Randomized Trial of Supplemental Synchronous and Asynchronous Telehealth to Improve Glycemic Control for Pediatric Patients With Type 1 Diabetes

Design: Randomized, controlled trial of two different telehealth modalities as supplements to usual diabetes care over a six-month period for pediatric patients with T1D and suboptimal glycemic control.

Recruitment: Patients will be recruited either: 1) during regularly scheduled in-person visits to the University of California, Davis (UCD) Pediatric Diabetes Clinic, or 2) via phone for patients receiving telehealth rather than in-person care from a UCD Pediatric Diabetes Clinic provider during the enrollment period. Inclusion criteria will be 1) age 5-18 years, 2) diagnosis of type 1 diabetes with a duration of >12 months, 3) suboptimal glycemic control at time of enrollment, defined as a hemoglobin A1c (HbA1c) level > 8% (64 mmol/mol) or a 14-day glucose management index (GMI) of >8% from continuous glucose monitoring (CGM) data or 14-day average blood glucose of >200 mg/dl in the preceding month, 4) intention to receive diabetes care at the UCD Pediatric Diabetes Clinic during the following year, 5) access to the internet via a device with video and audio capability, and 6) ability to connect the patient's home blood glucose meter or CGM device - as well as insulin pump, if applicable - to an internet-capable device via Bluetooth or physical cable. Patients will be excluded from participation if their home addresses are not in California, due to current physician licensing restrictions that apply to inter-state home-to-clinic telemedicine services. Written informed consent and assent will be obtained from participants and their parents/guardians according to IRB regulations.

Test-Run, Randomization, and Baseline Data Collection: Prior to randomized allocation, each patient enrolled in the study will complete a test-run of remote data-sharing and a home-to-clinic video encounter with research staff to resolve any technical issues. These procedures will help to ensure that patients' and their families' use of the necessary programs and follow-through with scheduled tasks is successful. In my pilot study of home-based video visits, 23% of enrolled patients failed to complete their first video visits, effectively withdrawing from the study before receiving any of the intervention. This test-run is designed to minimize similar attrition after randomization during the randomized trial. Participants who successfully complete their test-run will be allocated 1:1:1 to usual care, supplemental synchronous telehealth visits, or supplemental asynchronous remote monitoring via a computer-generated, permuted block randomization scheme with variable block sizes. The allocation schedule will be produced by a study statistician. Allocation concealment will be achieved by using enumerated opaque envelopes (or a functional equivalent) to prevent participant assignments being revealed to research staff prematurely. After randomization, all participants will be asked to provide several types of baseline data, including a HbA1c level - measured via home collection kit - and completion of several validated survey instruments to assess patient-centered outcomes such as diabetes-related distress and self-efficacy.

Synchronous Telehealth Arm: Participants in the synchronous telehealth arm will receive home-based video visits with a pediatric endocrinologist every month for a duration of six months. Video visits will be conducted using a HIPAA-compliant video conferencing platform. A single pediatric endocrinologist will conduct all video visits, eliminating any inter-provider variability in clinical approach. One day prior to scheduled video visits the participants will receive reminders to upload data from their diabetes devices using secure internet platforms that are compatible with their home computers, tablets, or mobile phones. The research team will assist with initial setup of these platforms, and provide support for any technical issues during the study. Each video visit will include discussion of interval health events and patient or family concerns, review of shared glucose data and insulin dose information, and provision of recommendations by the physician. This content is equivalent to the physician portion of in-person diabetes clinic visits, except that a detailed physical exam cannot be performed.

