High flow nasal cannula (HFNC) is a valuable respiratory support for patients with acute hypoxemic respiratory failure (AHRF) and may serve as a less-invasive alternative to mechanical ventilation in certain contexts. [1,2] Early identification of patients treated with HFNC who are at risk for escalation in respiratory support can mitigate excess mortality and other adverse outcomes associated with delays in intubation. [[3], [4], [5]] Towards this end, the ROX (Respiratory rate-OXygenation) index, a calculated metric derived from the ratio of SpO2/FiO2 to RR, has shown to be a strong predictive marker for HFNC failure. [[3], [4], [5], [6], [7], [8], [9], [10]]

The ROX index was initially developed for patients with pneumonia in France and Spain, with further validation in the USA, China, and South Africa while treating patients with severe COVID-19. [5,10,11] Data from these studies, along with a subsequent meta-regression analysis, have revealed that a ROX index of 4.88, measured at 2, 6, or 12 h after initiation of HFNC, can reliably predict HFNC outcomes. [4,5] Another meta-analysis of COVID-19 patients in predominately high-resource settings with relatively small sample sizes (median 64.5 patients) proposed variable cut offs depending on the outcome of interest (4.20 and 7.06 for an 80% specificity and sensitivity for risk of intubation, respectively). [7]

The validity and generalizability of the ROX index to low- and middle-income countries (LMICs) has not been delineated. In healthcare landscapes characterized by technology and equipment constraints, the ROX index may uniquely support clinical decision making and facilitate judicious allocation of sparse resources. Substantial differences in patient comorbidities, monitoring, and critical care capacity may impact the performance of the ROX index, highlighting the need for adaptation to diverse global contexts where the etiology of AHRF may not be easy to ascertain.

The main aim of this analysis was to evaluate whether the ROX index might function as a clinical decision support tool in India to support both early referral among centers without the capacity for non-invasive ventilation (NIV) or invasive mechanical ventilation (IMV) and escalation to early intubation among more advanced facilities. We describe a pragmatic approach and re-calibration of the ROX index in a large, multicentre cohort of ICU patients requiring HFNC in India.