To the Editor.
—Dr Saudek and colleagues1 have published a well-designed study to assess the efficacy and safety of the implantable insulin pump. Following published criteria to assess the quality of an article on therapy,2,3 this study seems to have an excellent level of scientific evidence: it was a randomized controlled trial, the sample size was calculated a priori, the population was well defined, and assessment parameters were robust. Nevertheless, the study results are not clear.First, the results on efficacy are the means of blood glucose level: "Both groups showed significant decreases in average blood glucose levels over time (P<.001).... Fasting plasma glucose level was 8.18±1.5 mmol/L (147.5±27.3 mg/dL) for implantable insulin pump vs 9.1 ±2.0 mmol/L (163.9±36.2 mg/ dL) for multiple-dose insulin (P=.57 treatment effect)." Nevertheless, the authors have defined their treatment objectives for each study group in the introduction "to achieve fasting and
Pazart LH. Treatment of Patients With Non-insulin-dependent Diabetes With the Implantable Insulin Pump. JAMA. 1997;277(7):529. doi:10.1001/jama.1997.03540310027017