Physician Perspectives on Telemedicine at an Integrated Health System During the COVID-19 Pandemic

Introduction

Telemedicine (TM) is the use of two-way communication technologies to provide remote health care through a variety of modalities. ¹ TM has been conceptualized and implemented for more than 100 years. ² However, due to the COVID-19 pandemic, TM has been more widely used to provide patient access to health care without the risk of transmission from face-to-face encounters. ^{3 –5} By the week of March 27, 2020, the number of telehealth visits across all specialties had already increased by 150% in comparison to the year prior. ⁶ One institution noted a 400% increase in telephone visits from February 2020 to March 2020, and then, an additional 400% increase from March 2020 to the middle of April 2020. Over the same time frame, the same institution cited a 900% increase in video clinic visits. ⁷

As TMs use has become more widespread, so has its acceptance by both physicians and patients. ^8,9 Prior studies have described the necessary framework, infrastructure, and execution of their systems' transformations during the initial COVID pandemic to handle TMs increased use, ^10,11 as well as compared its pre-COVID and post-COVID usage among different specialties. ^12,13 However, to the best of the authors' knowledge, no studies have attempted to compare the different opinions and uses of TM across a combination of different physician subspecialists, experience levels, and those in different practice environments. This is critical to understand because different physicians may experience varied success in completing an impactful patient evaluation and treatment by telephone or video. Our study provided physicians across our large integrated health network with a survey designed to better understand their perspectives and experiences with TM during the initial COVID pandemic in the hopes that it could help project further use and resource allocations.

Methods

A 27-component questionnaire approved by our institutional review board (Supplementary Data) was created and distributed across our health system during the Spring 2020 during the height of the COVID-19 pandemic. The survey was virtually distributed using the REDCap survey software to 1,794 physicians who practice across all specialties within 4 regional locations. There was no formal consent process for the distribution of the survey, and participants either freely chose to participate or declined by not completing the survey. Specialists were divided into seven groups, based on the level of patient interaction in the specialist's standard, pre-pandemic practice (Supplementary Data).

Physician perspectives were categorized by their opinions on different aspects of TM visits, including overall satisfaction, expectations of compensation, and frequency of facing technological challenges. This included perceptions of all different variations of TM visits, including those with and without video and using home versus office computer. Descriptive statistics were generated for physician characteristics, such as specialty grouping, primary practice environment, primary regional campus, and years in practice. Frequency and percentages are reported for categorical variables. Chi-square and Fisher's exact tests were used to compare the differences in perspectives among specialty groups, years in practice, and primary practice environment.

Weighted and unweighted descriptive statistics were generated for physician characteristics, such as years at institution, specialty group, title, and employment status. Inverse probability weighting was used for weighted analyses of the data. Frequencies and percentages are reported for categorical variables, whereas the means, standard deviations, medians, and interquartile ranges are reported for continuous variables. Finally, chi-square test and/or independent t-test were conducted to compare unweighted and weighted differences in physician characteristics between physicians who responded and physicians who did not respond to the survey. Cronbach's alpha was used to assess the reliability of survey questions.

All analyses were performed in SAS Enterprise Guide v8.2 (SAS, Inc., Cary, NC) and R package version 64 4.0.3, with contrasts of p < 0.05 considered statistically significant.

Results

Of the 1,794 physicians, 365 responded to the study, yielding a 20.3% response rate. The characteristics of the study population are described in Table 1.

There were a total of five sets of question items, which measured different components of the physician's attitude. Reliability was assessed with Cronbach's alpha (Supplementary Data). Questions 21–23 reliably measured physician's attitude on provider compensation (α = 0.78). Questions 24–26 reliably measured physician's attitude on expected difference in patient payment between an in-person visit and TM visit (α = 0.81). Questions 23 and 26 reliably measured physician's attitude on expected difference in patient payment and provider compensation between a TM visit with video and TM visit without video (α = 0.74).

