If women with unrecognized preexisting diabetes were overrepresented in the control group, this could lead to bias toward a lower risk of gestational diabetes in the bariatric-surgery group as compared with the control group.
If women with unrecognized preexisting diabeteswere overrepresented in the control group, thiscould lead to bias toward a lower risk of gestationaldiabetes in the bariatric-surgery group ascompared with the control group.