Every health choice trades benefit for risk, including doing nothing

Key takeaways

• An IBS drug was pulled in 2000 after serious bowel injuries, then allowed back in 2002 for a narrower group of women with tight safeguards, after patients pushed for its return.
• The FDA weighs benefit against risk with a structured judgment, not a formula, and accepts more risk when a disease is serious and options are few.
• Waiting is a choice too: in a 15-year prostate cancer trial, monitoring led to similarly low prostate cancer death rates but about twice as many cancers that spread.
• The same facts worded as lives saved or lives lost change what people choose, and a 36-lab replication confirmed the effect.
• Green tea is considered safe as a drink for adults, but concentrated extracts in pills have been linked to uncommon liver injury and drug interactions.

In February 2000, the Food and Drug Administration (FDA) approved Lotronex, a pill for irritable bowel syndrome, or IBS¹. Within months, reports arrived of ischemic colitis, a loss of blood flow to the bowel, and of severe complications from constipation¹.

The FDA counted at least 49 cases of ischemic colitis and 21 serious constipation complications. Some patients needed surgery or blood transfusions, and some died¹. In November 2000, the maker withdrew the drug¹.

That could have been the end. Instead, patients who said Lotronex was the only treatment that had relieved their symptoms lobbied the FDA and the manufacturer, GlaxoSmithKline, to bring it back².

In June 2002, the FDA agreed, with conditions. The drug returned only for women with severe diarrhea-predominant IBS that other treatments hadn’t helped. It came with a boxed warning, a low starting dose, a doctor sign-up program and a patient agreement form².

Not everyone was convinced. Public Citizen, a consumer advocacy group, said the safeguards were too weak and should not be run by the company that profits from the drug. Larry Sasich, a pharmacist with the group, called the plan “a classic case of the fox guarding the hen house”³.

The pill was the same in 2002 as in 2000. What changed was who could get it, how it was started and watched, and how much weight regulators gave to relief that some patients valued deeply. That weighing sits inside every health choice you make, including choosing a natural remedy or choosing to wait.

The FDA weighs benefit and risk with a structured judgment, not a formula

People often imagine regulators plugging numbers into an equation. The FDA’s own guidance, finalized in October 2023, says otherwise. It calls its benefit-risk framework a structured, qualitative approach that involves judgment, decided case by case⁴.

The framework asks five sets of questions⁴:

• The condition: How serious is it, and which symptoms or outcomes matter most to patients?
• Current options: What treatments already exist, and what needs are still unmet?
• Benefit: How large is the effect, how certain, how long does it last, and does it matter clinically?
• Risk and risk management: How severe, likely and reversible are the harms, and can they be predicted, monitored or prevented?
• Conclusion: Taken together, with the remaining uncertainty, do the benefits outweigh the risks, and what safeguards are needed?

Context changes the answer. The guidance says greater risk may be acceptable when a disease is serious and there are no good alternatives⁴. For conditions with many safer options, or for preventive drugs given to healthy people, the agency tolerates less serious risk⁴.

Patients’ voices count. Their input can shape which benefits matter most and how much risk and uncertainty are acceptable⁴. But the FDA says it will not approve a drug that would do more harm than good across the whole population it is meant for⁴.

Lotronex’s safeguards loosened over time. In 2016, the FDA made prescriber training voluntary, and it later decided the special safety program was no longer needed, though a boxed warning remains¹.

Your own risk can flip the answer

Daily low-dose aspirin shows how the balance can tip for one person and not another.

In 2016, the U.S. Preventive Services Task Force, an independent expert panel, recommended aspirin to prevent a first heart attack or stroke for some adults in their 50s at higher risk, if their bleeding risk was not raised⁵. Then three large trials published in 2018 found little or no benefit and more bleeding⁵.

In 2022, the Task Force narrowed its advice. For adults 40 to 59 with a 10% or higher chance of heart attack or stroke over 10 years, and no added bleeding risk, it found a small net benefit and said the choice should be individual⁵. For people 60 and older, it recommended against starting aspirin for prevention, because bleeding risk rises with age and outweighs the benefit⁵.

