Week 6: Lecture 27: Food Frequency Questionnaires (FFQ)

Hello and welcome dear learners to the NPL course on nutritional epidemology a way towards a healthy life. In the module three for dietary assessment, nutritional status and measurement error for the chapter 27, we'll be taking discussion on food frequency questionnaires. It is being even called as FFQs. So we'll try to understand the structure of FFQs, the development and cultural adaptation of food frequency questionnaires, the components, uses, validity and reproducibility of the FFQs, the strength and limitations of the FFQs will be discussed followed by case discussion. So what is food frequency questionnaires or FFQ? has FFQs are the most commonly used dietary assessment instruments in epidemological research. They are being designed to assess the habitual diet by asking about the frequency with which food items or specific food groups are consumed over a specific period of time. So if you understand the major objective of FFQ is to assess the diet over long period of time making them more particularly valuable for research on epidemology and non-communicable disease. So there is an association of this with NCDS. Chronic diseases like cardiovascular diseases, cancer and diabetes develop over decades an individual's long-term average diet is generally more relevant to the disease ideology that what they ate yesterday. So you can understand the estimate relation of chronic diseases in research particularly for cardiovascular diseases, cancer and diabetes. They develop over a period of time. In FFQs, they essentially ask respondents about over the past year, how often on average did you consume this specific food? For example, how often did you consume apple over the past years? As you all know, an apple a day keeps doctor away. This is something which is there as being taken some sort of an assumption. So if someone says that one or three times in a month or one time in a week or two to four times in a week or five to six in a week or once daily or even two or more in a day. So it is like a habit of consuming something which is high rich in antioxidants, higher rich in iron as a fruit as something which is being seen as more appetizer. So by aggregating these responses the researcher can get a calculation about intake of food and nutrients over the long term. The researcher can rank the individuals based on the nutrients and the food intake levels where it can be like a pyramid food that you are making it up. Also, it identifies the links in association with diet and the disease condition which can give you the health outcomes. So where it is highlighted to states that the FFQs has a valuable for research on epidemology and non-communicable diseases. Nutritional epidemology as is an association between diet and diseases cardiovascular diseases cancer diabetes obesity and maternal and child health outcomes. public health in association with this as in monitoring dietary trends of evaluating interventions and also risk assessment. The larger cohort studies you can get with examples of the nurses health study which we have stated in the previous sections also health professional follow-up study on their health conditions and a study on European prospective investigation into cancer and nutrition. The positions of FFQs among dietary assessment method. So there is a position of FFQS in the 24 hours recall the food records and you may have a sort of an dietary history where you can put the association if you take the time frame 24-hour recalls take of previous day 24 hours food record is of several days FFQ is for months and years and decade dietary history is of months if you take the burden 24-hour recall it is very low because the time period food record it is very high and FFQ is extremely low because of a questionnaire and dietary history is extremely high as it takes more time because of buns. If you consider the cost 24 hours recall the cost is moderate. In the food record it is very high because you take it for months days. FFQ it is low comparatively because of the administration the mode of administration and in diet history it is very high. What is the habitual intake? You can understand it is very poor food record it is moderate FFQ and diet history is excellent. So you can understand the assessment of dietary assessment method FFQ stands among the position to give you more and moderate position. What is the structure of an FFQ? It has a qualitative and a semi-quantitative FFQ. So in a qualitative you can understand there are a lot of questions actually which gives us more reason as an individual to react more subjectively. The qualitative FQ they focus exclusively on the frequency of consumption. It is like how often it is being taken like how often in times in a week you have ate apple as an example. But do not ask about the amount eaten per person. How many times is like an proportion. The advantages is like simpler, faster to complete and lower respondent burden. Limitation is like lacks proportion. Proportion is something which is missing over here. So cannot accurately estimate the total calories and the nutrient which is to be finded as an constraint which is being taken in this in the semiquantitative FFQs they ask about the frequency and proportion size which is missed in the first part. So these ask actually both frequency and proportion size. Frequency which has been taken in the first part and added of the proportion size. So they do this earlier by providing a standard reference portion or by asking respondent to select their typical portion size. So you can say never or one or three times per month or one or two times per week adding with what is your usual proportion like you may say small, medium, large, too large. So the both the things are covered over here. What is the advantage? It provides data on both frequency and proportion size which was missing in the previous case. It also allows the estimation of total energy and nutrition intake and this is a very standard in modern research. The nutrition calculation is done with respect to food frequency, portion size and nutrition composition. So this allows the estimation of energy, protein, fat, vitamins and minerals. You can understand the food frequency. You can understand the portion size and the nutrient composition taking the energy, protein, fat, carbohydrate, calcium, iron and other micronutrient intake. The estimation is