#197 - The science of obesity & how to improve nutritional epidemiology | David Allison, Ph.D.

Feb 28, 2022 Episode Page ↗
Overview

Peter Atiyah hosts David Allison, Dean of the Indiana University School of Public Health and an expert in obesity research. They delve into the complexities of obesity science, the pitfalls of nutritional epidemiology, and the critical need to distinguish between scientific rigor and advocacy to rebuild public trust in science.

At a Glance
15 Insights
2h 14m Duration
15 Topics
6 Concepts

Deep Dive Analysis

David Allison's Background and Approach to Science

The Emergence of the Obesity Crisis and Early Scientific Approaches

Genetic Component of Obesity: Twin Studies

Historical Approaches to Obesity Treatment and Pharmaceutical Development

Bariatric Surgery's Evolution and Impact on Mortality

The 'Obesity Paradox' and BMI Interpretation

Confounding Factors and Body Composition in Obesity Data

Limitations of BMI and Future Tools for Measuring Obesity

Challenges of Public Health Interventions for Obesity

Evolutionary Reasons for Difficulty in Weight Loss Maintenance

Critique and Reformation of Nutritional Epidemiology

The Mouse Study Illustrating Observational vs. Causal Effects

Addressing the Obesity Epidemic: Promising Interventions

Reproducibility in Science: Normative and Non-Normative Errors

Rebuilding Trust in Science and Differentiating Science from Advocacy

Energy Balance Statement

This statement posits that changes in energy stores equal energy in minus energy out. It is considered a tautology, akin to a law of thermodynamics or a geometric definition, rather than an explanatory model for the causes of obesity.

Obesity Paradox

This refers to two observations: first, that mortality rates are U-shaped, meaning both very thin and very heavy individuals have higher mortality than those in the middle BMI range. Second, that heavier individuals often have better outcomes when facing acute illnesses or injuries, potentially due to greater energy reserves.

Thrifty Gene Hypothesis

This evolutionary theory suggests that humans and animals historically faced starvation, leading to selection for genes that promote energy preservation and efficient eating when food is available. However, David Allison suggests this hypothesis is simplistic and not fully supported by evidence.

Red Queen Hypothesis (Evolution)

Applied to the evolution of sexual reproduction, this hypothesis suggests that organisms must constantly evolve and adapt (like the Red Queen running to stay in place) to keep up with co-evolving pathogens and parasites. Sexual reproduction, by mixing DNA, provides a mechanism to 'reset the locks' against rapidly evolving microbes.

Cluster Randomized Trial

A type of clinical trial where entire groups (clusters) like classrooms or neighborhoods are randomized to different interventions, rather than individuals. While valid in theory, it requires specific analytical methods to account for clustering, which are often incorrectly applied, leading to flawed results.

Normative vs. Non-Normative Errors

Normative errors are mistakes made at the edge of current scientific understanding or technological capability, like Galileo's incorrect measurement of light speed. Non-normative errors are mistakes that could have been avoided with existing knowledge and methods, such as flawed sampling or statistical analysis that was known to be incorrect at the time.

?
What is the genetic concordance for obesity in identical twins separated at birth?

For identical twins separated at or nearly after birth, the correlation coefficient for BMI is approximately 0.9, which is nearly the same as for twins reared together, indicating a very strong genetic component.

?
How effective is bariatric surgery in prolonging life for obese individuals?

Studies, including the Swedish Obese Subject Study, have shown that bariatric surgery can reduce all-cause mortality rates by approximately 50% or more, making it a life-saving treatment.

?
Does the optimal BMI for lowest mortality change with age?

Yes, the BMI associated with the lowest mortality rate (the nadir of the U-shaped curve) tends to move to the right (higher BMI) as individuals age. For example, a 20-year-old might have an optimal BMI near 20, while an 80-year-old's optimal BMI might be in the low 30s.

?
Are BMI classifications universally applicable across all ethnicities?

No, while the general U-shaped curve for BMI and mortality holds, its shape varies by age, race, and sex. For people of Middle Eastern and East Asian descent, the risk associated with higher BMI appears at lower BMI values, and for Hispanic Americans, research has struggled to find an association between elevated BMI and mortality.

?
Why is it so difficult to maintain weight loss?

There is no single explanation, but evolutionary theories suggest that humans may have been selected for mechanisms that promote energy preservation and eating when food is available, making it challenging to sustain reduced energy intake in a modern environment with unlimited food.

?
What are the major limitations of nutritional epidemiology?

Nutritional epidemiology faces significant challenges with confounding variables and severe measurement error, particularly with self-reported food intake. Many findings from observational nutritional studies have not been replicated in randomized controlled trials.

?
What are the most promising interventions for addressing obesity and improving public health?

David Allison believes the most promising interventions are bariatric surgery and pharmaceuticals (like GLP-1 agonists), along with general education (especially for girls and women) and efforts to reduce socioeconomic disparities. He also advocates for increased funding for basic research into areas like senolytics.

?
Is science currently facing a crisis of reproducibility?

While there is 'no crisis' in the sense of an imminent collapse of science, there is 'no cause for complacency.' Science is likely more rigorous than ever, but flaws are more visible, and there's a strong societal push for continuous improvement and greater transparency.

1. Distinguish Science & Advocacy

Always be self-aware of whether you are speaking as a scientist (delivering nuanced, hard truth without regard for listener’s feelings) or as an advocate (aiming to change behavior for perceived best interest). This clarity is crucial for maintaining public trust and fostering honest dialogue.

