"But for the hurricane": measuring natural disaster mortality over the long-term.

Dear Editor: Impact assessments of disasters caused by natural hazards usually measure mortality for only a few weeks following an event, and studies that count death totals for one year or more are rare. 1–3 Limited surveillance periods are problematic and may result in an underestimation of mortality since the delayed, indirect deaths are not included. 4 A lack of resources may explain the limited, longer-term surveillance. Additional barriers to conducting more thorough, extended assessments include the challenges of attributing a delayed death to a specific event and the misguided belief that a few weeks of follow-up are sufficient to identify relevant impacts. 5 There is no universally accepted, standard definition for a disaster-related death, and the definition of " indirect death " remains particularly vague and subject to inconsistent application. 4,6 Researchers, first responders, and public health planners must use common definitions in order to effectively describe and communicate the impact of an event. They also can use such definitions to inform post-event surveillance activities (e.g., content and duration). The aim of this comment is to capture Lew and Wetli's approach to their study of Hurricane Andrew-related mortality, to record those deaths that, " but for the hurricane, would not likely have occurred. " 7 Limited examples of research on delayed, disaster-related mortality are available in the literature. However, it seems that certain hazards (e.g., those linked to diseases with known latency periods) are more likely to receive ongoing attention than others. After decades of ongoing surveillance, experts continue to attribute deaths to radiation events, such as Chernobyl and the atomic bombings in Japan. 8 The Indian Council of Medical Research followed a cohort of people exposed to methyl isocyanate gas in Bhopal, India from 1985–1994 and counted delayed deaths. 9 Also, systems are in place to monitor the residents and recovery workers exposed to toxic dust during and after the terrorist attacks of 11 September 2001.The perception may be that disasters caused by natural hazards, such as floods or tornados, do not appear to be enduring threats. Thus, experts seldom count deaths displaced in time from the event of interest. The experience from Hurricane Katrina demonstrates how inappropriate surveillance and classification has clear implications for preparedness and response. Hurricane Katrina, which made landfall in southeastern Louisiana on 29 August 2005 and quickly resulted in flooding the New Orleans metropolitan area, is among the five deadliest hurricanes in US …