Children with medical complexity (CMC) account for <1% of all children but approximately 40% of all pediatric deaths and inpatient care spending in the U.S.1 Optimizing their outcomes requires a comprehensive approach to augmenting care in all settings: clinic, hospital, and home. The clinic component of the comprehensive care (CC) program provides 24/7 access to an experienced team of primary care providers and subspecialists and reduced their serious illnesses and hospital and ICU days by 47-69% and health-system costs by >$10,000 per child-year.2,3 The hospital component (inpatient consultation service) is further improving outcomes. Having improved both inpatient and outpatient care, the investigators now propose to complete a 360 degree approach by developing and rigorously assessing an integrated telemedicine-home-visitation program (THVP) to augment care for CMC in their homes to reduce the need for clinic visits as well hospitalizations. Building on prior experience in using telemedicine for children at UTH and evidence of benefits in other populations, 4,5 the providers will use a convenient, inexpensive, HIPAA-compliant telemedicine platform to make observations in the home to augment care, help address acute problems remotely at any hour, better coordinate care with healthcare personnel, and thereby reduce clinic visits, ED visits, and hospitalizations. Home visits will be conducted by a nurse home visitor whenever considered likely to be beneficial for any of the CMC and at least once by the primary care providers (PCPs) immediately following enrollment of children with chronic respiratory failure requiring mechanical ventilation at home. To promote reimbursements and further grant funding, the investigators will test the integrated THVP in a randomized quality improvement (QI) pilot study to verify its effectiveness in reducing total days of care outside the home.

Background Children with medical complexity (CMC) have one or more chronic illnesses, require treatment from multiple specialists, and often depend on medical technology for respiratory and nutritional support.6 Their care is fragmented, ineffective, and inefficient,2,7 and while they represent only 0.4% of all children in the US, they account for approximately 40% of pediatric deaths and hospital charges.6

Optimizing their outcomes is likely to require a comprehensive approach to their care in all settings: clinic, hospital, and home. The investigators have developed effective programs for their clinic and hospital care and now propose to add an innovative program to improve care in the home and verify its incremental cost effectiveness.

The clinic component is provided in the High-Risk Children's Clinic (HRCC), an enhanced medial home for CMC staffed by a highly experienced team of nurse practitioners, pediatricians, and pediatric subspecialists. The multiple features to promote prompt effective care at all hours include 24/7 cell-phone access to the PCPs. As shown in the prior randomized trial (funded by a CMS Health Care Innovation Award), this program reduced ED visits, hospital days, and ICU days by 47-69%, and health system costs by $10,258 per child-year below that with usual care. This trial was published in JAMA,2 highlighted by CMS, and attracted national attention as the most rigorous evidence to date supporting medical homes for patient group.8,9 The remarkable cost effectiveness of this program was found in later cohort analyses to have been well maintained or enhanced when the program was expanded to include all eligible CMC, including prior control children and new enrollees.3

The hospital component is an inpatient consultation service in which the HRCC staff consult and follow their patients admitted to Children's Memorial Hermann Hospital (CMHH) to make treatment recommendations, coordinate care, and plan their discharge. The investigators randomized half of HRCC patients to receive this service in pilot testing. Results indicate a major reduction in inpatient days for patients given the service vs. those not given the service: 289 vs. 615 total hospital days/100 child-years with a 95% (posterior) probability of reduced hospital days in Bayesian analysis (rate ratio [RR], 0.62 [0.36, 1.09]). Having improved both inpatient and outpatient care, the investigators now propose to develop and rigorously assess a novel integrated telemedicine-home-visitation program (THVP) to also augment care for CMC in their home. The investigators will test the program in a clinical trial to verify its effectiveness and cost-effectiveness in reducing total days of care outside the home and to augment clinical practice and health policy.

Rationale and Evidence Base

As emphasized by Hoffman and Emanuel10:

"Clinicians need to abandon their long-established approach of caring for patients in the hospital or the office…Patients spend most of their time away from the health care system and the focus has to be one of managing their health literally where they live with much more wireless monitoring, electronic and phone visits, at-home care, and patient engagement."

