Feasibility and Acceptability of Virtual Rounds on an Academic Inpatient Pediatric Unit

We performed a survey-based cross-sectional study of pediatric residents and attending hospitalist physicians on a single pediatric hospital medicine service from September to December 2020 within a freestanding, 259-bed, university-affiliated children’s hospital. Residents and attending physicians on the inpatient service were approached by a study team coordinator and informed of the study design and inherent risks. Individuals choosing to participate in the study provided verbal consent. This study was reviewed and approved by our institutional review board (IRB No. 00259624).

On each day of service, the inpatient medical team consisted of 3 PGY-1 (Post-Graduate Year 1) residents (hereafter referred to as “interns”), 1 PGY-3 (“senior”) resident, and an attending physician. The team typically cares for 12 patients, with each intern responsible for one-third of the patients with supervision from the senior resident and attending. During daily rounds, a single intern, the senior resident, and the attending physician would be present at the bedside. The remaining 2 interns would be virtually participating from a separate workroom. Participants were instructed to continue rounding utilizing our institution’s standard family-centric rounding procedures, with the intern presenting the clinically relevant information and leading the clinical discussion with the caregiver(s) at the bedside while supported by the senior resident and attending physician. The team would round on each patient for whom the intern was caring before transitioning to a different intern.

The three members of the rounding team brought a computer on wheels to access the electronic medical record as well as a second custom-built videoconferencing solution consisting of an iPad tablet with built-in camera, a conference-style combination speaker-microphone system, and a portable battery, all on a rolling stand that allowed for adjustment of camera angle. The stand was brought into each patient’s room and generally oriented to face the patient and/or caregiver, with the senior resident responsible for transporting it and positioning it appropriately in each room. The equipment was cleaned between rooms according to hospital infection prevention protocol. In the remote workroom, a large computer monitor equipped with a webcam, microphone, and external speaker was available. The Zoom videoconferencing platform (Zoom Video Communications, Inc; San Jose, CA) was used to securely conference between the 2 locations to allow real-time, bidirectional communication between team members at the bedside and those in the workroom.

A 6-item survey using 5-point Likert-type ratings was developed for an attending physician observer to determine which technical issues may have occurred during the encounter as well as to measure overall efficiency of the encounter, perception that remote team members contributed to the encounter, and impact on teaching (Supplemental Information). The attending observer did not enter patient rooms but was present immediately outside, observing either just outside the doorway or via the virtual platform on their own portable device. The attending observer role was created to objectively assess encounters in addition to distributing surveys to the trainees after each encounter, while allowing the attending on service to focus on patient care and education. Brief 3- to 4-item surveys were developed for the bedside intern and senior resident to measure impact on learning, efficiency of the encounter, and the perception that remote team members contributed to the encounter. An 8-item survey was developed for the virtually participating interns to measure their ability to clearly view the examination and hear the participants, as well as comfort with participation, engagement in the encounter, tasks performed during the encounter, and perceived educational value. At the end of each encounter, once discussion of the patient had ceased entirely, the surveys were distributed by the attending observer to the bedside participants and then returned upon completion, before advancing to the next patient. For the encounters in which they would be participating virtually, the interns were given a packet of surveys before rounds. These surveys were completed in the workroom after each corresponding encounter and returned to the attending observer after rounds.

Surveys were developed by the research team and designed to be completed in <1 minute. Multiple residents and faculty took the surveys and provided feedback on content and structure. Based on this feedback, changes were made and subsequently reviewed by faculty with expertise in survey development to create the final versions. Surveys were distributed 1 day each week based on the availability of the attending observers and did not include weekends. All rounding encounters, defined as the interaction with that patient/family plus any team discussions pertaining to that patient before and/or after that interaction, were considered eligible for survey distribution on data collection days. For the purposes of this descriptive study, all survey responses were included for analysis even if only partial responses.

The frequencies of technical issues were reported as a proportion of total encounters. Individual survey responses were summarized using medians and interquartile ranges. Data normalcy was assessed using the Shapiro-Wilk test. Categorical comparisons of survey responses between interns, seniors, and observing attending physicians were assessed using Fisher’s exact test. In an exploratory analysis, the association between reported engagement of workroom residents on rounds and dependent variables was determined using multivariable ordinal logistic regression. Given a paucity of existing data to guide derivation of a multivariable model, associations with P <.15 in univariate analysis were included in the final model. Significance for all statistical comparisons was set to an α of .05.

