Implementing a Post-Discharge Telemedicine Service Pilot to Enhance the Hospital to Home Transition Free

Our institution is a 323-bed tertiary care freestanding urban children’s hospital in the Mid-Atlantic United States. This study protocol was deemed exempt by our hospital’s Institutional Review Board.

We used the Template for Intervention Description and Replication (TIDieR) checklist to describe our program (Table 1).¹⁷  TIDieR was developed with the objective of improving the completeness of reporting and replicability of interventions or programs.

Our program was developed to ensure postdischarge follow-up in a time of limited in-person pediatrician evaluation.

Patients were admitted to the PHM service. The discharging attending identified appropriate patients if they met any of the following criteria, which were determined to reflect risk factors for discharge care failures:

1. Diagnostic or medication complexity: defined as ≥2 new medications, history of difficulty with medication adherence, ≥2 specialty follow-up appointments, or diagnostic complexity and/or uncertainty as determined by the discharging team.

2. Perceived poor health literacy: uncertain comprehension of discharge plan or perceived poor family engagement as determined by the discharging team.

3. Lack of timely PCP follow-up or inability to identify a PCP.

PHM faculty and pediatric residents were provided institutional resources on conducting telemedicine visits and the referral criteria, as well as American College of Physicians Telemedicine Modules.¹⁸  For patients referred because of inability to identify a PCP, a list of potential options was provided. Discharging teams discussed and referred patients for a postdischarge telemedicine visit using the standard process for requesting all follow-up appointments. Call center staff scheduled the appointments and notified families via an e-mail with instructions for using the telemedicine platform and appointment link. This e-mail communication was consistent with institutional practice for scheduling telemedicine visits in the ambulatory setting.

Because of the decreased inpatient census, hospital medicine teams were consolidated, leaving some residents without a clinical assignment. Partnering with our residency program, these residents were assigned to the new role of leading telemedicine visits under the supervision of a PHM attending. PHM attending physicians on nonclinical assignments were used as visit preceptors.

Visits were scheduled and completed within 1 week of discharge over the Zoom for Healthcare platform. Appointments were scheduled for 30 minutes in the afternoons between 1 and 4 pm, 2 to 4 days per week depending on resident and attending availability. Providers were able to conduct visits with patients across state lines because of temporary license reciprocity enacted during the pandemic. A video interpretation service was available for conducting visits in the caregiver’s preferred language. Caregivers verbally consented to telemedicine treatment at the onset of the encounter, which included an explanation of potential financial responsibility pending insurance coverage of telemedicine.

Each visit was structured as follows:

1. Review visit purpose and reinforce that the visit was an adjunct to PCP follow up.

2. Elicit caregiver and/or patient questions.

3. Summarize key aspects of hospital stay to ensure caregiver and/or patient understanding.

4. Discuss postdischarge course and conduct a limited visual physical examination.

5. Assess postdischarge issues (PDI) related to clinical concerns, medications, follow-up visits, laboratory testing, imaging, and equipment and determine steps for resolution (Appendix 1).

After visit completion, telemedicine providers initiated actions to resolve PDI, then documented a visit summary in the EHR. An EHR-based notification was sent to the PCP, but telemedicine providers provided direct handoff if urgent communication was warranted.

Demographics and clinical factors for our study population were collected. For implementation evaluation, we applied taxonomy as described by Proctor et al to evaluate implementation and health service outcomes. Our outcomes included adoption, feasibility, and effectiveness (Table 2).¹⁹ 

Adoption was defined as percentage of referred patients who were scheduled for a postdischarge telemedicine visit.

Feasibility was defined as percentage of scheduled patients who completed a postdischarge telemedicine visit.

For program effectiveness, we reviewed postdischarge issues (PDI) and health system reutilization.²⁰  PDI were defined as concerns that required additional steps to enact the discharge plan or unexpected issues discussed during the visit. Health system reutilization included 7- and 30-day readmissions and rerouting. Rerouting was defined as urgent PCP, emergency department, or specialty visit because of clinical information obtained from the postdischarge encounter.

Retrospective chart review was conducted for all patients who were referred for a postdischarge telemedicine encounter between April 30 through December 17, 2020. For each patient, the EHR was reviewed; deidentified data were collected and managed using Research Electronic Data Capture (REDCap) electronic data capture tool.^21,22  This tool was developed and reviewed by study team members to ensure the ability to capture referrals, completed visits, demographic information, healthcare reutilization, and PDI. Demographic information included gender, age, race and ethnicity, payor, and primary diagnosis.

Implementation Outcomes: For adoption, referrals and scheduled appointments were tracked. For feasibility, scheduled appointments were reviewed to assess for completion. If a scheduled patient did not present for the visit, they were categorized as a no-show.

Health Service Outcomes: For effectiveness, a study team member reviewed all completed visit documentation to identify PDI. PDI were categorized as related to medications, follow-up appointments, postdischarge services, laboratory testing, imaging, or clinical change. Within these categories, details were recorded to understand the scope of problems addressed. If there was uncertainty regarding recognition or categorization of PDI, study team members discussed to obtain consensus. For each scheduled patient, the chart was reviewed to evaluate for 7 and 30-day readmissions. Institutional data for readmissions within the PHM division were obtained for the same timeframe.

