A Year of Inpatient Telehealth: Lessons From the COVID-19 Pandemic

As previously described, an interdisciplinary team developed, implemented, and supported an ITH program at an academic freestanding pediatric hospital as part of a quality improvement effort.⁴  Using a vendor video conferencing system, we created 414 ITH accounts for specific hospital beds to be used by clinicians and other hospital staff to communicate with patients. Throughout the project we made iterative improvements based on the results of brief surveys and rounds with patients, providers, and medical interpreters.⁵  One such improvement was the addition of 17 accounts for our inpatient psychiatry unit ∼9 months into the initiative because of a COVID-19 outbreak on the unit. This project was deemed quality improvement by our Department of Pediatrics Performance Excellence Group and exempt from institutional review.

We used our electronic health record data to identify all hospitalized patients from March 23, 2020 through March 22, 2021, including demographic and professional charge data. We matched patients from our query with ITH sessions during that year-long period based on video conference reporting, including hospital bed and the date and time of the session. Because we designed our ITH system for ease of use, we do not require a user to login and thus have limited data on participants. In an effort to better understand the use cases for ITH, we identified charged ITH sessions based on a telehealth modifier code added by physicians.

As part of a secondary analysis, we sought to identify team and family meetings created ad hoc by users, instead of using our established ITH system. Because of vendor data retention policies, we were only able to obtain these hospital-wide video conference account data for 30 weeks out of the year-long study period. These data include any video conferences hosted at our institution during the time period, including for nonclinical use. We excluded sessions that used our dedicated ITH accounts. We then filtered meeting topics that included the word “team” or “family.” We reviewed the remaining meetings and excluded those with meeting topics that were either clearly nonclinical (eg, weekly admin team check-in) or hosted by users who did not have a clinical role (eg, information technology staff).

We analyzed our data using descriptive statistics with median and interquartile range (IQR) for non-normally distributed data. We used Wilcoxon rank sum test to compare sample medians. We used χ² analysis with posthoc residuals and Bonferroni correction to compare categorical data.

There were 14 090 unique patients hospitalized during the project period, including 1648 (11.7%) that participated in an ITH session based on our patient matching. Patients who participated in an ITH session were older and had longer length of stay compared with those who did not (median age 12 vs 8 years, P < .001; median length of stay 9.03 vs 2.03 days, P < .001). Table 1 shows the demographics for hospitalized patients based on whether they participated in an ITH session.

Over the course of the year, patients participated in a total of 6931 dedicated ITH sessions during 1879 unique hospitalizations (Table 2, Supplemental Fig 1). The median number of ITH sessions for a given inpatient hospitalization was 2 (IQR 1–4). The median session duration was 25 minutes (IQR 12–49), with a median of 2 users per session (IQR 2–3). Most sessions occurred while the patient was on a pediatric medical service.

During the observed period there were 2006 charges submitted using the telehealth modifier code for 836 patients. One hundred fifty-eight unique providers charged for ITH services across 24 different clinical services. The majority of charges were for patients hospitalized on inpatient wards (n = 1831, 91.3%), including 847 sessions for psychiatry consultation or other mental health care. A small fraction were for patients on the inpatient psychiatry unit, added later in the study (n = 161, 8%). Overall, 50% of charges for ITH were from Psychiatry providers (n = 1002), with the next most from Genetics or Metabolism (n = 369, 18.4%) and Endocrine (n = 138, 6.9%). Billing codes used for ITH sessions were significantly different than those used for in person patient care (Χ² = 2567, P < .001) (Table 3). Notably, 4928 ITH sessions did not have a corresponding professional charge. Likewise, 513 charges did not match an ITH session.

Secondary analysis of video conferencing data from the available 30 weeks out of the year-long project period identified 579 787 total meetings, including 11 937 where the meeting topic contained the word “Team” or “Family.” After exclusion of nonclinical meeting topics and hosts there were 1132 meetings (37 meetings per week) with a median duration of 49 minutes (IQR 36–62). The median number of users was 5 (IQR 3.5–8).

