Fragmentation of Healthcare Services as a Possible Determinant of the Low Completion for the Tuberculosis Cascade of Prevention among Asylum Seekers: Results from a Prospective Study with Historical Comparison

Abstract Background: Effective screening for tuberculosis infection (TBI) among asylum seekers (AS) is crucial for tuberculosis (TB) elimination in low incidence countries. Methods: We assessed the proportion of completion of the screening for TBI among asylum seekers with a centralized delivery method compared to the decentralized model previously adopted in the study area (historical control). In the historical model (January 2017 to May 2018) screening of AS was performed at the arrival offering TBI testing (TST followed by IGRA among those positive), radiological investigation, treatment initiation and hospital referral, if needed, at three sites: migrant health clinic, pneumology clinic and infectious diseases department for active disease (decentralized model). In the study model (June 2018, centralized) all steps of screening were performed at a single site, at a minimum of 6 months after arrival. Multivariable Poisson regression analysis, with robust variance, was used to assess variables associated with the completion of screening for infection. Multivariable logistic regression was used to identify factors associated with the diagnosis of TB infection. Results: The intervention approach was offered to 144 AS with an overall 98.6% proportion of completion (98.7% for those with a positive TST). In the historical screening model, 1192 AS were candidates for screening, which was completed by 74.5% of those who started it (44.7% for those resulted TST positive). Major losses (55%) were detected in the TST/CXR-IGRA sequential step, followed by the execution of TST test (25%). The ratio of screening completion was significantly higher in the intervention period (aIRR 1.78, 95% CI 1.68–1.88) and for AS coming from high incidence TB countries (aIRR 1.14, 95% CI 1.04–1.25). Screening after 6 months from arrival and age were associated with TB infection (2.09, 95% CI 1.36–3.2 and 1.14, 95% CI 1.01–1.29). Conclusions: Screening for TBI can be improved by a centralized approach. Higher prevalence of TBI 6 months after arrival could reflect recent (either during travel or in Italy) acquisition of the infection.


Introduction
Migration has been constantly increasing worldwide in recent decades [1]. Particularly since the late 2000s, the WHO European Region has experienced an influx of refugees and migrants, with a rapid increase occurring in 2015 [2]. Migrant pressure remained especially high in the Mediterranean area, although declining in Italy during the last 3 years [3].
Migration is associated with an increased risk of tuberculosis (TB) infection/disease, with higher rates of detection in many Western European Countries [4] and should be faced with tailored interventions to achieve TB elimination in these areas [5]. The risk of TB in migrants depends on several factors, such as the incidence of the disease in the country of origin, the higher HIV/TB coinfection prevalence, the type of migration journey and the socio-economic disadvantaged status in the host country [6]. Post-arrival reactivation of previous infection is one of the main drivers of TB among refugees and migrants [7].
Migrants often live and travel in crowded conditions, in an environment that can favors the transmission of tubercle bacilli or the progression from infection to active disease. Forced migrants, as asylum seekers (AS) and refugees, represent a special category with higher rates of active TB, due to the precarious living conditions in refugee camps and the interruption of healthcare services in the country of origin because of socio-political instability or war [8].
Nevertheless, screening and treatment completion ratios for tuberculosis infection (TBI) remain low [4], and novel strategies are required to improve both effectiveness and cost effectiveness.
Two systematic reviews were performed to assess effectiveness and cost effectiveness of screening for both active TB and TBI among migrants in EU/EEA [9,10]. For active TB screening, the highest yield and cost effectiveness was reached in programs targeting migrants from high-TB-incidence countries, although the best threshold of TB incidence still needs to be determined. Symptom-based screening has low sensitivity and should be limited to emergency settings, where a chest X-ray (CXR) cannot be performed or in case of a high influx of migrants (as at first arrival countries in Southern Europe in recent years, or in transit camps) [10]. For TBI, the effectiveness of testing and treatment programs are limited by the use of poorly predictive tests, long-length treatments and a weak care cascade [9]. Increasing screening uptake and treatment completion are required to provide individual and public health benefits.
In European Union, the European Centre for Disease Control (ECDC) recommends screening all migrants coming from high-TB-incidence country for active TB with CXR, and for TBI, either using tuberculin skin test (TST) or interferon-gamma release assay (IGRA). However, certainty of evidence remains low and large heterogeneity of implementation still exists within countries [11].
Since 2016, the University of Brescia has been a partner of an international multicenter research project funded by the European Union's Health Programme 2014-2020: E-DETECT TB project, "early detection and integrated management of tuberculosis in Europe" [12]. The project aims to implement TB prevention, diagnosis and treatment in six different European countries (Italy, Romania, Bulgaria, Sweden, United Kingdom and The Netherlands).
Within the project, we previously performed a retrospective analysis on screening for both active TB and TB infection among AS in Brescia in 2015-2016 [13]. We detected high TB prevalence and incidence and a poor performance of the symptom-based screening. The cascade of care for prevention was affected by significant losses: screening for infection was completed by 49.0% of the eligible. Among those who completed the screening, 47.9% initiated the treatment, which was completed by 63.6% of them eventually [13]. The complex and fragmented organization of health services was identified as a possible determinant of the poor performance of the cascade.
Here, we report, in the same population and setting, the effect of a centralized "onestop-shop" service organization on the ratio of screening completion.