In 2022, the same pill and the same evidence produced different verdicts depending on age and risk. If you already take aspirin, talk with your clinician before changing anything. To read the size of a benefit like this one, see our companion article on what meaningful benefits look like.

Waiting is a choice with its own odds

When a treatment carries real harms, doing nothing can feel like the safe default. Sometimes it is a strong option. It is never risk-free.

A British trial called ProtecT shows both sides. Researchers randomly assigned 1,643 men aged 50 to 69 with prostate cancer still confined to the prostate to one of three paths: active monitoring with regular checks, surgery or radiation⁶⁷.

After about 15 years, prostate cancer deaths were low in all three groups: 17 men in monitoring, 12 after surgery and 16 after radiation⁸. By our arithmetic, that’s roughly 2% to 3% in each group. The University of Bristol reported that monitoring matched the high survival of surgery and radiation⁶.

Waiting had a cost. The cancer spread in 9.4% of the monitoring group, about twice the rate after surgery (4.7%) or radiation (5.0%)⁸. Treatment had a cost too: harm to urinary and sexual function from surgery and radiation lasted up to 12 years⁶. About a quarter of the monitored men were alive at the end without ever having treatment⁸.

So neither path was safe. Each traded one set of risks for another, and the right choice depends on what a man fears more.

Psychologists have also found that people tend to judge harm from acting more harshly than harm from not acting. In a 1990 study, Ilana Ritov and Jonathan Baron described a hypothetical disease that killed 10 of every 10,000 children and a vaccine whose side effects could also kill⁹. Asked how much risk from the vaccine they would tolerate, 23% said none, even though the vaccine would prevent the disease in everyone⁹. Any vaccine less deadly than the disease would have saved lives overall.

The researchers called this omission bias⁹. If you value letting the body heal, “wait and see” can feel like the gentle path. It deserves the same honest odds as any pill. And when a serious diagnosis is involved, waiting should be a plan made with a clinician, not a decision to stop looking.

The same facts in different words change what we choose

In 1981, psychologists Amos Tversky and Daniel Kahneman asked people to imagine an outbreak expected to kill 600 people¹⁰. When a sure-thing program was described as saving 200 lives, 72% chose it. When the identical program was described as letting 400 people die, only 22% did¹⁰.

That flip is called a framing effect, and it holds up. In a 2014 replication across 36 labs on three continents with more than 6,000 people, the reversal appeared in 31 labs¹⁰.

Doctors are not immune. In a 1982 study in the New England Journal of Medicine, researchers asked 238 patients, 491 graduate students and 424 physicians to imagine having lung cancer and to choose between surgery and radiation¹¹. In all three groups, surgery looked more attractive when outcomes were described as chances of living rather than chances of dying¹¹.

The format of a number matters too. In one study described by risk researchers Paul Slovic and Ellen Peters, experienced forensic psychologists and psychiatrists judged whether to release a hospitalized mental patient¹². When his risk of violence was described as “20 out of every 100” similar patients, 41% refused to release him. When it was described as a 20% chance, 21% refused¹².

The fix is simple. Ask to hear every risk both ways: how many live and how many die, how many are helped and how many are not.

Feelings slip into the math

Slovic and Peters separate two ways of sizing up danger: fast, intuitive reactions, which they call risk as feelings, and slower reasoning from logic and evidence, which they call risk as analysis¹².

When we like something, we tend to rate it as both more helpful and less risky. That runs against the real world, where activities with big payoffs often carry big risks too¹². In experiments, telling people a technology had large benefits made them judge it safer, and time pressure strengthened the effect¹².

Dread matters too. The public sees nuclear power’s radiation as far more dangerous than the radiation in a medical X-ray, a view risk experts don’t share¹².

Our reading of this research is that the warm feeling around “natural” can make a remedy seem both more helpful and safer than the evidence shows. The cold feeling around “drug” can do the reverse.