Over 10% of hospitalized patients participated in an ITH session and over half of those used the system more than once. Participating patients were older and had considerably longer lengths of stay compared with those who did not participate in an ITH session. These findings may reflect the increase in number and length of stay for patients admitted to the hospital awaiting inpatient psychiatric care²⁹  amid the concurrent mental health crisis.^30–32  This is also supported by the relative increased use of ITH for female patients, in keeping with Ibeziako et al’s findings on psychiatric boarders.²⁹  Though smaller in magnitude, the increased use of ITH for minority patients and those with a preferred language other than English is reassuring that these populations were not selectively disadvantaged in the use of telehealth. Indeed, our prior evaluation of satisfaction found patients who prefer Spanish to have high satisfaction with ITH.⁵  Most ITH sessions were for patients on a pediatric medical service; however, the system was also used for patients on pediatric surgical, critical care, and psychiatry teams. Notably, nearly 80% of our ITH sessions did not have a corresponding charge, which suggests the encounter was either unbillable, the provider chose not to charge for the encounter, or the provider did not use the telehealth modifier. This suggests important opportunities for the use of ITH for nonphysician services (eg, social work, case management, child life therapy) or as an adjunct to in person care by the medical team.

Half of the charges for ITH sessions were from psychiatry providers and all the top billing specialties were from disciplines where much of a patient evaluation depends on the patient’s history and a careful visual exam. The majority of charges from psychiatry providers was for consultation and follow-up on patients admitted to medical, surgical, or critical care teams. In addition, although our inpatient psychiatry unit was added during the last quarter of the study and represented only 4% of available ITH accounts, they accounted for nearly 10% of charged visits. This underscores the potential for use of telehealth for mental health care, including in group and community settings, especially in the face of an infectious outbreak. Notably, we could not match 25% of charges from ITH sessions to one of our ITH accounts. This likely indicates use of ad hoc video conference sessions, such as those identified in our secondary analysis, as well as meetings using other video conferencing software or telephone calls. Although likely done in good faith, these sessions emphasize the need for robust training and straightforward workflows to facilitate use of approved ITH systems with appropriate security and business agreements in place.

During the 30 weeks of the year-long study period for which we had ad hoc meeting data, we identified over 1000 additional video conference sessions created by users for the purposes of team and family meetings. These meetings lasted longer than most ITH sessions and had more than twice as many participating users as would be expected when discussing complex diagnostic, treatment, and care coordination decisions. Possible reasons for creating ad hoc meetings instead of using our ITH system include primarily usability issues and inadequate education and training.

Our findings suggest important ongoing use cases for ITH. High utilization by psychiatrists suggests ongoing opportunities to support behavioral health care with a limited workforce. This population may explain the disproportionate use of ITH sessions for inpatient consults, psychotherapy, and biopsychosocial assessment. Legislative efforts to support mental health care, including the expanded access to telehealth,³³  are critical to managing this crisis. Although the department of health and human services has developed a telehealth toolkit to support providers,³⁴  state-by-state variation in Medicaid reimbursement means pediatric providers must work closely with their compliance officers to understand local policy. The use of ITH for multidisciplinary team and family meetings provides the opportunity to achieve greater engagement from key stakeholders. Although these meetings occurred before the pandemic, telehealth provides the opportunity to bring together multiple specialists as well as family members, primary care providers, and other caregivers (eg, school nurses) in complex medical decision-making from disparate locations. The importance of this use case has been previously reported³⁵  and is further supported by our secondary analysis, which showed nearly 40 additional ad hoc sessions per week for team and family meetings. Finally, the preponderance of unbilled ITH sessions suggests ongoing opportunities to support patient care remotely with medical interpreters, social workers, case managers, and others, which may help address growing labor shortages in health care.

Our evaluation of ITH usage has limitations, most notably the fact that most sessions were not charged for and thus even though we were able to identify the patient in many cases, we have limited data about who conducted the ITH session and why. Charges for ITH sessions may also be under-represented because of the need for physicians to add the telehealth modifier manually during charge capture. Moreover, because all employees can create ad hoc meetings, the data from our ITH system accounts likely underrepresent the full use of telehealth for inpatients at our hospital during the COVID-19 pandemic. Our attempt to identify team and family meetings hosted ad hoc was further limited by loss of data for this secondary analysis. However, additional data on ad hoc team and family meetings would only underscore the importance of supporting ITH systems for multidisciplinary care in the future. Though psychiatry consultations were well represented based on our billing data, inpatient psychiatry visits may be under-represented since we did not create ITH accounts for our psychiatry unit until late in the project. Still, given the nature of behavioral health care on this unit it is unlikely ITH would have been used extensively in the absence of the infectious outbreak use case.

Clinicians used telehealth to support care of hospitalized pediatric patients during the COVID-19 pandemic. Our findings suggest ongoing applications for ITH beyond the pandemic, especially for mental health and multidisciplinary care.

The authors would like to acknowledge Mark Hourigan, Manager for Network Services at Boston Children’s Hospital, for his work configuring the videoconferencing software.