The fire that broke out on September 26, 2019, at the Lubrizol and NL Logistique sites located in the Rouen port area caused significant environmental impacts (a plume of black smoke, along with soot deposits observed on the ground and roof debris, diffuse and odorous emissions of volatile organic compounds, flames, and explosions). These exposures, the nature and intensity of which varied depending on the distance from the fire and the plume, may have generated a sense of insecurity and psychological stress, and may have consequences for the health of the various exposed populations. As part of the study designed to assess the short-, medium-, and long-term health impacts, a self-reported health survey was conducted to estimate the nature and frequency of the population’s exposure to various nuisances resulting from the fire, as well as the symptoms and health issues experienced during and after the event; it also allowed for an assessment of physical and mental health status and quality of life several months after the event. Within an exposed area (Rouen and its surroundings) and a control area (Le Havre and its surroundings), a sample of 10,777 people was randomly selected by INSEE’s Survey Division from the 2019 edition of the Demographic File of Tax Origin on Housing and Individuals (Fidéli). Data collection for the main component of the survey took place from September 4 to November 12, 2020, using a multimodal collection protocol (online or by telephone). Ultimately, 4,779 people responded to this component, representing 95% of the target of 5,000 respondents, corresponding to a response rate of 47.6%. Nearly two-thirds of respondents answered via the Internet (64.3%). In addition, 1,306 questionnaires regarding a child were collected out of the 1,400 targeted. This main section was supplemented by a specific section aimed at assessing the impact on various mental health disorders. It was offered as a follow-up to those who agreed to be contacted again after completing the main component. Data collection for the mental health component took place from November 12 to December 21, 2020, and was conducted exclusively via self-administered questionnaires (online or paper). Of the 4,779 respondents to the main survey, 2,885 (60.4%) consented to be contacted again for the mental health survey, of whom 1,968 (68.2%) responded to the survey. The use of Fidéli proved to be of great importance: availability of geographic coordinates for housing units, which allowed for the selection of individuals based on the proximity of their residence to the fire; richness of available contact information; and availability of socioeconomic data. The representativeness of the sample of survey respondents, in light of the statistical criteria used, was satisfactory. The multimodal protocol chosen for this survey encouraged participation by offering selected individuals multiple opportunities to participate. However, it is also associated with the potential existence of a mode effect. Indeed, the specific analysis conducted in this study suggests the existence of a mode effect associated with a measurement bias (online respondents differ from telephone respondents because responses vary depending on the data collection method). For future surveys, identifying and addressing this bias will require an adapted protocol that provides a variable capable of explaining participation independently of interest in the survey.