Green tea is a good test case. The National Center for Complementary and Integrative Health (NCCIH) reports no safety concerns for adults drinking green tea, though it contains caffeine¹³. Its natural compounds, catechins and caffeine, might nudge weight down slightly¹³.

Concentrated extracts in pills are different. Cases of liver damage are uncommon, but they have turned up, mostly in people taking those extracts. People with a particular immune-related gene variant, carried by 5% to 15% of Americans, seem more susceptible¹³.

High doses of green tea may also lower blood levels of nadolol, a blood pressure drug, and extracts may lower levels of atorvastatin, a cholesterol drug¹³. Same plant, different dose and form, different balance.

A one-page benefit-risk worksheet

This worksheet borrows the FDA’s structure and puts it in your hands. Copy the eight questions onto a sheet of paper, make one column for each option you’re considering, and always include waiting as an option.

Question	Why it matters	Where to find the answer
1. How serious is the problem, and how fast is it changing?	Urgent or serious problems justify accepting more risk	Your clinician; reliable condition pages from NIH or CDC
2. What are all my options, including waiting and lifestyle changes?	You can’t weigh an option you never listed	Clinician, pharmacist, a decision aid
3. Out of 100 people like me, how many are helped, and over what time?	Turns a percentage into people	Study results or your clinician’s risk estimate
4. Out of 100, how many are harmed, how badly, and can it be reversed?	A rare, permanent harm weighs more than a common, mild one	Drug label, NCCIH or NIH Office of Dietary Supplements fact sheets
5. Does it interact with anything I already take?	Herbs and drugs can change each other’s effects	Pharmacist, with your full list of drugs and supplements
6. What will it cost me in money, time and daily effort?	A plan you can’t keep up delivers no benefit	Your budget and your calendar
7. How will I know it’s working, and when will I recheck?	Builds in a fair test and an exit point	Agree on a symptom or measure and a date
8. Which outcomes matter most to me?	Two people can rightly choose differently	Rank them: living longer, comfort, avoiding side effects, cost, staying natural

Then ask one last question: what would make me change course? Write the answer down before you start.

Decision aids help people see the trade-offs

You don’t have to do this alone. Decision aids are booklets, videos or online tools that lay out options, benefits and harms side by side.

A 2024 Cochrane review pooled 209 trials with 107,698 adults facing 71 kinds of decisions¹⁴. People who used decision aids were nearly twice as likely to have an accurate sense of their risks as people who got usual care, a finding rated high certainty¹⁴. Their knowledge scores rose about 12 points out of 100, and they felt less uninformed, with no rise in regret¹⁴.

Ask your clinician whether a decision aid exists for your choice. Our articles on shared decision-making and on when to defer to experts go deeper.

What the evidence shows

Claim	Evidence rating	What it rests on
Wording a choice as gains or losses changes what people pick	Strong evidence	Classic experiment replicated in 31 of 36 labs10
Doctors, too, are swayed by survival vs. death wording	Promising but early	One classic 1982 study of 424 physicians, plus patients and students11
Liking something makes people rate it as lower risk	Strong evidence	Repeated experiments on the affect heuristic12
People overweight harms from acting compared with harms from not acting	Promising but early	Hypothetical-choice experiments; real-world effect less certain9
Active monitoring for localized prostate cancer gives survival similar to treatment at 15 years	Strong evidence	One large randomized trial678
Starting daily aspirin at 60 or older to prevent a first heart attack	Evidence says no, or unsafe	Task Force found harms outweigh benefits5
Concentrated green tea extract is risk-free because tea is natural	Evidence says no, or unsafe	Reported liver injury and drug interactions13
Decision aids improve understanding of risks and benefits	Strong evidence	209 randomized trials14

When not to weigh it alone

• Call 911 for emergency symptoms such as chest pain, trouble breathing or signs of a stroke. Waiting is not an option then.
• Call your prescriber or pharmacist before stopping a prescribed medicine or adding a supplement to one, since some combinations interact¹³.
• Report problems. Report side effects of drugs to the FDA’s MedWatch program, and problems with supplements to the federal Safety Reporting Portal at safetyreporting.hhs.gov¹⁵.

