Week 10 : Lecture 46:Diet Quality Indices and Their Applications in Population Health Research

Hello and welcome dear learners to the NPL course on nutritional epidemology a way towards a healthy life for the module 4 diet nutrients dietary patterns and disease epidemology for the chapter 46 we'll be talking about diet quality indices and their applications in population health research. In this topic particularly we will be discussing about diet quality assessment deconstructing the core global indices construction and interpretation of indices applications in cohort studies and applications in research and policy. The concept of diet quality assessment why measure it? For decades you have been seen that nutritional epidemology focused on identifying a single nutrient deficiencies for example vitamin C and scurvy vitamin D and ricketetts. However in the modern populations chronic non-communicable diseases have replaced acute deficiency states as the primary public health threat. Single nutrient reductionism face to adequately model these conditions because human diet consists of high complex intercorrelated combinations of food. The evolution of nutrition analysis is like single nutrient reductionism and diet quality indices. In single nutrient reductionism we find the focus is more on deficiency diseases. metric is isolated daily milligrams where versus in the diet quality indices we have the focus more on chronic non-communicable disease prevention. The metric is cumulative food matrices. The diet quality indices solve this analytical problem by functioning as a standardized evidence-based scoring algorithms that convert a participant's entire dietary profile into a single numerical value. So what it is measuring the diet quality is essential because it accounts for food metric effects. The reality that nutrients wrapped inside a whole food exhibits a complex biological synergies or you can say the antagonism that alter their absorption, metabolic pathways and systemic inflammatory property. By evaluating the overall architecture of a diet, the diet quality indices allow public health researchers to quantify cumulative lifestyle exposures and assess how closely a population's behavior aligns with established nutritional science. So where we try to understand the food matrix effect. Now in the deconstructing the core global indices the healthy eating index that is HI. The healthy eating index is a guidelinedriven validation tool designed to measure a population's address to the official dietary guidelines for Americans. The index is updated every 5 years by the USDA and National Cancer Institute to reflect changes in federal policy with health healthy eating index 2020 serving as the current reference standard. The healthy eating index operates on a scale of 100 points split across 30 distinct components. It assesses the dietary quality rather than quantity by using a density based scoring system typically per 1,000 calories. The components are divided into two distinct group. The first group is adequacy component that is having a nine categories. Whole fruits, total fruits, whole grains, total vegetables, greens and beans, daily total protein foods, seafoods, plant proteins and fatty acid ratio which are expressed in PUFFA and MUFA saturated fatty acids and higher consumption in these category awards a higher score capped at five or 10 points for the components. The second category is moderation components which has four subcategories that is refined grains, sodium, added sugar and saturated fats. For these categories, higher consumption lowers the score penalizing the diets that exceeds the federal threshold. So where you can understand from the components how it is being derived. Now we will go for the Meditaran diet score that is called as MDS. In the Meditaran diet score, it evaluates the address to traditional meditarian eating pattern. It is widely used in epidemological studies to evaluate the diet's proactive effects against the systemic inflammation, cardiovascular decline, and cellular aging. The traditional index often referred to as triricopolio MDS using a simple 0 to9 point absolute scale based on nine core components. What are those nine core components? Vegetables, legumes, fruits, nuts, whole grains, fish, red and processed meats, dairy and alcohol, and the monossaturated to saturated fatty acid ratio. The median cutoff scoring mechanism. In the median cutoff scoring mechanism, we use participants intake. We compare it and then we put it to the study population median where it is further being seen with protective foods and non-protective foods. The key mathematical feature of the MDS is its median cutoff mechanism. Unlike the HI which uses the fixed absolute point targets, the MDSA scores participants relative to actual consumption of the population being studied. For the protective food groups, a participant receives one point of its daily consumption, which sits above the study's population's median and zero points if it sits below. So for non-protective food groups like red meat and dairy the scoring is inverted. One point is awarded if it takes its below the median and for alcohol one point is awarded only if the consumption falls within a specific moderate window. Typically that is 10 to 50 g per day for a man. Now we'll try to see what is dash dietary score. DASH dietary score is dietary approach to stop the hypertension. It is more an initiative to be taken towards seeing how hypertension should be controlled. The DASH diet score was developed to track eating pattern that lowers blood pressure and it protects the cardiovascular events. While other indexes focus heavily on broad food groups, the DASH core targets the specific mineral matrix which is known to regulate the blood pressure. It rewards the diet high in potassium. It gives more in magnesium and calcium while penalizing excess sodium intake. So you can understand over here what is the target. The scoring system evaluates eight specific food categories. So participant receives maximum points for high intakes of fruits, vegetables, nuts, legumes, whole