Table 4.

Sensitivity, specificity and likelihood ratio for detecting hypocalcemia and agreement of equations to classify calcemia status

Equation	Sensitivity	Specificity	Positive likelihood ratio	Negative likelihood ratio	Agreementa	p value
Equation 1	0.564	0.797	2.780	0.547	0.303 (0.180 – 0.427)	<0.001
Equation 2	0.697	0.780	3.168	0.389	0.424 (0.296 – 0.551)	<0.001
Equation 3	0.552	0.831	3.266	0.539	0.310 (0.184 – 0.435)	<0.001
Published general predictive equations
McLean-Hastings	0.175	0.949	3.439	0.869	not calculableb	–
Zeisler	0.977	0.136	1.131	0.169	0.142 (0.030 – 0.254)	0.192
Zeisler simplified	0.977	0.153	1.153	0.150	0.176 (0.06 – 0.291)	0.046
Pottgen	0.423	0.831	2.503	0.694	0.290 (0.164 – 0.416)	<0.001
Siggaard-Andersen	0.683	0.661	2.014	0.480	0.387 (0.245 – 0.528)	<0.001
Hanna	0.089	1.000	–	0.911	not calculableb	–
Butler	1.000	0.051	1.054	0.000	0.071 (0.004 – 0.138)	0.035
General predictive equations for corrected calcium
Payne	0.259	0.949	5.074	0.781	0.247 (0.134 – 0.359)	<0.001
James	0.477	0.881	4.008	0.594	0.363 (0.236 – 0.489)	<0.001

Figures in parentheses are 95s% CI. The sensitivity value of 1 would denote perfect ability to detect hypocalcemia when hypocalcemia was present; the specificity value of 1 would denote perfect ability to exclude hypocalcemia when hypocalcemia was not present; a positive likelihood ratio value of <1 increases the probability that the equation confirmed hypocalcemia. Greater values increased the probability; a negative likelihood ratio value of <1 increases the probability that the equation discards hypocalcemia. Smaller values increased the probability.

^aWeighted kappa, a value of 1 denotes perfect agreement.

^bWeighted kappa smaller than mean chance concordance.