Evidence-based priority setting for dietary policies. Ashkan Afshin, MD MPH MSc ScD November 17, 2016 Acting Assistant Professor of Global Health

Transcription

1 Evidence-based priority setting for dietary policies Ashkan Afshin, MD MPH MSc ScD November 17, 2016 Acting Assistant Professor of Global Health

2 1. Defining optimal nutrition Agenda Setting Policy Formulation 2. Quantifying the burden of disease due to malnutrition 3. Evaluating the effectiveness of policies to improve nutrition 4. Evaluating the cost-effectiveness of nutrition policies 5. Evaluating the political/legal feasibility of nutrition policies Policy Evaluation Policy Adoption 6. Evaluating the intensity of implementation of nutrition policies 7. Evaluating the short/long term effects of nutrition policies Policy Implementation

3 1. Defining optimal nutrition Agenda Setting Policy Formulation 2. Quantifying the burden of disease due to malnutrition 3. Evaluating the effectiveness of policies to improve nutrition 4. Evaluating the cost-effectiveness of nutrition policies 5. Evaluating the political/legal feasibility of nutrition policies Policy Evaluation Policy Adoption 6. Evaluating the intensity of implementation of nutrition policies 7. Evaluating the short/long term effects of nutrition policies Policy Implementation

4 Malnutrition 4

5 5

6

7 Defining diet Absolute Intake Nutrients Foods Dietary patterns Biological mechanism No biological knowledge Intercorrelations not a problem Statistical Power No food composition data needed Between-food interactions Supplementation Use in dietary advice Health outcome Disease endpoints (CVD, diabetes, cancer) Intermediate outcomes (obesity, blood pressure) Relative Intake 7

8 Evidence RCTs of disease endpoint Description Number of independent RCTs evaluating the effect of the risk on the disease endpoint % of independent RCTs showing significant effect in the opposite direction % of independent RCTs showing no effect Prospective observational studies of disease endpoint Number of independent prospective observational studies evaluating the association of the risk with the disease endpoint % of independent prospective observational studies with significant association in the opposite direction Strength Dose response Biologic plausibility Analogy Lower Limit of RR in observational studies> 1.5 (Yes/No) Evidence of the dose-response relationship between the risk and the outcome(yes/no) Potential biologic mechanism that could explain the effect of the risk on the disease endpoint (Yes/No) Evidence on the relationship between the risk factor and a disease endpoint from the same category (Yes/No) 8

9 Epidemiologic evidence supporting causality between dietary risk-outcome pairs Risk Outcome Diet low in fruits Lip and oral cavity cancer Diet low in fruits Nasopharynx cancer Diet low in fruits Other pharynx cancer Diet low in fruits Larynx cancer Diet low in fruits Oesophageal cancer Diet low in fruits Tracheal, bronchus, and lung cancer Diet low in fruits Ischaemic heart disease Diet low in fruits Ischaemic stroke Diet low in fruits Hemorrhagic stroke Diet low in fruits Diabetes mellitus Diet low in vegetables Oesophageal cancer Diet low in vegetables Ischaemic heart disease Diet low in vegetables Ischaemic stroke Diet low in vegetables Hemorrhagic stroke Diet low in whole grains Ischaemic heart disease Diet low in whole grains Ischemic stroke Diet low in whole grains Hemorrhagic stroke Diet low in whole grains Diabetes mellitus Diet low in nuts and seeds Ischaemic heart disease Diet low in nuts and seeds RCTs (Number) RCTs with significant effect in the opposite direction (%) RCTs with null findings (%) Prospective observational studies (Number) Diabetes mellitus Prospective observational studies with significant Lower limit of RR > 1.5 Dose-response relationship Biologic plausibility Analogy GBD 2015

10 Epidemiologic evidence supporting causality between dietary risk-outcome pairs Risk Outcome Diet low in milk Colon and rectum cancer cancer Diet high in red meats Colon and rectum cancer cancer Diet high in red meats Diabetes mellitus Diet high in processed meats Colon and rectum cancer cancer Diet high in processed meats Ischaemic heart disease Diet high in processed meats Diabetes mellitus Diet high in sugar sweetened beverages Body mass index Diet low in fibre Colon and rectum cancer cancer Diet low in fibre Ischaemic heart disease Diet low in calcium Diet low in seafood omega-3 fatty acids Diet low in polyunsaturated fatty acids Diet high in trans fatty acids Diet high in sodium Diet high in sodium Colon and rectum cancer cancer Ischaemic heart disease Ischaemic heart disease Ischaemic heart disease Systolic blood pressure Stomach cancer RCTs (Number) RCTs with significant effect in the opposite direction (%) RCTs with null findings (%) Prospective observational studies (Number) Prospective observational studies with significant Lower limit of RR > 1.5 Dose-response relationship Biologic plausibility Analogy GBD 2015

11 1. Defining optimal nutrition Agenda Setting Policy Formulation 2. Quantifying the burden of disease due to malnutrition 3. Evaluating the effectiveness of policies to improve nutrition 4. Evaluating the cost-effectiveness of nutrition policies 5. Evaluating the political/legal feasibility of nutrition policies Policy Evaluation Policy Adoption 6. Evaluating the intensity of implementation of nutrition policies 7. Evaluating the short/long term effects of nutrition policies Policy Implementation

