ObjectiveTo evaluate the effectiveness of E-learning in continuing medical education (CME) programs for Tuberculosis (TB) health workers in terms of knowledge improvement and behaviour change, and provide policy recommendations based on the evidence from programs implemented in China.MethodsTB health professionals survey and in-depth interviews were conducted at provincial, prefectural, county, and primary care levels in the three project provinces. Difference-in-difference (DID) and multivariate linear regression models was employed to capture the effectiveness of knowledge improvement. Behaviour change was analysed using both qualitative and quantitative data.ResultsFor clinicians, pre-post data showed no significant difference in knowledge increases between pilot and non-pilot areas (pilot areas: +17 points vs. non-pilot areas: +18 points, p = 0.066). However, clinicians who attended the synchronous E-learning had an average score 7.3 points higher (p = 0.026) and an average knowledge score increase of 0.3 points (p = 0.028) for each event attended. Asynchronous E-learning did not have a significant impact among clinical doctors. For public health physicians, E-learning intervention had no significant effect on their knowledge score (DID model: pilot: +4 points vs. non-pilot: +9 points, p= 0.389, multivariate linear regression model: p = 0.161). For primary care workers, pre-post data showed a significantly higher score increase in the pilot areas (+10 points) than the non-pilot areas (+4 points, p = 0.005). Primary care workers who attended the E-learning activities had an average score 10.9 points higher (p<0.001) and an average increase of 1.8 points per training module attended (p = 0.002). More than half of participants (56.0% - 57.3%) agreed that they could apply what they learned to their practice, especially among well-educated health personnel from provincial level institutions. Traditional face-to-face training has no significant impact on all types of medical staff.ConclusionsE-learning interventions for TB clinical and primary care health workers were associated with higher TB knowledge level, and some participants have already started applying what they learned in their practice. However, public health physicians didn’t get much benefits from E-learning activities, probably due to the mismatch of training supply of & demand for training and environmental factors.