Reducing the need for care outside the home is likely to be especially beneficial for CMC; >90% are Medicaid beneficiaries. Their parents have limited resources and many find it difficult or costly to miss work and travel to Texas Medical Center. Moreover, any time spent in a medical setting imposes a risk of acquiring serious, even life-threating infections for CMC.

A systematic review by AHRQ in 201611 reported that: 1) telemedicine promotes positive outcomes for chronically ill adults but has received little study for CMC, and 2) research to identify cost-effective models of telehealth deserves high priority.

Home visits have been widely recommended to improve care and reduce clinic visits and hospitalizations.12-15 However, systematic reviews of home visitation have emphasized the need for further study, particularly for socially high-risk children.13

To maximize the potential benefits, the investigators will provide both interventions in an integrated program. The investigators will randomize the CMC in the HRCC to receive either the usual complex care (UCC [including comprehensive outpatient care and inpatient consultation] or THVP (UCC plus this telemedicine-home-visitation program)

Study Hypotheses

Primary Hypothesis: THVP will reduce days of care outside the home (in the hospital, ED, or clinic [excluding well-child checks]).

Secondary Hypotheses:

1. THVP will reduce the rate of serious illness (death, PICU admission, or prolonged hospitalization >7 days);
2. THVP will reduce the total number of ER visits;
3. THVP will reduce the total number of admissions;
4. THVP will reduce the total number of PICU admission
5. THVP will reduce the total number of readmissions within 30 days of discharge;
6. THVP will reduce health system costs;
7. THVP will not reduce the total number of well-child checks;
8. THVP will increase maternal rating on pre-selected questions of The Consumer Assessment of Healthcare Providers and Systems (CAHPS) Child 12-Month Survey;

Methods

Enrollment and Randomization: The UTH institutional review board has reviewed this study and determined that it does not qualify as human subject research. As a result, the study was registered within the UTH QI Project database. Partly because of differences in the number of routine visits, all eligible CMC will be stratified by age (< 2 years or older), and by estimated baseline risk (risk level 1 [mechanical ventilation], risk level 2 [equal to or above the expected median risk but not ventilator-dependent], and risk level 3 [below the expected median risk]), as judged by the clinic's medical director [R. Mosquera] based on patient's diagnoses, prior clinical course, current medical status, and socioeconomic risk factors. Patients will be randomized to either THVP or UCC using a computer-generated algorithm in REDCap with variable block sizes. Randomization will occur at baseline for the existing HRCC patients, before hospital discharge for newly enrolled patients with chronic respiratory failure requiring mechanical ventilation at home, and during the first clinic visit for any eligible new non-ventilator dependent children. Patient enrollment will occur between July 2018 and June 2020.

Integrated Telemedicine-Home-Visitation Program

The program is based on: 1) published findings of telemedicine or home visits in chronically ill children with diabetes, hemophilia, or leukemia;5,13-15 2) the collaboration with Drs. Harting and Mary Austin in successfully using telemedicine for postoperative follow-up visits for approximately 40 children in the past 14 months; 3) consultation from IT personnel at both MHH (Brian Thyer, Lead Programmer Analyst), and UTH (Andrew Streckfuss, MBA, Manager of IT Projects and Research) who have setup Zoom platform at UTH. In refining the program during the project, the investigators expect to interact with centers advancing telemedicine use for children with asthma, diabetes, hemophilia, or leukemia.4,5

Planned Telemedicine Features: The investigators will use the Zoom platform for telemedicine. Because the clinic is already equipped with a smart television, web cameras, and an emergency IPhone, no additional equipment will be required. With the help of HRCC staff, the families randomized to THVP will download a free Zoom application to any Android or iPhone, which almost all the patients already have. Telemedicine will be utilized systematically following calls received by the PCPs during clinic hours from parents seeking medical advice or trying to schedule same-day appointments for their sick child and only when considered likely to be beneficial for phone calls to the HRCC cell phone on weeknights and weekends. It will also be used for scheduled follow-up appointments as judged needed by the PCPs.

