Surveillance of tuberculosis.

Surveillance of tuberculosis is one of the oldest disease surveillance systems in the world. This article briefly reviews its history, describes its methods and main results, with a specific focus on low- and middle-income countries, and underlines its main challenges and future prospects. Surveillance of tuberculosis started more than two centuries ago with the recording of tuberculosis mortality in England and Wales. After Koch's discovery of the tubercle bacillus, repeated tuberculin surveys were implemented to monitor infection, and case notification of active disease was progressively established during the 20th century. Because treatment of tuberculosis disease is the key intervention to stop transmission, monitoring of treatment outcome and more recently of drug resistance became integral parts of tuberculosis surveillance. At global level, the World Health Organization (WHO) is publishing a global TB report annually since 1997. Reports present data notified by each country as well as global estimates. Estimates of tuberculosis incidence are based on case notification adjusted by various correction factors, or on derivation from results of tuberculosis prevalence surveys, e.g. in India. In 2018, 10.1 million cases of tuberculosis are estimated to have occurred, among which about 0.5 million were resistant to rifampicin, and an estimated 1.5 million patients died. While global estimates are useful to raise public awareness and attract public and private funding, their uncertainty make them less useful to guide national policies. The backbone of tuberculosis surveillance at national and subnational level is the case notification. Newly diagnosed cases of active disease, whether new or recurrent, are reported with their key characteristics (age, sex, case category, HIV co-infection, drug resistance), and treatment outcomes are reported after scheduled treatment termination. All countries follow internationally standardized definitions. Incidence rates are compared by time to observe trends, by place to compare geographical areas, and by patient's characteristics to identify high-risk groups. The laboratory plays an essential role, since the surveillance of bacteriologically confirmed cases allows the most reliable comparisons, and because of the importance of the identification of drug resistance. The number of patients examined for tuberculosis diagnosis is also a key indicator to monitor case-finding activities. Tuberculosis surveillance is today among the most performant surveillance systems for infectious diseases. The two major changes currently observed are the move from paper-based registers to individual computerized surveillance databases and the multiplication of indicators for documenting progress towards tuberculosis elimination. There is a risk that implementation of these changes be followed by a loss in data quality. All efforts should be made to accompany these changes with adequate quality control. This will only be possible if health care workers are actively involved in the process of data production and analysis.