Table 3.

Surrogate category analysis study comparison table.

Study	Study population	BPA exposure characterization	n samplesa	Sensitivity	Specificity	PPV
HOPE study, analysis 1b,c	83 women (subset of HOPE study)	Serial daily first-morning collection during pre/conception cycles	1 2	0.59 0.65	0.76 0.79	0.52 0.59
HOPE study, analysis 2b,d	Same as above	Same as above	1 2	0.55 0.61	0.74 0.78	0.45 0.45
Braun et al. 2012	91 women from subset of EARTH study	3 or more spot samples collected from preconception to delivery	1	1st trimester, 0.70, 2nd trimester, 0.60, 3rd trimester, 0.67	1st trimester, 0.85, 2nd trimester, 0.80, 3rd trimester, 0.84	1st trimester, 0.70, 2nd trimester, 0.60, 3rd trimester, 0.67
Mahalingaiah et al. 2008e	Couples seeking infertility treatment in Boston	149 samples from 31 subjects with at least 3 repeat samplesf	1	0.64	0.76	0.63
Mahalingaiah et al. 2008e	Same as above	67 samples from 8 subjects with at least 6 repeat samplesf	2	0.67	0.84	0.85
Fisher et al. 2015g	80 pregnant women (P4 Study)	5 repeat samples from early pregnancy to deliveryh	1	0.65	0.66	0.61
Abbreviations: BPA, bisphenol A; PPV, positive predictive value. aNumber of samples used to calculate predictive statistics (sensitivity, specificity, and PPV). bResults for 1 and 2 repeat samples to predict high tertile among females presented for comparison. cMethods most directly comparable to those of Braun et al. (2012) and Mahalingaiah et al. (2008). dMethods most directly comparable to those of Fisher et al. (2015). eMen and women included in analyses together. fSpot samples collected at recruitment, at clinic visits during treatment, and during post-treatment clinical appointments. Women also collected samples during the trimesters of pregnancy. gStatistics calculated with slightly different methods. hTwo samples were 24-hr collected before 20 weeks gestation, and three were spot samples collected during the second and third trimesters and then 2–3 months postpartum.