Real-Time Carbon Monoxide (CO) Poisoning Surveillance during an Ice Storm Event in Michigan – December 21-29, 2013

BACKGROUND:   The Michigan Syndromic Surveillance System (MSSS) is a voluntary reporting system used since 2003 as a component of routine communicable disease surveillance for the state.  It collects chief complaint and limited demographic information on Emergency Department (ED) visits that are securely transmitted to the Michigan Department of Community Health in real time. 90/139 (65%) hospital EDs currently participate in MSSS, covering over 80% of Michigan’s population.  MSSS was used to monitor syndromic presentation to local EDs for potential carbon monoxide (CO) exposure/poisoning after a December ice storm caused over 600,000 utility customers in mid and southern Michigan to lose power for an extended time period. During a power outage, risk of CO poisoning increases with use of alternative sources of heat or electricity such generators, grills, camp stoves, or other fuel burning devices in or near homes.

METHODS: MSSS was queried to identify ED visits with primary complaints of: “carbon monoxide” or “CO” including word derivatives and misspellings. A descriptive analysis was performed on ED data from the 9-day period in which sustained power outages occurred, including visits representing household clusters of CO exposure/poisoning, and compared to baseline data from the previous time period.  Visits and household clusters were also analyzed by county of residence and Michigan Public Health Preparedness region.

RESULTS:   During the December 21-29 power outage there were 81 ED visits in MSSS representing approximately 45 households where CO was included in the chief complaint. This was a 360% increase in visits and a 150% increase in households compared to baseline.   The majority of visits were in counties and regions where prolonged power outages occurred.  A statistically significant positive correlation was found between the rate of household power outages by county and the rate of CO exposure/poisoning ED visits by county.

CONCLUSIONS: Syndromic surveillance using ED chief complaints data were useful in providing situational awareness in rapidly detecting ‘outbreaks’ in CO exposure/poisoning after a severe weather event caused widespread power outages unlike data such as death certificates which are not timely. Frigid temperatures during this time period increased the probability that households experiencing extended power outages would use risky alternative energy sources for heating. Stratifying analysis of ED visits by region allows targeting public health messaging to areas at highest risk.  Syndromic data were limited because cases were not confirmed using medical records and did not include CO poisoning visits where the chief complaint did not include the query key words.