Although high rates of uncontrolled blood glucose values have been reported in Puerto
Ricans with diabetes (Lipton et al., 1996), baseline HbA1c levels were near normal in this
study. There was a baseline HbA1c covariate by group interaction, suggesting that the
impact of the intervention varied according to initial HbA1c values. Although not reported
here, secondary analyses did show the intervention’s impact was strongest for those with
the highest baseline HbA1c values. This is consistent with studies showing improved HbA1c
is much more likely when initial HbA1c values are >10% (Sarkisian et al., 2003).