A total of 82 encounters were eligible for survey distribution during the study period. Video conferencing was used in 75 of 82 (91%) of the encounters. Of the 7 encounters not utilizing video conferencing, 3 did not have a family member or caregiver present at the bedside, 1 was seen urgently before the start of work rounds, 1 had a sensitive diagnosis and discussion, 1 had the attending round separately from the resident team, and 1 did not have bedside rounds performed. Of the 75 included encounters, 75 (100%) included an intern and attending at bedside and 66 (88%) included a senior resident. Over the course of the study, 16 different residents (9 interns and 7 senior residents) provided responses as well as 6 different faculty observers.

Overall, 29% (22/75) of encounters experienced technical issues of any kind (Table 1). 45% of those issues were attributable to a poor-quality stand not being able to hold and properly position the tablet (reported as “Other” in Table 1). Otherwise, 10% (8/75) of all encounters involved audio issues and 3% (4/75) video issues. An issue with connectivity was reported in 1 encounter (Table 1).

Among bedside participants, the response rates were 97% (64/66) for senior residents, 99% (74/75) for interns, and 97% (73/75) for attending observers. Generally, both bedside interns and senior residents reported a neutral or positive effect of virtually observed rounds on their own learning experience (Table 2). Attending observers similarly reported neutral or positive effects of the virtual rounding experience on the rounding attending’s ability to teach (89%; 60/67) and the education provided by senior residents (84%; 69/73) for most encounters. The perception that the virtually observing residents contributed to the clinical encounter (defined as indicating “Agree” or “Strongly Agree”) differed between attending (54%; 39/73), senior resident (28%; 18/64), and intern (44%; 33/74; P <.001). The impact of the virtual rounding process on efficiency was reported as negative in 6% (12/212) of total survey responses, with some variation among respondent categories: 1% (1/74) for attending observers, 6% (4/64) for senior residents, and 9% (7/74) for interns (P = .045).

The response rate for interns participating virtually from the workroom was 83% (125/150). Overall, responses from virtual participants indicated that they felt engaged with the patient’s care during the video encounter (70%; 92/132) and found educational value (65%; 85/131). Those who indicated feeling comfortable asking questions in the room over the virtual platform had greater odds of feeling engaged (odds ratio [OR] 3.34, 95% confidence interval [CI] 1.97–5.66). Performing tasks pertaining to other patients (entering orders, speaking on the phone, or reviewing/documenting in the EMR) was indicated in nearly half of responses (48%; 72/150) (Table 3) and was associated with lesser odds of perceiving engagement (OR 0.42, 95% CI 0.18–0.99). There was not, however, a significant association between perceived engagement and perceived educational value (OR 0.99, 95% CI 0.99–1.00).

This study demonstrates that incorporating video conferencing into pediatric bedside rounds is technically feasible, can maintain perceived educational value for trainees in the majority of encounters, and can engage learners in encounters from which they may otherwise have been excluded. Although restrictions due to COVID-19 are currently being lifted, there is ample evidence that virtual clinical encounters will continue to play a significant role in a multitude of medical settings even after the pandemic.^9–13  In our study, virtual participation did not result in any self-reported negative effects on education. An important strength of this study is that all participants were surveyed simultaneously and immediately after each patient encounter. This adds an important perspective to the existing literature, which has focused on attitudes reported by participants reflecting on the experience at a later time.

Technical reliability (high-quality audio, clear video, and lack of technical malfunction) and efficiency were keys to feasibility and acceptability, and our study showed that they both could be achieved in virtual rounds. The technical feasibility demonstrated in our study was consistent with Rosenthal’s study using streaming with families in a NICU, which showed that 95% of virtual encounters conducted with similar equipment had no technical issues.¹⁴  Many of our reported technical issues were related to a poor-quality tablet stand that did not reliably hold the position of the iPad; we have subsequently corrected this problem by investing in hospital-grade videoconference carts identical to our video language interpreter carts. Efficiency had been expressed as a priority among our faculty and trainees, particularly after experiencing inefficiencies with other approaches such as “table rounding” as a team followed by abbreviated bedside rounds limited to a single attending and intern. Our study suggests that efficiency is not negatively affected by incorporating virtual rounds to allow wider participation, with only 6% of survey responses indicating a negative effect. While we did not formally measure the length of the rounds, it was anecdotally noted that rounds were completed before noon conference time, which is typically the goal for our resident teams.