like energy is calculation proportion of protein in terms of grams, fats in terms of grams, carbohydrates in grams, calcium in milligram and iron in millig. So estimation is also there, proportion is also there, timing is also there and the intake actually which is taking the whole lot of formula to be completed for calculating the nutrition. Now we will try to understand the development and cultural adaptation of food frequency questionnaires. An FFQ cannot simply be translated into new languages and it is used in different countries. So it must undergo the rigorous cultural adaptation. It cannot be simply be a translation which is done with the help of only changing the words and languages. No, it is not. The diets are highly specific to geography. The diet are very specific to the culture, soio economic status and even on festives and cultural needs. So what is the step? The step is define the purpose. For example, maternal nutrition, child feeding, diabetes research, cancer research, AIDS research, cardiovascular disease. So on the purpose, the step one is being completed. Generate a food list. Out of all the total whether it is a 24-hour recall, food records, national nutrition survey, market survey, expert consultation, the goal should be food contributing most for energy, macronutrients and micronutrients. So there should be a purpose actually to generate a food list which is very important and this solves the major reasons to conduct it. Then comes the group similar foods. There are many foods which are similar in patterns taking fruits like apple, banana, mango or papaya. They can be grouped in tropical fruits, citrus fruits, sweet fruits and all those. Then comes the step four that is select the frequency categories. Example you may say never taken monthly taken weekly I can take daily I can say or twice a day twice a week. So need enough categories to capture the variation to make it more frequency based then comes the very important part that is proportion. proportion in terms of size whether it is a household measures that you are using or you're using some calibrated cups and all those or food photographs or models or standard servings. There are two things which comes not as unconstraint but as considerations to be taken as that is cultural adaptation and adaptation process. So in cultural adaptation an FFQ developed in one country may not work elsewhere. FFQ must be adapted and validated from country to country because food list are culturally specific. In different countries they are very different. Different regions have a very unique dietary patterns and food availability even within a country. I mean to say even in country like India in regions and in states you may have a very different dietary pattern according to the availability also it changes the diet. A single FFQ cannot capture a dietary intake or sometimes if you even try to make it as a single FFQ which can capture a dietary intake across multiple regions when properly validated. So there are two variations which may comes that sometimes validation can make it more adjustable versus if the validation is not done it cannot really capture. Western foods are more like a examples like beon or blueberries or bangales or Indian foods are more specific with the chapati, upma, raggi, dsha or idli where the cultural adaptation ensures that the FFQ is relevant, understandable and appropriate for the target population that you are trying to assess. It also leads to the accurate assessment of the dietary intake which is the purpose of the food frequency questionnaire. Now we will try to understand what are the components of food frequency questionnaire. In components there are three divisions that are being made that is food list. Second is the proportion size questions and the frequency categories. The food list is something which is very important here to be understood where usually it contains more than 80 to 100 food items or nearly you can say it takes more than uh closer to 200 food items. The food list should capture common foods, seasonal foods, nutrient-dense food or even fish items and all those according to the categories and the region specific food availability or staple diets. The examples can be cereals, pulses, millets, vegetables, fruits, milk products, fish, eggs, meat, pork, snacks, beverages and so on. In terms of proportion size food for the FFQ, you can estimate the amount usually in terms of the consumption. You can say for the methods a standard proportion is one serving assumed for everyone which usually many times the Indian family follows where the FFQ the proportion is being given in the questionnaire multiple portion categories you may say small medium large to large and all that can state that there's a multiple portion categories which is being used open-ended portions where you can say number of cups or spoons or grams which is somewhere or the other you move on to the complexity where you have to derive according to the proportion in terms of measurement you have to state this. The third and very important part is the frequency categories. These frequencies are converted into daily intake estimates. Category wise stating that never is equal to zero or 1 to 3 months you can say 07 per day 1 week4 per day 2 to 4 weeks 43 per day or daily that is one per day or within 2 days you can say two per day. So what what does it mean by there should be some sort of categorization and the frequency can be made according to that. So these frequencies are converted into daily intake estimation. Frequency multiplied by portion size multiplied by nutrient composition. What does it gives? It gives equal to daily intake estimate. The daily intake estimate is the purpose of understanding the food frequency questionnaire. Now the use of food frequency question where it is being used. It is used for diet pattern score. A diet pattern score which is very important where the foods are consumed together. Interaction occurs between the nutrient foods and cooking methods. There are three trajectories the food has to go through. One is the nutrient what it is being there. What is the sort of food that is to be prepared where you can say the meal or what are the cooking methods which makes one meal different from the other meal might be a rice which is being prepared in the morning might be very different for the rice which is being prepared in the evening or at night. What does it makes