Risk Assessment - The Illusion of Certainty

Guy Heynen, MD
January 24, 2003 (Reviewed: January 21, 2005)

Introduction

Every day brings a checklist of things to accomplish, read or discuss, and decisions must be made. Whether we like it or not, decision inevitably involves a certain degree of uncertainty, or risk. Consider, for example, how you feel when you are about to go to your physician, because you feel more tired then usual and have had some abdominal bloating. If you were to list the first 2 questions that you would like to see answered, what would they be?

Research shows that despite all the health information that can easily be accessed through the Internet, many patients visit their physicians totally unprepared - they don't know what type of questions to ask. However, they still expect to be taken care of and get reassurance from their physician. These expectations are entirely legitimate, but they should not prevent you from asking specific questions. Very often, information given to you by health care professionals, especially on what to do or how to do it for your health, involves uncertainty (or risk) about the likely outcome. In some cases, your participation in the decision process may not be critical; for instance whether you chose hot tea or honey and syrup to relieve your sore throat isn't critical to the outcome. However, your doctor may propose a surgical procedure; in that case, your active understanding of the various risks involved, such as the likelihood of a full recovery and the risk of dying, is vital, to help you and your physician to reach the "right" decision for you. Very often also your active participation in an exercise program is bound to be better managed by yourself if you grasp how big the effect is likely to be on your health.

How good are you in assessing the risk, chance or degree of uncertainty? Do you know what information you need to participate intelligently in the decision process?

A series of short quizzes and summaries based on Gerd Gigerenzer's book "Calculated Risks" will guide you through the maze of "risk evaluation", hopefully helping you better understand, assess and feel (more) at ease to ask additional questions when confronted with statements (like diagnosis, or prognosis) or with treatment options regarding your health.

How good are you at assessing risk or degree of uncertainty?

The first law of uncertainty (Franklin's law) states: "In this world nothing can be said to be certain, except death and taxes." However logical this quote sounds, beliefs and expectations, along with miscommunications, invade us so pervasively that one often concludes that a lab result, a diagnosis or a prognosis is an absolute certainty. Let's lean back a little and consider several examples. First, test your skills in percentages.

Example # 1 - Depression and sexual problem

David D, a 40-year-old truck driver, has recently been diagnosed with depression, and prescribed Prozac. His physician has told him that there was a 30% to 40% chance of his developing a sexual problem, such as impotence or loss of sexual interest.
How do you understand the 30% to 40% chance here?

1. In 10 attempts at intercourse, he will have problems in 3 or 4 of them.
2. After 6 successful acts at the first attempt, subsequent attempts are highly likely to go awry.
3. 3 or 4 patients out of 10 experience some kind of sexual problem.

Only answer number 3 truly reflects the published observations. Please note that information given as a frequency (i.e. "3 or 4 patients out of 10") increases the clarity of the information when compared to quotations in percentages. In other words, when you ask the doctor to clarify, specify exactly what you want: "could you put that as a frequency, please?" You may also try to suggest to your physician a way of phrasing that you understand, such as: "Out of 10 patients like me, Doctor, who are suffering from depression, how many experienced this problem when treated with Prozac?"

By specifying "patients like me with the same diagnosis and the same treatment", you can better imagine yourself as a member of a group of people similar to yourself, and realize the likelihood of your falling in the group with sexual problems, or with those without problems.

Example # 2 - A decision to make: which surgeon?

Laura T is a 65-year-old housewife suffering from severe, debilitating, knee osteoarthritis, and is contemplating surgery. Wisely, she consulted 2 surgical specialists before taking a decision. Surgeon A has a very good reputation and has told Laura she has a 90% chance of full recovery after the operation. Surgeon B also has a very good reputation and has told Laura that 70 out of 80 women in her age-range who he has operated on have shown functional improvement and a better quality of life one year after surgery. How should Laura decide on the basis of this information?

First, the 90% chance (or risk) of full recovery of surgeon A sound very promising and slightly better than the 70/80 quoted by surgeon B. However, the 90% quote by surgeon A is probably only a degree of his belief, and sounds like a subjective figure; no class reference has been given - what are these 90% of what? As such, the statement is no more reliable that a neighbor's degrees of belief that it will rain tomorrow.

By contrast, surgeon B has provided evidence from experience (at least 80 patients) with success (70 out of 80) after a defined period of time (one year after intervention). Laura can now better decide with the help of her physician: she should either ask surgeon A to give her the same information as that given by surgeon B, or she can go ahead with surgeon B. One would not expect Laura to go ahead with surgeon A on the basis on available information, unless Laura (and her physician) misunderstands percentages.

Example # 3 - Why shouldn't I take this drug?

This new drug is wonderful and has no side effects. True or false?

If you ever think that this could be true, then start reading again at the top until you come to Franklin's law. Then think again.

Conclusions

Percentages are not always easily understood by every one. They are very popular in all aspect of daily life and news: our modern society finds it the best way to communicate successes and failures. Percentages are often used to quantify uncertainty (a quantified uncertainty is also called a risk) for success or failure, say of a surgical intervention or medication. Sadly, these percentages often come without any specifying information, which commonly leads to misunderstandings. Frequencies, instead, are easier to understand and should be used more often, particularly when one tries to convey information forming the basis for a decision or consent to an intervention. Franklin's law (. . . nothing certain but death and taxes) is a tool to help us reflect on the degree of uncertainty.

Future articles in this short series "Risk Assessment" will cover these topics:

• Test and review your knowledge (or ignorance) of risk. Examples on risk of breast cancer, car accident, prostate cancer and colorectal cancer, smoking, risk of adverse events to antidepressant agents and the risk of treating these adverse events, risk of adverse events to herbal medicines; risk of significant gastrointestinal toxicity with non-steroidal anti-inflammatory agents and the possible advantage of COX-2 inhibitors...
• Significance of screening for diseases, DNA testing, fingerprints.

Vioxx (rofecoxib), a COX-2 anti-inflammatory drug, has recently been withdrawn from all world markets by Merck, its manufacturer. Merck found that there was an increased risk of heart attack or stroke for people taking a standard dose (25 mg per day) for 18 months or more. The risk of such problems was twice that of people taking a placebo. However, someone's risk of having a heart attack or stroke is relatively low; double the risk means the chance of such an event is still "small," according to the USA Food and Drug Administration (FDA). Celebrex (celecoxib) and Bextra (valdecoxib) are other prescribed COX-2 inhibitors. Studies have so far indicated that only Vioxx poses a risk of heart problems. The FDA will examine the other two drugs to see if there is any cause for concern. Robert Griffith, Editor.

Source

• Gerd Gigerenzer (2002). Calculated Risks: How To Know When Numbers Deceive You. Simon & Schuster, New York, NY 10020. ISBN 0-7432-0556-1. You can buy this book at Amazon, just click here

Related Links
Health Risk Factors
RiskFactor Roulette
Do You Intend To Be a 'Responsible' Patient?

Please take a moment to give us your comments. For questions about Health matters you may check our "Questions & Answers" Portal and Service.