Telehealth Coverage for an Inpatient Pediatric Gastroenterology Consult Service at a Regional Children’s Hospital: An accepted, useful and sustainable model of care

Abstract

Objective:

To evaluate a novel telehealth inpatient pediatric gastroenterology (GI) consult service at a regional children’s hospital in regard to acceptance, utility, quality, sustainability and provider resiliency.

Study design:

Patients requiring GI care at a regional children’s hospital between July 2020 and June 2021 were treated by an in-person or telehealth physician with physician assistant support, randomly assigned based on a weekly pre-set staffing schedule. Retrospective, multi-domain program evaluation was performed based on the RE-AIM (reach, effectiveness, adoption, implementation and maintenance) and STEM (SPROUT Telehealth Evaluation and Management) frameworks, utilizing statistical analysis to compare the patient cohorts and anonymous surveys to assess provider perceptions.

Results:

In total, 1051 patient-days of GI care were provided for 348 patients, 17% by telehealth and 83% in-person. There were no significant differences in diagnosis, transfer or readmission rates between the cohorts. No transfers occurred for reasons other than need to access specialized services not available at the regional hospital. Daily consult workload was slightly higher for telehealth physicians. Primary and consult team providers accepted the practice. The model continued beyond the first year.75% of local GI physicians reported greater Brief Resilience Scores in the context of shifting 20% of their inpatient call weeks to another campus’s physicians.

Conclusion:

Episodic pediatric GI consult service coverage via telehealth at a regional hospital was well-accepted, useful and sustainable, with improved physician resilience and no adverse outcomes seen. Telehealth holds promise for leveraging pediatric subspecialty physicians across hospitals, allowing complex patients to be admitted closer to home while reducing inpatient coverage requirements for smaller physician groups.

Introduction:

Successful technological innovation in health care enhances the delivery of care with comparable or better outcomes. Telehealth (defined here as live audio-video patient care) is one example that has grown exponentially, and concurrent with its growth, there has been increasing attention paid to where it can be best utilized in terms of economic value ¹, safety, effectiveness and efficiency across the spectrum of care. ² Successful subspecialty telehealth in pediatrics has included outpatient management of specific chronic conditions such as obesity, celiac disease, and eosinophilic esophagitis ³ and more broadly, for children with medical complexity. ⁴ On the inpatient side, telehealth has supported intensive care of neonates ^5,6 and older children ⁷, but there is less evidence for telehealth outside of neonatology and critical care. This contrasts with more extensive data supporting subspecialty tele-consults for adult in the emergency department and inpatient settings (e.g. acute stroke care ⁸).

Given the geographic distribution imbalance between patient and physician locations for many pediatric subspecialties ⁹, the potential benefit of virtual subspecialist access for pediatric inpatients may be even higher than for adult care. Similarly, virtual care has the potential to distribute inpatient call coverage requirements among more physicians in a geographically spread practice group, as they do not need to be physically close to the hospitals that they are covering.

In the Children’s Hospital Colorado system, the Denver metropolitan hospital supports a smaller hospital in Colorado Springs (COS), 60 miles south over a mountain pass that can limit winter transport. This regional pediatric hospital has 115 medical/ surgical, pediatric intensive care unit and Level III neonatal intensive care unit beds with 2850 admissions in 2021. Most patients are admitted to hospital medicine, surgery, neonatology and critical care services, with other specialties available on a consult basis by small provider groups who balance busy ambulatory practices with 24 hour/day call coverage. In the Children’s Hospital Colorado GI section, only 5 of 36 physicians and 3 of 10 advanced practice providers (APPs) are based in COS.

To mitigate burnout risk for this GI group, a plan was developed to provide 10 weeks/year (Monday to Friday) of inpatient consult service coverage using a virtual GI physician in Denver paired with an in-person COS GI APP. Hospitalists and intensivists participated in planning to increase buy-in.

Here we evaluate the first year of this novel telehealth program, with its objectives: acceptance by GI and primary service providers; no change in patient volume or diagnoses; no transfers or readmissions that occurred due to receipt of telehealth care; and improved COS GI physician resilience.

