Dietary assessment - Recalls
A dietary recall is a retrospective method of dietary assessment where an individual is interviewed about their food and beverage consumption during a defined period of time, typically the previous day or the preceding 24 hours. Recall of intake over a longer time period is problematic due to the limitations of memory. Several national surveys use the 24-hour recall method because of its high response rate and its ability to obtain detailed information. The interview can be carried out in person, by telephone or increasingly via the Internet. In the Norwegian arm of the EPIC study no significant differences in the dietary data obtained were found when face-to-face 24-hour recalls were compared to telephone 24-hour recalls (Brustad et al, 2003).
A single 24-hour recall is not considered to be representative of habitual diet at an individual level but is adequate for surveying intake in a large group and estimating group mean intakes. In a preliminary study to decide the method for the UK Low Income Diet and Nutrition Survey (LIDNS), four repeat 24-hour recalls were recommended as the most appropriate method of dietary assessment in this group (Holmes et al, 2008). Repeat 24-hour recalls can be employed to assess a typical diet at an individual level; these are also known as multiple recalls. In a recent Australian study in adults, eight repeat 24-hour recalls were recommended to capture the variation in macronutrient intake (Jackson et al, 2008); Sunday was the day of greatest variation in diet. Refer to section on ‘diet variation’ for more information on the number of dietary assessment days required.
Traditionally the 24-hour recall is undertaken in chronological order of consumption. A more recent development has been the multiple pass recall and this protocol is used increasingly in national surveys and research studies. The multiple pass recall (MPR) is a staged approach to the dietary recall. Although the exact stages or passes may vary between protocols they all follow the pattern of a free and uninterrupted recall of intake, followed by detailed and probing questions about intake (including quantities consumed) and concluding with a review of everything that was previously recalled, allowing for the addition of any items not remembered up to this point, and often also the location of the consumption.
The stages followed in the UK LIDNS study are detailed below (Nelson et al, personal communication, 2008).
1. A quick list of foods eaten or drunk
Respondents are asked to report everything that they had to eat or drink (e.g. on the previous day between midnight and midnight) in an uninterrupted free flowing list.
2. Collection of detailed information
For each item of food or drink in the quick list, respondents are asked to provide additional detail, including:
• The time at which the food or drink was consumed
• A full description of the food or drink, including brand name where available
• Any foods likely to be eaten in combination e.g. milk in coffee
• Recipes and other combinations of foods e.g. sandwiches
• The quantity consumed, based on household measures, photographs of different portion sizes of foods, or actual weights from labels or packets
• Any leftovers or second helpings
3. A recall review
The interviewer reviews all of the food eaten and drunk in chronological order, prompting for any additional eating or drinking occasions and foods or drinks consumed, and clarifying any ambiguities regarding the type of food or drink consumed and portion size. Finally the interviewer asks the respondent to select a place name for each eating and drinking occasion.
It is thought that this method is more tailored to individuals’ cognition rather than the chronological approach. In the US, an interviewer administered automated MPR (AMPR) is the dietary assessment method used in national surveys (Raper et al, 2004; Blanton et al, 2006; Subar et al, 2007). Currently a web-based AMPR is being developed by Dr Amy Subar and in association with the National Cancer Institute and the US Department of Agriculture.
24-hour recalls are particularly challenging in young children and the elderly. The cognitive processes involved in the recall of food have been described in children as (Baranowski & Domel, 1994):
· Attention
· Perception
· Organisation
· Retention
· Retrieval
· Response formulation
The ability of a child to provide a dietary recall increases markedly after the age of 8 years (Livingstone & Robson, 2003). It has been demonstrated that the ability of younger children (aged 5-7 years) to recall a lunch meal a short time after consumption varies widely and preference and familiarity increased recall; leftover food was not reported (Warren et al, 2003). One researcher in particular has looked at the use of prompts in small numbers of children aged 10 years. Specific prompts were found to hinder recall compared to free recall (Baxter et al, 2000); an open interview format was preferable to a meal interview format (Baxter et al, 2003). It is not clear what order of recall i.e. reverse order versus forward order, is preferable to use in children (Baxter et al, 2003). Recency was found to have a positive effect on children’s recall, which may mean it is preferable to inquire about intake during the previous 24-hours rather than the previous day (Baxter et al, 2004).
One study in UK preschoolers has shown that repeat (x3 days) MPR is not valid at an individual level to measure energy intake when compared to energy expenditure using doubly labelled water; over-reporting was apparent (Reilly et al, 2001). A similar study was repeated in young school children and although inaccuracies were reduced, intakes were still only valid at a group level (Montgomery et al, 2005). This concurred with an earlier US study in young children (Johnson et al, 1996). These studies all relied on surrogate reporters, like parents and carers.