12 Comparative Risk Assessment 12

13 Global Nutrient Database Calcium Other nutrients Sugars (Glucose, Sucrose, Starch) Protein Vitamins (A, B, D, E, K) Folates Iron Zinc Magnesium Phosphorus Potassium Selenium GBD 2015

14 Standardizing the modelling approach Sex Suboptimal metric Nationally Representativeness Data from FFQ Data from HBS Data from FAO Country level covariate Diet low in fruits - Diet low in vegetables - Diet low in whole grains - - Diet low in nuts and seeds - Diet low in milk - Diet high in red meat - Diet high in processed meat - National availability of red meat and pig meat Diet high in sugar-sweetened beverages - National availability of sugar Diet low in fiber - Diet suboptimal in calcium - Diet low in seafood omega-3 fatty acids Landlocked nation (Yes,/No) Diet low in polyunsaturated fatty acids - Diet high in trans fatty acids - National availability of hydrogenated oil Diet high in sodium GBD 2015

15 Comparative Risk Assessment 15

16 Age curve of relative risks Body Mass Index Total Serum Cholesterol Fasting Plasma Glucose Systolic Blood Pressure Diet low in fruits Diet low in vegetables Diet low in whole grains - Diet low in nuts and seeds Diet high in red meats - - Diet high in processed meats - Diet low in fiber Diet low in seafood omega-3 fatty acids - - Diet low in polyunsaturated fatty acids - - Diet high in trans fatty acids - - GBD

17 Measurement Error Subar (2001) 17

18 Publication Bias SSBs & Diabetes Nuts & IHD Trans fat & IHD Red meat & Diabetes Processed meat & Diabetes 18

19 Definition of dietary factors Wang Huang Johnsen Wu Jacobs Jensen Steffen Liu added bran added wheat germ bagels bran breakfast cereals brown rice brown rice flour buckwheat bulgur cooked cereal cooked oatmeal corn meal dumplings corn meal flat cakes corn meal porridge corn meal steamed bread non-white bread oats other grains pancakes pizza popcorn psyllium Aune (2016) 19

20 Covariates Age Sex Race Education Merital Status Smoking Alcohol Physical Excerise Engery Fruit and Vegetables SFA Fish/ Seafood PUFA MUFA Trans- FA Red Meat Sucrose Coffee Sodium Soy Dairy BMI Waist/hip Vitamin Supplements Oral Contraceptives HRT Diabetes Hypertension Hypercholesterolemia Mental Stress Menopausal Status Sleep Duration Atkins (2014) Eshak(2011) Eshak (2014) Jacobs(2001) Jensen(2004) Johnsen (2015) Liu (1999) Mink (2007) Muraki (2014) Muraki (2014) Pietinen (1996) Rautiainen (2012) Steffen(2003) Tognon (2014) Wang(2016) Yu (2013) Whole grains and Ischemic Heart Disease Aune (2016) 20

21 Correlation between dietary factors Veg Fruit Proc Meat Red Meat Nuts/seeds Whole grains Veg 1.00 Fruit Proc Meat Red Meat Nuts/seeds Whole grains SSB Milk Sodium Omega PUFA SFA Fiber Calcium SSB Milk Sodium Omega-3 PUFA SFA Fiber Calcium NHANES

22 Afshin 2014 Luo 2014 Zhou 2014 RR for CHD per 1 serving (28.4g)/week of nuts

23 Comparative Risk Assessment 23

24 New approach to determine TMREL GBD 2015 Fruits gr/day Vegetables Whole grains Nuts gr/day gr/day gr/day Free sugars<10% E/d Adiposity Salt <5 g/d Sodium< 2g/d Cardiovascular disease Red meats gr/day Processed meats 0-4 gr/day Milk gr/day Sugar sweetened beverages 0-5 gr/day Polyunsaturated fatty acids Seafood omega-3 fatty acids 9-13% of total daily energy mg/day Total Fat < 30% E/d Cardiovascular disease Trans fatty acids 0-1%E Dietary fiber gr/day Dietary calcium gr/day Sodium 1-5 gr/day 24

25 Reducing intake of free sugars and body fatness Increasing intake of free sugars and body fatness Adults Children Morenga (BMJ, 2012)

26 Isoenergetic exchanges of free sugars with other carbohydrates Free sugars<10% E/d Morenga (BMJ, 2012)

27 Sodium excretion and risk of cardiovascular disease Sodium<5 g/d Sodium<2 g/d Mente (Lancet 2016) Mozaffarian (NEJM 2014)

28 Citation network graph with 269 reports and 2165 citations. Ludovic Trinquart et al. Int. J. Epidemiol. 2016

29 Co-authorship network graph with 643 authors. Ludovic Trinquart et al. Int. J. Epidemiol. 2016

30 Howard (2006) Total Fat < 30% E/d

31 Comparative Risk Assessment 31

32 Ranking of dietary risks based on the disease burden attributable to them in the most populous countries GBD

33 Acknowledgment GBD Collaborators 33