Methods:

This retrospective, multidomain program evaluation examined all pediatric patients undergoing a GI consult who received care at Children’s Hospital Colorado’s Colorado Springs regional hospital from July 1, 2020 to June 30, 2021. Patients received in-person or telehealth GI consult care based on a preset weekly call schedule, and as outlined in the hospital admission agreement, patients/ parents could opt out of telehealth care at any time (the primary alternates were transfer to the Denver metro campus to access in-person care or forego the GI consult). During telehealth weeks, there was a back-up call schedule for urgent endoscopy if needed.

The Denver-based telehealth group comprised 4 physicians, with a mean of 23 years in clinical practice since fellowship. These physicians volunteered for the service from a group of 7 physicians who were preselected by leadership based on telehealth experience and schedule flexibility. The COS-based in-person group consisted of 5 physicians with a mean of 10 years of clinical practice since fellowship. All physicians were board-certified in pediatric gastroenterology. The 2 COS-based pediatric GI APPs supported both in-person and telehealth weeks and had a mean of 4 years of inpatient pediatric GI experience and a mean of 13 years of total practice. (Note: one COS-based APP worked exclusively in outpatient care, so she was not involved in this program). All GI providers had ambulatory telehealth experience.

Telehealth was already well-established in the COS hospital for virtual psychiatric evaluation and occasionally used by other consult services, including wound care and child protection. Infrastructure consisted of a 10 gigabyte fiberoptic internet connection and Vidyo^™ video-conferencing software, accessed on the patient end via a workstation on wheels with a high-resolution, wide-angle camera (Logitech C930e 1080p Business Webcam) and on the physician end by a standard desktop or laptop computer with video capabilities. Both endpoints logged into a secure internal network via Ethernet or wireless routers. All Health Insurance Portability and Accountability Act requirements for virtual care were met. As a result of the COVID-19 pandemic, online telehealth training modules were already completed. GI telehealth consult physicians were provided with department-specific telehealth resources, including billing guidelines and electronic health record note templates to maximize ease and effectiveness.

Before going live, a robust process for this new tele-GI service was developed to standardize care. At the beginning of the workday, the COS-based APP pre-rounds on all patients then contacts the on-call Denver-based telehealth physician to discuss patients on the consult list. They subsequently develop a rounding plan based on patient acuity and physician availability. While the APP is prerounding, the telehealth physician typically has a half-day ambulatory telehealth clinic scheduled to minimize this service’s impact on the Denver outpatient workload.

During family-centered rounds, the APP is present at the bedside and the virtual physician is present via the workstation on wheels managed by the APP. After rounds, the APP and telehealth physician jointly communicate recommendations to the primary inpatient service team, with the consult physician participating via telehealth. In the event of a telehealth technology failure that cannot be resolved quickly, the GI consult team’s contingency plan is switching to an alternate video platform or completing the encounter with an audio-only telephone call. (Figure 1 describes the workflow details.) Virtual physicians were required to obtain hospital privileges at the regional hospital before the service’s launch, but the GI practice group spans both campuses from an administrative perspective so new contracting and billing processes were not required. New physician licenses were not necessary because both hospitals are in the same state.

Figure 1: Telehealth Consult Team Rounding Workflow.

Note:

APP pre-rounding typically consists of chart review, a check-in with the primary service provider managing the patient (e.g. a hospitalist), and a check-in with the patient/ family. Occasionally, the APP will also check-in with the nurse before rounds to clarify chart data or defer check-in with the patient/ family to rounds if time is limited.