the traditionality for the nutritionbased food approaches focusing on individuals nutrition needs in modern epidemology? It is increasing the studying these dietary pattern which makes a variation to understand the health outcomes. There is a dash score over here. There is a meditian diet score and a plant-based diet index. So we'll try to understand in detail what those are. So you can say the dashboard score is an assessment aligned with the dietary approaches to stop the hypertension rewarding high intakes of fruits, vegetables and low-fat diets and penalizing high sodium and saturated fats. You can understand there is a very noble purpose of taking this. So you have a higher score for fruits, you can have a higher score for veggies, whole grains and low fat dairy. You can say there's a boost. Higher dash scores indicates better alignment with heart healthy dietary guidelines and you may have your heart more good in functions. The lowest scores for sodium, red meat, sugary beverage and alcohol. So the dash scores are more giving proportion towards higher scores and lower scores for food which are indicating according to the health dietary guidelines. Now we'll try to understand a meditarian diet score. In Meditarian diet score it assigns points for above median consumption of the protective food. Protective food like which are lowering the inflammation olive oil, legumes, fish, cesum oil and also below median consumption of detrimental foods. Detrimental foods which are more something which can say increase the inflammation, lowers the immunity or increase the risk of cardiovascular disease condition. Red meat and sweets particularly. So you can say higher intake is necessary from olive oil to fruits, veggies, legumes and fish. Moderate intake of dietary or higher fatty dairy. Lower intake of red meat. So higher Mediterranean diet score are associated with lower risk of chronic diseases. Then comes the plant-based diet mix that is PDI. The differentiations between a healthful plant-based diet that is whole grains, nuts and unhealthy full plant-based diet that is refined grains, sugary beverages. So you can say the measure ad to healthy plant food, whole grains, fruits, veggies and legumes and lower score goes on animal foods, refined plant foods and sugar beverages. So a higher PDS score reflects a healthier plant-based dietary patterns. So why FFQs are ideal for pattern analysis? The reason is long-term intake. It captures diet over months to years reflecting a very habitual behavior which is a benefit of taking this. Then comes the usual habit. It reduces the day-to-day variation and random errors. The food combination gives more inclusions of a wide variety of foods and meal eaten together and no doubt making them highly suitable for dietary pattern research. So dietary pattern research is somewhere or the other the ideal way to be dealt by FFQs. The validity and reproducibility of FFQs. A dietary instrument must be valid and reliable. These are the two important things which is very important to be stated to be of remembrance if you are trying to take FFQs which makes dietary instrument a very standard way of dealing with the challenges that overcomes to these FFQs if they are validated and reliable. No doubt you may get a very standard health outcome. The validity is based on the ability that the FFQ measure what it is intended to do. It does the FFQ actually measure what the person eats. So actualness is very important rather than assumption. The common validation statistics like Pearson correlation which is very important to be taken as an validification method, Pearson correlation, deatuated correlation, bland analysis and crossclassification. The typical correlation comes like energy, protein, fat, fruits and vegetable in terms of the correlations which are being stated on the right side. Reproducibility, the ability to obtain similar results when the FFQ is reported. Example, administering the FFQ today and again after 1 month. So there's a sort of an comparison actually which is being stated in terms of the similarity of the results statistical measures intraclass correlation coefficient IC Pearson correlation and kapa statistics because of the repetition you have to do this level examples of validated FFQs you may see the examples being stated the first is NCI diet history questionnaire that is DHQ. It is being taken in US National Cancer Institute. The characteristics are developed by the US National Cancer Institute and it is one of the most rigorously validated FFQs. It consists of 80 to 120 food for a total diet assessment which is very large. The purpose of the DHQ is to obtain frequency and portion size information over specified period that is past months or years. The second examples of validated FFQ comes as Harvard will vlet FFQs. It is taken by Harvard Thchan School of Public Health. The characteristics are it is widely used in the famous nurses health study and it has been popularized with this study. Then comes block FFQs developed by Glatty block. It is heavily focused on accurately capturing the nutrient intakes based on NHES data. Then comes a study which is being stated that it captures from India that is semi-quantitative FFQs. The characteristics are showed acceptable validity across multiple regions. As you can understand the constraints which comes into different Indian regions actually to give it for the FFQs. Other India specific FFQs can be region specific, population specific. It can also include the local foods, mixed dishes, seasonal and traditional items extensively validated using recalls and biomarkers. What are the strength of FAQs? Long-term intake assessment which is already stated. It has a seasonal variation, habitual diet, usual eating patterns are also being taken. Cost effective estimated to hundreds to thousands and even over a mail and online. So there is a massive prospective cohort studies also which are making it economically feasible. Low participant burden where it is a reduction of time from 30 to 45 minutes where a quick completion time versus weighing and recording food is being something as a constraint which comes to the respondents. Excellent for ranking you can say it comes more than the moderate intake to higher intake on options to take for a small, medium and large in terms of proportion. You can say what are the higher quantiles versus to the bottom quantiles accurately. It classifies as the