A recently published map entitled "The Contiguous United States Visualized by distance to the nearest McDonald's" correlates somewhat but overall not so well with the 2008 map published by the CDC of obesity trends across the United States.
For example, the New England states show close proximity to the restaurants but some of the lowest percentages of obese individuals among the population. In South Carolina, over 30% of the population is considered obese (BMI > 30 kg/m^2) and greater than its neighboring states, but the proximity to McDonald's seems no different than in Georgia, North Carolina and other states in the region. Alabama and Mississippi have some of the highest obesity trends in the USA, but lower prximity to McDonald's. Of note, there is a high concentration of McDonald's franchises in the Chicagoland area because this is the location of the company's headquarters and test kitchen. Lastly, there are high concentrations of the restaurants in the main population centers of Utah (Salt Lake City) and Colorado (Denver), but both these states show rather low population trends of obesity.
Where there is one, you will find many more. In many instances the presence of one type of establishment acts as a seed for others. Thus, the above in essence uses the location of and priximity to McDonald's as a proxy for other fast food restaurants. The National Health and Nutrition Examination Survey (NHANES), with data from 2007 - 2008, can give some information on meals eaten in the home, meals eaten with the family and money spent on meals outside the home. Those data can be found here. I have not analyzed these data nor have seen a published analysis. I'll have to take a look.