**4. Discussion**

In our study, adherence with the screening of active TB and LTBI and with pulmonary TB therapy was very high compared to other recent investigations conducted in the USA, Switzerland, and Italy [13–17]. However, a direct comparison between these studies is difficult because study design, countries of origin and sample size were different.

In our study, we identified few fragile groups on which to focus to improve the performance of the screening program [18]. In particular, immigrants from America, Europe, Nigeria, Tunisia, and Georgia showed very low rates of compliance with screening, highlighting the need to target these subgroups with tailored procedures. Immigrants who were employed had lower overall compliance, probably due to the absence of health protection plans and the difficulties in acquiring health permits for these irregular workers. Conversely, knowledge of the Italian language was not significantly associated with compliance. The systematic presence of cultural mediators may have played a key role in determining this outcome, leading the immigrants to fully understand the significance of the screening program and the treatments they underwent, and to freely express questions and concerns [9,19,20].

Our study has a few strengths. First of all, the decision to include in the study the whole population of illegal immigrants should be mentioned. Had we not included eligible immigrants from countries with low TB incidence, we would have excluded about 48% of the subjects and missed the 80% of TB cases. Therefore the model adopted in our centre is innovative because it is based on the presence of a stable team of specialists and a nurse dedicated to the periodic monitoring of compliance and to the recovery of the subjects that don't show up for appointments. Finally, the fact that all the investigations were for free and that there was a direct delivery of the therapy are important elements in ensuring such a high adherence to screening and treatment.

Our study also has limitation. To begin with, the immigrants' self-reported data on education, employment status and symptoms are likely to imply an information bias that may have limited the accuracy of such data. Furthermore, we have not studied the correlation between adherence to screening and characteristics such drinking habits and drug abuse, smoking, or possible congenital or acquired immunodeficiency disorders. Moreover, the small size of the study population led to a lack of statistical precision in our results, which need to be confirmed in larger populations. In addition, in patients with negative TST, QFT was not systematically performed and as reported in recent studies [21], QFT might increase the identification of LTBI cases in recent immigrants. Finally, the findings of this study and the investigated setting may not be easily extended to other contexts, since the model adopted in the study centre was based on the fact that all procedures were for free, and were characterized by a stable health care team. This team also included a nurse dedicated to the periodic monitoring of compliance and to making contact with subjects who failed to show up for appointments.
