Less-energy-dense diets of low-income women in California are associated with higher energy-adjusted costs but not with higher daily diet costs

I have concerns about the relevance of the research and the accuracy of Townsend et al’s conclusion in their recent article (1). The article failed to show the usefulness or policy relevance of an inverse association between energy density and energy costs while promoting the perception that healthier diets cost more, when their own data show that the highest-energy-dense diets (‘‘least healthy’’) are actually associated with the highest daily diet costs. Because both energy density and energy costs use calories in their definitions, the ‘‘finding’’ of an inverse association can be easily explained mathematically. The energy density measure uses calories in the numerator, whereas the energy cost uses calories in the denominator. Mathematically, then, higher calories would increase the energy density measure while simultaneously decreasing the energy cost. To illustrate, last night I priced a 2-lb container of strawberries and a 32-oz bag of tortilla chips. Both cost the same, $3.99. The strawberries provided about 280 kcal (US Department of Agriculture nutrient database; www.nal.usda.gov/fnic/foodcomp/search/), with an energy cost of $28.50/2000 kcal; the tortilla chips provided about 4480 kcal (nutrition label), with an energy cost of $1.78/2000 kcal. Although both products cost the same, the higher-energy-dense chips have a lower energy cost, whereas the lower-energy-dense strawberries have a higher energy cost. It is not clear how useful it is to know that the calories provided by the strawberries ‘‘cost’’ 16 times as much as the calories provided by the chips, especially because both packages cost the same amount of money. Rather, it seems to me that the question of whether low-income consumers can afford a healthy diet requires a comparison of the monetary cost of different diets. Unfortunately, the article often misleads readers into thinking the authors are comparing the monetary cost of different diets. For example, consider the misleading conclusion that ‘‘healthier diets, as selected by participants, did indeed cost more,’’ rather than the more correctly worded title, which states that ‘‘less-energy-dense diets have higher dietary energy costs.’’ Fortunately, the authors’ Table 3 allows us to estimate the daily diet costs for each of the 3 energy-density tertiles. Dividing the average dietary energy cost ($/2000 kcal) by 2000 provides the cost of each kilocalorie; multiplying that by the average energy content of the diet then provides the average daily diet cost for each tertile: $6.25/d for the lowest tertile (lowest energy density), $6.07 for the middle tertile, and $6.34 for the highest tertile. That is, among these low-income women in California, the highest-energy-dense diets are actually associated with the highest daily diet costs. Although these differences are likely not statistically significant, they clearly refute the Townsend et al’s conclusion that ‘‘healthier diets, as selected by participants, did indeed cost more.’’ Furthermore, to the extent that consumers are taught to use fruit and vegetables to replace energy-dense foods, and not just as an addition to usual food intake, the reduced spending on the energy-dense foods would offset—perhaps entirely—the increased spending on fruit and vegetables. The anticipated nutritional benefits of such food substitutions would be increased intake of fruit and vegetables; increased intake of important vitamins, minerals, and fiber; and reduced intake of calories, fat, saturated fat, and added sugars—at little, if any, additional cost to the consumer.