Description

The standard model for initiating Advanced Hybrid Closed-Loop (AHCL) systems in Spain is primarily hospital-centric, consuming significant healthcare resources and creating potential bottlenecks in patient access. In the Region of Murcia, high diabetes-related hospitalisation rates suggest an opportunity to optimise care delivery. The HOME-AHCL study evaluates an alternative care-delivery pathway that shifts the initiation of the Tandem Control-IQ system to the patient's home, supported by structured online education, 14-day telemonitoring with automated alerts, and shared follow-up between hospital endocrinology teams and a specialised diabetes technology care team (Novalab).

Study Pathway and Procedures:

Study activities initiated in April 2026 following ethics committee approval. The first participant signed written informed consent on 22 April 2026, and the first device was activated (Day 0) on 14 May 2026, marking the start of a 12-month staggered recruitment period planned to conclude in May 2027 (14 May 2027).

Eligible adults with type 1 diabetes undergo online technical training provided by a specialised nurse prior to system activation. The nurse subsequently conducts a home visit on Day 0 for device setup according to parameters prescribed by the hospital endocrinologist. Participants are monitored remotely every 14 days throughout the 12-month follow-up. Continuous glucose monitoring (CGM) data are reviewed fortnightly to classify automated alerts based on predefined criteria: High alerts (TIR ≤50% or TBR \<70 mg/dL ≥8%), Medium alerts (TIR \>50% to \<70% or TBR \<70 mg/dL \>4% to \<8%), and No alert (TIR ≥70% and TBR \<70 mg/dL ≤4%). Medium or High alerts trigger a protocolised telephone triage by specialised nurses. Technical issues are resolved directly, while persistent clinical issues are referred back to the hospital team. Scheduled data collection occurs at baseline, and at months 1, 3, 6, and 12.

Healthcare Professionals Sub-study:

Participating physicians and nurses will complete two study-specific, ad hoc questionnaires when approximately 50% of participants complete 12-month follow-up. Questionnaire 1 assesses model acceptance and feasibility; Questionnaire 2 evaluates healthcare professionals' perceptions regarding clinical time allocation and estimated resource impacts. These estimates are interpreted strictly as professional perceptions rather than objectively recorded comparative healthcare utilization data.

Data Management and Quality Assurance:

Data are prospectively collected using an electronic Case Report Form (e-CRF) hosted on the secure cloud platform ReseaArch® (Persei Vivarium, S.L.) under joint-controller GDPR compliance. Re-identification keys remain under the exclusive custody of site investigators. Missing data are expected to be minimal due to automated CGM cloud synchronization and monthly database audits.

Sample Size Assessment:

The planned sample comprises 80 participants. Assuming up to a 15% attrition rate, approximately 68 participants are expected to contribute complete 12-month primary outcome data. Given the descriptive nature of this single-arm study, statistical adequacy is based on estimation precision for the primary safety outcome: Time Below Range (TBR \<54 mg/dL at Month 12). Assuming a conservative standard deviation of 0.65 percentage points (based on real-world PLGS baseline variability), N=68 provides a 95% confidence interval precision (half-width) of approximately +/-0.15 percentage points around the mean TBR \<54 mg/dL.

Statistical Analysis Plan:

Baseline characteristics will be summarized using descriptive statistics. Longitudinal outcomes across follow-up time points (M1, M3, M6, M12) will be evaluated using mixed-effects models under the Missing at Random (MAR) assumption: Linear Mixed Models (LMM) with AR(1) or Compound Symmetry covariance structures (selected by AIC) for continuous CGM and PROM/PREM metrics, and Generalized Linear Mixed Models (GLMM) with a logit link function for dichotomous consensus targets. Centre, sex, age, duration of diabetes, and baseline BMI will be considered as prespecified potential fixed-effect covariates in adjusted analyses; their inclusion in individual models will depend on their clinical relevance to the outcome and on whether they can be meaningfully estimated given the available data. Because all participants transition systematically from FreeStyle Libre 2 at baseline to Dexcom G7 upon AHCL initiation, baseline-to-follow-up CGM comparisons are considered exploratory due to platform-time confounding; post-initiation longitudinal analyses utilize Dexcom G7 metrics exclusively. Accumulated QALYs (derived from EQ-5D-5L) will be reported descriptively within the cohort. Post-hoc pairwise comparisons will be adjusted using Bonferroni correction.

Cost Analysis:

Resource utilization (consultations, emergency visits, admissions, consumables) will be measured and valued using official regional tariffs. This allows to conduct a cost analysis from the public healthcare system perspective (Servicio Murciano de Salud), considering direct medical costs over a 1-year time horizon, and thus without discounting. In the absence of a concurrent comparator, no incremental cost analysis or economic evaluation (such as cost/effectiveness analysis) will be performed.