Preoperative Hospitalist Telehealth Visits for Medically Complex Children During the COVID-19 Pandemic

Our hospital, a 200-bed, tertiary care children’s hospital, has a large pediatric orthopedic department (14 surgeons). One pediatric hospitalist is assigned to surgical comanagement responsibilities for 1 week at a time, typically focusing on postoperative inpatients, but also seeing 1 to 5 preoperative patients per week.

We began providing preoperative hospitalist evaluations for medically complex children in 2009 as described in a previous report.⁸  These patients included any child with an underlying diagnosis of cerebral palsy, myopathy, skeletal dysplasia, or syndrome such as neurofibromatosis scheduled to undergo major orthopedic surgery. Goals of preoperative visits include medical preparation for surgery and coordination of care. Preoperative visits typically occur several weeks before surgery to ensure that interventions such as medication changes or nutrition optimization can be implemented before surgery. Before March 2020, all preoperative visits were in-person; subsequently, patients were offered preoperative in-person or telehealth hospitalist visits per caregiver preference.

In our preoperative program, patients are first seen in Orthopedics clinic. Once the surgeon confirms the need for surgery in a medically complex child using described criteria,⁸  eligibility for telehealth is determined on the basis of hospitalist licensing and patient state of residence (Fig 1). Our hospitalists are licensed for telemedicine in several surrounding states, including Maryland, Pennsylvania, and New Jersey; they could not offer telehealth to patients residing elsewhere. The hospitalist also confirms that telehealth is medically appropriate via chart review. For those deemed appropriate, the coordinator contacts the family to offer telehealth as an alternative to an in-person visit. Potential reasons for requiring an in-person visit included needing an accurate weight or the need for a specific aspect of the physical exam.

Our telehealth platform, Nemours Care Connect, is provided by Amwell (Boston, MA). It includes an iOS app that integrates into our electronic medical record (EMR; Epic, Verona, WI) in several ways, including appointment scheduling and patient instructions.

After receiving institutional review board approval, we collected data from preoperative hospitalist visits from April 2020 through October 2020 by individual chart review (H.G., S.B., E.F., D.R.). We collected 3 types of data: 1) patient demographic and clinical data, 2) data related to the visit itself and 3) data related to the visit outcome, such as number of recommendations made.

Patient demographic data included age, sex, state of residence, and primary insurance provider (private, public, other, or none). Clinical data included the patient’s underlying medical condition (cerebral palsy, chromosomal problem, or other), type of planned surgery, number of standing daily medications, dependence on medical technology (yes or no), and gross motor function classification system (GMFCS) score (1–5, with 5 indicating the highest level of assistance required). When a patient had >1 GMFCS score listed, we used the score determined by the orthopedic surgeon closest to the surgery date. Data related to the visit itself included visit type (telehealth or in-person) and number of days from preoperative visit to surgery. Data related to visit outcome included number of hospitalist recommendations and progression to surgery status.

We recognized that patients seen preoperatively fell into 3 major clinical groups: patients with 1) cerebral palsy (CP) and CP-like conditions, 2) those with chromosomal abnormalities, and 3) others.

We used SPSS Statistics for Windows, version 27 (IBM, Armonk, N.Y.) to compare the telehealth and in-person groups. We used the Fischer-Freeman-Halt exact test to identify significant differences in categorical variables. Shapiro-Wilk and Kolmogorov-Smirnov tests of normality were performed on numerical data. All numerical data were determined to be nonparametric, and these variables were analyzed by using a Mann-Whitney U test. We used an α value of .05 to establish statistical significance.

We reviewed a total of 89 visits; we excluded 21 patients with skeletal dysplasia. We excluded patients with skeletal dysplasia as they are specialized group with medical (fewer comorbidities) and demographic differences (most were from outside our local area) from other patients. We analyzed the remaining 68 visits seen April through October 2020 of which 34 (50%) occurred via telehealth and 34 (50%) were in person. No patients were found inappropriate for telehealth except for those excluded based on physician licensing restrictions. Patients’ state of residency in the 2 groups is listed in Table 1.