Asynchronous Remote Monitoring Arm: Participants in the remote monitoring arm will receive monthly outreach from a pediatric endocrinologist based on the physician's asynchronous review of remotely shared data from their diabetes devices (blood glucose monitors, CGM devices, and/or insulin pumps) for a duration of six months. Recommendations from the endocrinologist will be delivered to the parent/guardian for participants <18 years of age and will also be sent to the pediatric participant if they have a phone or messaging account and the parent/guardian consents to this contact. The recommendations provided may include but will not be limited to advice about insulin dose changes, adjustments to insulin pump settings, and discussion of strategies for diabetes-related behaviors and family communication. In cases where glucose data suggests that glycemic targets are being met and no adjustments are needed, the endocrinologist will provide feedback to this effect. Participants' diabetes data will be shared with the research team via one or more secure platforms that are selected based on the participant's specific device(s) and method of uploading to the cloud (e.g., mobile versus computer with WiFi connection). In most cases, the data will continue to share passively once the necessary link is established, but in some cases monthly action may be needed by the participants. For this reason, the research team will not only facilitate the initial setup of data-sharing but will also monitor on a monthly basis for gaps in data and provide reminders, outreach, and assistance to participants as needed to ensure data sharing continues.

Usual Care (Control) Arm: Participants in this arm will receive the same initial assistance with setting up secure platforms for diabetes data-sharing but will not be scheduled to receive monthly video visits or asynchronous messages with recommendations based on remote monitoring. They will receive usual care at the UCD Pediatric Diabetes Clinic, which typically includes at least quarterly in-person and/or telehealth visits, as well as the option to share data with the diabetes clinic team at any point between visits and contact providers for feedback or assistance as needed. It is possible that the assistance with establishing data-sharing at time of enrollment will increase control participants' engagement with diabetes self-care and/or their likelihood to contact the clinic team for help. However, this is a strength of the study design because it will reduce any difference in outcomes between groups that are due to these factors and will help isolate the effects of the telehealth interventions themselves.

Post-Study Data Collection:

After completion of the six-month study period, all participants will be asked again to provide a HbA1c value via a mail-in home kit, and to complete the same survey instruments about patient-centered outcomes (diabetes-related distress and self-efficacy) that were assessed at baseline, as well as to report any incidents of severe hypoglycemia (defined as involving loss of consciousness, seizure, or need for emergency glucagon administration) during the study period for safety monitoring purposes. In addition, participants in the two intervention arms will be surveyed regarding the perceived burdens and benefits of the intervention they received, including estimated time investment and technical problems encountered - see manuscripts from the pilot study for details about these survey questions.

Data Collection and Study Variables:

1. Patient characteristics: Demographic information will be extracted from participants' medical records at the time of enrollment including age, sex, race, ethnicity, insurance, and home address (for calculation of geographic distance between home and clinic).
2. Healthcare utilization: Dates of encounters of the following types will be extracted from participants' medical records for the 6 months prior to and 6 months during study participation: diabetes clinic visits (in-person and telehealth), ED visits, and hospitalizations. Participants' use of diabetes technology - including insulin pumps and CGMs - will also be extracted from the medical record at enrollment and at study completion. In addition, the provider time spent to deliver the telehealth intervention each month will be recorded for participants in the intervention arms in order to facilitate analysis of provider time as a secondary outcome and/or an effect modifier for the primary outcome. Categories of discussion - e.g., insulin dose adjustments, diabetes skills/behaviors, diabetes technology use - will also be recorded each month for participants in the intervention arms, along with the provider's assessment of whether all, some, or none of the prior month's recommendations were implemented. This data will assist with secondary analysis of what and when various types of T1D management advice are most utilized by the participant cohort.
3. Patient experience: Responses to surveys administered to all participants will be collected at baseline and at study completion. In addition, responses to the final survey about perceived burdens and benefits of the interventions will be collected at study completion from participants in the two intervention arms. Survey responses for all study arms will be presented and evaluated qualitatively to gain insight into patient experience.
4. Glycemic control: HbA1c values will be collected via home kits from all participants at baseline and again at study completion. The primary outcome will be change in HbA1c from baseline to study completion. All HbA1c values measured as >14% will be treated as 14.5%. This "trimming" of higher outlying values will tend to underestimate any improvement in HbA1c during the study, so should not introduce type 1 error.