DETERMINING THE IMPACT OF A NATIONAL PUBLIC HEALTH EDUCATION PROGRAM THAT TEACHES THE IMPORTANCE OF FAMILY HISTORY OF DIABETES: CAN GOOD NEWS DO BAD?

Purpose: Family history (FH) is an important risk factor for diabetes and is used by clinicians as a screening tool for diabetes prevention and promoting healthy behavior. A national public health program teaching the importance of diabetes FH has been hypothesized to benefit people with a FH of diabetes by increasing their likelihood of healthy behavior that can reduce their risk of developing diabetes and by increasing screening to reduce the duration of undetected diabetes.  However, such a program could also decrease the likelihood of healthy behavior and screening among those without a FH of diabetes, producing the opposite effects for those patients. It is unknown whether the potential benefits of a program of this type are greater or less than the potential harms.
   Methods: Using a first order Monte Carlo simulation, we model the effect of knowledge of importance of diabetes FH (KIFH) on the individual’s annual diabetes screening behavior and level of physical activity.  We stratify the population using a fixed distribution of diabetes FH and account for baseline diabetes prevalence and KIFH in each FH strata.  The population progresses through annual Markov health states defined by diabetes, KIFH, screening, and physical activity. Population parameters are drawn from the CDC’s HealthStyles and the NHANES survey data.
   Findings: We find that among subjects with diabetes FH, who account for a third of the population, a program to teach the importance of FH will decrease the average age of onset of diabetes by only 0.04 years but decrease the number of years with undiagnosed diabetes by 4.1 years. Effects among subjects without a FH of diabetes cannot be estimated precisely but, sensitivity analysis suggest a 1.7 year increase in the number of years with undiagnosed diabetes, again keeping the average age of onset unchanged.
   Conclusion:  While results suggest a harmful effect from exposing those without a FH of diabetes to the program, the current data lacks the precision to provide a conclusive estimate of the net effect.  Regardless, medical or public health models that target education of genetic risk factors to patients at higher risk (e.g. by promoting outreach by affected family members) may be preferred to non-selective education, especially in cases such as diabetes where patient behavior can mitigate the effects of favorable genetics.