[Prediction of a risk for drug resistance in the pathogen of pulmonary tuberculosis]

The authors studied the spread of drug resistance in Mycobacterium tuberculosis was studied in Armenia and its association with risk factors in 1114 patients with positive bacterial culture and developed a mathematical method for early detection of a risk of its developed, which had been tested in 210 patients. Over 3 years (2002-2004), there was a rise in multiple drug resistance from 8.83 to 10.03% and polyresistance from 5.61 to 7.67%, a reduction in monoresistance from 8.68 to 6.96%; among new cases, there was an increase in multiple resistance from 3.23 to 6.59%, polyresistance from 4.42 to 7.01% and a decrease in monoresistance from 8.73 to 6.74%. For estimation of the value of clinical, laboratory, and social risk factors, the authors complied a rapid diagnostic table (RDT) that reflected 16 signs. Fischer's angular transformation method and stage-by-stage computer-aided analysis were analyzed, by detecting the diagnostic coefficients and informative value of signs, which was based on Wald's analysis called a heterogeneous sequential statistical recognition procedure. Testing this procedure indicated that their data was concordant with the results of bacterial cultures. The use of the proposed procedure make is possible in any conditions to predict the possible development of drug resistance and to take measures of its prevention.