[Retracted] Prediction of Multidrug-Resistant Tuberculosis Using Machine Learning Algorithms in SWAT, Pakistan

Table 2

Variable selection via the logistic regression model. Small p value indicates the significance of a variable.

Variable	Coeff.	S. E.	value	OR	95% CI for OR
					Lower	Lower
Previous TB history	1.719	0.772	0.026	1.179	0.039	2.485
Improper treatment	1.376	0.564	0.015	3.960	1.312	7.951
Smoking	1.344	0.658	0.041	3.835	1.057	8.915
MDR close contact	−2.037	0.398	0.000	2.130	0.060	4.285
Break-in TB medication	−0.834	0.238	0.000	0.434	0.272	0.893
Malnutrition	2.479	1.067	0.001	6.431	4.006	12.55
Hemoptysis (Produce cough)	−1.731	0.668	0.010	0.177	0.048	0.656
Restless or fatigue	−1.056	0.374	0.005	0.348	0.167	0.724
Depression	−0.723	0.355	0.041	0.485	0.242	0.972
Chest pain	2.322	0.679	0.001	10.195	2.697	18.545
Weight loss	−1.003	0.436	0.021	0.367	0.156	0.861