• diabetes mellitus;
  • aging;
  • randomized clinical trial;
  • counseling;
  • physical activity;
  • veterans;
  • obesity


To determine whether a home-based multicomponent physical activity counseling (PAC) intervention is effective in reducing glycemic measures in older outpatients with prediabetes mellitus.


Controlled clinical trial.


Primary care clinics of the Durham Veterans Affairs (VA) Medical Center between September 29, 2008, and March 25, 2010.


Three hundred two overweight (body mass index 25–45 kg/m2), older (60–89) outpatients with impaired glucose tolerance (fasting blood glucose 100–125 mg/dL, glycosylated hemoglobin (HbA1c) <7%) randomly assigned to a PAC intervention group (n = 180) or a usual care control group (n = 122).


A 12-month, home-based multicomponent PAC program including one in-person baseline counseling session, regular telephone counseling, physician endorsement in clinic with monthly automated encouragement, and customized mailed materials. All study participants, including controls, received a consultation in a VA weight management program.


The primary outcome was a homeostasis model assessment of insulin resistance (HOMA-IR), calculated from fasting insulin and glucose levels at baseline and 3 and 12 months. HbA1c was the secondary indicator of glycemic control. Other secondary outcomes were anthropometric measures and self-reported physical activity, health-related quality of life, and physical function.


There were no significant differences between the PAC and control groups over time for any of the glycemic indicators. Both groups had small declines over time of approximately 6% in fasting blood glucose (P < .001), and other glycemic indicators remained stable. The declines in glucose were not sufficient to affect the change in HOMA-IR scores due to fluctuations in insulin over time. Endurance physical activity increased significantly in the PAC group (P < .001) and not in the usual care group.


Home-based telephone counseling increased physical activity levels but was insufficient to improve glycemic indicators in older outpatients with prediabetes mellitus.