Inpatient Consultation for High-Risk Chronically Ill Children Receiving Care in an Enhanced Medical Home

Background and Rationale: Comprehensive care (CC) provided in the High-Risk Children's Clinic (HRCC) at The University of Texas Health Science Center at Houston (UTHealth) involves care for acute and chronic conditions from a team of ethnically diverse pediatricians and nurse practitioners who are highly trained and experienced in treating medically complex and fragile children. This clinic serves as a novel medical home where both primary and specialty services are provided in the same place at the same visit. Acute problems presenting before 4 pm are seen the same day or if they occur over the weekend, on Monday morning. All parents have the cell phone number to directly reach 1 of the primary care clinicians at all hours. Multiple measures are used to promote and ensure the highest quality of care such as daily checks of emergency department and hospital logs to ensure prompt follow-up and coordination of care, weekly HRCC provider team meetings to review the care provided before every ED visit and hospitalization, high priority given to minimizing unnecessary ED visits and hospitalizations, and a relatively low patient-to-staff ratio of no greater than 100 (to allow for longer clinic visits, more telephone calls and e-mails, and extensive quality improvement measures).

Comprehensive care undoubtedly increases access to high-quality healthcare for high-risk disadvantaged children. The findings of our previous randomized trial of comprehensive care vs. usual care that were given accelerated publication in JAMA (2014) have attracted national attention as the best evidence to date supporting medical homes to reduce either adverse outcomes or costs (Mosquera et al., 2014). Comprehensive care resulted in major benefits (ED visits, hospital admissions \& days, pediatric ICU admissions \& days were all reduced by 47-69%) and savings of >$10,000/child/year from the health system perspective. Indeed, the improvements in outcomes and reduction in costs exceed those previously reported in prior studies of medical homes for patients of any age or condition (Mosquera et al., 2014; Homer et al., 2008; Jackson et al., 2013).

Despite these demonstrated benefits, high-risk patients seen in the High-Risk Children's Clinic still experience high rates of morbidity and often require hospitalization for treatment of acute exacerbations of their underlying chronic conditions. In 2015, 100 HRCC patients experienced a total of 175 admissions at Children's Memorial Hermann Hospital (CMHH) for a mean length of stay (LOS) of 7.1 days. As a result, a pilot quality improvement (QI) trial will be conducted to assess the impact of offering inpatient consultation (IC) at CMHH to further optimize coordination and integration of inpatient and outpatient care and reduce total hospital days per child-year.

Inclusion/Exclusion criteria: HRCC patients who require admission to CMHH will be included in the pilot trial. HRCC patients ≥ 18 years of age at study initiation will be excluded from the pilot trial and so will those with a Do-Not-Resuscitate (DNR) order, unrepaired congenital heart disease, active cancer, mitochondrial disorder, or seen in the HRCC clinic solely for compassionate care.

Enrollment period: All eligible HRCC patients who are admitted to CMHH between October 3, 2016 and October 2, 2019 will be enrolled in the pilot trial.

Design: All eligible HRCC patients will be randomized prior to study enrollment to either usual inpatient care or comprehensive care with inpatient consultation utilizing a computer-generated algorithm with variable block sizes. Patients will be stratified into two groups according to whether or not they are considered to have greater than median risk of repeated hospitalizations for patients in this study. Baseline hospitalization risk (high-risk or very high-risk) will be determined by the clinic's medical director based on the patient's diagnosis and clinical course.

Treatment Groups:

Usual Inpatient Care: Inpatient care at CMHH will be provided by the hospital inpatient team (residents, fellows, and attending faculty physicians) with usual occasional communication with HRCC providers. This care will not be modified by the study protocol.

Comprehensive care with Inpatient Consultation: Inpatient care at CMHH will include inpatient consultation by HRCC providers with input and recommendations conveyed to the hospital inpatient team on admission and at discharge at a minimum. The HRCC providers will review the inpatient care plan and make treatment and discharge recommendations with a focus on coordination and integration of inpatient and outpatient care. The inpatient consultations will consist of either face-to-face consultations, non-face-to-face consultations, or a phone call to the inpatient physician team.

Data Collection and Management: Research Electronic Data Capture (REDCap) software will be utilized to collect and manage predefined study variable data. REDCap is a secure, web-based application designed to support electronic data capture for research studies providing: 1) an intuitive interface for validated data entry; 2) audit trails for tracking data manipulation and export procedures; 3) automated export procedures for seamless data downloads to common statistical packages; and 4) procedures for importing data from external sources.

Cost Assessment: Total hospital costs for CMHH admission and any ER visit or re-admission within 30 days after CMHH discharge will be assessed from a health system perspective by multiplying total hospital charges by department-specific cost-to-charge ratios specified in the hospital's annual Medicare cost report. Hospital costs for the comprehensive care with inpatient consultation group will be augmented by the incremental HRCC personnel cost for providing inpatient consultation based on the HRCC staff salary and fringe data. All costs will be inflated to the year of analysis based on the Consumer Price Index for medical services. The investigators will consider the program to be highly cost-effective if it reduces the total number of hospital days per child-year without increasing costs, reduces costs without increasing the total number of hospital days per child-year, or reduces both.

Statistical and Economic Analyses: All intent-to-treat statistical and economic analyses will be performed using multilevel generalized estimating equations (GEE) models with exchangeable correlation to account for patient and family clustering, robust standard errors, and log link. The economic analyses will be performed according to current standards for such analyses (Drummond et al., 2005; Gold et al., 1996; Siegel et al., 1996; Sanders et al., 2016). Differences in hospital days and costs between treatment groups will be assessed using GEE models with gamma distribution. Rates of secondary outcomes will be assessed using binomial GEE models, whereas total number of secondary outcomes will be assessed with negative binomial GEE models. All the models will be adjusted for the trial's stratifying variable (baseline risk of repeated hospitalization) and for within-family correlation. In this small pilot study, some treatment effects that would be considered important by family members and clinicians (reduced hospital days) may not be statistically significant. As a result, Bayesian analyses will also be performed to estimate the probability of a benefit of a given magnitude and evaluate whether further study is justified. Neutral and skeptical priors will be used. The study will also allow us to obtain the experience and data needed to refine the process for providing inpatient consultations for high-risk chronically ill children. All frequentist statistical analyses will be performed using Stata version 13.1 (Stata Corp, College Station, TX). The Bayesian model will be implemented using OpenBUGS.

Stopping Rules. We plan to enroll patients in the pilot trial for 3 years. However, under predefined stopping rules, enrollment will cease if Bayesian analyses performed at the end of the first year identifies a 97.5% or greater probability that inpatient consultation results in at least a 10% reduction in the total number of hospital days per child-year relative to usual inpatient care or if at the end of the second year it identifies a 95% or greater probability of such a reduction.