Question 2
The participants reported their dietary intake (3 month recall of their food intake patterns) at baseline and again at year 3 of the study using a food frequency questionnaire (FFQ). It is well known that error is commonly associated with the use of FFQ due to self-reported data and misinterpretation of portion sizes/food consumed, however measurement error is typically associated with total energy intake and macronutrient intake. Do you think that the results of this study are accurate since the FFQ was examining primarily micronutrients not macronutrients and energy?
posted by MelanieP @ 10:36 AM
29 comments
29 Comments:
I believe that the results of this study are accurate, but I also realize that there are limitations to any study. While a FFQ may not always be an ideal way to report dietary intake due to the fact that it is self-reported data, the sample size of this study was extremely large. I think it would have been unrealistic for the researchers to obtain this data any other way since the sample included 93,676 women.
I do not think that it makes the results of this study any less valid by using a FFQ. You are going to have some self reporting error with any study that uses FFQ. The best way to eliminate some of those errors is to educate on portion sizes before participants have to complete a FFQ. I also do not think the results are any less accurate because the focus was on micronutrients. In this study, the importance is on the micronutrients and it would be a lot more data to include or focus more on macronutrients.
I agree with Liz in that using FFQ is really only way to get data from such a large sample size.
Liz, I agree with your response 100%. It may be unrealistic to obtain data utilizing a different method, considering the size of this study.
This is always a question to the many studies that utilize FFQ. There may be more appropriate methods that provide better accuracy for some studies. Given the large sample size the researchers used for the present observation, I think FFQ provided the best possibly accuracy.
While educating the participants on details related to reporting their consumption to improve the accuracy of reporting information may be beneficial, it would have taken an extensive amount of time and labor to educate the large number of participants involved in this observation.
It was previously said that portion sizes can be misinterpreted. Therefore this leads me to believe that this would not only affect macronutrients, but also micronutrient intake information. Even with this said, I do believe that the food frequency questionnaire in this situation was the best choice given the number and age of participants. The FFQ is very efficient and also allows us to see a more extended intake versus only 1 to 7 days like in other methods for collecting food intake records.
I do not think that the results are inaccurate due to the utilization of a FFQ. Self-reported data can be questionable, but for this study, it would be difficult to assess intake in any other manner. The FFQ provided a quick and reliable method of assessing intake for the 93,676 participants. I think it would be next to impossible to assess intake in any other manner with this large of a sample size. It could be possible for future studies to utilize diet histories or food records if they focus on more specific populations with smaller sample sizes.
This is a very interesting question, and something I hadn't considered when first reading the article. FFQ, like any method, has limitations. It would be so interesting to conduct a study testing the variability in macro vs micronutrients. Ultimately, the same error for macronutrient measure will apply to micronutrient estimates.
Liz, Amy, and Stephanie, I agree that despite the limitations, with a population this large, the convenience of FFQ outweighs the limitations.
Great question! In my opinion, I think the same inconsistencies would go for both micro and macronutrients when using a FFQ. The participants are still reporting on foods they consumed in the past 3 months and just like with the macros and energy, there is going to be human error. I definitely think that the FFQ was the best way to collect information on this vast of a sample and that the results are just as accurate if the researchers were only looking for calories consumed. Micronutrients are definitely much harder to measure than macros and energy but the researchers were still looking at the same general concept of what are these women eating and how much.
I am sure that using an FFQ research design did give less acurate results of micronutrient intake, but with 93,000 participants using FFQ in my opinion was the best option. In all studies with dietary recalls there will be some error. Without having a dietitian sit down with each individual, show them food models, talk about serving sizes etc. I don't think error could have been reduced and obviously this is not a realistic approach for this sample size.
Sarah, I agree that although the same measuring error would exist, testing the variability in macro vs micronutrients would be interesting and valuable.
Although dietitians presently lack an ideal form of nutrient data collection, FFQ are considered the best option for collected nutrition information for a large population. I think the FFQ was the most appropriate option in this study considering the sample population size, but assume that the information collected is not going to be perfect. As with most research concerning food consumption, gathering accurate data is a challenge, and self-reported information can only provide a general understanding of the nutrients consumed by the population.
