Teleneurology Expansion in Response to the COVID-19 Outbreak at a Tertiary Health System in New York City

Discussion

We found significant increases in synchronous TN visits, clinicians, subspecialties, and technology platforms in response to the COVID-19 crisis in a large academic health system in NYC. We found that patients generally reported high acceptability with synchronous TN, whereas clinicians reported numerically lower acceptability. Although we also found that MyChart logistics were the most significant reason for switching platforms, MyChart remained the most frequently used platform to complete synchronous TN visits, likely because of MyChart's position as MSHS′ primary telemedicine platform. The barriers to adoption that we identified involved lack of adaptability and malfunctions in technology platforms, administrative burdens due to frequently changing workflows, and inherent limitations of the remote neurologic examination, as well as the fact that several dozen of our neurology faculty and administrators were deployed to COVID medicine or other teams during May and part of June 2020, thereby limiting their availability to provide TN care during this period. State-wide and federal practice policy shifts that loosened restrictions on licensure and telemedicine reimbursement, and proactive ongoing small group meetings to obtain data on barriers and facilitators, as well as disseminate practice guidelines, helped increase successful expansion.

We found that patients and clinicians experienced MyChart logistical challenges with synchronous TN, which resulted in switching to non-MyChart platforms. These challenges included problems logging into the mobile application, electronically checking into the visit, and audio or video malfunctions. We also found that patients reported the least satisfaction with visit check-ins, which are consistent with the challenges reported in the clinician survey. Despite these logistical challenges, we observed that patients found TN more acceptable overall than did clinicians. It remains unclear whether this may have been related to the logistical impact of technical difficulties on the remainder of a clinician's workday or the effect of technological literacy. Although patients reported higher audio and video quality than clinicians, rapid increases in platform usage, which occurred across multiple medical specialties during the pandemic response, may have also contributed to audio and video interruptions.

Of interest, although the audio and video quality and rate of MyChart visits requiring a platform switch did not increase between April and May 2020, provider acceptability did significantly improve during this time frame, which suggests that at least some of the reported dissatisfaction with TN may have been related to the initial troubleshooting and problems that frequently accompany the implementation of any technological system. The evaluation period was admittedly limited to only 2 months, so we cannot definitively conclude whether the technical dimensions of provider experience improved as providers became more accustomed to TN workflows.

Patients reported high acceptability (4.7 of 5) as defined by likelihood of recommending TN to others, whereas the numerical rating of providers regarding the likelihood of recommending TN to colleagues (3.4 of 5) was more than a full point lower on the same 5-point Likert scale, although that rating improved to a mean of 4.1 over the course of the study period. Technology issues were reported as a challenge some of the time by providers. It is also possible that provider and patient experiences could have been influenced by the diagnoses of patients who were evaluated by TN. However, because our experience data were anonymous and not linkable to any patient information, we could not measure and compare scores across patient diagnoses. We note that most patients who were evaluated by TN during the pandemic had stable, nonurgent clinical problems, whereas in-person visits were reserved for urgent neurologic evaluation with the objective of avoiding a visit to the ED.

Furthermore, we found that many asynchronous TN services, with the exclusion of e-consults, were not used in the immediate COVID response. The latter finding is most likely related to the recent introduction of remote monitoring and digital evaluation and management services in 2019 and 2020, respectively.^4,21,22 Although both asynchronous and synchronous TN services offer important advantages in the COVID-19 response by limiting patient and clinician exposure to the novel coronavirus, our findings are most likely secondary to the Department of Neurology's proactive approach to responding to the COVID-19 pandemic by primarily converting scheduled in-person visits to telephone or synchronous TN visits and the deployment of a majority of referring clinicians from the Department of Medicine to COVID teams during this period. In addition, the Centers for Medicare and Medicaid Services, which administers the US Medicare program and jointly administers Medicaid alongside individual US States, increased payments for telephone visits during the pandemic, and our Department also used prolonged non–face-to-face services by telephone to provide services during this time.

During the pandemic, faculty from all adult subspecialties (excluding stroke and neuro-oncology, which separately covered their separate inpatient and outpatient subspecialty services) were called to inpatient COVID medicine team service. While in theory, COVID-related redeployment of neurology clinicians could have affected the different types of patients that were evaluated by TN during our reported implementation, in practice, a coverage system was purposefully created for each subspecialty having deployed faculty so that outpatients in each subspecialty practice had one or more covering faculty members (not deployed to COVID service) to care for those subspecialty patients' clinical needs during the pandemic surge.

From our formative evaluation, we identified that state telehealth laws and reimbursement policies constituted a significant outer setting barrier to TN implementation during the prepandemic period. Many policy changes at the outset of the COVID-19 pandemic effectively addressed such barriers. For example, from March 13, 2020, onward, 51 of 55 US States and Territories issued proclamations outlining modifications to licensure requirements in response to COVID-19, which allowed health care professionals to practice TN more freely across state lines.³ Specifically, Connecticut, New Jersey, Pennsylvania, and Massachusetts offered waivers or temporary licenses for providers whose patients who were no longer located in New York State. Policy reforms also authorized the use of non–HIPAA-compliant platforms and loosened restrictions for Medicare beneficiaries receiving telehealth, allowing the distant site to be the patient's home.

Judging by the high acceptability of TN that we observed among patients, we anticipate that many patients will continue to opt for TN visits over in-person care as the COVID-19 crisis subsides and that providers will apply for full licenses in neighboring states to practice TN to meet this demand. However, it remains unclear to what extent any of these policy shifts will remain in place and potentially affect the sustained adoption of TN as the global pandemic ends.

