A Risky Perspective

Decision-Making Is More Than Science

There are no solutions. There are only trade-offs.
- Thomas Sowell [1]

A short scroll on social media reveals numerous fitness influencers demanding that you avoid certain foods because they’re “toxic”. Adding to the confusion are different demigods instructing you to avoid different foods. Worse: some fearmongers cite evidence that, they claim, demonstrates that their vilified food is “toxic”. It’s nigh-on impossible to eat anything these days without some level of fear. But is this fear justified?

The simple answer is “no”. Such a narrative misses the point. Everything – yes, everything – has the potential to be toxic at a certain quantity. Take one of the most fundamental life-giving substances on the planet: water. It is possible to consume too much of it. Water intoxication can lead to a potentially fatal disturbance in brain function due to the normal balance of electrolytes in the body being pushed outside safe limits by excessive intake. Hyponatremia – a condition caused by dangerously low sodium levels – is another example. Granted, these cases are rare and only occur when individuals consume between 6 to 10 litres of water within a few hours, overwhelming the kidneys’ ability to excrete the excess [2]. But however rare, they remind us that even water, when consumed in excess, can become a hazard.

Some substances have been assigned safe upper limits, above which caution is advised. Backed by decades of research, these levels are usually extremely conservative. Where research has shown a certain dose of a substance to be associated with a statistically significant higher risk of a disease in, say, rodents, a safe upper limit has been assigned that’s 10 or even 100 times lower than this potentially toxic dose. This means that you’d have to consume an unrealistically high amount of the substance to come close to the already conservative level. In other words, the chance of illness occurring is unfathomably remote. Because everything has the potential to be poisonous, there’s a useful phrase that’s worth remembering: “It’s the dose that makes the poison.” Rather than asking, “Is something poisonous?” we should instead enquire, “At what level?”

Does Science Have the Answer?

We can look to science to tell us the potential toxic amount of a substance. But risk assessment isn’t black and white; it involves nuance and perspective. Of course, a fair-minded science is crucial to helping us resolve issues of debate. But risk must also be considered at a population level. To do this effectively, scientists often rely on an approach that integrates findings from multiple studies and disciplines. The term “weight of evidence” infers that there’s been a full and critical appraisal of the available evidence on a topic and might even imply that there’s a scientific result. However, does the weight of evidence actually support a clearly defined real-world conclusion?

When it comes to foods and ingredients, the study and assessment of risk is largely relegated to technical fields such as risk analysis, decision analysis and toxicology. Performing a risk analysis requires an understanding of the potential hazards that could result from a particular product or technology. Paramount to any thorough risk assessment is empirical investigation and logical reasoning involving heuristics and calculations.

However, such an analysis is of limited real-world use without asking, “What risks are acceptable and at what cost?” This part of risk assessment is much less scientific and involves a considerable degree of subjectivity. Naturally, there will be diverse opinions because people have different beliefs, worldviews and life experiences, and these differences will be especially prevalent between different population groups. This means that we will often have to make informed decisions ourselves. I decide whether I should take a drug with its attendant and possibly unpredictable side effects in exchange for its potential benefits. If the issue in question requires a policy, the decision should be made by informed individuals and institutions. A policy sector or regulatory body will decide whether to approve a product given its level of risk.

This helps to explain why there’s considerable disagreement when it comes to novel food ingredients. If risk assessments were purely objective science, we could resolve most issues relatively easily. The empirical arm of risk assessment requires leaving the acceptance arm to the representative institutions of democratic societies.

Fundamentally, the evaluation of risk doesn’t fall neatly into either the domains of pure science or policy. Data is crucial, but intuition can be important too (see Data vs. Intuition in Nutrition Science). Moreover, there will, nearly always, be some data lacking, mandating scientists to make informed guesses about the outcome or probability that certain events will occur. However, these assumptions may not be universally accepted, nor will they always be subject to empirical validation. The result? Further disagreement. This is where cognitive tools like Bayesian reasoning come in. As described in my article, Bayesian Nutrition, this method of evaluation tells us that, as we gradually get closer to the truth, we should be continually updating our knowledge in proportion to the weight of any new evidence we encounter. Nothing in life is certain, no matter how strong the evidence. By keeping our minds open, we can adjust and adapt to new evidence as and when it becomes available. Viewing knowledge this way allows us to update our beliefs to be better aligned with the available evidence. Bayesian reasoning allows risk assessors to base their decisions on the most likely scenarios.

Some individuals may decide what they feel is an acceptable level of risk based on a worst-case analysis. Here the probabilities are lower but the outcome potentially more hazardous. Conversely, others may base their assessment on the odds of something happening, especially if the potential consequences are relatively low. Science can’t resolve the framing of the risk problem because it is not a scientific issue, and understanding this allows us to be more sympathetic to others’ views when they’re different from our own.

