Statistical analysis of various risk factors of tuberculosis in district Mardan, Pakistan

In this study, an effort has been made to determine the most important risk factors of tuberculosis (TB) in district Mardan. A total of 645 cases were examined, and their personal and medical data were collected. For each case, the phenomenon of TB was studied in relation to different risk factors. Statistical techniques of logistic regression and backward elimination procedure were used to analyze the data and to determine a parsimonious model. For both male and female cases, the final selected logistic regression model contain the risk factors: sex, residence, household population, diet, medical care, and close contact with infectious patients as well as a joint effect of two factors and three factors, namely, medical care and marital status; and economic status, and medical care and marital status. Separate logistic regression was then fitted for each sex using the same procedure. For male cases, the final selected logistic regression model contains risk factors: residence, diet, and close contact with infectious patients as well as a combined effect of two factors, namely, economic status and diet, medical care and diet. For female cases, the final selected logistic regression model contains the risk factors: household population, economic status, diet, and close contact with infectious patients as well as a combined effect of two factors, namely, medical care and close contact with an infectious patient.