Comment on Knowler et al. Preventing Diabetes in American Indian Communities. Diabetes Care 2013;36:1820–1822

Knowler and Ackermann (1) provided a commentary regarding our article reporting the Special Diabetes Program for Indians Diabetes Prevention (SDPI-DP) demonstration project (2). We deeply appreciate the efforts of the editorial staff in recruiting such a knowledgeable group of diabetes prevention experts to comment on this article. We are especially grateful to the authors for underscoring the contribution that the Indian Health Service has made, which includes the SDPI-DP, to lessen the daunting diabetes disparities that face American Indian and Alaska Natives.

The commentary correctly pointed out difficulties in quantifying the benefits of the intervention among SDPI-DP participants absent a control group. However, as we sought to communicate, the primary purpose of the SDPI-DP was to demonstrate the feasibility and impact of translating a proven intervention into a diverse array of American Indian and Alaska Native communities, not to assess the magnitude of program benefits in comparison with a control group.

Historical control data from the Diabetes Prevention Program (DPP) (3), along with diabetes incidence among various high-risk populations, were referenced in our Conclusions to contextualize the findings. We acknowledge further that interpretation of the data was constrained by our inability to directly compare with those studies. Knowler and Ackermann observed that the SDPI-DP employed broader eligibility criteria than the DPP and speculated that SDPI-DP participants thus were probably at much lower risk of diabetes. We agree partially. Yet, American Indians in the DPP also were subject to broader eligibility criteria (only elevated postload blood glucose required) than the other participating populations (both elevated fasting and postload blood glucose required) (4). Even so, American Indian participants in the DPP placebo group exhibited the highest diabetes incidence of all race/ethnic groups (12.9 vs. 11.0% overall rate). This suggests to us that American Indians may be at higher risk of diabetes with similar or even broader eligibility criteria.

When comparing SDPI-DP results to those of other translational projects, Knowler and Ackermann calculated weight changes among completers at each annual assessment for SDPI-DP participants using data in Table 2 and Supplementary Table 1 of our article (2). Unfortunately, those tables do not provide the data appropriate to their point. Table 2 presented the estimated mean weights for all SDPI-DP participants, which likely differ from those of the completers at each time point. For clarification purposes, we calculated the average weight changes among completers using paired t tests. The 1-year weight change among the 1,503 year-1 completers was −5.6/215.5 lb = −2.6%. The 3-year change among the 834 completers was −2.9/212.6 lb = −1.4%. In a subgroup of DPP lifestyle intervention participants, the −5.7 lb mean weight loss over a year was significantly and independently associated with lower diabetes risk (5), indicating that the magnitude of year-1 weight loss among SDPI-DP participants could significantly decrease their diabetes risk.

Again, we thank Knowler and Ackermann (1) for their thoughtful commentary. We hope this letter provides additional clarification regarding the ultimate goals and achievements of SDPI-DP, which highlight important implications for future similar translational efforts.