Population Risk Does Not Equal Individual Risk

Risk In Perspective: Population Risk Does Not Equal Individual Risk – April 4, 2018 – This series is a collaboration between neuroscientist Alison Bernstein and biologist Iida Ruishalme.

Errors in risk perception are at the core of so many issues in science communication that we think this is a critical topic to explore in detail

Population risk is not the same as individual risk

We tend to think in very small sample sizes (after all, what happens to me and my family must be most important, right?) and not in terms of populations (which is how epidemiological statistics are calculated).

However, scientists measure population risks

We, as consumers, assume that reports of population risk translates to the same amount of risk for individuals. But it is not that simple and this part of risk assessment is not intuitive at all.  

Individual risk is a flawed concept

If something is reported to increase risk of disease in a population, does that mean that you or me will have a higher risk from exposure to this substance? Not necessarily.

In fact, risk is, by definition, a population-based measure.

The very idea of extrapolating a population level-risk down to a personal one is highly flawed, as described by Dr John W. McEvoy in Risk and the Physics of Clinical Prediction. In an interview about predicting individual risk of heart attacks based on population-derived risk factors, Dr. McEvoy said:

the concept of individual risk applied — truly applied — to any given person is an oxymoron.

Risk for an individual is like a square peg for a round hole. We can never know, or estimate, one person’s risk.

In fact, if you do the math, the confidence interval for a given risk estimate in one person would range from a 0% to a 100% chance of a cardiac event.

Thus, risk is not ‘personalized’ and I think of risk as a ‘group-phenomenon’.

Calculating population risk

The graphic above illustrates that many factors influence risk for disease.

Genetic variations may have a positive, negative, or neutral effects on development of disease.

Exposures (including lifestyle factors) can also have positive, negative or neutral effects.

By looking at large groups of people, epidemiologists can detect differences in the percent of people in various subgroups (by genetics or exposure) that develop a disease.

However, from that population level data, it is not possible to predict an outcome for any specific individual within that population.

One of my pet peeves is that individual pain patients are treated as though the population risk of opioid addiction applied directly to them.

What should be a medical decision for a specific individual is handled as though the patient were a perfect example of the population at large, as though their individual “risk” were exactly representational of the whole population.

But it’s not even possible for an individual to be an exact average of the population. A simple way to see the absurdity of this is that the “average American” has one breast, one testicle, and a family with 1.87 children. 

Epidemiology is more complicated than this simplified example because many diseases and traits are influenced by multiple genetic factors and multiple exposures, resulting in a complicated combination of positive, negative and neutral factors. Teasing apart those factors is a very important part of epidemiology.

What’s an individual in a population to do?

On an individual level, it is important to remember that many risks can be mitigated by encouraging potentially beneficial exposures that counteract the effect of potentially harmful exposures.

If we focus on these big beneficial choices that we can control, we can mitigate many of the possible risks from exposures that we cannot control.

Focusing on these ‘big picture’ risks, at both the individual and population level, that are within our control can actually go a long way towards reducing and mitigating a wide variety of risks, both known and unknown.

If you would like to read more about different aspects of risk perception, please see the other parts of the series, which this article belongs to:

Risk In Perspective: Introduction

1) The difference between hazard and risk is a critical distinction.

2) All hazards are not equal.

3) Zero risk and zero exposure are impossible expectations.

Author: My name is Iida Ruishalme. I am a biologist specialising in biomedical research, a science communicator, and a fiction writer. My pieces have been published at Food and Farm Discussion Lab, Genetic Literacy Project, and Biofortified, as well as the cultural journal The Woolf and the Finnish newspaper Aamulehti.

6 thoughts on “Population Risk Does Not Equal Individual Risk

  1. canarensis

    VERY important point, and one that not even most physicians really grasp….which leads to devastating consequences in patient (& pain) management.

    “One of my pet peeves is that individual pain patients are treated as though the population risk of opioid addiction applied directly to them.” YES!! Me too, me too!!

    Liked by 1 person

    Reply
      1. canarensis

        Indeed! Tho I guess it could be said that if you’re surprised, it means you still have some faith in the group as a whole (i’m trying to exercise optimism here…that was my allotment for the day ;-)

        Seriously, it is startling & terrifying how poorly so many of the so-called researchers understand what they’re doing, what it means, & how to interpret stats. One of the really odious papers I reviewed for the Oregon cptf was a review about all the dire harms that opioids cause. Two on the list were impotence & increased auto accidents. In the results section, we learned that a paper showed that men on opioids took ED meds at a higher rate…and (1) the rate was not statistically significant, & (2) that they had no idea if the men were on ED meds before starting the opioids (!!). As for the auto accidents, that “higher” rate also was not statistically significant. In Conclusions, though, the authors shrieked that opioids caused impotence & auto accidents (among other things). This is, of course, one of the papers that the Oregon task farce is using to justify their insanity.

        In this case, the question is maybe more whether the incredibly flawed conclusions were a result of extreme bias than inability to grasp stats. What is even more appalling is that this utter crap passed editorial review, not to mention the clear fact that nobody involved on any level groks that whole “correlation does not equal causation” thing, which is about as basic a scientific & statistical idea as there is. Yet it continues to elude even “researchers,” much less the MSM. It is to weep.

        Liked by 1 person

        Reply
  2. Pingback: HHS Report on Pain Mgmt Best Practices – part 1 | EDS and Chronic Pain News & Info

Other thoughts?

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.