Home Visitation Features: Home visits will be conducted by a nurse home visitor when considered likely to be beneficial for any of the CMC and at least once by the assigned PCP shortly after discharge home from the hospital following enrollment of a ventilator-dependent child. Visits will be scheduled at a time convenient for the family. When needed for clinical or safety concerns, the nurse visitor or the PCP will be joined by a respiratory therapist, dietician, social worker, or medical assistant. Home visits will not be conducted whenever the visitor(s) feel unsafe. In these circumstances, the investigators will rely instead on telemedicine.

Blinding: While the families and HRCC staff cannot be blinded, the healthcare economist and the statistician will remain blinded to treatment group when performing the analyses.

Economic Evaluation and Statistical Analyses: The incremental costs of THVP vs. UCC will be assessed from a health system perspective. Inpatient costs will be based on hospital charges multiplied by MHH department-specific cost-to-charge ratios. The personnel time for providing office visits at the HRCC will be assessed by time motion studies and will be multiplied by personnel unit costs and the observed number of HRCC visits occurring in each treatment group during the study. The personnel time cost devoted to providing inpatient consultation will be assessed by multiplying the mean time spent on hospital consultations by the number of hospital days at CMHH occurring in each treatment group and by personnel unit costs. For the THVP group, the investigators will add the time spent providing telemedicine consultation by tabulating data on overall Zoom usage and multiplying it by personnel unit costs. THVP costs will also be augmented by the monthly fees for the Zoom licenses, and the HRCC staff time spent assisting patients with Zoom App installation and use. The remaining clinic costs for compressive care will be estimated based on the HRCC's total expenditures (including personnel salary, benefits, and overhead costs) and will be allocated to patients based on each patient length of follow-up during the study.

All economic and statistical analyses will be conducted using multilevel generalized estimating equation (GEE) models. Number of days of care in a medical setting in the THVP and UCC groups will be compared using a negative binomial GEE model with log link. Differences in costs between treatment groups will be assessed using a GEE model with log-link and gamma distribution. All the models will be adjusted for age (< 2 years or older), baseline risk (risk levels 1,2, or 3), within-family correlation, length of follow-up, and any important differences in treatment with inpatient consultation.

The investigators will perform Bayesian analyses to assess the effectiveness of the THVP in reducing treatment days outside the home relative to UCC (primary outcome) using a neutral prior probability. Given the favorable prior evidence of benefit from THVP in other conditions,4,11 a skeptical prior is considered unnecessary. Based on the current number of patients and enrollment rates in the HRCC, the investigators expect to randomize ~400 patients during the 2-year study for a total of ~800 child-years of follow-up. THVP will be considered beneficial if: a) Bayesian analyses indicate it has a >70% probability of reducing treatment days outside the home; and b) there is no evidence of an increase in adverse secondary outcomes. If so, provision of THVP could be recommended simply to increase access to care for vulnerable, disadvantaged children.

Inclusion Criteria:

• Children attending the High-Risk Children's Clinic
• 1 or more chronic conditions
• High healthcare utilization in the year prior to enrollment (of ≥3 ED visits, ≥2 hospitalizations, or ≥1 pediatric ICU admissions)
• >50% estimated risk of hospitalization in the year after enrollment (as judged by Program's Director [Dr. R. Mosquera] based on patient's diagnosis, clinical course, and socioeconomic risk factor).

Exclusion Criteria:

• Unrepaired congenital heart disease
• Mitochondrial disorders
• Active cancer
• Do-Not-Resuscitate (DNR) order
• Patients receiving compassionate care
• No Internet access

While the families and HRCC staff cannot be blinded, the healthcare economist and the statistician will remain blinded to treatment group when performing the analyses.