Table 3.

Common biases in typhoid fever surveillance and potential solutions

	Under estimation biases in typhoid fever surveillance	Potential solutions
1	All the people in the target community do not visit index facility used for surveillance	a. Conduct active surveillance by making house to house visit which is resource intensive but more precise
		b. Conduct a census and health care utilization survey, and apply a correction factor for the underutilization of health facility
2	All people visiting surveillance site and meeting inclusion criteria are not included in sampling	Estimate what proportion of people with inclusion criteria were not recruited and apply a correction factor
3	Febrile syndrome does not capture all typhoid fever infected people because some may not have symptoms severe enough to be captured and others may be asymptomatic	Broaden the inclusion criteria, particularly for younger children. This will be resource intensive.
4	Blood samples could not be collected from all eligible cases	Document blood sample collection failure along with reasons such as consent issues and apply a correction factor
5	Could not be included in data analysis for various reasons such as incomplete data, blood sample contamination	Document dropouts and apply a correction factor it
6	Blood culture does not detect all typhoid fever cases	a. Document history of antimicrobial intake prior to blood sample collection and estimate its relation to culture positivity.
		b. Apply a correction factor for blood culture sensitivity based on best applicable evidence for that settings (e.g. based on empirical research findings, literature review)