Solving Diabetes

Or The Futility of Thinking in Buckets

An individual diagnosed with diabetes (specifically, the common type—Type 2 diabetes) stands before you¹. Now, tell me:

What caused their diabetes?

What’s the first answer that came to you? I’ll bet that answer came rather quickly. Chances are, diabetes is a (at least somewhat) familiar phenomenon to you.

But ok, if the answer came so easily, why have we not solved it already?

Well, you might say in a patient lecturing tone, just because we understand the cause of something doesn’t mean we have the solution for it. Having the solution doesn’t mean we can implement it. Being able to implement it doesn’t mean we can do it in a sustained, effective manner. Yes, of course, you are completely right. But I want to take our argument one step back today. Underlying all these difficulties is a common denominator. My proposition is this—we do not understand the causes of diabetes, at least not well enough.

Why I’m Wrong

I suspect at this point, some of you are up in arms.

Hol’ up, you’ve got to be kidding, right? What gives you the authority to say that?

Diabetes is one of the most prevalent and well understood diseases. Ask Pete from the street, and he’ll probably be able to tell you that you are much more likely to develop diabetes if both Uncle Georgie and Aunty Georgette (and the rest of the family tree) have it, walking to the car constitutes your daily exercise, and McDonald’s tops the most ordered list in your food delivery app.

And that’s not just folk wisdom. There is plenty of science to back that up. Just look at the number of articles showing up on a Google Scholar search:

The sheer number of articles tells only part of the story. Diabetes has been studied by scholars from a multitude of disciplines, who have uncovered multitudes of causal factors ranging from genes to behaviours to the environments and systems we are embedded within.

An individual diagnosed with diabetes stands before a group of scientists. Here are some of their responses:

1. Their diabetes is the result of their poor dietary habits, physical inactivity, poor sleep, and stress.

2. It is a function of systemic forces that promote poor dietary habits, physical inactivity, poor sleep, and stress.

3. It is caused by a cascade of biological events resulting in fat accumulation in the liver and pancreas, leading to insulin resistance and impaired insulin production.

4. They have a specific variant of the TCF7L2 gene, and a specific variant of the SLC30A8 gene.

5. They are from an impoverished neighbourhood and experience healthcare inequalities.

6. It is rooted in a fundamental mismatch between our evolutionary history and our current environments.

7. Pah… pay those myopic geneticists and sociologists no heed. Of course it is caused by a complex interaction between their genes and environment.

Now, which of these correspond most closely with your own answer? Each of these answers is, of course, correct. Some of those scientists may say that some answers are more correct than others (their proclivity to do so likely correlates with how much time they’ve spent chipping away at a particular answer), but we can find robust evidence for each of these views among the 3.8 million articles scattered throughout the research cosmos.

Which is great! We truly do have a deep, multifaceted understanding of the disease. So what on God’s obesogenic earth am I talking about?

Why I’m Right

If you look closely at each scientist, you will see the faint outline of a barely visible adornment over their head. What’s that, you ask. Why, a bucket, of course. Painted on each bucket are some words. This one says “genetic risks”; that one says “liver and pancreatic function”; another one says “ultra-processed food consumption”. You may find some scientists donning two buckets—“lifestyle behaviours” and “health equity”, or “liver and pancreatic function and “genetic risks”. With luck, you may even find one who dons three or four buckets.

The point is this—when I say we do not understand the causes of diabetes, I mean that the causes are deeply complex, and that each of us individually, even the foremost authorities on this subject, only understand some aspect of the problem².

Thinking in Buckets

Part of that can be attributed to our finite capacity for acquiring knowledge. Unless you’re Terence Tao or Neo from the Matrix, there are architectural limitations to how much knowledge you can acquire on a particular subject over your lifetime³.

Tenor

Another part can be attributed to the structural limitations of our scientific institutions. Our institutions tend to reward depth, and disincentivise breadth. As a knowledge base develops (i.e., we realise the nature of the beast), scientists begin to specialise, and get rewarded for specialising. Entire careers can be made on the study of a single gene, hormone, or class of behaviours. Paradoxically, in the words of the psychologist Henry Roediger, we study more and more about less and less. Our buckets get deeper, but narrower.

Three Bucket Problems

This is important. Our buckets dictate how we view problems and their solutions. Thinking in buckets helps us make sense of phenomena, but they present us with additional problems:

1. We can’t agree on a solution. We need more drugs! No, we need better targeted drugs based on genomic information! No, we should prioritise better eating and physical activity habits! No, we need to fix access to healthy foods and physical activity promoting environments! No, we need all of the above, and it should be done this way! No, it should be done that way!

2. We design flawed solutions. We design solutions based on what we know, without incorporating information from other knowledge bases. Thus, we develop behaviour change plans around random physical activity targets (e.g., 10,000 daily steps) without understanding the physiological effects (or non-effects) of meeting this target. Thus, we focus on a single metric (blood glucose) without considering contextual mechanisms and lived experiences of the individual.

3. We implement solutions imperfectly. We develop and push drugs onto patients without considering if they will take them, or recommend specific diets (e.g., Mediterranean diets) for populations that have fundamentally different food practices and access. The prescribers will say, I’ve provided recommendations based on the best available knowledge. It’s up to them to follow my recommendations, or not. But the fact of the matter is, getting individuals to take their prescribed medications or change their lifestyle behaviours continues to be a huge challenge.