Setting and Study Population
We conducted a prospective cohort analysis of the cascade of secondary prevention for TB in AS in the province of Brescia, Northern Italy, an urban area with a population of about 1,200,000 inhabitants. At the beginning of 2021, the city registered 19.1% of foreign-born residents, one of the higher urban percentages in Italy after Prato (25.3%) and Milan (20.1%). The whole province registered 12.4% which is still higher than national average percentage (8.5%) [14].
The study population consisted of the AS arriving at the second level reception centers (those hosting AS after the first arrival on borders while the request for asylum visa is processed). In June 2018, a novel centralized model was implemented for all AS who were in the study area for at least 6 months and had not yet been screened for TBI. Historical comparison (from January 2017 to May 2018) was used to assess the effectiveness of the novel strategy. The ratio of screening completion was compared between the study and control periods.

Screening Procedures
According to local health policies, all new migrants were offered a voluntary and free-of-charge screening for active TB and TBI [15].
At the first medical evaluation, all AS were investigated for TB through a clinical examination and a standardized verbal questionnaire to look for any signs and symptoms suggestive of TB, including persistent fever (more than one week), cough lasting ≥ twoweeks, night sweats, weight loss and haemoptysis. Screening took place in a first level structure, the Migrant Health Clinic, or in the hosting center using a mobile unit, depending on the recipient's preference. Screen-positive individuals were referred to a second level structure, the ASST Spedali Civili Hospital, where a CXR was systematically performed. In case of radiological abnormalities, further investigations included at least microscopic examination, Xpert MTB-RIF and culture of the sputum. TB cases were then managed in the same second level structure until treatment completion.

Centralized Screening for TBI (June 2018)
The centralized procedure was implemented in line with updated regional policies [15]. According to the policy, to increase its effectiveness, screening for infection was delayed of six months after arrival.
All the AS present in the area in June 2018 who were still unscreened for TBI and negative for active TB, were tested and treated using a centralized procedure, at the Infectious Diseases Department of Spedali Civili Hospital, were a dedicated multidisciplinary team was created. A TST was performed and on the same day of the TST reading, positive patients underwent an IGRA test and CXR. The team scheduled an appointment two weeks later for the discussion of results. Positivity of both tests without clinical and radiological signs of the disease defined TBI and indicated evaluation for treatment.

Decentralized Screening for TBI (January 2017 to May 2018)-Historical Control
Screen-negative individuals for active TB were addressed for testing for infection. As a first step, a tuberculin skin test (TST) was performed following the Mantoux technique, TST was considered positive with an induration of 10 mm or greater. Positive TST patients were prescribed IGRA test and CXR, to be performed in any available laboratory or clinic and followed by a specialist consultation at the Pneumology Outpatient Clinic of the ASST Spedali Civili of Brescia, where preventive therapy was offered if appropriate.

Preventive Therapy
Three options were available for preventive therapy, the choice being a responsibility of the health care provider: rifampicin 600 mg plus isoniazid 300 mg per day for a duration of three months, rifampicin 600 mg daily for 4 months or isoniazid 300 mg daily for 6 months. Follow-up was performed monthly in the same structure through a medical visit and a blood test examination (blood cell count and transaminases).

