America’s power grid is aging. Questions about the vulnerability of the grid abound and the frequency of power blackouts is expected to increase.
Novel research from the Yale School of Forestry and Environmental Studies and Johns Hopkins University examines the relationship between power outages and mortality rates. The study, published in Toxicology, finds that, in August of 2003, during the largest U.S. blackout to date, both accidental and disease-related deaths increased significantly in New York, New York.
Using mortality data from 1987 to 2005, the researchers constructed a statistical model to calculate expected mortality rates had the blackout not occurred. They considered confounding effects, like temperature, day of the week, air pollution, and mortality trends over time.
When the researchers compared their expected mortality rates to the actual number of deaths that occurred during and immediately following the power outage, they attribute 90 deaths to the blackout. New York reported the total death count at six.
The majority of these 90 deaths were non-accidental or, in other words, disease-related. They accounted for 78 of the 90 deaths and increased by 25% when compared to non-blackout periods. Accidental deaths accounted for 12 of the 90 deaths and spiked 122% when compared to non-blackout periods.
Currently, death tolls from disasters only include accidents, like drowning, which can be directly connected to the tragedy. Yet, this approach greatly discounts the severe stress on health that blackouts induce.
People were trapped in subways and elevators. Residents in high-rises needed to climb flights of stairs. No air conditioning was to be found and cell phone service was disrupted. Stores and pharmacies closed. Ambulance response time increased and hospitals became overcrowded.
Until non-accidental deaths are attributed to power outages, the actual costs of blackouts are not being considered. As governments plan for natural disasters and seek to build more resilient infrastructures, these true costs should not be ignored.