Questions readers ask

The bottom line

No health choice is free of risk, including the choice to do nothing or to try something natural first. Regulators, clinicians and patients all weigh the same kinds of questions: how serious the problem is, how big the benefit is, how bad the harms are and what the alternatives look like.

The difference is that you get to add your own values. Hear every number both ways, notice when a feeling is doing the deciding, list waiting as a real option with real odds, and write your trade-offs down before you choose.

Keep exploring: absolute vs relative risk, miracle cure language and AI-generated health misinformation.

Related reading

• A 22% risk cut can mean helping 2 or 3 people in 1,000
• How medical decisions are supposed to get made with you
• When expert advice deserves weight, now that patients have the right to decide
• What happens when you don’t tell your doctor about supplements
• How to judge a cup of herbal tea, which is gentle but not automatically harmless

Sources

1. U.S. Food and Drug Administration. “Lotronex (alosetron hydrochloride) Information.” Postmarket Drug Safety Information for Patients and Providers.
2. Drug Topics. “FDA Approves Restricted Marketing of Alosetron.” July 1, 2002. drugtopics.com
3. Public Citizen. “Allowing Lotronex Back on the Market Will Endanger Patients.” Press release, June 7, 2002. citizen.org
4. U.S. Food and Drug Administration. “Benefit-Risk Assessment for New Drug and Biological Products.” Guidance for Industry, October 2023. fda.gov
5. Mallick S, Shroff GR, Linzer M. “Aspirin for Primary Prevention of Cardiovascular Disease: What Do the Current USPSTF Guidelines Say?” Cleveland Clinic Journal of Medicine, 2023;90(5):287-291. ccjm.org
6. University of Bristol. “Delaying Treatment for Localised Prostate Cancer Does Not Increase Mortality Risk, Trial Shows” (on Hamdy FC, et al., and Donovan JL, et al., New England Journal of Medicine, 2023). March 13, 2023. bristol.ac.uk
7. The ASCO Post. “Delaying Treatment for Patients With Localized Prostate Cancer May Not Increase Mortality Risk, ProtecT Trial Shows.” American Society of Clinical Oncology, March 2023. ascopost.com
8. American Journal of Managed Care. “Prostate Cancer Deaths in PSA-Detected Disease Remained Low at 15 Years” (on Hamdy FC, et al., New England Journal of Medicine, 2023). AJMC, 2023. ajmc.com
9. Ritov I, Baron J. “Reluctance to Vaccinate: Omission Bias and Ambiguity.” Journal of Behavioral Decision Making, 1990;3:263-277 (author’s copy). sas.upenn.edu
10. Brohmer H, von Götz S. “Framing Effect.” In Rahal RM, ed., Open Social Psychology, FORRT (on Tversky A, Kahneman D, Science, 1981, and Klein RA, et al., Social Psychology, 2014). forrt.org
11. McNeil BJ, Pauker SG, Sox HC Jr, Tversky A. “On the Elicitation of Preferences for Alternative Therapies.” New England Journal of Medicine, 1982 (Harvard Medical School publication record). hcp.hms.harvard.edu
12. Slovic P, Peters E. “Risk Perception and Affect.” Current Directions in Psychological Science, 2006;15(6):322-325. courses.washington.edu
13. National Center for Complementary and Integrative Health. “Green Tea.” NIH, updated February 2025. nccih.nih.gov
14. Stacey D, Lewis KB, Smith M, et al. “Decision Aids for People Facing Health Treatment or Screening Decisions.” Cochrane Database of Systematic Reviews, 2024 (plain-language summary). cochrane.org
15. U.S. Food and Drug Administration. “How to Report a Non-Emergency.” FDA. fda.gov