Pattern Of Tuberculosis Transmission And Model Of Active Case Finding By Conventional And Molecular Epidemiological Method In Rural Areas Of China

China has the second highest burden of tuberculosis (TB) worldwide. The endemic of TB is higher in poor rural counties than in more developed urban areas, and around 80% of TB patients are from rural areas. The increasing incidence and unbalance is largely due to acceleration of migration, HIV/AIDS epidemic, spread of drug resistance TB and extending gap between the poor and the rich.The National TB Control Programme (NTP) in China has adopted the directly observed treatment, short course (DOTS) since 1992. Thanks to the expanding coverage of DOTS, case detection and cure rate of TB are significantly improved. Despite of the achievements, there are some substantial challenges in providing effective TB control for rural patients, especially in access to TB care, adherence to anti-TB regimen and prevention of relapse.A low proportion of rural TB patients have been detected and those who have got TB diagnosis experience a long delay both by patients and health providers. Patients' financial burden, slow diagnosis process, low referral rates, and social factors and migrant status are the main reasons that people experience delays between onset of first symptoms and an accurate TB diagnosis.Our study, therefore, is designed to study the socioeconomic, demographical, and molecular characteristics of patients and/or their close contacts, using a combination of traditional epidemiological method and molecular technique. We conducted a cross-sectional study on the behavior of contacts regarding TB, and analysis the behavior and pathways of health-seeking of TB patients diagnosed in county TB dispensaries; we used IS6110 and MIRU combined genetic makers to describe the genotyping of TB isolates from the patients of two counties; we conducted a cluster analysis to determine the proportion of the patients who are infected due to recent transmission or reactivated from remote infection; we re-investigate the patients in the clusters to trace the possible transmission routes and places; and determinates of the pathways of health-seeking and clustering are also analysed.Specifically,1. This cross-sectional study examined health-care seeking pathways for patients with tuberculosis (TB) and barriers related to the pathways in two counties under the National TB Control Program in rural China in 2004-2005. A total of 557 TB patients were recruited and interviewed by physicians at the time of TB diagnosis. Of 557 participants, 13.3% had a direct visit to specialised county TB dispensaries after onset of symptoms; 31.4% initiated their health-care seeking in village health stations, and 51.2% visited township or county hospitals first. The proportion of recommendation to an up level hospital or county TB dispensary was the highest in county level hospitals (76.9%) and the lowest in village health stations (51.4%). The most prompt pathway from first health-care seeking to TB diagnosis was to visit a county TB dispensary directly with only one day of provider's delay in median. There was an increase in provider's delay when more health facilities were involved.2. A house-to-house screening was implemented every two months within 7 months in Deqing County. Three villages with ten thousands general population each were sampled for the screening to detect suspects. A suspect was defined as who coughed for more than 3 weeks or hemoptysis. Then the suspect was referred to further diagnosis in county TB dispensary by Chest X ray. Of 32,582 community population screened, 21 TB patients were found. And 7 of those patients were smear-positive. The ratio of effectiveness vs cost decreased at the second screening, but increased slightly at the third screening. The direct costs for screening were 1096 RMB for per new case identified. Of total direct cost, 5.9% was taken on by TB patients, whereas 35.9% by county financing. The house-to-house screening can achieve high case detection rate than passive case finding method, especially suited in low case detection and richer areas in China.3. A contact screening was implemented in Deqing County. The contacts of the patients diagnosed in the county TB dispensary were the screening subjects. The contact investigation was used with a tool of purified protein derivative (PPD) test(diluted to 1:10000). PPD positive contacts were referred to the county TB dispensary for diagnosis. The results showed 2.5 contacts were identified from every patient. PPD positive rate of close contacts of smear-positive patients was 44.2%, and it was 41.7% among the contacts of smear-negative patients. Finally, 10 close contacts developed TB disease, with an incidence rate of 1.8%. Direct cost for contactscreening was 1432 for per new case identified.4. A cross-sectional investigation and qualitative study was designed to describe the factors associated with TB infection among close contacts of TB patients. Participants of quantitative study were the patients diagnosed during 8 months in Deqing and their close contacts. 24 TB patients and 6 close contacts from two counties were purposively selected for in-depth individual interview; 22 health-care providers were selected into 4 groups for focus group discussion. Quantitative analysis showed that relationship between index case and contact and age of the contacts were significantly associated with purified protein derivative (PPD) positive. The result from qualitative study showed contacts between couples and parents-kids/in-law were important to TB transmission.5. IS6110 and MIRU combined genetic makers were used to describe the genotyping of TB isolates from the patients of two counties in one year. The clustering characteristic was analysed. Of all 210 isolates, 179 (86.9%) had unique genotyping patterns, whereas 27 isolates (13.1%) involved into 13 clusters (identical genotyping patterns). Number of patients included in the clusters varied from 2 to 4 isolates from the patients. A minimum estimation of the proportion of patients from recent transmission vs. reactivation was 7.3% (Deqing 8.2%, Guanyun 6.4%). Both of results from univariate analysis and multivariate model showed a significant association between MDR-TB and clustering (OR=2.779, P <0.05).6. Based on the results of RFLP, the patients in the clusters was investigated again by case and contact investigation. The patients were separated into 13 groups according to their cluster information. Relationship, contacts place, contacts patterns, contacts period and potential other patients contacted among the patients in the clusters was recorded and interpreted. The results showed most of the clusters can be supported by the epidemiological links, whereas no evidence of epidemiological links can be identified in some of the clusters, showing an impact of causal contacts in rural areas. Floating patients were more likely to be an index case.