After the first year, provider survey data and quantitative data comparing the patients who received care from a telehealth physician (i.e., the intervention group) and patients who received traditional in-person-only care (i.e., the control group) were collected to determine program acceptability, quality, and sustainability as well as COS GI physician resiliency. Feasibility was assumed due to prior ambulatory telehealth experience within the GI section and other inpatient telehealth services in the hospital and health system. The RE-AIM (reach, effectiveness, adoption, implementation, and maintenance) and SPROUT Network STEM (SPROUT Telehealth Evaluation and Management) frameworks were referenced when choosing measures to ensure a balanced and multifaceted evaluation approach. ^{10, 11}

Specific measures collected for the intervention and control groups were patient volume and demographic stratifiers, billing and diagnosis codes, inpatient transfer rates to the quaternary care campus in Denver, 30-day readmission rates, length of stay, mortality, technical failure, and provider satisfaction. Patient data were obtained from provider coding and electronic medical record (EMR) databases, with manual chart review done by 2 independent physicians (1 gastroenterologist, 1 hospitalist) for any encounter that included a transfer or readmission to identify the reason.

There were some longer-stay patients who received both telehealth and in-person GI care because their hospitalization crossed over 2 or more call schedule weeks; for encounter-level measures that could only be assessed for an entire hospitalization (e.g. length of stay, transfer, readmission), these patients were assigned to the intervention group if they received any GI telehealth care. In addition to patients with incomplete hospital data, patients admitted for less than 24 hours and those admitted to the NICU were excluded from length of stay, transfer and readmission evaluation because chart review indicated that factors other than the modality of GI care directed the majority of clinical decision-making for these patients. Planned admissions and unplanned admissions for complaints completely unrelated to the index hospitalization were excluded from readmission evaluation. Multiple readmissions for the same patient were counted as one readmission following chart review showing that recurrent admissions were related to disease and/or patient factors common to all the patient’s hospitalizations.

Diagnosis data were grouped when similar diagnosis codes were identified (i.e. – right upper quadrant, epigastric and generalized abdominal pain were grouped into a single abdominal pain category), and diagnosis rates were calculated by dividing the number of times a diagnosis was coded by the GI physician by the total patient-days for each modality of GI care (in-person vs. telehealth). Diagnosis rates rather than absolute numbers were used for comparison since the intervention and control groups were different sizes.

Quantitative analysis included Chi-Square, Fisher’s exact and Wilcoxon rank-sum test to determine significance of differences between the intervention and control groups. Statistical analyses were performed using RStudio v3.

To assess the program qualitatively, anonymous survey data were collected from GI and primary service (hospital medicine, neonatology and critical care) providers via REDCap^™ at the end of the year-long evaluation period to capture the experience of telehealth care compared to in-person care. Additionally, a survey containing the Brief Resilience Scale (BRS) ¹² and a question asking about the impact of their telehealth program on individual resilience was sent to the COS GI physicians. Lastly, COS GI provider data from the 2020 and 2021 medical staff satisfaction surveys were compared; although survey questions were not specific to this program, the implementation of episodic inpatient telehealth coverage was one of the most significant changes that the respondents experienced between these 2 surveys. The 2020 satisfaction survey sent the same month that the telehealth model was implemented, and the 2021 survey was sent near the end of the program’s evaluation year. Resiliency was not separately assessed among the Denver-based physician group because the impact on their individual workloads was negligible. Most of the COS telehealth weeks were covered by GI physicians who were not in the Denver inpatient call group, and no one was required to add clinics to compensate for this new service.

Process assessment and fine-tuning occurred via open discussion with the in-person APP and telehealth GI physicians, both iteratively during the first year of service and at the end of the evaluation period. These discussions identified benefits, technical failures, other challenges, and tips for successful implementation. Standardized patient/ family satisfaction data specific to the tele-GI service was unavailable since the hospital’s patient experience survey encompasses all aspects of a hospitalization, but requests for transfer to access in-person care and patient complaints related to telehealth were monitored.

The Children’s Hospital Colorado’s Organizational Research Risk and Quality Improvement Review Panel determined that human subject research protection requirements were not applicable for this program evaluation.

