Table 2.

The results of logistic regression (estimator value and its significance)^∗.

	Estimate	Std. error	z value	Pr(>|z|)
Hair model
(Intercept)	6.5	2.58	2.52	0.0119
Zn.H	−0.088	0.0274	−3.21	0.00131
Cu.H	0.204	0.0884	2.3	0.0213
Mn.H	1.71	0.692	2.47	0.0135
Co.H	−19.7	8.71	−2.26	0.0235
Nail model
(Intercept)	−6.65	3.52	−1.89	0.0591
Fe.N	0.112	0.0368	3.03	0.00247
Zn.N	−0.101	0.0338	−3	0.00274
Cu.N	0.38	0.141	2.69	0.00719
Total model
(Intercept)	−3.5	4.43	−0.791	0.429
Zn.H	−0.122	0.0408	−2.98	0.00286
Fe.N	0.0909	0.0293	3.1	0.00193
Cu.H	0.361	0.124	2.9	0.00373
Pb.N	2.7	1.08	2.51	0.012

^∗The modelled equation computes depression probability from the most significant elements determined in hair and nails.