A cross-sectional population survey supplemented by laboratory tests was conducted in 2007 to assess the effectiveness of carbon monoxide detectors in a public housing complex in Morbihan, where 1,316 units had been equipped with them in 2003. Seventy-four households with detectors (19.5% [16.5–22.9]) reported an alarm activation. The cumulative incidence of days with an alarm activation, estimated at 0.29% [0.22–0.36], increased by a factor of more than 8 between 2004 and 2006. The activations were consistent with known causes of carbon monoxide exposure, and a highly dangerous situation was described. Clinical signs consistent with symptoms of poisoning were reported for 7.7% [3.1–18.0] of days with an alarm activation. Lack of knowledge about the detectors and appropriate actions affected tenants’ ability to respond to an alarm activation. The impact of the detectors on risk awareness was low. Acceptance of the devices, although generally good, was lower among tenants who had experienced an alarm activation. A decline in device performance was observed, likely due to drift or even possible sensor saturation. The study’s results, based on a specific case, complement the work of the National Metrology and Testing Laboratory, which identified insufficient performance in eight detector models. Action should be taken to identify reliable carbon monoxide detectors whose performance, lifespan, and maintenance procedures are well-documented. The use of the National Carbon Monoxide Poisoning Surveillance System to assess the effectiveness of CO detectors should be considered. (R.A.)