Table 2

Bias mitigation strategies

Bias classification	Bias type	Mitigation strategy
Information bias	Over surveillance	Use equal surveillance that is, evaluate both groups equally
Information bias	Knowledge of the exposure	For the interviewer—blind the interviewer to the hypothesis so their recording will not be influenced
Selection Bias	Confounding	Limit participation in study to individuals who are similar with respect to confounders. During analysis apply stratification to classify participants according to confounder or use multivariate regression analysis
	Participation of subjects	Using a regression model including several covariates
	Selection of subjects	Using a computer-generated random selection list for selecting CKT groups and clinics
	Loss to follow-up	Recruitment of participants within primary healthcare catchment area Verification of contact details at recruitment Request alternative contact details and possible permission for lay health workers to perform follow-up home visits

CKT, Circle Kubatana Tose.