For a research study I'm planning on conducting, we want to give participants the option of whether to provide information about their nutritional intake either via telephone, in a written diary form, or via online survey. In the case of telephone data, this would need to take the form of a 24 hour recall - whereas in the written forms it could take the form of a food diary. These are different methodologies, but both trying to access the same underlying reality - what the participant what a participant ate or drank over a given 24 hour period. So is this ok? Or if we try to conflate the results of the two methods into a single 'nutritional intake' variable are we introducing some sort of systematic bias or other problem? Is this sort of multi-method nutritional assessment used by anyone else?
I'm agree, they are two different methodologies, so you might introduce a bias, because, usually in a 24 hour record the subject use the memory, while in food diary the subject write the foods as he consumes. Another bias, might be the form of application, the record is administered for one investigator by telephone, and the Diary is self-administered. In our experience in the UCV, we don't combine several methods of nutritional assessment.
My recomendation is: study well the objetive of your reserach and choose the most appropriate.
Hi Pablo, Thanks for your response. In our study, nutrition is just one of the factors we are trying to collect data on, and for a variety of logistical reasons we'd prefer to have multiple options to collect data. That being said we don't want to compromise the validity of the findings. I think thats one vote in favor of using just one method tho - so thanks!
Agree with Pablo- less variability due to differing methods and more likely to see a treatment effect of whatever you are studying. If you want to publish it stick to one Method- or use both Methods in all individuals. If you use a food diary you should give them an instruction session with some measuring and or weighing devices that you give them to take home. If you have limited access to your participants you might also want to consider using a food frequency questionnaire.
The problem is that the two methods should be used for different purposes. A single 24-h recall is really only suitable for group comparisons, as individual's intakes vary over a day, and to aggregate this out you need to create groups (e.g. lean and obese) where the individual deviations from normal are important (one would expect more deviations towards the excess end from usual diet in the obese group, for example). So 24-h recall is good for comparing lean and obese. Check out Willet's Nutritional Epidemiology text book. Food diaries are good for individual - level data e.g. associations between reported intake and an outcome. Recalls and food diaries have been validated as usable in the same study, but only when the recalls are collected over several days. So - you should decide what purpose your dietary data is for (group comparisons vs. within-person associations) and choose a methodology that fits your purpose.
Hi Pablo and colleagues.
I agree with the fundamentals of all answers that you have provided, but in regards the questionnaires, it is very important to have them validated as well, one against another (such as 24-hr vs 3 day vs FFQ) for selective macronutrients (if your study is looking at energy levels) or against some of the micronutrients (folate, vit b12, etc) if your study is looking at mechanistic pathways. Also, once that is done, validate the questionnaires against some blood/urine biomarkers. Then you are certain that data that you have collected is valid and representative of what are you looking for. I am often surprised how many questionnaires are used in studies without any mention of validation.
In other words, rather than just selecting type of questionnaire, use the data that these questionnaires will provide you with...
Warm regards
Nenad
I think the two methods are different. It is well known both of them have limitations. Based on my experience, FFQ might be a better way to esitmate the nutritional status of population. problem is why you use them at the same time, of course, they can be used.
You would probably need to have a proportion of the population completing both the 24-h dietary recall and the food diary to calibrate the methods against each other. There may be differences in the sub-population that chooses to take the dietary recall vs. the food diary that affects how well they recall their diet that you would want to try and capture. Perhaps some questions about health motivation and educational background in addition to standard demographics would help here. I would also recommend conducting at least two 24-hr dietary recalls.
Wow, Thanks everyone for all your thoughts, its really useful as the nutrition element of this study is not my area of expertise. Seems to be a clear consensus that using both should be avoided - or we need to give some participants both to validate against each other. Think the reality is we will just go with one method and run with it. Everyones comments also gives me some greater confidence in the other aspects of the study as we are already including a FFQ at baseline for habitual intake and collecting 24 hour recalls or Diaries (whichever we opt for) on multiple days over a 6 week period. So thanks everyone, really appreciate it!
I think methods are important , but is is more important to gather data accurately. Only one method you can apply, but the possibility of exact information will be more, if face to face interview will be conducted.
By using two different methods, you risk introducing bias. I would not give participants an option. Decide on one method and use it for all participants. Alternatively, you can run two separate studies: online and in-person food intake. I think the written form should closely match the 24-hr recall. Also, you would need 4-5 24 hr recalls for a limited sample size. Using one 24 hr recall can work for large, representative samples.
.If your objective were to compare results from two methods of data collection then you could consider using the two. So I agree to the use of only one method. Good luck
Strongly not recommended to use two different methods. No matter how hard you try to harmonise the outputs there will always be some error associated with the measurement method that will add noise to your results.
I depends what you want to measure, what your age range is etc. i would use one method only and stick to it, 24 MPR is a good method but time consuming (from a research point of view) and should be measured over 3 days 2 weekdays on weekend. You need to make sure the MPR has been validated for your population so you know the bias to expect. I wouldn't use an online tool unless you know it has been validated properly. I think you would find it very hard to publish if you had used multiple methods in your study.