individual in groups. So it is stated to be very useful for calculating the disease risk ratio also self-administered which gives us a benefit of getting to a literate population and it also minimizes the data collection expenses which is very important constraints where we talk about the economy. Taking the limitations social desiraability we are always over repeating is more for healthy food as you can understand and see the picture unhealthy foods usually under reportporting when you are taking alcohol or beverages sweeteners you really don't report recall bias usually it is a strain on the memory to give the participants may not remember accurately for the food sometimes they have missed in terms of a small proportion or small collect proportion size is something which is difficult to estimate food which is fixed in the list are only the one and the missing which are not there are something which seasonal for example strawberries or even we can say dragon fruit which are not listed over there are being missed nutrient estimation like proportion assumption food consumption tables and estimated nutrition there might be generic databases or generic data size there can be some sort of an errors or sometimes it may be missed In terms of the calculation measurement error consequence what it is it is something as a diet disease association where the weaker apparent association is being stated where observed weak relationship is being underestimated. So the attunated versus weaker apparent association is something which estimates or give a true relationship. Now there is a case distinction uh case discussion which I want to take ahead. So this is the India specific food frequency questionnaire. So India presents unique challenges for dietary assessment because of its enormous diversity in terms of geography, culture, religion, a soioeconomic patterns, food availability and cooking practices. So you can say a single universal FFQ really captures dietary intake accurately across all Indian population. Most Indian FFQs are region specific. They are population specific or they undergo extensive validation. Might be it's a time-consuming job. They have to take step one as dietary survey where the researcher conduct it in a multiple 24-hour recall taking in consideration all the variations. Step two selection of the food items which are there on the macronutrient size for the micronutrients. Understanding the energy, protein, fat, carboids in the macro. Grouping the categories where you can say the major categories are four cereals, pulses, vegetables and fruits. Step four towards the proportion size where you can say in terms where Indians usually come up with rice taking in form of a proportion for bowl or kori or dal in terms of kori or milk in terms of glass or chapati in numbers. So number versus the usual measurement part. The typical structure of an Indian FFQ is a comprehensive Indian FFQs may contain 100 or more than that or closer to 50 food items which can be grouped from cereals, meallets, pulses, legumes, veggies, fruits, dairy products, fish, eggs, meat, oils, snacks, sweets and beverages. The important FFQs which are being taken in India are the National Institute Nutrition FQS which is developed by ICMR and National Institute of Nutrition. The characteristics were adapted to Indian foods used in nutrition studies and it includes regional dietary variations. The application was in nutrition surveillance and public health research. Very important. The CARES FFQ which is developed by center for cardio metabolic risk reduction in South Asia which is very uh approximately covering 160 to 200 food items capture intakes of across India, Pakistan and neighboring countries validated against a very repeated recalls and used for diabetes, obesity and cardiovascular disease. The EP new India FFQ that is developed by nutritional epidemological studies and it is very region specific designed to estimate and long-term in intake including mixed dishes common in India which is very very difficult actually the Kerala specific FFQ is specially designed to understand the coconut based food preparations the fish consumptions and traditional foods very interesting to understand this which is useful for cardio iovascular disease studies which are very common in Kerala and metabolic syndrome research which is very important to take ahead the responses of understanding the hormones. Urban Indian FFQ is designed for metropolitan food and nutrition transition studies which are understanding the burden of food over here which includes food which is fast food, processed food, sugar and sweetened beverages in additional to the traditional additional foods. The special challenges in FFQs in India are particularly mixed dish and it gives a variations seasonal foods which are more uh clubbed with lot of consumptions and festive foods which gives lot of variations in terms of the annual intake and contribution. Regional diversity is something which is a complexity and sometimes it has been seen in many places from southern states to even northern states. Therefore many FFQs in India the core nutritional food list and the region specific leads are modulated. So with this we come an end to this chapter. In the summary I can state that FFQs are widely used a dietary assessment tool which are designed to measure an individual's usual food and nutrient intake over a extended period typically months and years. They collect information on how food often the specific food are consumed in specific semicontitive versions and the usual portion size. FFQs are particularly valuable in nutritional epidemology because they capture a habitual dietary pattern relevant to chronic disase development. They are very cost effective, easier to administer in the larger population using for ranking individuals according to the dietary intake. However, the FFQs rely on memory and self-reporting. There comes sort of an individual bias making them susceptible to recall errors and systematic biases which we have studied in the previous sections. Despite this limitations, FFQ remains one of the most common used method for investigating the diet disease relationship which is the purpose of nutritional epidemology in the population based research. So with this we come and end this chapter. These are few important references which you can take and help for. Thank you. Thank you very much. [music] [bell] [music] >> [music]