Food records, food recalls, and food-frequency reports all have been used historically to assess dietary intake. One problem inherent with dietary assessment instruments is the underreporting of food intake by research subjects. 1-3 Underestimation of energy intake is reported to be approximately 20% in free-living subjects who are normal to overweight, 1,2,4 but may be higher for subjects who are obese 5 or during weight loss. 6

Jonnalagadda et al 6 found underestimation of energy intake to be approximately 12% when comparing 24-hour recalls with energy intake for weight maintenance among adults self-selecting diets in an uncontrolled environment. When food was self-selected in a controlled environment, men continued to underreport energy intake by approximately 13%, whereas women overreported energy intake by 1.3%. 6 In a controlled metabolic study, 3-day diet records provided significantly better estimates of actual intake than a food-frequency questionnaire. 1 Food-frequency questionnaires may be particularly unreliable and inadequate for assessing absolute and relative macronutrient intakes, especially when fat intakes are of interest. 1

The Heart Fit Rx Diet Habits Survey (HFD) presented by Watson et al 7 is a short, self-administered tool that is easy to score and cost effective for large-scale studies. However, the question remains whether this tool effectively quantifies the types and amounts of foods that are of greatest concern for cardiac rehabilitation participants. The HFD shows considerable variance in 3-day food record fat percentage, with only 42% of the variance accounted for by the fat subscore. Only 34% of the variance in estimated fat intake, using the fat subscore, could be accounted for by the 3-day food record fat percentage. Scores assigned to the various components of the HFD fat subscore are questionable. Previous study has shown that the HFD compared favorably with the Oregon Health Sciences University Dietary Habit Survey (DHS). 8 However, the scores for the DHS questions comprising the saturated fat-cholesterol subscore are based on the Cholesterol-Saturated Fat Index (CSI). 9,10 Foods with low CSI scores have a lower content of cholesterol-saturated fat than foods with high CSI scores, whereas the HFD scoring appears to be arbitrary.

Even if dietary fat intake were adequately approximated with the HFD, the questions that comprise the fat subscore also are of concern. Each question in the HFD is given the same weight, which might be appropriate if total fat intake were the only assessment desired. However, in cardiac rehabilitation, use of the American Heart Association (AHA) dietary guidelines 11 would seem prudent. These guidelines encourage individuals to maintain a healthy weight and to limit their intake of foods high in saturated fat and cholesterol by replacing them with grains unsaturated fats from vegetables, legumes, nuts, and fish. 11 Questions on the HFD, however, classify all types of fat together, with fish, nuts, and oils counting the same as red meat, eggs, and dairy products. The authors state that the HFD is limited in its ability to categorize dietary fats. 7 Therefore, should these questions be used to assess dietary fat intake for patients in cardiac rehabilitation?

The carbohydrate questions on the HFD were not provided to determine the type and amount of carbohydrate or soluble fiber intake. Total carbohydrate intake exceeding 60% of energy intake, without weight loss, can lead to elevated triglycerides and reduced high-density lipoprotein cholesterol levels. 12,13 Because hypertriglyceridemia also can be precipitated by a high intake of simple carbohydrates 14, it is important to assess this in patients undergoing cardiac rehabilitation. Soluble fiber has been shown to reduce cholesterol levels below the levels achieved by a diet low in saturated fat and cholesterol alone. 15 In individuals with hypercholesterolemia, soluble fiber reduced low-density lipoprotein cholesterol an average of 2.2 mg/dL for every gram increase in soluble fiber. 16

Although the authors' suggest the usefulness of the HFD in clinical community wellness programs, there may be less efficacy in such settings. The participants in the current study were phases 2 and 3 cardiac rehabilitation subjects. This population tends to be more motivated to make dietary changes and more educated about dietary recommendations, especially fat intake, than the general population. It is unlikely that similar results would be found in the general population.

Dietary assessment tools are most effective when they quantify both the amount and types of foods consumed. In addition, scoring of food items should be based on acceptable standards. Finally, the questions must reflect food items that are of greatest interest for a particular population. The HFD would be a more effective tool for the target population if the questions were based on the aforementioned principles.

References