Descriptive statistics were used for sociodemographic information and PDI. Fisher’s exact test was used to compare demographics among those that completed visits and no-show visits. χ² test was used to evaluate readmission rates between those that completed visits, no-show visits, and PHM patients not referred to the program who did not satisfy inclusion criteria.

A total of 107 patients were referred for a postdischarge telemedicine visit. Among patients who participated in a telemedicine visit (n = 62), sociodemographic characteristics demonstrated relatively equal distribution of sex and age groups, and majority insured by Medicaid. Those who disclosed their race most often identified as Black or African American. The most common diagnoses were COVID-19 and multisystem inflammatory syndrome in children (Table 3).

Among patients who were referred but did not participate in a postdischarge telemedicine visit (n = 45), there were no statistically significant differences in patient sociodemographic information when compared with those with completed appointments (Table 3).

All 107 referred patients agreed to the intervention (100%) and were scheduled (100%). Referral volumes were variable throughout the pilot (Fig 1A).

Of the 107 patients scheduled, 62 patients (58%) completed a visit. Although monthly show rate varied, the trend remained relatively stable throughout the pilot (Fig 1B).

Among patients who participated in a postdischarge telemedicine visit, 1 was readmitted within 7 days (2%); 3 were readmitted within 30 days of discharge (5%). In the no-show group, 2 patients were readmitted within 7 days (4%); 3 patients were readmitted within 30 days (7%). By using baseline readmission data for all PHM patients during the same time period, we determined that among PHM patients not referred for a postdischarge telemedicine visit, 6% were readmitted within 7 days, and 8% were readmitted within 30 days. Differences in readmission rates between these 3 distinct groups for 7- and 30-day were not statistically significant (P = .30 and P = .70, respectively).

Three patients (6%) were rerouted from their visit to another provider. Of these, 1 was sent to the emergency department, 1 to their PCP, and 1 to a subspecialist because of new or worsening clinical symptoms.

Among those who completed a telemedicine visit, 51 (82%) identified 1 or more PDI related to execution of the discharge plan. Of these patients, 71% reported a concern related to follow-up appointments and 25% reported a concern related to medications. Among issues related to follow-up appointments, 52% had not yet scheduled follow-up with their PCP, whereas 50% identified problems related to subspecialty appointments. Fifty-three percent of patients with any medication-related concern specifically had difficulty filling prescriptions, whereas others identified problems related to adverse effects, administration, or adherence. PDI related to postdischarge services, imaging, and laboratory testing were also identified (Fig 2).

This study details a pilot telemedicine discharge follow-up program for general pediatric patients, along with key implementation and health service outcomes. Our program started in April 2020 amid the early COVID-19 pandemic. Our pilot reached a variety of patients with diagnostic variability and identified many PDI. Additionally, our pilot demonstrated a method for improving the hospital-to-home transition, while providing a foundation for program infrastructure and a new opportunity for resident education.

As telemedicine evolves the ability to rigorously study its use and outcomes is paramount. Through Supporting Pediatric Research on Outcomes and Utilization of Telehealth (SPROUT) researchers can evaluate telehealth using the SPROUT Telehealth Evaluation and Measurement framework.²³  This framework considers work by health quality organizations in the creation of 4 telehealth measurement domains: health outcomes, health delivery, experience, and program implementation or key performance indicators. Assessment of these domains will frame future evaluation of our program beyond the pilot stage.

Our results demonstrate that scheduled visits showed a variable trend, although show rates generally remained consistent. One likely barrier to show rate was lack of appointment details at time of discharge, which prevented inclusion of appointment details in the discharge paperwork. Furthermore, appointment attendance required the caregiver to have an e-mail address to receive instructions and access to devices with internet and video capabilities. Additionally, families may have been deterred by the possibility of being billed for telemedicine services.

Our analysis showed no significant difference in any demographic category among patients who attended visits as compared with those who did not. This differs from a study of pediatric patients in the outpatient setting, which demonstrated that racial and ethnic minorities were less likely to have participated in telemedicine visits; however, given large percentage of unreported or unknown race in our analysis, it is challenging to say if this same pattern exists.²⁴  It is unclear whether this reflects differences in telemedicine referrals or barriers to completing visits. Additionally, research in an adult population showed non-English preferred language was independently associated with fewer completed telemedicine visits.²⁵  Though our analysis did not include language preference, appointment reminders and visit instructions were sent in both English and Spanish. Additionally, the platform permits integration of video interpreters into the visit. Given the concern that telemedicine can exacerbate inequities regarding technology access and literacy, further research is necessary to determine the impact of social determinants of health on accessibility and utilization of telemedicine services.^26,27 

