Table 4. Methods to measure progress in TB control: recommendations of a WHO task force (June 2006)1
Routine TB surveillance and monitoring
• Routine surveillance (all reported cases) and monitoring (treatment outcomes) should be considered the ultimate method of evaluating TB epidemiology and control.
• All national TB control programmes (NTPs) should strengthen and evaluate the performance of systems for reporting TB cases so that the data reflect, to a close approximation, the true incidence of TB and its time trend. The process of evaluation should be supported by appropriate operational research studies.
• The analysis of disaggregated surveillance data should be encouraged (e.g. clinic, district, province; by age, sex, etc.) so as to draw out the maximum information on the TB epidemic and the impact of control measures.
• Appropriate computer software should be developed and implemented to improve routine recording and reporting.
Surveys of disease prevalence
• Countries with high and intermediate TB burdens are encouraged to carry out one or a series of disease prevalence surveys if these are likely to be beneficial in assessing prevalence and trends, and/or optimizing planning for TB control. The decision to carry out a prevalence survey in any country should be guided by criteria (to be further defined), which will include:
- Poor information on burden and trends of TB disease
- Functional TB control programme that can utilize survey results to guide implementation of control activities
- High HIV burden
- Weak or poorly informative surveillance system
- Available experience and expertise (national and/or international)
- Willingness of the NTP to support national prevalence surveys
- Full participation of population to be surveyed
- Logistic feasibility and security for field staff
Surveys of infection prevalence
• Acknowledging the importance of measuring infection, but understanding the limitations of the tuberculin technique, tuberculin skin test surveys (TSTs) are recommended only in settings where they are likely to be informative about the prevalence and risk of infection and its trend. A TST is not guaranteed to give interpretable results in any setting, but is more likely to be useful for measuring trends, and where there is:
- data on infection prevalence from previous surveys
- a firm plan to repeat surveys
- a high risk of infection
- capacity to ensure strict adherence to standardized methodology
Evaluating TB mortality
• The accuracy of the current cohort monitoring system in correctly capturing deaths among TB patients should be reviewed and optimized.
• The study of TB mortality in the general population (i.e. outside treatment cohorts) should be undertaken in the context of studies of all causes of death.
- Vital registration. NTPs should ensure linkages and cross-referencing of data from cohort monitoring with data from available and developing death registration systems, thereby improving vital statistics.
- Verbal autopsy. Further evaluations are needed to establish the reliability and validity of verbal autopsies as a way of evaluating TB deaths in the general population, and their feasibility within general cause-of-death surveys.