P.s. We are running an advanced course on dietary methods in May
http://www.abdn.ac.uk/acdam/
Although the diary method is more accurate, the 24 hour recall method generates good and useful data if the sample is adequate. I think the data collected by the two methods can be merged for use. However, if you wish to compare intake of the respondents of the two groups, it may not work. If you are in doubt, or anticipate query from the reviewer, you may also use only 24 h recall data for all. Your 'Diary Method' subjects will probably agree to give 24 h recall data instead of a intake diary. Weakness of the 24h recall method is that it relies on subject's memory. Hence, it should not be used in subjects with poor memory or poor ability to convey accurate estimate of portion sizes (e.g., old, sick,children). Another point is if you do it on telephone, you may not be able to use the conventional recall aids to assist the respondents used in face to face interviews. You may need to find a way to offset it. If you plan to work in urban Britons, you should keep in mind the day-of-the week effect, and, hence, holidays should be proportinately represented for a good average intake picture. If you wish to look into any correlation or do regressinon analysis, or categorize your subjects on the basis of intake, you should do multiple 24h recalls/diaries on the same subject to get an estimate of his/her usual intake. Food frequency is not a substitute for either of your methods. Both methods are vulnerable to willful distortion of reprt by the respondents and flat slope syndrome.
I think sending diagrams for food models could offset the conventional research aids used in the face-to-face dietary recall. It is true that results from multiple recalls for the same subjects would be a better way than to use a single or double recall. Yes, the diary recall subjects would probably agree to doing a 24 hr recall or several of them as opposed to the diary method.
I think 24 h recall was designed for face to face interview, be careful to use it in other ways!.
A couple of issues. First, regarding number of recalls and sample size: One recall will give you a group mean that you can use to compare, for example, intervention and control. Sample size is not an issue in that regard. For that purpose, one recall is as good as many. The issue of number of recalls arises if you want more than a group mean, and instead you want some information at the individual level. Because people's intake varies from day to day, one recall does a poor job of estimating an individual's long-term usual intake. For that, you need multiple recalls, or a food frequency questionnaire.
Of course, if you really do care exactly about the previous day's intake for some reason in your research, rather than usual intake, then one recall is fine.
The second issue, the question you asked. I agree with others that optimally you would use only a single method. Reviewers would chastise you otherwise. And diet records, unless subjects are carefully trained, will tend to be incomplete and thus produce not only lower estimates but poorer estimates. Mailed portion size pictures or diagrams can help with portion size estimation, for telephone assessment by recalls or food frequency questionnaires.
I agree with Vaccaro that using two different methods will introduce bias unless if you want to do a comparative study. Choose one method and stick to it. 24hr recall is the best for food intake measurements, however, it should not be done once....use at least two weekdays and one weekend day giving a total of 3 days (minimum).
Why two techniques for the same study? It you alreday have two sets of data for the same population, try to compare each other for agreement with your hypothesis, what of them fits better?.
Hi,
Both dietary assessment techniques are considered as "gold standard" methods in terms of estimating daily intake (energy and macro-nutrients), despite their limitations (e.g. recall bias, mis-estimation of consumed quantity). Of course multiple records are required to accurately capture this information. To conclude, I think energy intake and/ or macro-nutrient intake will be adequately measured with either tool. The main problem is that the diary will be available on-line thus the population that will have access to this will be different (especially younger) to the one that will have to fill in the 24HR. So you need to pay special attention at the recruitment and sampling procedure.
Food diaries are more reliable compared to 24 hour recall. You can mix but you have to mention,that the differences in the intakes by two methods were not significant
When you are using 2 methods to assess the same outcome.......you most likely want to compare which is more appropriate or better than the other. 24HR in itself is tedious, time consuming & expensive, so why would you want to combine it with another one. The participant will definitely be fatigued (remembering all the foods & measurements). Remember that in the 24hr, it depends on the number of nutrients you are measuring......the more the nutrients the more the repeats. How is any of these 2 going to improve your results?
I agree with previous answers. If you look for recommended methodology I would recommend you the one prepared by EFSA for EU countries.
Repeated 24h recall (not correlated days) for adults and record for very young children has been recommended.
Combination with food frequency/propensity questionnares give you a chance to recognize true consumers/non-consumers which is important for some models of intakes.
There is a growing debate over the validity and value of data derived from memory-based dietary assessment methods (M-BMs) such as FFQs and 24HRs. My empirically supported position is that all forms of M-BMs are pseudo-scientific, produce physiologically implausible data, and should not be used in nutrition research.
Please see the recent 'Controversy & Debate' Series in the Journal of Clinical Epidemiology for details:
Archer E, Marlow ML, Lavie CJ. The Fatal Flaws of Food Frequency Questionnaires and Other Memory-Based Dietary Assessment Methods. Journal of Clinical Epidemiology. 2018.
Link: Article Controversy and Debate: Memory based Methods Paper 1: The Fa...
Archer E, Marlow ML, Lavie CJ. Controversy and Debate: Memory-Based Dietary Assessment Methods Paper #3. Journal of Clinical Epidemiology. 2018.
Link: Article Controversy and Debate: Memory-Based Dietary Assessment Meth...
- Badges
- Science topic