Breslow-Day test

Breslow-Day test

A test for homogeneity of odds ratio, which is used to evaluate changes in the degree of difference between 2 datasets being analysed in 2 different periods.
References in periodicals archive ?
The Breslow-Day test was used to examine the homogeneity of the hazard ratios between concordant and discordant groups.
The Mantel-Haenszel statistic and the Breslow-Day test of trend in odds ratio heterogeneity are chi-square techniques; the former mainly to detect uniform DIF and the latter non-uniform DIF (Angoff, 1993; Bertrand & Boiteau, 2003; Dorans & Holland, 1993; Fidalgo & Madeira, 2008; Kristjansson, Aylesworth, McDowell, & Zumbo, 2005; Narayanan & Swaminathan, 1994; Penfield, 2003).
Applying the Breslow-Day test of trend in odds ratio heterogeneity to the analysis of non-uniform DIF.
The homogeneity of odds ratios was tested using the Breslow-Day test and used to determine whether the estimated odds ratio is different between the three treatment groups.
Application of the Breslow-Day test of trend in odds ratio heterogeneity to the detection of nonuniform DIF.
The homogeneity of the odds ratios for code assignment in VA versus Medicare-reimbursed facilities was evaluated with the Breslow-Day test. Alpha error was considered significant at less than 0.05, two-tailed.
Table 4: Accuracy of Inpatient Coding for Chronic Renal Failure (Group A Codes) by Medicare Provider versus VA Provider Medicare All CKD * (%) No CKD (%) Coded as 4,461 4,228 233 group A (39.6) (43.3) (15.6) Coded as 6,792 5,530 1,262 groups B-G (60.4) (56.7) (84.4) Total 11,253 9,758 1,495 Odds ratio 4.1 (3.6-4.8) p < .0001 Breslow-Day test for homogeneity of odds ratios p = .0330 VA All CKD * (%) No CKD (%) Coded as 6,166 6,050 116 group A (29.0) (30.6) (7.6) Coded as 15,110 13,700 1,410 groups B-G (71.0) (69.4) (92.4) Total 21,276 19,750 1,526 Odds ratio 5.4 (4.4-6.5) p <.
Effect modification (interaction) was assessed by the Breslow-Day test of homogeneity[15] with a p-value of [less than] 0.20 used as suggestive of meaningful heterogeneity of risk across the levels of each covariate.
National estimates were obtained by employing the sampling weights assigned to each observation, according to NAMCS guidelines.[13] Tests of significance, when applied, utilized the chi-square statistic for frequency tables and the Breslow-Day test for homogeneity to assess the significance of statistical interaction in stratified analyses.