which visits are best conducted via telemedicine versus in person. Although some telepsychiatry guidelines and best practices have been developed (7–9), we could not identify any literature describing how psychiatrists decide on different care modalities and how they assess the appropriateness of telemedicine for a given patient.

Understanding how clinicians make decisions about telemedicine versus in-person care is key to informing clinical practice guidelines and reimbursement policies that mandate in-person care with some frequency. To fill this knowledge gap, we conducted semistructured interviews with psychiatrists across the United States to explore what influenced their decisions about offering telemedicine.

Methods

Study Participants and Sampling Strategy

From June 25 to August 4, 2021, we conducted semistructured interviews with 20 psychiatrists practicing in outpatient settings. We worked with a research firm with an online panel of 730,000 physicians to recruit participants. The panel comprises physicians who have joined the platform to access clinical content and continuing medical education activities and has been used in multiple research studies (10–12).

Psychiatrists in the panel were sent an eight-item screener survey to assess eligibility for study participation, and those deemed eligible were invited to participate in a 60-minute videoconferencing interview with the study team. Eligible psychiatrists needed to have conducted both telemedicine and in-person visits in the previous month, and no one modality could represent >90% of all visits. Moreover, they needed to work in an outpatient setting (providing direct clinical care for ≥8 hours per week) and treat adults. We conducted criterion sampling, but we also sampled for heterogeneity to ensure that the final sample of psychiatrists varied along the following dimensions: practice setting, U.S. region, and rurality.

Interviews

At the start of each interview, we defined telemedicine visits as synchronous video visits. We included questions about audio-only visits; however, we did not include findings unique to that modality because participants reported providing relatively few audio-only visits. Interviews followed a semistructured protocol that incorporated critical incident technique and clinical vignettes. Questions were designed to uncover both conscious and unconscious factors that influenced the decision to offer telemedicine visits.

Critical Incident Technique

We first used the critical incident technique to uncover unconscious factors that influence the decision to offer telemedicine. This method asks participants to reflect on real-life experiences and, as such, “obtains a record of specific behaviors” (13). The technique has been used in previous studies to explore prescribing and referral behaviors (14–16). We asked psychiatrists to describe an example of a telemedicine visit in which the quality of the visit was equivalent to that of an in-person interaction and an example in which they felt uneasy or uncomfortable about the quality of the visit.

Clinical Vignettes

Second, we presented four clinical vignettes developed by two psychiatrists on the study team (Box 1). Clinical vignettes are a research method in which a story or scenario is presented to research participants, and participants are prompted to reflect on it (17). The literature on vignettes notes that this research method is particularly valuable in exploring an unobservable phenomenon (e.g., clinical decision making) (18).

In the interviews, participants were asked to read each vignette and discuss their thoughts on the appropriateness of telemedicine in that particular case. Interviewers probed what factors participants considered in their decision.

Three members of the study team trained in qualitative research conducted the interviews. Interviews were recorded and transcribed. Participants were given a $225 gift card for their participation and provided verbal informed consent. This study was approved by RAND’s Institutional Review Board.

Data Analysis

We analyzed interview data by using inductive thematic analysis (19). We first conducted open coding of interview transcripts, followed by axial coding to establish connections among themes. We treated decision-making factors that participants reported consciously considering as well as those revealed through thinking aloud (in vignette or critical incident technique questions) as equally important and triangulated findings across the two types of responses.

The lead author (L.U.-P.) developed the initial codebook by reviewing eight transcripts. The codebook was then discussed, refined, and finalized in group meetings among three members of the study team (L.U.-P., A.M.P., J.S.). The lead author then coded all transcripts using NVivo (version 12) data analysis software (20). Analyses began after three interviews had occurred, and saturation was reached (i.e., no new themes emerged) after 16 interviews.

Results

A total of 20 psychiatrists representing 13 different U.S. states participated in the interviews. On average, participants reported that they had provided 59% of their visits via telemedicine (Table 1).

Decision-Making Context

Impact of spring 2020.

Experiences converting in-person visits to telemedicine visits in the spring of 2020 changed attitudes about telemedicine. At the time of the interviews, participating psychiatrists were comfortable conducting telemedicine visits with a wide variety of patients. Most psychiatrists in the sample provided all of their visits via telemedicine for a period of 2–12 months starting in March 2020, and because this transition was viewed positively, psychiatrists reported becoming comfortable with this care modality. As a psychiatrist from a hospital-based outpatient clinic in Georgia explained, “For the most part, I don’t think that 5 years ago anybody would have thought that seeing 100% of your patients via telemedicine would work, but of course, it has.”

Decision making without constraints.

Psychiatrists could decide on telemedicine use without significant constraints and reported few limits on their decision making by their organization or practice setting. As such, the decision to offer telemedicine was largely their own, and they were free to use their clinical judgment. Many psychiatrists commented that the decision to conduct telemedicine visits requires clinical discretion and that any move to limit clinician autonomy (e.g., by limiting reimbursement to certain types of visits) could discourage legitimate uses of telemedicine. A psychiatrist in private practice in New York said, “For every example of something that’s usually a bad idea [for telemedicine], I can come up with one scenario where it might be a good idea.”

