Tuberculosis screening of migrants to low-burden nations: insights from evaluation of UK practice

RESULTS

DISCUSSION

This nationwide evaluation of the provision of new entrant screening by local primary care organisations in the UK has revealed that although screening for active TB is consistently undertaken, screening for LTBI is highly variable, deviates from national guidance and is inversely related to regional TB burden. Our work suggests that heterogeneity particularly exists in the selection of new entrant subgroups to screen for LTBI and the specific methods used.

Migration and infectious diseases, particularly immigrant TB, are gaining increasing importance as a Europe-wide health policy issue 1, suggesting that our findings have wider implications for most European nations. Our study methodology provides a basic template from which European nations can evaluate their own new entrant screening programmes to gain objective insights into how screening is undertaken at the front line, whether national guidance is being adhered to and whether screening relates to regional heterogeneity of TB burden within and across European Union member states.

We found that all PCOs screened new entrants issued with an abnormal CXR forms, presumably as they are suspected to have active TB and thus are perceived as the greatest threat to public health 9. However, abnormal CXR forms comprise only a fraction of all port forms issued. More often, port forms indicate that the CXR is either inconclusive or has not been undertaken, or is normal, but our findings show that fewer PCOs (75.7 and 54.2%, respectively) attempt to undertake further screening action for these port notifications. Only 35% of PCOs reported routinely screening new entrants identified through primary care registrations. Therefore, the port-of-entry system remains the main method by which new entrants are identified and referred for further assessment by local TB services.

Few cases of active TB were diagnosed through new entrant screening (median 0.0%). Although only 65 out of 177 PCOs provided these data, the yields were similar to previous UK Health Protection Agency (HPA) estimates (0.12%) and local experience from port screening 19–21. This reinforces the view that there is little pulmonary TB to identify amongst new entrants arriving in the UK, suggesting that the current emphasis on CXR for initial screening may be misplaced 22. Indeed, a recent HPA review recommended urgently reassessing the benefits of continuing with the CXR as the initial diagnostic test for new entrants 19.

Low yields of active TB and the fact that TB in the foreign-born results largely from reactivation of LTBI 6, 23 reinforce the potential of tackling LTBI in new entrants 10, 24, 25. However, we found only 60.4% of PCOs screen new entrants for LTBI, despite the yield being higher than for active TB. Thus, a substantial proportion of UK PCOs are not implementing NICE guidance on LTBI screening 10.

In addition, our study has revealed, for the first time, that high-burden PCOs are significantly less likely to attempt to screen new entrants with normal CXRs for LTBI. This important finding suggests that high TB burden areas in the UK, which account for most foreign-born TB, are the most ethnically diverse (with individuals who have migrated from TB-endemic regions, such as the Indian subcontinent and sub-Saharan Africa) 26, 27 and also have the highest prevalence of LTBI, are actually following up and screening the lowest proportion of new entrants. This potentially undermines national policy. Remedying this disparity between current practice and actual need requires a regionally targetted increase in resources.

Although it was beyond the scope of this study to identify reasons for the disparity in screening for LTBI, it is likely that PCOs determine which new entrants to screen based on pragmatic considerations, such as limited service capacity, a lack of dedicated funds for undertaking organised screening and a feeling that screening asymptomatic persons with normal CXRs provides little benefit in preventing TB 19, 28. It is possible that auditing immigrant screening practices may reduce the heterogeneity that our study has uncovered.

Amongst PCOs that screen for LTBI, most were relatively consistent, and in line with NICE guidance, in which <16-yr-old (from countries with TB incidence >40 cases per 100,000 p.a.) and >35-yr-old (screening not recommended) new entrant subgroups they screened. However, most immigrants are 16–35 yrs of age and in this group, PCOs vary considerably in whom they screen for LTBI 4. Whilst almost all PCOs screen individuals from countries with a TB incidence >500 per 100,000 p.a. and sub-Saharan Africa, in accordance with NICE, nearly half of PCOs still screened individuals from countries with a TB incidence >40 cases per 100,000 p.a. 10. This heterogeneity probably reflects uncertainty about current NICE guidance, which has resulted in targetting only immigrants from sub-Saharan African countries, even though >60% of foreign-born cases occur in immigrants from non-sub-Saharan African regions, including the Indian Subcontinent, South-East Asia and Latin America, where TB incidence is 40–500 cases per 100,000 3, 10, 29.

In addition, PCOs used variable screening methods for adult new entrants. Although 72.2% of PCOs used IGRAs to identify LTBI, over a quarter still used the TST alone. Amongst PCOs that do employ IGRAs, most still used them to confirm a positive TST as per NICE guidance, though a few areas have now moved to single-step IGRA testing. This shift may have been driven by recent evidence suggesting IGRAs are cost-effective and, if positive, can predict progression to active TB 30, 31.

Previous work in this area has been small scale, hampered by poor response rates, not focused on comparing screening practices by TB burden and, often, conducted prior to NICE guidance. Nonetheless, a prior smaller scale survey undertaken by the British Thoracic Society found that a low proportion of TB clinicians undertook new entrant screening 32. A previous survey also found that less than half of public health consultants would act on port forms that indicate the CXR is either inconclusive or has not been undertaken, or is normal, with areas receiving most notifications actually screening the fewest new entrants 33.

Our study has several limitations. The information was gathered through a questionnaire with the potential for recall/responder bias, especially for the small proportions that provided estimates of the yield for active and latent TB. In addition, this cross-sectional survey only provides a snapshot at one time-point. Although we focused on legal migrants, undocumented migrants are also a high-risk population who are likely to benefit from TB screening, although they are often difficult to identify.

Future work should consider emergent data highlighting the high prevalence of LTBI in new entrants and their elevated rate of progression to active TB 20, 30, 34, 35 to inform revised health economic models. The resultant cost-effectiveness analyses should, in turn, clarify the optimal threshold of TB incidence in immigrants’ respective countries of origin at which to screen for LTBI and whether screening programmes should move from port-of-entry screening to a primary care-based model, which may facilitate wider migrant health programmes, such as blood-borne virus screening 20, 24, 36. An urgent reappraisal of screening policy has recently been called for in the UK 37, where the recently initiated review of NICE guidance provides a timely opportunity to prioritise screening for LTBI in a wider spectrum of new entrants to include those from the Indian subcontinent, especially in high-burden regions.