Table 3.

Summary estimates of sensitivity and specificity, using a bivariate random effects model (10³)

SYMPTOM	No. of studies	No. of Patient	Sensitivity	(95% CI)	Variance Logit (sensitivity)	Specificity	(95% CI)	Variance Logit (specificity)
Dysuriaa	12	2845	0.79	(0.72-0.85)	0.39	0.39	(0.31-0.49)	0.40
Frequencyb	11	2246	0.88	(0.82-0.92)	0.43	0.21	(0.14-0.31)	0.64
Back painc	4	635	0.38	(0.26-0.52)	0.22	0.57	(0.40-0.73)	0.43
Feverd	6	926	0.12	(0.05-0.26)	1.15	0.91	(0.80-0.97)	1.16
Flank paine	4	783	0.29	(0.18-0.43)	0.31	0.65	(0.59-0.70)	0.01
Hematuria	6	854	0.22	(0.18-0.27)	0.02	0.87	(0.81-0.91)	0.11
Lower abdominal painf	5	914	0.44	(0.26-0.64)	0.78	0.58	(0.37-0.77)	0.86
Nocturia	5	1492	0.59	(0.48-0.70)	0.22	0.57	(0.51-0.62)	0.04
Urgencyg	7	1739	0.62	(0.46-0.76)	0.72	0.51	(0.35-0.68)	0.78

^a One study reported dysuria as painful voiding (Medina-Bombardo 2003)

^b Two studies reported frequency/dysuria (Dans&Klaus, Wigton: training and validation set)

^c One study reported back pain as 'back or groin pain' (O'Brien)

^d One study reported fever as pyrexia (Lawson 1973)

^e One study reported flank pain as loin pain (Lawson 1973)

^f Different definitions were used: 'Suprapubic pain' (Hummer-Pradier 2005), Suprapubic pressure' (Baerheim), 'abdominal pain' (Fahey 2003)

^a One study reported urgency as urgency/frequency (Hummers-Pradier 2005)