Role of patient preference.

Patient preferences drove a significant fraction of in-person care. Participating psychiatrists repeatedly mentioned that when deciding whether to offer telemedicine or in-person visits, they generally tried to accommodate patient preferences. A psychiatrist in private practice in California explained, “I believe a patient’s comfort, a patient’s preference comes first. So, for me it definitely makes a big difference. If they want to come in, then I try.” Multiple participants pointed out that most of their in-person visits were driven by patient demand and not by their requests. Participants provided several reasons why certain patients prefer in-person visits, including better connection and rapport and lack of comfort with technology.

General Beliefs About and Approach to Hybrid Care

Candidates for telemedicine.

Psychiatrists believed that almost all patients are good candidates for telemedicine and that telemedicine is appropriate for most patients. As a participant from a community mental health center in Washington State noted, “I really haven’t run into anything where I feel like it can’t be done with telemedicine versus in person.”

Beliefs about hybrid care.

Participants disagreed on the appropriateness of telemedicine alone versus hybrid models in the long term. Some participants thought that some patients could be treated exclusively through telemedicine; however, others reported that for provision of high-quality care, it is necessary to offer hybrid models and mix telemedicine with in-person care. For example, one concern was that telemedicine on its own can lead to deterioration in a patient’s condition over the long term. A psychiatrist in private practice in New York explained, “Behavioral activation is part of the treatment for depression. Letting you sit at home [with telemedicine] and do nothing is actually worsening depression.”

Telemedicine as part of longitudinal care.

In selecting the care modality, psychiatrists did not focus on whether a particular “visit” was appropriate for telemedicine. Rather, they considered whether telemedicine was appropriate for an individual as a component of their longitudinal care, and they reported experimenting and adjusting over time. Psychiatrists explained that patient engagement as well as various nonclinical factors are critical to the success of telemedicine and that these factors are not immediately evident. As a result, psychiatrists typically experiment with telemedicine and adjust the treatment plan depending on “how it goes.” Participants also noted that if telemedicine is not successful, they can correct the course of treatment by requesting that the patient attend an in-person visit. A psychiatrist from a hospital-based outpatient clinic in Texas explained, “I’d be willing to try telemedicine once [with this patient], and then if I thought it didn’t go well, it’s like, ‘Hey, I need to see you in person maybe tomorrow or within the next week.’”

Visit frequency.

Psychiatrists argued that the frequency of visits is more important than the modality and that some telemedicine care (if that is the only option) is preferable to no care. Participants pointed out that because telemedicine removes certain treatment access barriers, it can allow patients to be seen more often. Some noted that ensuring frequent contacts trumps other considerations. As a psychiatrist in private practice in Washington noted,

A different psychiatrist in private practice in California said,

Several participants acknowledged that telemedicine visits may not be equivalent to in-person visits for certain scenarios (e.g., a patient with a psychotic disorder and low function) but that telemedicine can still play a role. A psychiatrist in private practice in New York explained,

Facilitating engagement.

Telemedicine was viewed as a tool to “meet patients where they are,” and psychiatrists leveraged it to facilitate engagement in care. Psychiatrists felt that telemedicine provided flexibility to patients and that offering telemedicine can improve attendance. According to a psychiatrist from a community health center in Kansas, referring to a patient who had missed some in-person visits, “I think that a flexible solution would be telemedicine, because we need to meet the patient where they are. So, this offers a flexible alternative.” Furthermore, interviewees suggested that after patients are seen more regularly with telemedicine, they can more easily be transitioned to in-person visits when needed. A psychiatrist in private practice in Michigan noted, “If a patient is struggling to make [it] in person, telemedicine can offer a path to reengagement.” A different psychiatrist in private practice in Rhode Island said,

Telemedicine as a bargaining chip.

Psychiatrists framed telemedicine as a privilege and used the offer of telemedicine as a bargaining chip in negotiations with patients. Multiple participants discussed actively negotiating with patients and offering telemedicine to improve attendance or adherence with treatment. A psychiatrist in private practice in Maryland described a hypothetical conversation with a patient:

Several participants also spoke about requiring patients to demonstrate proper telemedicine etiquette and to comply with the rules of the visit. According to a psychiatrist from a hospital-based outpatient clinic in New York,

Factors Influencing the Decision to Offer Telemedicine to Individual Patients

Required elements for therapeutic rapport.

Psychiatrists pointed out that for a patient to be a good candidate for telemedicine, several elements were required to support therapeutic rapport. Clinicians felt these elements predicted success with telemedicine better than did any diagnosis or fixed patient characteristic. For telemedicine to work well, the patient must be willing and able to engage (i.e., the patient and psychiatrist are open and forthright via videoconferencing), be able to take directions (e.g., adjust the camera angle and follow instructions in a mental status examination), have good Internet connection and technical setup, be respectful of telemedicine etiquette (which several psychiatrists suggested can be particularly challenging for patients with borderline personality disorder or oppositional defiant disorder), and be able to secure a private place to meet. The participants also noted that what contributes to complexity of telemedicine visits is that the presence of these elements is not clear to the psychiatrist in advance, so it is necessary to attempt telemedicine with each patient to determine whether the required elements are present. (Illustrative quotes on these points, as well as on the themes presented below, are included as an online supplement to this article.)