Risk Is Bi-Directional

Any thorough risk assessment – either on an individual level or when devising policy – involves having to accept that there will also be risks involved in not accepting a technology, drug, ingredient or food. There are no solutions, only trade-offs. An assessment should ask, “What are the risks?”, “Who do they affect?” and “Who will benefit from the technology?”

The GLP-1 agonist weight-loss drugs provide a useful example. We shouldn’t think of these drugs as “good” or “bad” as we know there are side effects and we don’t know if there will be harmful long-term effects. But the lives of many people who have used them have been changed for the better, with improvements to their health and self-image. In the UK alone, tens of thousands of people have used them*, despite there being a small number of deaths**. How many people would have gone on to die prematurely if they hadn’t been able to reap the advantages of the drugs? Any deaths linked to their use is, of course, something we should definitely pay attention to. But, if you’re someone who’s struggled to lose weight your entire life, despite dieting time and time again, fully aware that your health could be in jeopardy, then, together with your doctor, you can make a credible assessment of the risks and make an informed decision as to whether you should take GLP-1 drugs to help you lose weight.

When it comes to novel food technologies, we should take a broader assessment of the risk level and the populations who may benefit from them. This is where things get really tricky. People have different perspectives, and risks to more vulnerable groups might not be the same as to those who are more affluent. Ironically, it’s usually the latter who make decisions on behalf of the former. Take Golden Rice. Rice is one of the most commonly consumed foods; more than 3.5 billion people depend on it for more than 20 percent of their daily calories [4]. When scientists bred rice with a gene found in the daffodil, they massively increased the beta-carotene content of the grain [5]. Beta-carotene is a red pigment and a precursor to active vitamin A, and it gives this rice its “golden” colour. Making one of the world’s most consumed staples rich in this sight-protecting vitamin has the potential to help reduce the number of vitamin-A-deficient children in the world, half a million of whom become blind every year [6]. Moreover, Golden Rice is more drought-resistant than other strains: a huge boon to sustainable agriculture. Yet, GMOs – genetically modified organisms: crops that have been altered at the molecular level to improve attributes such as the nutritional value, hardiness, spoilage and ease of farming – have been subject to much skepticism, and not always for invalid reasons. Policies concerning technologies like GMOs, such as Golden Rice, require more than mere scientific validation. Social, political, financial and democratic considerations are similarly important.

Unintended Consequences

A thorough risk analysis should also acknowledge that there might be unintended consequences of a food, ingredient or technology. By definition, these can’t be articulated or measured: it’s simply not possible to consider in advance what’s not known. With any new technology, we must test the product in sufficiently varied sessions at extreme conditions in order to increase the odds of revealing any potential adverse consequences.

We can also ensure that there’s suitable labelling and traceability of supply chains, possibly mandated by regulation. For new food technologies, this is important as it allows any resulting damage from unintended consequences to be traced, and, if necessary, products can be stopped and recalled. The company and relevant agencies can learn from the incident and take action to mitigate any future consequences.

Risk analysis, done properly, involves both data and perspective, and what’s agreed to be a reasonable level of risk. Science can tell us only so much. Pondering this, I’m reminded of Donald Rumsfeld’s famous quote: “There are known knowns; there are things we know we know. We also know there are known unknowns; that is to say we know there are some things we do not know. But there are also unknown unknowns – the ones we don’t know we don’t know.”

* Due to unregulated online pharmacies and the prevalence of GLP-1 agonists on the black market, this is an estimate.
** As of February 2025, 82 deaths in the UK have been directly linked to their use [3].

References:
1. Sowell, T. (2007) A Conflict of Visions: Ideological Origins of Political Struggles. Rev edn. New York: Basic Books.
2. Rangan, G. K. et al. (2021) ‘Clinical Characteristics and Outcomes of Hyponatraemia Associated with Oral Water Intake in Adults: A Systematic Review’, BMJ Open, 11(12), e046539.
3. Iacobucci, G. (2025) ‘GLP-1 Agonists: 82 Deaths Linked to Adverse Reactions, UK Data Show’, BMJ, 388.
4. National Geographic (n.d.) Food Staple. Available at: https://education.nationalgeographic.org/resource/food-staple/ (Accessed: 22 April 2025).
5. Tang, G. et al. (2009) ‘Golden Rice Is an Effective Source of Vitamin A’, American Journal of Clinical Nutrition, 89(6), 1776-83.
6. World Health Organization (n.d.) Vitamin A Deficiency. Available at: https://www.who.int/data/nutrition/nlis/info/vitamin-a-deficiency (Accessed: 22 April 2025).