OBJECTIVE: To evaluate the impact of an active case-finding programme on tuberculosis (TB) transmission in homeless shelters in Paris, France. DESIGN: Between 1994 and 1997, an active case-finding programme was implemented in homeless shelters using a mobile radiological screening unit, and continued from 1997 to 2007. During these periods, the strains isolated from TB cases diagnosed in the homeless were genotyped by restriction fragment length polymorphism analysis using the insertion sequence IS6110 as a probe. RESULTS: Between 1994 and 2007, 313 new TB cases were diagnosed among the homeless population: 179 through the programme among shelter users, and 134 among homeless people not using shelters. Half of the strains were clustered in 35 distinct patterns (2-48 cases/cluster). The clustering of TB cases steadily decreased in shelters during the 13 years of the survey, from 14.3 to 2.7 related cases per year (P < 0.01) and from 75% to 30% of related cases among all TB cases (P < 0.01). In contrast, there was only a slight trend towards a decrease in homeless people not using shelters. CONCLUSION: Active case finding in homeless shelters resulted in a decrease in case clustering, mainly in shelter users. Genotyping contributed to confirming the positive impact of the intervention.
OBJECTIVE: To evaluate the impact of an active case-finding programme on tuberculosis (TB) transmission in homeless shelters in Paris, France. DESIGN: Between 1994 and 1997, an active case-finding programme was implemented in homeless shelters using a mobile radiological screening unit, and continued from 1997 to 2007. During these periods, the strains isolated from TB cases diagnosed in the homeless were genotyped by restriction fragment length polymorphism analysis using the insertion sequence IS6110 as a probe. RESULTS: Between 1994 and 2007, 313 new TB cases were diagnosed among the homeless population: 179 through the programme among shelter users, and 134 among homeless people not using shelters. Half of the strains were clustered in 35 distinct patterns (2-48 cases/cluster). The clustering of TB cases steadily decreased in shelters during the 13 years of the survey, from 14.3 to 2.7 related cases per year (P < 0.01) and from 75% to 30% of related cases among all TB cases (P < 0.01). In contrast, there was only a slight trend towards a decrease in homeless people not using shelters. CONCLUSION: Active case finding in homeless shelters resulted in a decrease in case clustering, mainly in shelter users. Genotyping contributed to confirming the positive impact of the intervention.