2. Practice Epistemic Humility

When presenting scientific findings, honestly acknowledge all the ways your conclusions might be wrong and suggest further tests. Avoid systematically dismissing competing explanations to genuinely advance knowledge.

3. Prioritize Causal Evidence

When evaluating claims, especially about health interventions, prioritize evidence from randomized controlled trials over observational studies. Recognize that even pristine observational studies can yield associations opposite to causal effects due to confounding.

4. Demand Evidence for Claims

When presented with a claim about a treatment or intervention, reliably ask critical questions: Was there a study? Was it in humans? Was it randomized? Did it measure the actual outcome? Was it long enough? Was the result statistically significant and large enough to matter? Was the dose realistic?

5. Avoid Body Shaming

Refrain from shaming individuals about their body habitus, regardless of the direction (e.g., too thin, too heavy). This is a moral issue of denigrating people, not an empirical one tied to whether it causes weight loss or gain.

6. Utilize Effective Obesity Treatments

For individuals suffering from obesity, consider bariatric surgery and pharmaceutical interventions (like GLP-1 agonists or SGLT2 inhibitors) as they are the most effective, life-saving, and life-changing treatments available, with demonstrated benefits on longevity.

7. Invest in General Education

To improve public health outcomes related to obesity and diabetes, prioritize general education, especially for girls and women. Provocative data suggests this leads to lower BMIs and reduced rates of these conditions.

8. Reduce Wealth Disparities

Consider policies that reduce disparities in wealth and living environments, as studies suggest moving families to less poor neighborhoods can lead to less obesity and diabetes. Confucius’s wisdom highlights concern for disparity, not just absence of wealth.

9. Fund Key Obesity Research Areas

Allocate resources strategically across four areas: providing and studying surgery, providing and studying pharmaceuticals, investing in general education for well-being and reducing disparities, and funding basic research into senescence (e.g., senolytics, microchimerism) related to metabolism.

10. Question Everything Critically

Cultivate a habit of asking ‘Are you sure? How do you know? Where did that come from? What makes you think that’s true?’ to foster critical thinking and avoid taking information at face value.

11. Apply Rigorous Scientific Thinking

Approach all variables, whether in human behavior or physical phenomena, with the same laws of probability, physics, and mathematics. Avoid sloppy thinking that separates behavior from biology or applies different standards to human-related fields.

12. Consider Multiple Angles

When studying complex problems like obesity, examine them from many different angles (e.g., physiology, genetics, anatomy, clinical medicine, behavior, culture, environment). No single factor or perspective is likely to provide a complete solution.

13. Reform Nutritional Epidemiology

Demand and implement significant reformation in nutritional epidemiology to address issues like confounding, measurement error (especially with food frequency questionnaires), and the overstatement of weak associations. The status quo is unacceptable and requires greater honesty and rigor.

14. Analyze Cluster Trials Correctly

For researchers conducting cluster randomized trials, ensure data analysis correctly accounts for clustering and nesting effects. Failure to do so leads to invalid results and distorted evidence.

15. Embrace Scientific Non-Complacency

Recognize that while science is likely more rigorous than ever, there is no cause for complacency. Continuously strive for improvement, addressing flaws and non-normative errors to elevate the quality and trustworthiness of scientific endeavors.

I don't really know why I turned out the way I turned out. But even as a kid, I was always the kid asking questions. Sometimes that would interest people. Sometimes that would annoy people and that's true today. But I was always one saying, are you sure? How do you know? Where did that come from? What makes you think that's true?

David Allison

The same person who would never question a diabetologist on what the beta cell of a pancreas is doing, unless they themselves are a diabetologist studying the beta cells of the pancreas, will say to an obesity researcher, well, this is how obesity works.

David Allison

Everybody knows that surgery will cause weight loss and help people live longer. There's no point in doing this study. And Lars said, you may be right that it does, but we have to do this study because everybody doesn't accept and believe that it hasn't been shown. We know over and over again, you've got to do the experiment.

David Allison

To know what you know and to know what you do not know, that is true knowledge.

David Allison

The science is the data, the methods, and the logic connecting the data to conclusions, not whether I trust you or not.

David Allison

Shaming people about their body habitus is not good in either direction. It's not a directional thing. It's just shaming people about their body habitus is not okay.

David Allison

If what you mean by science is trust in individual elements of the scientific community, then I'm not sure it's down either, but it's spread around. So some people think Fauci is trustworthy, and some people think Gwyneth Paltrow is trustworthy.

David Allison
0.9-ish
Correlation coefficient for BMI in identical twins (monozygotic) separated at birth Indicates a very strong genetic component to BMI, similar to twins reared together.
1997-ish
Approximate year fen-fen was withdrawn due to valvulopathy Fenfluramine component was found to cause a peculiar valvulopathy.
50% or more
Reduction in all-cause mortality rate from bariatric surgery Varies across studies but is a significant reduction.
16 to 17
BMI range for a supermodel (Kate Moss) Considered underweight.
18.5
Beginning of normal weight BMI Below this is considered underweight.
30
Beginning of obesity BMI A common threshold for diagnosing obesity.
43 to 44
BMI for a top-class sumo wrestler An example of extreme BMI.
Near 20
Approximate BMI nadir for 20-year-olds BMI associated with the lowest mortality rate.
Well above 30, possibly low 30s
Approximate BMI nadir for 80-year-olds BMI associated with the lowest mortality rate, shifts rightward with age.