Contextual considerations.

In addition to assessing the presence of required elements for therapeutic rapport, psychiatrists considered several factors when deciding whether a particular visit could occur via telemedicine. They considered whether the patient had major barriers (e.g., transportation or scheduling) to coming in person and whether the in-person visit provided an important therapeutic benefit (e.g., if remaining at home could inadvertently contribute to decline in a patient’s conditions). They also took into account the relative importance of conducting a general medical examination and taking vital signs (and whether the patient had the ability to monitor vitals independently with home monitoring devices). Finally, they considered whether the patient was routinely meeting in person with other providers (such as a primary care provider) and whether they had supportive caregivers who could provide collateral information or technical support during the telemedicine visit.

Psychiatrists also reported requesting telemedicine visits with specific patients to leverage telemedicine’s unique benefits. For example, they sometimes recommended a telemedicine visit to obtain insight into patients’ home lives and how they were functioning at home, to involve family members who could not accompany patients in person, or to ensure the safety of the psychiatrist and staff (e.g., in cases where a patient may become aggressive or violent).

Challenging conditions or situations.

No consensus was reached among the interviewed clinicians on conditions or situations that were less appropriate for telemedicine. A subset of respondents felt that telemedicine was inappropriate or less appropriate for the following conditions: substance use disorders, suicidal ideation, psychotic disorders, eating disorders, trauma, greater functional impairment, and intimate partner violence. However, others disagreed, suggesting that telemedicine offered unique benefits even in these cases (e.g., improved patient safety and reduced likelihood of escalation).

Discussion and Conclusions

In semistructured interviews, 20 psychiatrists reported that almost all of their patients were good candidates for telemedicine visits in the context of hybrid care models. What drove telemedicine versus in-person care was patient preference and situational factors such as access to a private space, rather than any particular diagnosis or patient demographic characteristic. Psychiatrists described numerous factors affecting their decision to offer telemedicine, and they were generally driven to try telemedicine and adjust modalities as needed to “meet patients where they are” and improve engagement in care.

A 2021 Substance Abuse and Mental Health Services Administration–supported guide on telemedicine for mental health professionals echoed some of the same factors discussed by psychiatrists in our sample (9). For example, the guide suggested that when assessing whether telemedicine is appropriate, it may be necessary to consider the patient’s history of cooperativeness with the provider, substance use, violence, and self-injurious behavior. It also suggested that psychiatrists trial telemedicine before committing to ongoing telemedicine care (21), suggesting a need to “experiment” with telemedicine to assess appropriateness. We identified additional factors (e.g., the possible therapeutic benefit of coming in person and the safety of psychiatrists and staff). Notably, we also found that the offer of telemedicine can be leveraged in negotiations to improve treatment engagement, attendance, and adherence. If telemedicine is successfully negotiated in this way, and assuming patients feel more supported and engaged in care, hybrid care models could lead to higher-quality care. Furthermore, offering the option of telemedicine, with its added convenience and potential for privacy, could help individuals expressing ambivalence about treatment or behavior change overcome their personal activation barriers to engagement.

Our findings have implications for future reimbursement policy. At present, the Centers for Medicare and Medicaid Services is considering how often to require in-person visits in hybrid care models. The current proposal is that behavioral health providers using telemedicine with Medicare beneficiaries should be required to have an in-person visit within 6 months of initiating telemedicine and then every 6 months thereafter (22). Our findings suggest that clinicians offered in-person care because of patient preference, but few felt that any specific scenarios required only in-person care from a clinical standpoint. Rather, many felt that telemedicine increased patients’ engagement in treatment and that decisions about telemedicine use remained fluid. Although there may be arguments for requiring periodic in-person visits (e.g., to reduce the potential for fraud and deterring overuse of care and increased spending), it does not appear that psychiatrists perceive intermittent in-person visits to be essential for high-quality care.

This study had some limitations. First, participants’ responses may have been influenced by social desirability bias. Second, we used hypothetical case vignettes, which may have not exactly mimicked real-life clinical practice, and we presented only a few of many possible case scenarios. Third, because of time constraints, we did not ask participants to differentiate between different visit types in their comments (e.g., medication management and psychotherapy sessions). It is likely that the visit type influences decisions about the appropriateness of telemedicine.

To our knowledge, this study is the first that has explored how psychiatrists make decisions about the appropriateness of telemedicine. Future research should further investigate how hybrid care models with different doses and timing of telemedicine affect care quality. This research is especially important for clinical scenarios described as more challenging for telemedicine (e.g., active suicidal ideation and eating disorders).