Objective: Based on an epidemiological survey of NIDDM in the general population, two sub-population groups were selected to develop a screening method on risk factors to identify people at increased risk for undiagnosed diabetes.
Methods: Logistic regression analysis on the original data of population A was carried out to screen the main risk factors of diabetes. The score of the variable was determined based on the OR value and the aggregate score was used to predict the risk of undiagnosed diabetes. Both validity and effectiveness of the method were evaluated in population B.
Results: Results showed that the risk of having diabetes in the population increased along with aggregate scores of the method. Trend chi 2 test showed statistical significance (P < 0.01). When the threshold value was set up at 7, both sensitivity and specificity for identifying undiagnosed diabetes were 74.3% and 63.2%, respectively making the positive predictive value 4.2% and negative predictive value 99.2%.
Conclusion: The benefits of this screen method seemed to be simple, economical and helpful to obtain a satisfactory response rate. The method could be used for health education and to identify people (community) at high risk for potential diabetes. It was predicted that this screening method could serve as an effective and useful tool for mass screening of NIDDM.