Pandemic influenza-implications for critical care resources in Australia and New Zealand

Objectives: To quantify resource requirements (additional beds and ventilator capacity), for critical care services in the event of pandemic influenza. Materials and Methods: Cross-sectional survey about existing and potential critical care resources. Participants comprised 156 of the 176 Australasian (Australia and New Zealand) critical care units on the database of the Australian and New Zealand Intensive Care Society (ANZICS) Research Centre for Critical Care Resources. The Meltzer, Cox and Fukuda model was adapted to map a range of influenza attack rate estimates for hospitalisation and episodes likely to require intensive care and to predict critical care admission rates and bed day requirements. Estimations of ventilation rates were based on those for community-acquired pneumonia. Results: The estimated extra number of persons requiring hospitalisation ranged from 8,455 (10% attack rate) to 150,087 (45% attack rate). The estimated number of additional admissions to critical care units ranged from 423 (5% admission rate, 10% attack rate) to 37,522 (25% admission rate, 45% attack rate). The potential number of required intensive care bed days ranged from 846 bed days (2 day length of stay, 10% attack rate) to 375,220 bed days (10 day length of stay, 45% attack rate). The number of persons likely to require mechanical ventilation ranged from 106 (25% of projected critical care admissions, 10% attack rate) to 28,142 (75% of projected critical care admissions, 45% attack rate). An additional 1,195 emergency ventilator beds were identified in public sector and 248 in private sector hospitals. Cancellation of elective surgery could release a potential 76,402 intensive care bed days (per annum), but in the event of pandemic influenza, 31,150 bed days could be required over an 8- to 12-week period. Conclusion: Australasian critical care services would be overwhelmed in the event of pandemic influenza. More work is required in relation to modelling, contingency plans, and resource allocation. (C) 2003 Elsevier Inc. All rights reserved.