Results:

The pediatric GI consult service completed 1051 patient-days of care for 348 unique patients between July 1, 2020 and June 30, 2021 (Table 1), with 4 telehealth connections switched to audio-only telephone care due to technical failures that could not be resolved in a timely manner by the consult team. In-person teams provided 868 patient-days of care to 267 patients, and telehealth teams provided 183 patient-days of care to 81 patients. Of note, 54 patients were seen by both in-person and telehealth consult teams due to having hospitalizations that encompassed more than one call schedule week. Patients receiving in-person-only care had an average of 2.5 GI team visits per hospitalization, patients receiving telehealth-only care had an average of 1.5 GI visits/hospitalization, and patients receiving a combination of in-person and telehealth care had an average of 9 GI visits/hospitalization. Fifty-four patients (16%) accounted for 46% (488 patient-days) of the GI consult care captured. There was no noticeable downward trend in telehealth utilization throughout the year-long evaluation period, no requests by GI team providers to stop participating in the telehealth model, and no requests from primary service team members to suspend the service. In addition, telehealth coverage of the GI consult service has continued beyond the evaluation period without change. Staffing this new service resulted in the cancellation of 10 half-day ambulatory clinics in Denver during the first year, out of a total of 987 clinics.

Table:

Patient, Process and Quality Data

	In-Person	Telehealth
Patient Demographics
Mean age in years (range)	7.8 (0–19.5)	8.4 (0–19.9)	p=0.48
Female sex	45.3%	30.2%	p=0.02
Ethnicity/ race			p=0.12
Non-Hispanic white	50.3%	44.2%
Hispanic	34.2%	36.0%
Black/ African American	4.3%	1.2%
Other	6.8%	16.3%
Unknown or not reported	4.3%	2.3%
Payer type			p=0.25
Commercial/ Contract	22.9%	22.1%
Medicaid	50.9%	61.6%
Tricare	25.5%	16.3%
Self-Pay	0.6%	0%
GI Consult Service Process Measures
Patient–days of care provided (%)	868 (83%)	183 (17%)
Mean number of patients seen per day	3.7	4.1	p=0.004
Mean evaluation and management code level*	2.8	2.8	p=1
Patient service type			p=0.39
New consult care	30%	33%
Consult follow-up care	70%	67%
GI Consult Service Quality Measures %
Total inpatient transfers to another hospital #	10	2
Rate of inpatient transfer to another hospital	6.2%	2.4%	p=0.23
Total unplanned readmissions <30 days, excluding readmission for a completely unrelated complaint	10	5
Rate of unplanned readmissions <30 days, excluding readmission for a completely unrelated complaint	6.2%	5.9%	p=0.99
Deaths during hospitalization	0	0	p=1
Patient/ family complaints or requests for transfer related to delivery of care via telehealth	N/A	0

Notes:

^*Inpatient consult E/M codes (i.e. 99221–99223 for new consults and 99231–99233 for consult follow-up care) were assigned by the GI consult service physician for each day of care. Therefore, a mean E/M code level of 2.8 corresponds to an average coding level just below 99223/99233.

^%For quality measure data, the in-person column represents hospitalizations for which all GI care was provided in-person, and the telehealth column represents hospitalizations that included GI care via telehealth only or a combination of telehealth and in-person care (due to the hospitalization spanning more than one call schedule week).

^#All transfers occurred to access a specialized service unavailable at the regional hospital via in-person or telehealth care. No transfers occurred due to patient/family or primary service preference for in-person rather than telehealth GI consult care.

There were no significant differences in patient age, ethnicity/ race or payer type between the in-person and telehealth groups. There was, however, a statistically significant sex difference with more days of care provided to female patients by the in-person teams compared to the telehealth teams (45.3% female vs. 30.2% female, p=0.02).

Using coding data, the mean number of consult patients seen per day was greater for the telehealth group than for the in-person group (4.1 vs 3.7 patients/day, p=0.004). The proportion of new vs. follow-up patients and the mean evaluation/management code level were consistent between the groups, with a high level of coding (mean evaluation/management consult code level just below 99223/ 99233) seen for both groups. Both in-person and telehealth teams saw a variety of diagnoses consistent with a standard inpatient pediatric GI practice, and there were no significant differences in the 5 most common primary diagnosis codes between the two groups (Figure 2).

Figure 2: Primary Diagnoses.

