Computer-administered screening of reproductive-aged women for diabetes risk in primary care settings, feasibility and acceptability of such screening, and validity of risk assessments based on self-reported weight

Introduction Obesity, a major public health problem, is the key modifiable component of diabetes risk. Addressing obesity and diabetes risk during primary care visits is recommended but, because of time constraints, is often difficult for health care providers to do. The purpose of this study was to determine whether technology can streamline risk assessment and leave more time to educate patients. We also tested the validity of self-reported weight in assessing diabetes risk. Methods We recruited English-speaking women aged 18 to 44 years who came to a clinic for medical appointments from July through October 2003. Study participants completed a self-administered computer questionnaire that collected the following data: Weight, height, family history of diabetes, level of exercise, amount of television time, and daily servings of fruits and vegetables. Self-reported and scale-measured weights were compared to determine the effect of self-reported weight on results of the American Diabetes Association's Diabetes Risk Test (DRT). In determining the sensitivity and specificity of self-reported weight, we used scale measurements as the standard. Results Complete data were collected on 231 women, including 214 women without a history of a diabetes diagnosis. Compared with DRT results (determined by scale-measured weight), questionnaire results (determined by self-reported weight) had sensitivities of 93.9% (95% confidence interval [CI], 85.2%-97.6%) for high risk for diabetes and 90.4% (95% CI, 83.3%-94.7%) for moderate risk. The specificity of the self-administered DRT for any diabetes risk was 97.8% (95% CI, 88.4%-99.6%). About half the women reported discussing nutrition and exercise with their health care providers. Conclusion Health care professionals can provide personalized diabetes education and counseling on the basis of information collected by self-administered computerized questionnaires. In general, patients provided a self-reported weight that did not substantially bias estimates of diabetes risk.