grains and low fatty diet. Conversely, they are penalized for high consumption of sodium, sugarsw sweetened beverages and red or processed meat. It's very very clear to be seen over here. In population studies, the DASH score is highly effective at predicting the hypertensive disase pathway. stroke index and early markers of chronic renal failure or kidney failure. The alternate healthy eating index that is called as AHI. The alternate healthy eating index was developed by researchers at Harvard Thchan School of Public Health to address the structural limitations in early federal guidelines. The original HI treated all fats and proteins similarly, which meant a diet high in processed meat and transfer could still receive a high score if it meets the basic macronutrient targets. The H AI was designed to fix this by focusing on the specific dietary drivers of chronic NCDS. The AHI operates on a 110 point scale across 11 components which are introduced with a higher biochemical precision. Fat quality differentiation. The AHI reward the complete elimination of trans fats and scores polysaturated fatty acids that is PUFAS relative to saturated fats reflecting their differing impacts on LDL cholesterol particles. protein source selection. Rather than using a generic total protein category, the AHI penalizes processed red meats due to their links to colurectal cancer and cardiovascular disease inflammation while rewarding intake of nuts, legumes, fish and poultry. Grain refinement filters. So it isolates whole grains from refined carbohydrate to track the impact on insulin resistance and glycemic load. Moderate alcohol inclusion. It awards points for consistent moderate alcohol intake that is 0.5 to two drinks per day reflecting the J-shaped curve observed in cardiovascular mortality data. Now we'll try to understand the construction and interpretation of the indicasis. Stepby-step construction framework. Step one, we see that dietary assessment input which is being seen that the research process begins by collecting a raw food intake logs from participant. This is a typical done using a validated food frequency questionnaire that is FFQ multiple 24-hour dietary recalls or weighed food records. In the step two, we will see the food group conversion. Here the absolute way of the consumed food are extracted from these logs. Complex meals are broken down into their individual ingredients. Example, separating a meat pizza into refined flour or cheese, tomato sauce and processed meat using a national nutrient database. Here the database is national nutrient databases. Now the step three is energy standardization. To isolate dietary quality from total protein sizes, raw food ways must be standardized to an energy adjusted format. Most indexes apply a standard density model. Okay. So dietary component density is equal to grams of target food groups divided by 1,000 total daily kilo calories. Now we'll see step four that is point allocation and summation. What it is? The standardized density values are evaluated against the index specific scoring criteria. Points are assigned to each component and summed to generate the final global DQI score. Methodological interpretation thresholds. Once the continuous scores are calculated, epidemologists interpret them using two primary statistical approaches. Categorical classification scores are grouped into fixed health brackets. For example, an HI score above 80 indicates a good diet. That is high compliance and 51 to 80 signals a diet that needs improvement. Okay? And 50 or below reflects a poor diet. So you can understand over here how the categorization is done. Now there is a quantile segmentation. In prospective cohort studies, researchers rank the continuous score of the entire population and divide them into equal tiers. Typically tiles that is three groups, cile four groups and quintile satisfy groups. This allows the statisticians to directly compare the disease risk of the highest adurance group against the lowest adurance group. That is called as the reference category. Then we will see the applications in cohort studies. You know all the prospective cohort studies from the bedrock of nutritional epidemology because researchers cannot easily randomize thousands of people to follow specific diets for decades. They must track free living population over long periods using the DQIs to transform the complex dietary histories to clear exposure data. So what is a prospective cohort epidemological pipeline? It is followed by baseline dietary assessment that is FFQs. Then followed by converting it into DQI contiles that is 1 to5 and then later in 20 years followup track the HR for the NCDS. Okay. So here you can understand how it uh fragments over there to go to the next step. Exposure modeling and risk prediction. In a typical cohort study, a individual's DQS score serves as the primary exposure variable and statisticians use a multivariate COX proportional hazard regression models to calculate the hazardous ratio for a specific clinical endpoints by adjusting for non-dietary confounders such as smoking status, physical activity, soio economic bracket, genetic history. researchers can isolate the independent effect of overall diet quality on disease development. Okay. So it's very clear over here how we try to take the applications in cohort studies which makes us more important for to get the clarity for this subject. Now tracking dietary trajectories over time. So advanced cohort studies do not rely too slowly on a single baseline dietary assessment. Eating habits change over time due to aging, medical diagnosis and socioeconomic shifts by collecting dietary data at repeated interval that is every four years in a nurses health study that you have seen and researchers can calculate change score. This allow epidemologist to answer critical longitudinal public health questions. If a participant improves their HI score by 10 points over a decade, does their subsequent risk of developing cardiovascular disease drop? So this can be something a very interesting question. Now mitigating the measurement error. So dietary