Data Analysis
Descriptive analysis of the characteristics of the two cohorts at baseline was performed. Median values and interquartile ranges (IQR) were used to describe numerical variables, while counts and percentages were employed for qualitative variables. Chi-square (χ 2 ) test or Mann-Whitney non-parametric tests were used to compare groups for categorical or continuous variables, respectively.
We classified countries at high TB incidence according to WHO estimates for 2019 (threshold 150/100,000 population).
Asylum seekers were classified into high or low influx period according to the median number of arrivals per month during the study period, which was 35 migrants/month. Months with more than 35 arrivals were defined as "high influx" and the others were defined as "low".
TB prevalence was calculated as the ratio between the number of diagnosed TB cases occurring within 6 months from the arrival and the number of all asylum seekers/100,000. TB cumulative incidence was calculated as the ratio between the new TB cases diagnosed after 6 months from the arrival and the number of all asylum seekers at risk/100,000.
The 95% confidence interval (95% CI) for TB prevalence and TB cumulative incidence were computed assuming a binomial distribution. Multivariable Poisson regression analysis, with robust variance, was used to assess variables associated with screening completion for infection, calculating incidence rate ratios (IRR) and their 95% CIs. Gender, age at arrival, TB incidence in the country of origin, influx period and screening procedures (centralized screening vs. decentralized) were included in the multivariable model. Multivariable logistic regression was used to identify factors associated with the diagnosis of TB infection, reporting the odds ratio (OR) and relative 95% CIs. The multivariable model included gender, age at arrival, TB incidence in the country of origin and screening timing (over 6 months from arrival in Italy vs. within 6 months from arrival).
All statistical analyses were performed using STATA release 15 (StataCorp. 2017. StataCorp LLC, College Station, TX, USA). A p-value less than 0.05 indicated conventional statistical significance.

Ethical Aspects
The study protocol received ethical approval from the Ethics Committee of the province of Brescia (study number 2901, date of approval 21 November 2017). All data were collected and analyzed according to current Italian laws for management of sensitive data. Standard clinical practice at all health care services included thorough information and verbal consent to any of the offered practices. Communication was facilitated by cultural mediators, as appropriate.

Subjects Enrolled
Overall, from January 2017 to June 2018, 1356 AS were resettled in the province of Brescia.
Symptom-based screening for active TB tested positive in 57 subjects (4.2%). Nine TB cases were diagnosed and classified as prevalent (diagnosed within the first six months from the arrival). The susceptibility test revealed 1 multi-drug-resistant (MDR) TB. Estimated TB Pathogens 2022, 11, 613 5 of 11 prevalence was 664/100,000 (95% CI 304-1256). Four additional TB cases were diagnosed at least six months after arrival and considered as incident cases. None of them presented symptoms at arrival, one had a negative IGRA following TST positivity and developed symptoms 12 months later. Three had positive TST and diagnosis was achieved following the detection of abnormalities at CXR during the screening for TB infection (performed in centralized, delayed procedure). Estimated TB cumulative incidence was 297/100,000 persons (95% CI 81-759).
For TBI screening, we excluded AS with active TB and those with symptoms for whom a CXR was not available (lost to follow-up).

Screening Performance
During the intervention, 144 AS were screened and included in the study, compared to 1192 AS included in the control period. Table 1 summarizes the main demographic characteristics of the two cohorts. For the centralized procedure the median days of staying since the arrival in the study area to the execution of TST was 381 days (289-394), while it was 4 days (IQR 2-13) in the decentralized historical model. AS without a TST result or CXR/IGRA were lost to follow-up (2, 1.4% in the centralized model, 187, 20.9% in the historical control). Age, sex and TB incidence in country of origin did not differ between the two cohorts. People screened with centralized procedure were most likely to come from Sub-Saharan countries compared to those in decentralized procedure and arrived less frequently during the high influx period. In both cohorts, all AS with a diagnosis of either active TB or TBI underwent HIV test, none of them were positive.
Flow-charts reported in Figure 1 summarize infection screening results in centralized and decentralized procedures, respectively. TST reading was performed in 99.3% (143/144) in the centralized procedure and in 71.3% (850/1192) subjects in the decentralized one. CXR and IGRA tests were performed in 98.7% and 44.7%, respectively, of those who were TST positive. Screening completion ratio was 98.6% (142/144) in centralized screening and 55.9% (667/1192) in the decentralized procedure.
ods: months with a number of arrivals greater than the median 35 migrants/month. Age, sex and TB incidence in country of origin did not differ between the two cohorts. People screened with centralized procedure were most likely to come from Sub-Saharan countries compared to those in decentralized procedure and arrived less frequently during the high influx period. In both cohorts, all AS with a diagnosis of either active TB or TBI underwent HIV test, none of them were positive.

