1. Volume 81, Issue 6, Pages 516-519 (March 2012)
Multiple comparisons, interaction effects, and statistical inference: lessons from chronic kidney disease progression among blacks 
Rajiv Agarwal 
Kidney International 
Volume 81, Issue 6, Pages (March 2012)
DOI: /ki
Copyright © 2012 International Society of Nephrology Terms and Conditions

2. Figure 1
Probability of albuminuria by categories from baseline to post-treatment among blacks and non-blacks, derived from the multiplicative ordinal logistic regression model. The predictive terms were ethnicity, drug, and time and all possible interactions in this model. The outcome was the change in category of albuminuria. Blue lines represent the combination of benazepril and hydrochlorothiazide (B+H), whereas red lines show the results of the combination of benazepril and amlodipine (B+A). The drug × time interaction was significant, but the three-way interaction (ethnicity × drug × time) was not. Post-Rx, post-treatment.
Kidney International  , DOI: ( /ki )
Copyright © 2012 International Society of Nephrology Terms and Conditions

3. Figure 2
Lack of interaction effect between therapies by ethnicity. Blue bars represent the combination of benazepril and hydrochlorothiazide (B+H), whereas brown bars show the results of the combination of benazepril and amlodipine (B+A). The y axis shows the rate of decline in estimated glomerular filtration rate (eGFR). The differences between the differences (the interaction effects), although not reported, are unlikely to be statistically significant.
Kidney International  , DOI: ( /ki )
Copyright © 2012 International Society of Nephrology Terms and Conditions