C36-19 A Novel Medicare Claims-based Approach to Identifying Tele-ICU Capability and Utilization Across U.S. Hospitals

Rationale Tele-intensive care unit (tele-ICU) programs are expanding nationally, however their impact on patient and operational outcomes remains unclear due to the absence of validated datasets identifying where and how tele-ICU is deployed. The American Hospital Association Annual Survey Database (AHAASD) is commonly used for tele-ICU research but relies on self-report, is incomplete, and lacks utilization and clinical information. We aimed to develop and validate a Medicare claims-based method to identify tele-ICU capable hospitals and patients receiving tele-ICU care to enable future outcomes research. Methods We conducted a retrospective cross-sectional study using the 2021 Centers for Medicare & Medicaid Services (CMS) 5% Limited Data Set, including inpatient, outpatient, and carrier claims linked with the CMS Provider of Services dataset and AHAASD. Tele-ICU hospitals were defined as those with ≥1 tele-ICU claim during an inpatient or emergency department encounter. Tele-ICU hospitals were identified in the AHAASD by response to question 98b. We compared hospital characteristics and evaluated concordance between claims- and AHAASD-based identification using diagnostic test metrics. The ten hospitals with the highest claim volumes were manually validated using public websites. Results Among 7,655 adult hospitals, 344 (4.5%) had tele-ICU claims, while 947 (24.1%) reported tele-ICU presence in AHAASD. Only 110 hospitals (32.0%) overlapped between data sources. The claims-based method demonstrated a sensitivity of 11.6%, specificity of 95.0%, and positive predictive value of 42.6% relative to AHAASD. Most hospitals (69.5%) submitted only 1-2 claims, while 75 (21.8%) hospitals submitted five or more claims and 34 (9.9%) hospitals accounted for half of all tele-ICU claims. Notably, 234 hospitals were exclusively identified through claims; these hospitals were more likely rural, non-profit, non-critical access, and non-academic, with greater ICU and ventilator capacity but fewer intensivists. Among the ten highest claim-volume hospitals, seven indicated tele-ICU presence on their websites and six indicated tele-ICU presence in the AHAASD. Among 1,008,106 ICU patients, 680 (0.1%) had tele-ICU claims and exhibited higher illness severity. Conclusions Tele-ICU claims are rare, possibly reflecting the non-revenue generating orientation of tele-ICU programs and raising questions about financial sustainability. While these findings continue to demonstrate challenges in identifying tele-ICU capability and patient-level utilization, claims identify a distinct subset of hospitals actively engaged in tele-ICU care not recognized in national survey data. A hybrid claims- and survey-based approach may more comprehensively measure national tele-ICU adoption and establish a foundation for evaluating its clinical and operational impact.