Discussion

At the advent of COVID-19, there was a dramatic increase in the utilization of TM. This rapid increase in utilization was born of necessity both for patients to maintain access to their physicians and for physicians and health systems to continue to generate necessary revenue during periods of clinical shutdown. Many questions still remain, however, as to the utility of these visits and the expectations for continued use as normal clinical operations resume now years after the initial pandemic. The purpose of this survey was to assess the perceptions of TM utilization across specialties at a large integrated health system.

Over the course of the pandemic, we identified significant increases in utilization across all specialties (Table 2) but notable disparities in the perceptions of usefulness of a TM visit. For instance, psychiatrists had the highest satisfaction rate with TM at 96.7%, whereas surgeons had the lowest satisfaction rate at 61.3% (Table 7); 26.7% of psychiatrists “always” found TM visits useful, whereas only 9.3% of surgeons reported the same (Table 6). These findings were consistent with expectations for future use of TM among specialties with psychiatrists favoring utilizing more TM in the future compared with the other groups and surgeons being the least likely to want to use TM in the future (Table 5).

In addition, our data suggested that surgeons may have more difficulty in adequately completing treatment plans via TM visits alone and that they often need patients to return for an additional visit (Table 6). This contrasts other published literature that suggests that TM alone has been sufficient for preoperative evaluation in several different surgical specialties, including spine ¹⁴ and head and neck surgery. ¹⁵ Donnally et al. evaluated a surgeon's ability to identify a patient presenting with a surgical emergency such as cauda equina syndrome via TM. ¹⁴ Despite not being able to perform spine-specific tests, such as a muscle grading and checking for abnormal reflexes via TM, these tests rarely are the deciding factor of a true surgical emergency.

Rather an examiner can take a good history and evaluate relevant imaging as well as patient's gait, balance, and gross motor function adequately enough to ensure that these patients have not lost the ability to walk or urinate, or have suddenly become incontinent. ¹⁴ Additionally, Applegate et al. demonstrated that surgeons found TM to be sufficient for evaluating patients in the preoperative evaluations setting, notably, after the patient had already been physically examined and diagnosed at least once in the in-person setting. ¹⁵ The present study queried specialist's perspectives concerning TM across their entire practice, and not just for specific uses.

One of the biggest driving factors in different subspecialists' opinions regarding TMs utility was the ability, or lack thereof, to perform an adequate physical examination. Surgeons had an 86.7% response rate of feeling very limited in ability to perform adequate examinations via TM visits, whereas psychiatrists reported a 22.2% rate of being not limited at all and only a 37% rate of being very limited (Table 6). The differences highlighted between psychiatric practice and surgical practice suggest that specialists who rely on physical examination may struggle to adequately treat patients via TM only, whereas specialists who do not rely on physical examination are more willing and able to treat patients via TM.

This presents an opportunity to develop standardized physical examination techniques for TM visits. ¹⁶ There were also significant differences in the ability or lack thereof to adequately complete a physical examination based on practice environment, with those in blended and outpatient practices feeling more limited than those in inpatient practices. There was no significant difference found based on years in practice. This could suggest that even the most experienced physicians struggled with performing physical examinations as much as their less experienced counterparts via the TM platform, highlighting a learning curve associated with utilizing TM visits. In time, more familiarity with an optimized TM physical examination may change physician perspectives on TM.

In the data presented, psychiatry seemed most satisfied with the practice of TM. Several other studies comparing TM use among different subspecialities have also found psychiatrists to view TM favorably and be apt in its usage. ^12–13,17 Institutions should consider these data when allocating future TM resource and planning delivery of psychiatric care.

We found that plans for future TM directly coincided with personal satisfaction from TM use during the pandemic. As stated previously, psychiatrists most favored future TM use and surgeons were least likely to want to use TM of different specialists. In terms of different practice environments, inpatient physicians were the most likely to use TM in the future and had the highest satisfaction rates with its use during the pandemic. Years in practice did not show any significant difference in future TM use.

