Background In Canada, tuberculosis disproportionately affects the foreign-born population. The national tuberculosis medical surveillance programme aims to prevent these cases. Individuals referred for further in-country surveillance (referrals) have a history of active tuberculosis or have features of old, healed tuberculosis on chest radiograph; those not referred (non-referrals) do not undergo surveillance. We aimed to examine the risk of transmission arising from referrals versus non-referrals. Methods We did this population-based retrospective cohort study of foreign-born migrants (aged 15–64 years) to Alberta, Canada, between Jan 1, 2002, and Dec 31, 2013. We obtained information about year of arrival and country of citizenship from Immigration, Refugees and Citizenship Canada, and data for tuberculosis cases and their contacts from the Alberta Tuberculosis Registry. The outcome of interest was culture-positive pulmonary tuberculosis. We compared the incidence of pulmonary tuberculosis and the odds of transmission among referrals versus non-referrals. By use of conventional and molecular epidemiological techniques, we defined transmission as either a secondary case or a tuberculin skin-test (TST) conversion among close contacts. We used multivariate logistic regression to determine the independent association between referral for tuberculosis surveillance and transmission. Findings Between 2002 and 2013, there were 223 225 foreign-born migrants to Alberta, of whom 5500 (2%) were referrals and 217 657 (98%) were non-referrals. 3805 (69%) referrals and 115 226 (53%) non-referrals were from countries with a tuberculosis incidence of more than 150 per 100 000 populations, or sub-Saharan Africa. 234 foreign-born individuals were diagnosed with culture-positive pulmonary tuberculosis between Jan 1, 2004, and Dec 31, 2013. The incidence of culture-positive pulmonary disease was nine times higher in referrals (n=50) than all non-referrals (n=184; incidence rate ratio 9·1, 95% CI 6·7–12·5) and five times higher in referrals than non-referrals from high-risk countries (n=167; 5·0, 3·6–6·8). 71 total transmission events arose from the individuals with culture-positive pulmonary tuberculosis—three (4%) from referrals and 68 (96%) from non-referrals. No secondary cases were attributable to a referral source case, whereas 18 secondary cases were attributable to 11 different non-referral source cases. Three TST conversions were attributable to three different referral source cases compared with 50 conversions from 31 different non-referral source cases. That is, three (6%) referrals transmitted tuberculosis compared with 42 (22%) non-referrals (adjusted odds ratio of 0·19, 95% CI 0·054–0·66; p=0·009). Interpretation Despite a much higher incidence of pulmonary tuberculosis in referrals than non-referrals, referrals were 80% less likely to transmit tuberculosis. Rather than a focus on referrals, Canada could consider screening and treatment of latent tuberculosis in all migrants from high-risk countries—a group that accounted for 100% of secondary cases. Funding Canadian Institutes of Health Research. © 2017 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY-NC-ND license

[No Keywords available]

longitudinal study transmission sputum smear Population Surveillance human risk assessment middle aged statistics and numerical data interferon gamma release assay priority journal sensitivity analysis morbidity procedures drug surveillance program restriction fragment length polymorphism Young Adult Humans migrant lung tuberculosis Adolescent male Canada Tuberculosis, Pulmonary Emigrants and Immigrants patient referral tuberculosis female Referral and Consultation evaluation study population research Incidence Article disease transmission Retrospective Studies program evaluation tuberculin test adult major clinical study infection risk thorax radiography outcome assessment cohort analysis retrospective study clinical effectiveness health survey