What Ford and Sciamanna seem to believe is that our study participants were individually randomized to each of the experimental conditions (as would be the case in a clinical trial). However, this was not the case. Physician practices were randomized and patients were then to be treated within the randomization scheme. Therefore, the patients who were referred to the Nutrition Center Model by their physicians but did not ever enter therapy (ie, were never enrolled) cannot be included in the analyses. In a clinical trial, they would not have been randomized because they would have been excluded owing to disinterest. As we said in the article, we made considerable effort to enroll them, but they were not interested in joining the study.
We thought that we had been very conservative with regard to analyses by including all patients who had entered therapy regardless of whether their serum cholesterol levels were