Validation of digital photographs, as a tool in 24-h recall, for the improvement of dietary assessment among rural populations in developing countries

Abstract

Background: Improvement of traditional methods for dietary assessment is necessary, especially in rural areas where it is more difficult to succeed with self-reporting methods. This study presents and validates a method for improving accuracy when measuring food and nutrient intake of individuals in rural areas. It is called the "Food photography 24-h recall method" (FP 24-hR) and is a modified 24-h recall with the addition of a digital food photography record and a photo atlas.

Methods: The study was carried out in a rural area in the tropical region of Bolivia; 45 women participated. Validation of the method was made by comparing it with a reference method, the Weighed Food Record (WFR). During the FP 24-hR, digital photographs were taken by the subjects of all food consumed during a day and a 24-h recall questionnaire was conducted by an interviewer. An estimate of the amount of food consumed was made using a photo atlas and the photographs taken by the subjects. For validation, comparison was made between the calculations, by both methods, of the levels of food, and nutrient, intake.

Results: The comparison was made in 10 food categories; most of which were somewhat underestimated from -2.3% (cassava) to -6.8% (rice), except for beverages (+1.6%) and leafy vegetables (+8.7%), which were overestimated. Spearman's correlation coefficients were highly significant (r from 0.75 for eggs to 0.98 for potato and cassava). Nutrient intakes calculated with data from both methods showed small differences from -0.90% (vitamin C) to -5.98% (fat). Although all nutrients were somewhat underestimated, Pearson's coefficients are high (>0.93 for all) and statistically significant. Bland Altman analysis showed that differences between both methods were random and did not exhibit any systematic bias over levels of food and nutrient intake, with acceptable 95% limits of agreement.

Conclusion: The FP 24-hR exhibits acceptable differences when compared with a WFR, digital photos are useful as a memory aid for the subjects during 24-h recall and as an estimation tool. The method is suitable for assessing dietary intake among rural populations in developing countries.

Figures

Figure 1

Design of the study: Validation of the developed method FP 24-hR by comparing it with WFR.

Figure 2

Example of photographs from the photo atlas shows portion sizes of cooked and raw food. *Example of cooked food in different dishes and from different angles: the name of the food (in Spanish) and weight of the portion is shown on the top of each photograph. In cases where there is a change in the weight during cooking, both weights are shown (weight of cooked food/weight of raw food). **Example of raw food, the weight of the individual food is shown inside the photographs.

Figure 3

Photo kit: Digital camera, camera case and marked table mat.

Figure 4

Representative photographs of breakfast, lunch (from two different angles) and dinner taken by a subject.

Figure 5

Bland Altman plots for estimated and weighed food amount. Differences between the log amounts of food portions estimated and weighed against their mean values, the solid line represents the average difference between the log estimated and the log weighed food amount; the dotted lines show the 95% log limits of agreement which, when calculating the antilog, represent the range of proportional agreement between both methods. a) Plot for noodles amount, b) Plot for potatoes, c) Plot for meat and d) Plot for vegetables. Plots show not systematic bias and that the range of proportional agreement is narrow enough to be confident using the photo method.

Figure 6

Bland Altman plots for nutrient intakes calculated from FP 24-hR and WFR. Differences between the mean dietary intakes of nutrients calculated from FP 24-hR and WFR against their mean values. The solid line represents the average difference between the FP 24-hR and WFR; the dotted lines show the 95% limits of agreement for the differences. a) Plot for energy intake, b) Plot for calcium intake, c) Plot for vitamin C intake and d) Plot for iron intake. The ranges of proportional agreement are narrow enough to be confident using the photo method.