Along with caregiver barriers to telemedicine visits, there are provider challenges. During our pilot, common provider challenges included extra time spent calling families when they did not join the visit on time and variable comfort level with coaching caregivers through technological issues. Prior studies have found that other barriers to telemedicine implementation include identification of provider workspace that preserves patient privacy, internet reliability, and cost including physician reimbursement.^28–30  As our pilot allowed providers to conduct visits from home, we did not encounter workspace issues. Waivers initiated during the pandemic expanded reimbursement for telemedicine service through Medicaid and commercial payers, but advocacy efforts are likely necessary to ensure these changes become permanent.¹  As stated in an AAP Policy Statement on using telehealth to improve access to and quality of pediatric care, overcoming these factors is essential and will likely require telehealth reform.²⁷ 

There were many successful aspects of our program. To our knowledge, this is one of the first studies to review health care reutilization following a telemedicine postdischarge follow-up in a general pediatric population. There was no statistically significant difference in readmissions between those who completed a visit, those who did not, and our baseline PHM population. Given this preliminary finding, further analysis of these distinct groups is warranted going forward. However, the literature regarding how telemedicine impacts health care utilization is difficult to generalize, as most studies focus on specific populations. In a study of patients on home parenteral nutrition in which 39% of eligible patients participated in a postdischarge telemedicine program, there were higher readmission rates in the telemedicine group when compared with a historical comparison group.⁷  In contrast, a study in which medically complex children attending a comprehensive care clinic were randomized to receive adjunctive telemedicine versus clinic care alone found that 99% of eligible children received telemedicine services and that they were less likely to be admitted or have ER visits.³¹ 

Most visits (82%) identified and addressed multiple PDI. Another pilot program also found that most postdischarge video visits added or modified discharge information.⁵  A systematic review of parental management of discharge instructions showed that medication dosing and adherence errors were common, and that over 60% of families missed appointments postdischarge.¹²  Additionally, results from the first multicenter collaborative focused on the pediatric hospital discharge process found that providing postdischarge support was a change strategy used by many participating sites.¹⁰  Our program provides a modality for postdischarge support that can serve as a model for other hospitals to address common PDI.

Another success of the program was the provision of telemedicine experience to our resident physicians. During the pandemic, telemedicine was seen as a way for institutions to protect the health of resident staff while allowing continued medical care.³²  By pairing residents and attending physicians, our trainees were able to provide care via a new modality with access to support. This program provided another feedback opportunity for resident performance based on direct observation, a benefit highlighted by leaders in pediatric resident education.³³  Visits provided resident staff with a unique glimpse into postdischarge care, a perspective that could lay the foundation for future quality improvement initiatives. However, the rapid deployment of telemedicine has drawn attention to the lack of curricula and the need for competency-based evaluation methods for our trainees.³⁴  These issues will need to be addressed as residency programs integrate telemedicine.

Directions for future study include program impact analysis through ongoing implementation and sustainability efforts. Using our pilot to provide proof of concept, our telemedicine program is now an established service line with dedicated resident and attending physician staffing secondary to additional internal funding. This service has been sustained, even during high census periods. These visits serve the same purpose, as any identified PDI can be readily navigated by the PHM team in real-time, which may not be as easily resolved by a PCP outside the hospital system. We have expanded our referral pool to include children with medical complexity. With institutional support, providers have obtained additional state licensure to provide telemedicine services across state lines.

Considering the domains of the SPROUT Telehealth Evaluation and Measurement framework will be key to further program evaluation and refinement. Regarding health outcomes, research is indicated to compare effectiveness of this program to other postdischarge interventions, such as automated or provider-led phone calls. Addressing experience can include evaluating patient, family, and provider satisfaction. Incorporation of PCPs into these visits might also optimize the patient and provider experience. Additionally, identification of key program indicators may include measures to promote telehealth equity by ensuring that families can use their devices to access care remotely, addressing language barriers, and screening for health literacy. Last, as the Association of American Medical Colleges has established telehealth competencies across a range of provider experience, there is opportunity for formal curricular development for both pediatric residents and attending physicians.³⁵ 

Limitations of this study include retrospective data analysis based on chart documentation. This pilot was conducted at a single institution with overall small sample size, which may impact generalizability of findings. Our assessment of adoption as an implementation outcome was focused on organizational adoption as opposed to individual provider adoption, as described by Proctor.¹⁹  Additionally, although general referral criteria were established, these criteria may be subjective. The referring team assessed health literacy on an individual basis without the use of a screening tool. No data on the criteria-based rationale for each referral was collected. Thus, in this pilot, it is not possible to know how many nonreferred patients may have met referral criteria. Increased understanding of these factors may help revise referral criteria and create a framework for improved assessment of provider adoption.

Our findings indicate that implementing a telemedicine postdischarge follow-up pilot was achievable at our institution and identified multiple actionable PDI. Though this pilot was initiated at the start of the COVID-19 pandemic with unique circumstances leading to fewer in-person follow-up visits, our goal is that the program now becomes a sustainable service that can be assessed and improved via the use of rigorous telehealth program evaluation tools.

We thank Drs Karen Smith, Paul Manicone, and Padmaja Pavuluri for their support toward the operationalization of our post-discharge telemedicine service line. We would like to additionally acknowledge Drs. Aisha Barber, Priti Bhansali, Jeremy Kern, and Neha Shah for their early vision to use telemedicine within our division and their contributions to the conception of this study.