Between the telehealth and in-person visit groups, we compared patients’ medical conditions, characterized as either cerebral palsy, chromosomal abnormality, or other, as outlined in Table 1. One patient was classified as having both cerebral palsy and a chromosomal abnormality. We found that there was no statistically significant association between patients’ underlying medical conditions (P = 1.00) and their use of telehealth.

To assess patient complexity, we analyzed technology dependence, number of standing patient prescriptions, and GMFCS score. There was no difference in clinical and demographic characteristics, including medical complexity, between in-person and telehealth patients. There was no statistically significant association between patients’ median number of standing (not as-needed) patient prescriptions (4.0 vs 3.0, P = .553) and their use of telehealth nor between GMFCS score in cerebral palsy patients and their use of telehealth (P = .088). We also found no statistically significant difference in whether a patient progressed to surgery between the in-person and telehealth groups (32 vs 32, P = 1.00, odds ratio 1.250, confidence interval 0.133–7.540). There was also no statistically significant difference in median number of hospitalist recommendations between the 2 groups (4.0 vs 3.0, P = .553). Almost all (94% of telehealth visits and 100% on in-person visits) included at least 1 recommendation made by the hospitalist.

Preoperative visits may improve patient satisfaction, decrease postoperative hospital length of stays, and reduce unnecessary testing and consultations in adult orthopedic patients.⁹  During the COVID-19 pandemic, we implemented telehealth to continue providing preoperative visits for medically complex children before major surgery. In our study, the 2 groups, those seen for preoperative evaluation via telehealth and those seen in-person, were clinically similar enough for comparison. We found no significant difference in number of recommendations made by the hospitalist during the preoperative visit based on visit type.

We found no statistically significant difference in the use of telehealth between technology-dependent children and nontechnology-dependent children. We believed technology-dependent patients may be more difficult to transport and avoiding infectious exposure might be more important for these patients; therefore, we were somewhat surprised by this finding, although it may reflect a lack of power in our study. Interestingly, we found a median GMFCS score of 5 for the in-person group compared to a median score of 4 for the telehealth group. We recognize that these are relatively crude measures of patient complexity, and previous studies have shown significant heterogeneity within these GMFCS groups.¹⁰  Although we try to schedule patients’ preoperative appointments so that as many as possible occur on the same day, sometimes this is logistically impossible. Some visits, such as radiology and the laboratory, must be done in person.

We found no significant difference between the mean number of hospitalist recommendations or progression to surgery between the 2 groups. These recommendations included medication changes, nutritional management, or diagnostic tests. We wanted to ensure that telehealth visits included the same thorough approach as in-person visits. We found that the number of visits with at least 1 recommendation made by the hospitalist in both groups (94% of telehealth visits, 100% of in-person visits) exceeded the value noted (72%) in a previous study at our institution of in-person visits for neuromuscular scoliosis patients.⁸  Although number of recommendations may be a crude approximation of visit thoroughness, this figure suggests that patients’ histories were considered in a detailed way, again noting our study’s limited power. Given that studies have found that patients, families, and providers have relatively high satisfaction rates with telehealth visits,¹¹  our data suggest that preoperative visits could be a viable alternative for medically complex children undergoing surgery.

Our study has some important limitations. It was a single-site, retrospective study. Our hospital’s patient population may not be generalizable elsewhere. Because the selection of telehealth versus in-patient visit was primarily at the discretion of the caregiver, the study was not randomized. However, this lack of randomization reflects real-world caregiver preference regarding the use of telehealth for their children, as well as limitations such as physician licensing. We also did not perform power analysis, so it is unknown whether the number of patients was sufficient to detect a difference between groups.

A telehealth program can address the needs of medically complex children undergoing major surgery. Further studies will hopefully include multiple sites and investigate associations between preoperative telehealth visits and surgical outcomes, as well as explore patient, caregiver, and provider satisfaction with telehealth visits.