self-reporting is vulnerable to systemic measurement errors. It's very clear such a memory recall gaps and social disability bias which we have studied in pre previous chapter where participants overall report vegetable intake and underreport alcohol or sugar consumption. So the DQI help mitigate these errors in two ways. The first is error buffering through aggregation. By grouping individual foods into broad categories, small reporting errors on specific items. Example, confusing broccoli with spinach average out with the larger component score. And the second way is deep biasing via calibration studies. So researchers often validate large scale FFQs data by conducting a smaller detail calibration studies using objective biochemical biomarkers like urinary sodium or plasma kerotenides adjusting the final cohort models to account for self-reporting bias. Okay. So it this is very very clearly understood over here. We'll go for applications in research and policy landmark cohorts. Try to concentrate and understand that the predictive validity of diet quality indices have been confirmed and formally cited across several major global population studies and clinical trials. What are those? In the cohort study clinical trial, the very famous nurses health study NHS and HPFS that is Q's HL 2012 and the DQI metric evaluated with alternative healthy eating index that is AHI which observed the clinical and population healthy endpoints which gives the long-term prospective data showed that participants in the high quintil of AHI scores experienced a 20% of reduction in the cardiovascular disease. Very important and a 28% dropped in all cause mortality. Okay, this is very finely being seen over here which compared to those in the lowest quintile. So the comparison over here is very important to be understood. Now the cohort study clinical the DQI metric evaluated and observed clinical and population health endpoints followed with the predime trial that is byl. The meditarian diet score that is MDS which we have studied in this section. This landmark multic-entered randomized control trial confirmed that high adrance to a Meditaran dietary pattern uh directly caused a 30% reduction in the major adverse cardiovascular events. So Meditaran diet is connected with cardiovascular events that is stroke, mioardial inffection and cardiovascular death. The multi-ethnic study of etherox sclerosis that is messa which is by hi 2015 and dash core which were evaluated it gives an observation of the multi-ethnic cohort demonstrated that higher diet quality scores consistently predicted significant reduction in subclinical etherosclerosis measured via coronary artery classification scans and lowered the long-term hazard ratio for heart failure. and the risk across the diverse black, white, Hispanic and Chinese American demographics very importantly to be understood. Okay. So you can find the correlation between these diets and the studies which are mentioned. Now we'll see the public health policy translation. The diet quality indices serve as a vital link between observational data and public policy. There is a state correlation turning them into complex research into practical population level interventions. The DQI policy translation engine first is observational research. The cohort data shows low HCI scores are associated with diabetic risk. Clinical translation specific components example added sugar are isolated as drivers and population policy national guidelines set firm actionable programmatic thresholds. Okay. So it's very clear to be understood over here. Try to correlate the previous part and you can revise to get more clarity on this. Now we'll be taking the programmatic auditing. What is programmatic auditing? When the government uses DQIS to measure the nutritional impact of public food assistance program, for example, tracking the HI scores of participants in federal nutrition assistance program like SNAP and WIC, which helps the agencies evaluate whether these multi-billion dollar support programs successfully improve the overall diet quality among the vulnerables and low economic population. It is very important for the policy to see these gaps. Now we have to see the national guidelines formulation where the DQI is allowed to public health bodies to shift their communication strategy. How to shift that? So instead of issuing a complex single nutrient target that confuse the public. Okay. So like example limit the saturated fats to less than 10% of the total daily energy. Agencies can issue a clear very food based guidelines. Example, adopting a dash eating pattern which can be directly stated to the public. And what is the uh pattern to be followed? Increasing the whole fruit, vegetables and low fatty dairy while restricting on sodium. Okay, which is very very important. So without policy people will not be able to understand it. So national guidelines formulation becomes very important for this part. What are the key takeaways from this chapter? The key takeaways are the food matrix focus. The diet quality indices move nutritional epidemology away from the single nutrient reductionism offering a standardized method to study the health impacts of entire food matrices. A guideline versus relative scoring. The HI measures absolute compliance with national targets while the Mediterranean diet score evaluates the adurance relative to population specific medians. Okay. And the predictive validity long-term prospective cohort studies confirm that higher scores across all the major indexes consistently predict lower rates of cardiovascular disease. type two diabetes and all cause mortality where the predictive validity is more being seen over here. Evidence-based policy that DQIs translate complex nutritional data into clear actionable metrics. This helps the governments evaluate the food assistance program and design a practical food-based public health guidelines. Okay. So, we end over here. These are very important references that you can follow and revise the chapter to see with your project's alignment. Thank you very much. [music] >> [bell] [music]