Antimicrobial utilisation across intensive care units in India: a retrospective registry-based descriptive study

Abstract Background Antimicrobial resistance (AMR) contributes to substantial morbidity and mortality annually. Intensive care units (ICUs), by virtue of the patient population are environment highly susceptible to high prescribing rates and the development of AMR. Sequential surveillance data combining microbiological, prescribing, and clinical data is absent. Digital registry enabled data for surveillance may help address this evidence gap. Using retrospective data from the Indian Registry of IntenSive care (IRIS), we sought to describe antimicrobial utilisation across 9 ICUs and assess existing guideline availability. Methods We undertook a retrospective assessment of registry data collected over 3 months by the participating ICUs. The indicators of interest included days of therapy (DOT) for antimicrobial utilisation, antimicrobial prescribing rates (APR), and concordance to recommended therapy. We also assessed the availability and quality of the ICU-antimicrobial prescribing guidelines using the International Centre for Allied Health Evidence framework (iCAHE). Results Of the 2450 ICU admissions, 71.1% received at least one antimicrobial at any point during the ICU stay. Antimicrobial utilization over 3 months, reported as days of therapy (DOT) was 682.5 per 1000 patient days, and the antimicrobial prescribing rate (APR) was 1226.9 per 1000 patients. Use of broad-spectrum antimicrobials from day of admission and polypharmacy was common. Culture rates were likely underreported. Missing daily data hampered interpretation of prescribing appropriateness. Guidelines for antimicrobial prescribing in the ICU, were available at 6 of the 9 participating sites. The quality of these guidelines was appraised between 7% and 64% using the iCAHE- instrument. Conclusions This study demonstrates a high dependency on broad -spectrum antimicrobial management at ICUs in India. The practice of prescribing antimicrobials empirically at admission, is suggestive of decision-making behaviours influenced by risk/ benefit in the context of high community resistance patterns. The study findings are mirrored by other regional publications. Registry systems such as IRIS may provide a valuable resource in providing data for benchmarking utilisation, and in supporting stewardship interventions, providing missingness in data, can be addressed.