We conducted social network analysis to be able to detect more MDR-TB cases than through passive case finding. Enrolling 99 MDR-TB cases and their contacts did not reveal new MDR-TB cases. One child with (probable) MDR-TB meningitis was missed by our study. Links between MDR-TB cases were found in two instances but did not lead to the detection of new cases. Only one of seventeen high-risk places agreed to participate in the screening, resulting in one additional presumed MDR-TB case identified.
The likely reasons why we did not detect additional MDR-TB cases was limited participation of contacts in TB screening. Participation was reasonable (~ 80%) in the first screening, but dropped considerably to ~ 40% at the second screening. Observation from discussions with staff suggest that participation may have been poor due to the following elements:1) perceived stigma among patients, contacts and high risk places and reluctance to cooperate and to reveal correct contact information, 2) low awareness about TB and its transmission, especially among contacts with low levels of education or contacts belonging to vulnerable groups such as drug users, and 3) participation in the second screening may not have been perceived as in their interest if they were busy and not diagnosed with TB from the first screening.
Furthermore, the relatively short follow-up period in our study of 6 months may be another reason. Studies have shown that active TB usually develops within five years after initial infection [13,14,15], and predominantly (45%) especially in the first year . The median time from infection to symptoms in secondary cases is estimated to be 1.3 years . Household contacts of MDR-TB patients are considered to be at higher risk to get infected than household members of drug-susceptible TB cases [17, 18]. This is because, even though MDR-TB isolates are usually less transmissible , family members of MDR-TB cases tend to have been exposed for a longer duration due to delays in correct treatment initiation[17, 19]. Therefore, contact investigation is useful for early case detection and treatment to reduce transmission of MDR-TB [4, 18].
The pick-up rate for MDR-TB cases may also increase by improving the sensitivity of our diagnostic approach. Future diagnostic approaches should consider: (i) to ensure the quality of sputum and chest X-ray, (ii) expanding TB clinical assessment criteria to any cough and other tuberculosis-related symptoms like chest pain, weight loss, lack of appetite, weakness or fatigue, chills, fever and night sweats (iii) including MTB culture with higher sensitivity  as an add-on test following Xpert result, (iv) using multiple rather than a single specimen, to increase the diagnostic yield of Xpert MTB/ RIF . However, resources in low and middle-income countries (LMICs) are generally limited and therefore it may not be feasible to implement all these recommendations.
A limitation of our study using Xpert MTB/RIF is that only TB and rifampicin resistance is diagnosed as an indicator for MDR-TB . In Vietnam we generally also perform culture and additional sensitivity testing of drugs included in first and second line regimens to confirm MDR-TB and tailor treatment.
There is a need to develop a system to identify and manage contacts of MDR-TB cases better, including providing of adequate instructions, and possibly screening. We recommend to use a simpler questionnaire rather than a comprehensive social network approach. This is a more efficient and likely more cost-effective means for MDR-TB case detection in Vietnam and other low and middle-income countries. Information about household contacts and those who have the most frequent contact with patients such as close friends and colleagues should be collected. Depending on available resources, screening may start with a clinical assessment to determine if the person has TB-related symptoms, followed by chest X-ray and Gene Xpert MTB/RIF. This should be combined with health education, i.e. inform contacts with what symptoms they need to come for TB screening.
Health education about TB, MDR-TB and its transmission among the general population should be more focused, and results of this study may help in prioritizing risk groups. It is needed to enhance awareness among contacts of MDR-TB and their compliance with screening programmes. Particular attention should be paid to enhance screening of non-household contacts as some studies show the incidence of TB among these contacts to be higher compared with household contacts [5,6,7]. Furthermore, we found a lower screening participation of male contacts in our study, which is in line with findings from our national prevalence survey. Therefore, more efforts are needed to find male tuberculosis patients .
Currently, about 50% of the estimated MDR-TB cases in Vietnam have not been previously treated, reflecting significant transmission of MDR-TB among contacts [20,21,22,24,24]. However, the routine case finding strategy for detection of MDR-TB during our study period mainly focused on previously treated TB cases Additional file 1 , with only 14% of MDR-TB patients diagnosed being treatment-naive. It is important for Vietnam to pay more attention to management of MDR-TB among new cases including close monitoring of MDR-TB contacts. Given the low yield of MDR-TB case detection from our study, beyond improving contact investigation, other potential groups should be considered to address 50% of undetected MDR-TB burden in Vietnam. Furthermore, diagnostic screening strategies should be enhanced. Approaches can be applied depending on the resources available as follows: microbiological testing by Gene Xpert MTB/RIF for (i) all newly detected TB patients including smear positive and negative (ii) presumptive TB cases who had/have contact with MDR-TB patients. These contacts can be identified by healthcare workers through interviewing TB presumptive cases who come to their health facility for health check up, and (iii) all TB presumptive.
While MDR-TB can be cured, social barriers to MDR-TB treatment could be an important factor that needs to be taken into account when designing and implementing a contact tracing program . Home visits by contact investigators are an effective method for interviewing household contacts and encouraging them to be assessed for TB . By visiting index patients and their household contacts, the investigator is able to observe the housing conditions, perform an environmental assessment for infection control measures, and discuss and evaluate the risk of exposure, as well as provide counseling to household contacts on symptoms suggestive of TB and when and where to seek health care and social support .
Even though we did not find any new MDR-TB case directly through our social network analysis, this approach may still be worth consideration if the key limitations of our study are addressed. The screening process should be simplified, well organized, to increase the participation of contacts, extend the time of follow-up of contacts, and improve diagnostic screening strategy. Given that the low participation rate in our study may have limited case detection, it is recommended to expand health education on transmission of TB and MDR-TB among contacts, reduce stigma attached to TB, improve communication skills of health staff, and increase staff resources to trace contacts and get them involved in the screening.