Despite the lifting of telemedicine reimbursement restrictions for Medicare-insured patients during the pandemic, our analysis found no significant difference in patient payor mix between periods. Medicare patients who were not covered for telemedicine prepandemic were offered the option to pay out of pocket for their TN care. The majority of these patients opted to pay out of pocket, which likely explains our finding that 20% of TN patients prepandemic were Medicare insured. However, our findings are expected, considering that payment reforms such as waiving of copayments and removal of restrictions were implemented for commercial insurance carriers as well as Medicare and Medicaid.

A significant implementation barrier was the limited nature of the remote neurologic examination, which we addressed through dissemination of educational material (table e-2, links.lww.com/CPJ/A264) and ongoing education sessions to clinicians. Despite these solutions, some elements such as the fundoscopic, sensory, and detailed motor examinations remain difficult to perform remotely without a qualified telepresenter. These limitations are ripe for digital health and remote monitoring device innovations to address, as well as research to study what aspects of the in-person examination are needed for clinical decision making vs solely to meet documentation requirements.^4,14

For instance, digital health solutions that adapt an ophthalmoscope to a mobile device camera to perform fundoscopy and remotely transmit fundus images,²³ as well as a mobile device application that tests visual acuity²⁴ could supplement the existing TN examination.¹⁴ Finally, a remote patent monitoring solution such as automated video analysis software to detect and quantify bradykinesia and tremor in Parkinson disease from patient videos²⁵ could potentially overcome the relatively nongranular video resolution of standard TN video encounters and thereby improve an examiner's assessment of a patient's fine motor examination.

Our results also demonstrate that the postpandemic period was associated with a tripling of TN platforms. The platforms included multiple non–HIPAA-compliant (e.g., Apple FaceTime™ and WhatsApp Messenger™) as well as HIPAA-compliant solutions (e.g., Doximity® and Zoom®) comprising a variety of implementation characteristics and requirements for patients and clinicians (table 5). Although faculty were guided to use the institution's official MyChart platform whenever possible, they were also given free choice of platform to use for TN, which most directly explains the numerical increase in TN platforms.

More specifically, the use of several non–HIPAA-compliant platforms is consistent with the Office of Civil Rights' announcement of HIPAA compliance enforcement discretion in late March 2020. Among HIPAA-compliant platforms, the use of VSee, Zoom, and Doximity during the pandemic response is likely explained by several practical factors. VSee had been used for TN before the pandemic and remained favored by a small number of providers. Clinicians began to use Doximity for its video call functionality, which was discovered incidentally as clinicians searched for solutions to avoid disclosing their mobile phone numbers when contacting patients via telephone during the pandemic. Finally, most clinical faculty were already using Zoom for administrative videoconferencing by the time the COVID-19 pandemic response began, which made Zoom a straightforward alternative to the official MyChart platform.

In addition, Zoom was probably the most commonly used non-MyChart platform due to functionality that addressed adaptability barriers we identified during the formative evaluation of our TN implementation (table e-1, links.lww.com/CPJ/A263), including screen sharing, third-party conferencing, and desktop deployment through a hyperlink rather than mobile device application. Despite these advantages, in cases where clinicians required to switch from MyChart, FaceTime was the most commonly used non-MyChart platform. The most likely explanation for this finding is that the MyChart TN capability was only available to clinicians on iOS devices, such as the iPhone or the iPad, which have FaceTime as a built-in application (table 5). Alternatively, this finding may be due to the leading position occupied by iOS platforms in the US mobile device market,²⁶ or that FaceTime video calls can be made more easily than with Zoom, which requires creating a meeting and sending a Web link to the patient.

Our study was limited by several factors. First, we did not assess the reasons for which patients or providers chose to use TN over telephone visits or deferral of care and therefore could not ascertain and differentiate how much of the increase in TN volume was attributable to patient choice or departmental dissemination efforts. Moreover, we excluded telephone visits from our definition of TN, which may have allowed a better understanding of factors associated with TN utilization. Second, we relied on anonymous survey information for patient and clinician experience, which may have introduced response bias and limited our ability to adjust our results for multiple responses provided by the same respondents. We also were not able to ascertain the platforms used by patients responding to the survey. Third, the pandemic period in our study was restricted to 2 months, which limited our ability to understand the long-term adoption and effects of TN at our institution, particularly with the potential elimination of emergency policies, such as shelter-in-place orders and suspension of HIPAA enforcement. Fourth, we lacked clinician and patient experience data for the entire 2-month pandemic period, and a substantial proportion of our faculty were deployed to COVID service during this period and unable to engage in TN, although covered by their colleagues. Finally, the clinician-facing surveys we used only assessed clinicians' experience with technology, rather than specific aspects of their clinical practice in TN, such as clinical decision making or diagnostic accuracy.

In addition, we did not include visit data from hospital-based clinics, which are not operated directly by the Department and have a higher proportion of patients with Medicaid insurance than faculty practices. Our results on patient perspectives and barriers may therefore not reflect those of the subpopulation with significant socioeconomic barriers to technology, including Wi-Fi, broadband internet access, and smartphones. These data may have generated valuable and translatable insights to other institutions contemplating similar implementations and platforms to ours.

This analysis of an academic neurology department's response to the COVID-19 public health crisis illustrates significant expansion in utilization and technology platform usage associated with the pandemic. Patients reported high acceptability, whereas clinicians found TN to be relatively less acceptable due to technological challenges. Subsequent studies should investigate the optimal strategy for implementation, as well as the clinician, patient, and health system effects of TN use. Further studies should also investigate the relationships between TN utilization, socioeconomic factors, and technological literacy.