The two major issues addressed by Drs Jones and Russell are (1) the use of the term "over 100% compliance" and (2) the validation of interviews by serum digoxin level. Regarding the first issue, we agree that "over 100% compliance" is noncompliance. We think it was clear that the term referred to those cases in which patients stated they were taking more medicine than what they understood their physicians wanted them to take.
The second issue is far more complex. We stated in our article, and the accompanying editorial reinforced, that measurement of medication compliance is fraught with methodologic problems. We disagree that the correlation for the group of serum digoxin levels with interview assessment is probably fortuitous. First of all, the correlation was significant at P <.05. Second, others have found a similar correlation.1 Nevertheless, as pointed out by Jones and Russell, it is still difficult to know