Twelve patients required inpatient transfer to another hospital (all to the quaternary care children’s hospital in the Denver metro area), with 10 transferred patients seen by an in-person GI physician and 2 seen by a telehealth physician. There was no statistical difference in the transfer rate between the intervention and control groups, and all transfers were necessary to access studies or expertise not available in COS via telehealth or in-person care: motility subspecialist, pancreas subspecialist, transplant hepatologist, advanced/ interventional endoscopy, interventional radiology, nuclear medicine, multidisciplinary complex pain team care, interventional cardiology and cardiac anesthesiology (both patients transferred for cardiac reasons presented with concurrent GI and cardiac complaints). There were no patient/ parent requests for transfer to access in-person rather than telehealth GI care, no recorded complaints related to care delivery via telehealth, and no in-hospital deaths in either group. Chart review of the very few postdischarge deaths determined that all were due to progression of terminal illness.

Excluding readmissions that were unrelated to the index hospitalization’s clinical complaint(s), 15 patients experienced unplanned readmission to either the Colorado Springs hospital or the Denver metro hospital within the first 30 days after discharge. There was no statistical difference in the readmission rate for patients receiving in-person-only GI care compared to those receiving telehealth-only or telehealth and in-person GI care.

Mean length of stay did increase for patients who received telehealth care (10.1 days vs. 6.4 days, p=0.01), but this result was confounded by the way that patients were assigned to groups for statistical analysis. Patients receiving any GI telehealth care during their hospitalization were assigned to the intervention/ telehealth group even if they also received in-person GI care, which meant that the intervention group was biased toward patients with long hospitalizations spanning multiple call schedule weeks. It was not possible to determine if telehealth use prolonged length of stay in a causal sense, so length of stay was excluded from further evaluation.

Anonymous telehealth efficacy and effectiveness survey data (Figure 3) were collected from 6 telehealth GI team providers (4 physicians, 2 inpatient APPs) and 13 of 26 primary inpatient service (hospital medicine, neonatology and critical care) physicians. Provider experience and perceived acceptability and utility for this service were high. All GI providers and >80% of primary service providers reported that their overall experience with GI telehealth consults and patient/ consult team communication were equal to or better than for in-person GI consults; the same 2 primary service providers who reported an overall lower telehealth consult experience also reported less effective communication between the consult team and the patient. No GI providers and <10% of primary service providers reported that telephone consults could have been effectively substituted for video telehealth consults. Time spent on telehealth consultation compared to in-person consultation was the same or less for 60% of GI providers and 100% of primary service providers, and telehealth consultation availability prevented transfers according to 67% of GI providers and 27% of primary service providers. Zero primary service physicians reported limiting GI consultations due to concern about the telehealth model being insufficient to meet their patients’ need(s) or concern about patient/ family acceptance of telehealth.

Figure 3: Provider Survey Results Regarding Efficacy and Effectiveness of the Telehealth Consult Model.

Notes:

1. “Consult Team” respondents were telehealth GI physicians and in-person GI APPs working with the telehealth physicians. “Primary Team” respondents were physicians from the hospital medicine, neonatology and critical care services.
2. There is no bar shown for the Consult Team’s response to the question about phone-only consults being as effective or better than telehealth consults, because 100% of consult team respondents disagreed with this statement.

The provider survey also assessed barriers to telehealth consultation. Three GI telehealth team providers identified challenges related to the technology, 1 GI APP reported difficulty managing daily workload and timing of new consults, 1 GI physician reported more difficulty answering the consult question(s) compared to in-person care, and 1 GI physician reported that concurrent workload not related to the telehealth service per-se (e.g. phone calls from outside EDs) was a barrier to performing telehealth consults. Among primary service physicians, 8 identified no barriers, 1 reported technology barriers, 1 reported difficulty reaching the consult team, and 1 reported that consult service recommendations were limited by the telehealth model.