Amy, I agree with you. Although it would be wonderful if the researchers could educate all 93,676 women in the study, it just doesn't sound feasible or practical. Instead, we can take the results with a pinch of salt, understanding the exact numbers may not be perfect, but the ability of participants to accurately identify intake is hopefully similar throughout most groups.
I still believe that the tool used was possibly the best one in this case. Yes a FFQ is not perfect but with some foods being better sources of these micronutrient than others a FFQ would still be helpful, regardless of portion size. The reseacher would be able to see how often those nutrient rich foods were being consumed.
Liz brought up a good point that the FFQ was the only cost and time-efficient method for gathering data for the study. Since the general population's knowledge of micronutrient food sources is probably minimal, analyzing a general FFQ was still the best fit data collection method for this study.
I still think that the results of this study can be considered accurate. It would be far too expensive and inefficient to make sure a dietetics professional was with someone for 3 months for both recalls. Until there is a more accurate way to get a report of what people eat I think this was the best way to know what micro-nutrients were being consumed by participants.
I agree with Stephanie that the best way to reduce the amount of error is to educate participants on serving sizes before taking a FFQ. This may be difficult but it could possibly be done through a website created for all participants.
The researchers discussed that significant error is experienced when reporting energy and macronutrient results from a FFQ, but don't specify how much that error decreases when reporting on micronutrients. I think the results of micronutrient intake are accurate enough for the purpose of this study, but I also think that micronutrient data could be just as easily skewed by incorrect portion size reporting as energy or macronutrients.
Jamie, I love your idea of educating study participants on portion sizes through a website. A tool like that would be very cost-effective. I can see a lot of potential there!
I believe that the FFQ was the best data collection tool for this population size and accuracy is likely similar to a study using macronutrients. I would be interesting to test this data collection tool to determine if it is more or less accurate when studying micro and macro nutrients.
Michela,
you bring up a great point that it would be difficult to have an RD sit down with every participant and show food models to get more accurate serving sizes. This tool was excellent for the sample size.
Michela made a good point. There is a chance of error with any diet recall. In this case, FFQ was the best choice. The sample size was much too large for any other type of diet recall.
I liked what Anna said about assuming the information collected is not going to be perfect. I think it is sometimes hard to accept this, but this statement is very true for any method gathering dietary intakes. Each method has their imperfections. The point is that the researcher must decide which method best fits their own study. And in the case of this study, I believe the researchers did choose the appropriate dietary information collection method given their sample size and purpose.
I believe that this is generally accurate. There is always a margin of error with the food frequecy questionairre, however since this is focusing on a less misinterpreted area of recall.
Amy, I agree with you that there is often misinterpretation with food frequency questionairres. There are other recalls and reporting methods that may be useful in this situation, however I agree that in this instance and for this specific purpose that the FFQ is most appropriate.
I do believe the results of this study are accurate. With such large participation it would be difficult to collect data any other way. This study was looking at one specific aspect of the participants diet and the FFQ allows you to do that quickly but still effectively. There will be some error but I do not believe this reduces the credibility of the results.
I think that the results of this study should be accurate enough for the scope of this study. The results gathered should still be able to point out trends in the population even if they are not exact. The fact that measurement error is typically associated with total energy and macronutrient intake should merely be listed as a limitation.
I think that the results of this study are going to be more accurate as they were looking at micronutrients rather than macronutrients and energy. I still think there is going to be some error, which is unavoidable with self-reported data and in particular with a sample size of this size. However, using the FFQ was probably the only realistic way to determine the approximate dietary intake of these individuals.
As with most others, I agree that there is not another method that could realistically measure the dietary intake of 93,676 women more effectively and more efficiently than the FFQ. In an ideal world the participants would be able to receive education on portion sizes and what nutrients they need to be eating, but again with this sample that is unrealistic.
I think most everyone has touched on the overarching theme of human error in research and finding the best fit for gathering data. Once any study is done, there will be some limitations and suggestions for future research, but we have to consider how much, thought, time and effort that went into the methodology and completion of the present study. As we all know from conducting our own research, the procedure that this researcher chose was probably the most realistic and best fit for the situation.
Post a Comment
<< Home