The sample size was calculated using the following assumptions: prevalence of TB in rural community of 0.36% [
6], 95% CI, margin of error of 20% and non-response rate of 10%. This gave us a total sample of 28917. Since the total adult population in Gilgel Gibe field research centre was 30040 (almost equal to the sample size), all adults > = 15 years were screened for TB.
A one-page screening questionnaire was prepared in Amharic (local language) to identify all adult TB suspects in the study area. Trained skilled workers who had completed high school education visited each household. If the heads of the households were not available during the visit, the data collectors repeatedly visited the same household 3 times. All adult (age > = 15 years) with a cough of two weeks or more were considered as TB suspect and given a sputum cup to bring a morning sputum to the nearby school or clinic (appointment centre) the next day. TB suspects were also asked to provide a second "on the spot" sputum sample during their visit to the appointment centre. During their visit, the TB suspects were interviewed about the presence of other symptoms of TB such as fever, their socio-demographic characteristics and perception towards TB using a structured questionnaire. We asked the presence of fever using Affan Oromo (the local language). The study area is malaria endemic and fever is well-known by the local language. Every TB suspect was counseled and asked consent to be tested for HIV. Screening for HIV was done using the KHB test (Shanghai Kehua Bio-Engineering Ltd, Shanghai, China; 2008). A positive sample was retested using the STAT-PAK test (Chembio Diagnostic System Inc, Medford, NY, USA; 2008). Collected sputum samples were transported to Jimma University specialized hospital using a cold box. The same day, sputum smears were examined for the presence of acid-fast bacilli (AFB) by an experienced laboratory technician using the standard Ziehl-Neelsen method [
8]. Sputum samples were kept at -20°C till transported within a maximum of 5 days to Armauer Hanssen Research Institute (AHRI) in Addis Ababa for culture. At the AHRI, sputum specimens (2.5-10 ml) were processed by the standard N-Acetyl-L-Cysteine-Sodium Hydroxide (NALC-NaOH) method [
9] and concentrated at 4000 ×
g for 15 minutes. The sediment, irrespective of the original sample volume, was reconstituted to 2.5 ml with phosphate buffer pH 6.8, to make the inoculums for the smears and cultures. Two Lowenstein-Jensen slants, one containing 0.75% glycerol and the other containing 0.6% pyruvate, were inoculated with the sediment and incubated at 37°C. Cultures were considered negative when no colonies were seen after 8 weeks incubation period. Isolates were harvested, Deoxyribonucleic Acid (DNA) extracted using a standardized protocol [
10] and confirmed as
Mycobacterium tuberculosis complex by an in-house Polymerase Chain Reaction (PCR) [
11]. Standard spoligotyping method [
12] was done generally as described by Kamerbeek and colleagues using a commercially available kit (Isogen Bioscience BV, Maarssen, the Netherlands). The SpolDB4 database [
13] and a web-based computer algorithm, Spotclust
http://tbinsight.cs.rpi.edu/run_spotclust.html, were used to assign new isolates to families, subfamilies and variants. SpolDB4 assigned names (shared types) were used whenever a spoligo pattern was found in the database. Patterns not found in SpolDB4 were assigned to families and subfamilies by Spotclust. Spoligotypes described only once (non-clustered) in this study and in the SpolDB4 were designated as 'NA' (not assigned). A cluster was defined as two or more isolates from different patients with identical spoligotype patterns.