The annual faculty provider satisfaction survey demonstrated improvement in COS-based GI physician satisfaction subsequent to the implementation of this inpatient telehealth coverage model; the most significant increase was seen for the statement “I have the support I need to provide excellent care and services to patients and families.” In addition, 4 of the 5 COS-based GI physicians completed an anonymous survey based on the Brief Resilience Scale (BRS) ¹² at the end of the telehealth program’s evaluation year, with instructions to assess their “resilience as impacted by the addition of telehealth inpatient coverage”. The mean score was 3.5 (6 questions scored on a 5-point scale ranging from strongly disagree to strongly agree) with a standard deviation of 0.6 points, and the score range for each question was 3.0 to 4.0. Separate from the BRS-derived questions, 75% agreed and 25% strongly agreed with the statement “Since starting the telehealth program to reduce inpatient call coverage by 10 weeks, I feel that my resilience has improved.”

Implementation and technology lessons learned are summarized in Figure 4.

Figure 4: Lessons Learned.

Discussion:

This retrospective, multi-domain program evaluation demonstrates acceptability, utility and sustainability of an intermittent telehealth coverage model for a pediatric GI consult service at a regional hospital. Its successful implementation also positively impacted resiliency for a small GI physician group experiencing burnout without negatively affecting several markers of patient care quality.

Random assignment of patients to telehealth vs. in-person care based on a fixed, weekly call schedule and a total number of patient-days of care >1000 increases the strength of this evaluation. The intervention and control groups were also generally similar in terms of patient demographics, GI process measures and diagnoses managed. Interestingly, the telehealth teams saw a slightly higher volume of patients per day, which could be explained by their increased availability compared to the in-person COS GI physicians. Telehealth physicians held their normal morning telehealth or in-person clinics (6 patients) with 3 of the 4 having only 3 morning clinic per week while covering the COS consult service as opposed to COS physician having either morning procedure block or clinic by telehealth or in-person (6–8 patients) and then transitioning to the hospital. Regardless of the reason for this daily workload difference, these data clearly demonstrate that primary inpatient service physicians’ GI consultation patterns did not change during weeks covered by the telehealth model. This is supported by the experience survey data indicating that primary service providers did not reduce their GI consultation frequency due to concern about telehealth’s effectiveness or suitability for their patients.

Telehealth use did increase perceived time spent on consults for 2 of the 6 GI providers, but others reported equal or less time spent on telehealth consults compared to in-person care. Now that the learning curve period is over, this will be important to monitor because time spent per consult affects staffing availability and cost. From a technology perspective, there were reported challenges but almost no telehealth consults could not be completed, and no technical failures led to transfer or other adverse patient outcomes. Open discussion did indicate that GI telehealth team providers became adept over time at technical troubleshooting, so perception of technical barriers might change with ongoing telehealth use or if team members change in the future. Finally, one potential barrier that will need close attention in the future is payer reimbursement for inpatient telehealth care; this evaluation did not look at payer claims data because all care was provided during the COVID-19 public health emergency when restrictions on telehealth coverage were liberalized. If reimbursement for inpatient telehealth consults were to be reduced below parity for similar in-person consults or denied, this service model would become nonsustainable.

Difficulty accessing the consult provider and daily workload balancing were also reported infrequently as barriers, but these were likely not specific to the use of telehealth. Within the health system, there have been several instances where new telehealth services led people to re-examine challenges in care delivery as a whole by highlighting inefficiencies that had been previously mitigated through work-arounds specific to in-person care. These reported barriers represent opportunities to improve processes at the system-level, for example by increasing use of EMR-based secure messaging.

Most importantly, patient-level outcomes were reassuring with no negative impact on 30-day readmission or transfer rates and no deaths seen. The unplanned 30-day readmission rate seen in this evaluation was consistent with a national average of 6.5% ¹³, and chart review of all readmissions demonstrated that appropriate care was provided during index hospitalizations with the modality of GI consult care unlikely to have accounted for the readmission. All transfers in both the intervention and control groups were necessary and initiated to access care not available locally via telehealth or in-person care.

There were no complaints from patients/ families related to telehealth care or refusals of telehealth care and no deaths or other serious safety events seen in either group. The nature of this program evaluation precluded specific assessment of patient/ family satisfaction related to their experience with GI consult team care, and the study was not powered to detect a difference in serious safety events. However, there were multiple opportunities for patients and providers to express lower-level concerns throughout the evaluation period, in the context of a nationally recognized health system culture of safety that encourages reporting of concerns by any clinical team member.