Discussion
The "one-stop-shop" service implemented in the centralized procedure achieved a better ratio of screening completion for TB infection in AS.
Intersectoral communication and service organization are essential to ensure adequate access to health services for AS [16]. Usually, non-governmental and voluntary organizations bridge gaps in the public health system by supporting examination booking, administrative procedures and cultural mediation. After the legal changes regarding the migrant reception system introduced at the end of 2018 [17], this support was reduced in Lombardy. Moreover, large numbers of AS may overwhelm the capacity of the staff in hosting centers. Booking and performing health consultations may be challenging (i.e., risk of overlapping the appointments, lack of vehicles for transportation), reducing the effectiveness of the screening. We speculate that the decentralized screening structure amplified bureaucracy. In our centralized intervention, we adopted simplified booking procedures (performed by the healthcare personnel of the dedicated multidisciplinary team) and a coordinated organization of consultations "per center". This way, even a single social operator of the hosting center may provide assistance to different persons for which an appointment was booked on the same day. These strategies helped reduce the barriers to screening completion which are related to bureaucracy and logistics. In support of the major role played by logistics in the Italian setting, similar high losses in linkage from initial screening sites to diagnostic clinics were reported in an experience in Milan in the same period (2016-2017). In this study, Villa et al. identified a higher ratio of losses (41%) in screening completion rather than in the follow-up and treatment after diagnosis [18]. In the decentralized model, we found the highest dropout (55%) in the TST-IGRA/CXR step, the one with highest logistic and bureaucracy burden (i.e., booking procedures for blood tests, CRX and consultations at different sites, transport of AS to the clinics), also for the hosting center with dedicated personnel.
Barriers to accessing healthcare are well highlighted in the literature, such as the absence of cultural mediation, the lack of guidance on rights, cultural communication and expectations [19]. In our centralized experience, all steps of the screening were performed by the same multidisciplinary team, possibly contributing to an improved trusting in the system and a better relationship with the patients, both affecting the overall acceptance and completion of the screening.
The six-month delay for the centralized screening could have selected a more "stable" population. However, this limited time of permanence is unlikely to have counteracted all barriers to healthcare access for AS. Indeed, the language barrier seems to not be reduced by a six-month stay, as suggested by a study performed in US, in which no changes were seen in the probability of speaking well by the fourth year since migration [20]. On the other hand, there are scarce data about the factors affecting the attitude to treatment initiation for TBI among migrants, but in one study performed in the Netherlands, the intended duration of stay (>5 years) was associated with higher treatment initiation [21].
The lack of qualitative data is a limitation of this study; however, the idea of Italy as a transit country is well known among AS coming through the Central Mediterranean route [22], and the six-month period seems too short to define the intention of permanent stay [23]. The small size of the cohort screened with the centralized procedure is another limitation of the study. A possible explanation for these results could be the limited number of arrivals in the second part of the study period (those in which the procedure has been mostly implemented), for which the European measures adopted in 2017 to reduce irregular immigration through the Central Mediterranean route would have possibly played a major role, as seen at the Italian borders [24,25].
The association of high TB incidence in the country of origin and odds of completing the screening is also interesting. In other studies, the attitude to initiate treatment has been associated with knowledge of relatives affected by TB [26], but there is scarcity of studies on the correlation between the ratio of screening completion and TB incidence in the country of origin. The higher incidence TB rate in the country of origin could contribute to the awareness of the disease and to have a better attitude to prevent it. However, more research is needed to identify gaps and strategies to improve access to preventive services among migrants, which remains low not only for TBI [27].
At the same time, the choice of administering mainly short regimens for TB infection in the centralized procedure largely influenced the higher treatment completion ratio observed, as previously documented and suggested in other settings [4,28].
The higher prevalence of infection found in migrants screened 6 months after their arrival is possibly due to recent infection, either acquired during the travel or in Italy. Similar considerations apply for the diagnosis of active TB performed in asymptomatic or