In terms of effect on trainee education at the bedside, our study suggests that video conferencing does not have an overall perceived negative impact on the bedside senior resident’s or attending’s ability to provide education to their team, though it was notable that senior residents perceived a negative effect on their team’s learning more often than interns felt their learning had been negatively affected. This may be related to the fact that the senior residents had experienced more traditional rounds with the entire team participating in person during their previous years of residency, whereas the interns entered residency during the pandemic and therefore had not. However, survey responses from both groups still largely indicated either neutral or positive effect on education.

The experience of team members participating virtually was also positive in terms of both perceived engagement and educational value, which is consistent with findings in the few related studies that have been published during the pandemic.^6,8  The virtual platform provided an avenue for residents to participate in bedside rounds, which would otherwise not be feasible. One key finding, however, is that residents in the workroom were frequently performing tasks for other patients on the team during the patient encounters, which was associated with decreased perceived engagement. This finding is not unexpected given that previous studies have identified a perception among residents that while rounds can be an ideal setting for active learning and teaching, tensions from other obligations and time limitations detract from that experience.^3,15  For virtual rounds, one strategy to encourage trainees to focus on the patient on whom the team is currently rounding would be to assign specific tasks to remote participants prior to the encounter, such as researching a pertinent clinical question, confirming/entering orders, or providing updates to nursing staff following the discussion.

This study has limitations, including that it was a single institution study, which limits generalizability, and as such, it includes a relatively small number of total survey respondents. In addition, while there were multiple surveys per individual respondent, their responses were not linked across encounters, which does introduce potential for respondents participating in more encounters to have greater influence on the data, particularly if their perceptions of virtual rounds were extreme in either direction. However, very few extreme perceptions were reported (Table 2). There are also factors contributing to technical feasibility that could be institution-dependent, such as quality of wireless internet connections, funding availability, and institution-supported video conferencing software. Another potential limitation of our study is the fact that interns had not experienced rounds, except potentially as medical students, prior to the COVID-19 pandemic, which may have given them limited perspective on the efficiency of rounds compared with the senior residents and attending physicians. When entirely in-person rounding is permissible, it may be beneficial to present surveys similar to ours to both an in-person team and a virtually rounding team, to better assess comparative strengths and weaknesses of each. The survey development process did include multiple rounds of revisions based on feedback from residents and faculty, but further validation of the final surveys would strengthen future studies. In addition, while no family declined participation in virtual rounds over the course of our study, data regarding the perspectives of the patient, family, and attending on service on virtual rounds was not obtained and could be an area for future investigation.

Importantly, the applications of this rounding platform are numerous and are not limited to the response to physical distancing required during the pandemic. For example, it could make multidisciplinary rounds more feasible by enabling subspecialists, pharmacists, social workers, and ancillary staff members to actively participate in rounds from their respective locations. It could also enable physicians—including trainees seeking diverse educational experiences—at academic centers to participate remotely in rounds in community hospitals that may have limited access to pediatric care, or conversely could allow trainees at programs with less access to subspecialty care to rotate at tertiary care hospitals remotely. Literature evaluating telemedicine or teleconsultation approaches within pediatric critical care is increasing,^12,13  and the inevitable application to hospital medicine is an important emerging field of study.^9,11  While literature on the feasibility of virtually including trainees in these endeavors is lacking, our study suggests that it could be both technically possible and educationally valuable, and future studies of inpatient uses of telemedicine should seek to evaluate the trainee experience. This concept can be extended to outpatient visits and associated multidisciplinary discussions as well. Medical students could also “shadow” rounds virtually regardless of their location, enabling virtual rotations or experiences that otherwise might be logistically difficult to achieve. In addition, virtual rounds will continue to be useful for cases in which patients are immunocompromised or on isolation precautions while in the hospital, or for families who are unable to be physically present with their hospitalized loved ones.

The results of this survey-based cross-sectional study of pediatric residents indicate that use of video conferencing to allow remote participation in rounds is technically feasible, does not sacrifice efficiency, and has potential to maintain educational value and engagement for resident learners.