Supporting the quantitative data, experience survey results and open discussion indicate acceptance and utility of this telehealth consult service by both GI and primary inpatient service providers. Live video connections led to a perception of the physician being present in the room, which likely improved acceptance and communication in a way that would not have been possible with phone-only care. Perceptions differed regarding telehealth’s impact on transfers with the majority of the GI team members feeling that their telehealth consultations reduced the need for transfer. Primary service team physicians, in contrast, did not feel as strongly that telehealth consults reduced transfers, potentially reflecting a low overall transfer rate and high-functioning nature of the primary service teams at this regional hospital. It is important to note that respondents were likely aware that a return to 24/7 in-person consult coverage by the COS-based GI physicians was unlikely for staffing and resiliency reasons, which may have impacted their acceptance of episodic GI telehealth care either positively or negatively.

Sustainability evidenced by the all-volunteer GI telehealth team continuing with the program throughout the evaluation period, and successful scheduling of 10 weeks of telehealth support in the subsequent year without need to recruit new providers. Although not specifically assessed, leadership support for this model of care also remains strong in the GI section and at the COS campus. There were minor adaptations of the model over time - specifically, a focus on talking with the primary service care team as early as possible in the day to share recommendations and answer questions – but no major changes in the workflow were needed. Most notably, the success of this model within GI subsequently led pulmonology and hematology/oncology to implement nearly identical models of care to support their small COS-based physician groups. Early reports indicate that similar success and sustainability is being achieved for these specialties.

This telehealth model also yielded travel cost savings. Initially, the GI section considered having a Denver-based physician drive to Colorado Springs to provide in-person care 10 weeks of the year (Monday-Friday). Based on a roundtrip of 120 miles and an Internal Revenue Service reimbursement rate of $0.56/mile in 2021, the use of telehealth saved $3360 in driving costs during the evaluation period. In addition, there were no delays in care or hotel costs due to unexpectedly prolonged drive times or road closures; this is important due to winter storms that can severely affect driving between Denver and Colorado Springs.

Lastly, the over-arching objective driving implementation of this novel telehealth model of care was achieved with COS-based GI physicians demonstrating improvement in resilience following the addition of episodic telehealth consult coverage by Denver-based colleagues. Four of the 5 COS physicians completed the Brief Resilience Scale (BRS) at the end of the evaluation period with a mean score consistent with normal resilience. Although the BRS was not completed before the tele-GI program, anecdotal reports of resilience were much lower. All of them also reported an improvement in their resilience specifically attributed to the telehealth service’s implementation. Additional evaluation moving forward will determine if this improvement in resilience is sustained. Impact on the resiliency among the Denver-based physician group was assumed to be negligible based upon several factors. Two of the volunteer faculty were not able to take in-person call so assignment to the telehealth team provided a way for them to contribute to the group’s call workload. This factored with limited application to 10 weeks/year and inpatient call at Denver being divided among 29 physicians resulted in a negligible impact on individual physicians’ inpatient call schedules in Denver. On the outpatient side, participating physicians held their normal ambulatory clinics while covering the inpatient service at Colorado Springs and were expected to support ad hoc appointments if their primary patients developed urgent issues. As a result, no new clinics were assigned to other Denver-based physicians and only 10 Denver-based ambulatory half-day clinics were cancelled over the year out of 987 total due to the addition of this service.

Limitations include the evaluation’s retrospective nature, implementation of the service in a health system with extensive telehealth experience, and reliance on EMR and coding database data extraction rather than 100% manual chart review. In addition, in-depth clinical quality of care and disease-specific outcomes for each patient were not evaluated. However, published quality of care standards are very limited for inpatient pediatric GI care so this would be nearly impossible to assess across all the diagnoses seen. The most significant limitations were lack of patient/ family satisfaction data specific to the GI consult modality of care and the significant number of patients who received a combination of in-person and telehealth care, which could have confounded the results. In addition, other factors could have impacted resiliency data including the COVID-19 pandemic (this limitation was partially mitigated by performing the evaluation fully within the public health emergency period).