It is unclear why inpatient physicians had a much higher satisfaction rate than their peers. Immediate advantages of inpatient TM use, especially during the pandemic, include less person to person contact with the most at-risk patient population and less personal protective equipment use. ^18,19 However, TM also has the potential for improved patient triage before the patient presenting to the emergency room, increasing the efficiency of consultation services to improve patient access, and creating team-based visits to allow for a multidisciplinary treatment approach. ^18,19

Inpatient TM has been successful and popular with intensivists (tele-intensive care) ²⁰ and neurologists (telestroke) before the pandemic, particularly because it creates a platform for a provider to respond immediately to an urgent manner from a remote location and can eliminate some of the burden of standard in-hospital call requirements on providers. ²¹ TM has also been shown to improve patient outcomes for high-risk interhospital transport when the receiving hospital can interact with the transport team. ²² The pandemic may have served as a catalyst to introduce this technology to other inpatient physicians, who, based on our data, also felt that TM was helpful and could be expanded in the future.

Physician opinion on reimbursement regarding TM visits is critical to understand due to the impact of public policy on the viability of TM in medical practices. At the onset of the COVID-19 pandemic, the United States Centers for Medicare and Medicaid Services expanded reimbursement of TM visits, and these policies continue to evolve. Currently, any provider eligible to bill Medicare for TM or other telehealth services may do so regardless of where the patient or provider is located. ^1,23 State Medicaid programs and even private insurers have begun to make billing and reimbursement rules for TM visits as well. ²³ A critical consideration is that if a net increase in TM utilization results in a net revenue loss for medical businesses, then TM could prove unsustainable in the long term.

From a statistical standpoint, there was not a consensus on how TM visits should be billed or reimbursed. Several studies conducted before the pandemic found no significant differences in reimbursements or out-of-pocket patient expenses when comparing TM and traditional clinic visits among several surgical subspecialists. ^24,25 However, both groups found that they were billing for lower levels of visits due to the inability to perform adequate physical examinations. ^24,25 Benefits described include the ability to reallocate physical resources and decreased travel burden for patients and physicians. ²⁵ These data may be biased, however, if there were too few patients included in the studies to find the true effect that routinely billing for lower level visits has on practice revenue. The impact of recent changes in public policy on TM reimbursement is still to be determined.

There are several limitations to this study, many of which are related to it being a retrospective, survey response research study. One of the biggest weaknesses of this study was the relatively low (20.3%) response rate. This could undermine some of the data found, as it is unclear how having a larger response rate could affect our data. Additionally, the way the subspecialties were grouped was ultimately chosen by the authors and is otherwise relatively arbitrary. The data cannot be subdivided among a group because that would make sample sizes too small for any meaningful statistical analysis. Another major limitation is the assessment of physician satisfaction, as the survey used for the study was not able to tease out what aspects of a TM visit made it satisfactory or unsatisfactory for an individual. For example, Group 2 may have had satisfaction in postoperative TM visits but not new patient consultations. It is also possible that the data from our health system may not be consistent with others, or generalizable to others.

We did not analyze our data using weighted statistics because we only collected three defining characteristics of the physician respondents due to the anonymous nature of the survey and a weighted analysis may still have hidden confounding variables. Another limitation to the study is that this survey has not been repeated since it was initially distributed. Therefore, the responses from the physicians in this survey only reflect opinions on TM use during the height of the pandemic and may be different if repeated now. Additionally, there was no use of peripheral or robotic technologies in the health system where the study took place that may have added to the ability to perform physical examinations. These technologies were not made available to the study population during the time the study took place; and, to the best of the authors' knowledge, these technologies are not available or being used in the health system currently.

Conclusions

Discrepancies exist among physicians with respect to their satisfaction and expected future use of TM based on their experiences during the initial wave of the COVID pandemic. This appears to be affected by subspecialty, practice environment, and overall clinical experience. Consensus may be difficult to reach regarding reimbursement for these visits, but great care should be taken to ensure that the benefits of TM are not outweighed by fears of practice insolvency. Further work is needed to fully clarify the optimal practice setting for TM, especially considering its learning curve.