Resolving Our Problems

To summarise our earlier points, we collectively possess a comprehensive, multifaceted understanding but individually only possess deep knowledge of some aspect(s) of diabetes. As such, we land on solutions too simplistic for addressing such a complex phenomenon. What then is the solution?

The obvious answer is to form interdisciplinary teams. Each member of the team contributes their perspective from their area of expertise, and we arrive at a comprehensive solution together. However, as those of you who are involved in such collaborations can attest, it’s really hard to do interdisciplinary work!

In a recent article, the psychologist Paul Bloom says, put a group of scientists of different backgrounds together and you’re likely to observe them engaging in parallel play.

The only interdisciplinary conversations worth having

It’s like gathering a diverse group of musicians together and commissioning them to compose a coherent piece. One is trained in the Western tradition and speaks German; another trained in the Carnatic tradition and speaks Tamil; a third trained in the Gamelan tradition and speak Indonesian; and so forth⁴. Each musician believes their own musical tradition transcendental, and other musical forms to be a cacophony of sounds. Physically next to each other, but ideologically worlds apart. The outcome is likely not going to be all that good.

The solution, Bloom argues, can be summarised in a quote by the late philosopher Jerry Fodor:

The only interdisciplinary conversations worth having are those that go on inside a single head.

To Bloom, that means one should already be

engaged and interested in the ideas and methods of other adjacent fields. So a neuroscientist who studies colour vision, say, will be conversant with the philosophical literature on the nature of colour. A philosopher interested in the ontology of race will have thought deeply about the debates within social psychology regarding the nature and extent of implicit racial bias. And so on.

I agree, but I think some further unpacking is in order. There are many ways of collaborating, to different degrees, for different purposes. Some like to draw distinctions (yay, more buckets) between cross-disciplinary, multidisciplinary, interdisciplinary, and transdisciplinary approaches. I won’t bore you with the differences, since folks can’t seem to agree on what these terms mean anyway.

Suffice to say, the odds of a successful collaboration correlates with the degree to which a common level of understanding can be established. That understanding requires the team to have a common goal (e.g., solving diabetes, composing a musical work integrating ideas from various traditions), and be engaged and interested in how others can contribute to this goal. It also requires collaborative action over time. These are not easy conditions to achieve.

Catalysis

One way to reduce the time requirements and increase odds of a successful collaboration is, borrowing a metaphor from the evolutionary biologist David Sloan Wilson, to introduce a catalyst⁵.

The term itself is of course borrowed from chemistry. Here’s the relevant definition from Merriam-Webster:

a substance that enables a chemical reaction to proceed at a usually faster rate or under different conditions (as at a lower temperature) than otherwise possible

Catalysts exist on a continuum—on one end, perhaps some musicians speak and understand a smattering of English, or our classical Western-trained musician has previously briefly acquainted themself with Gamelan music while on holiday in Bali. On the other, we can have musicians who are effectively bi- or multi-lingual, and have spent an extensive amount of time exploring various musical traditions⁶. Their views and work transcend specific knowledge bases. These individuals, let’s call them transdisciplinarians, serve as bridging agents that bring conversations out of buckets. The best transdisciplinarians even manage to convince others that their arbitrary buckets don’t really exist.

There is no bucket. Created by the writer from a scene in The Matrix

What Fodor speaks of interdisciplinary conversations that go on inside a single head, I imagine those who do it best not only see how knowledge and practices from two or more disciplines fit together, but are able to envision how emergent knowledge that transcends each discipline can arise.

Solving Diabetes: Towards A Bucketless Approach

Not all problems require such a bucketless approach, but they are particularly important for wicked problems which, by their nature, resist simple solutions. Diabetes is one such problem—messy, indeterminate, and deeply entrenched in many aspects of society. In reality, I do not believe a true bucketless approach is possible, nor do I believe there are true transdisciplinarians in the ideal sense⁷. But we can hold that as an ideal which we (individually and our institutions) strive towards.

A society grappling with diabetes stands before you. You have to pick a team to solve the problem. You could turn to superstar scientists who shine in their own buckets but rarely venture beyond them. Me? I’ll place my bets on the teams populated by catalysts strong and weak.

1. For avoidance of doubt, we’re talking specifically about Type 2 diabetes today.

2. Using myself as an example: I know a lot about behaviours and their environmental influences, a little about systems and their properties, squat about genomics, metabolomics, abracadabromics, and other -omics.

3. Factoring the time we spend learning not to poop our pants, and surviving our teenage blunder years, considerably less time.

4. I had the pleasure of studying musics of the Carnatic and Gamelan traditions in Junior College. 10/10 would recommend.

5. Here’s one way to think of catalysis:

Successful collaborations require a common goal and engaged collaborators, and collaborative action over time. A catalyst shortens the time required, and increases the odds of success.

6. Recently, I listened to the data scientist Liu Nan discuss the idea of transdisciplinarity with the health services and systems researcher David Matchar. Liu argued that clinician-scientists who want to successfully design and implement data-driven solutions need some understanding of data science. Hearing random forest or Markov chain should not, at the very least, conjure thoughts of an assortment of trees or necklaces.

7. I’m biased, but I think evolutionary science, with its influence on fields as diverse as biology, medicine, psychology, economics, and many more, comes pretty darn close.