Your doctor glances up from the screen and says, “Let’s start you on something for the blood pressure.” You say sure. The visit moves on to your flu shot.
That exchange happens constantly, and most of the time it ends fine. But something got skipped. You never heard what else was reasonable, whether waiting was safe, or how much that pill is likely to do for you.
In the 1990s a research team led by C.H. Braddock put tape recorders in 1,057 office visits with primary care doctors and surgeons, then counted every clinical decision on the tapes. There were 3,552 of them. Nine percent met the study’s standard for a complete discussion. The report ran in JAMA in 1999 and the recordings date to 1993, so read it as a portrait of an era rather than today’s number. The scale is still worth sitting with.
Fixing this doesn’t take confrontation. It takes a few questions asked in an ordinary voice, and some idea of how to read the answer when it comes back as a percentage.
The first suggestion you hear is usually a starting point
Clinicians have defaults, and defaults are not a bad thing. A family doctor who has treated several thousand cases of high blood pressure has watched what tends to work, what people tend to tolerate, and what tends to get filled at the pharmacy. The first thing out of her mouth is the distilled version of all that.
It is still a default. It can be shaped by things that have nothing to do with you: what the practice is set up to deliver in-house, what the last conference covered, what the person in the previous exam room needed. A fifteen-minute slot also pushes toward the shortest route to a plan.
So the first option is a proposal, not a verdict. Treating it that way is not an insult to the person offering it. Most clinicians will tell you plainly that two or three approaches are defensible, if you give them the opening.
The opening is one sentence: “Before we go with that, what are the other reasonable choices here?” You are not asking her to defend herself. You are asking her to think out loud, which is the part that usually gets cut for time.
Waiting is a real option, and it has a name
In medicine, deliberate delay has formal labels: watchful waiting, active surveillance, observation. It is the option readers most often never hear, partly because it sounds like a non-answer and partly because it is harder to explain in two minutes than a prescription.
Look at what it means in practice. In a trial published in Pediatrics in 2005, researchers at the University of Texas Medical Branch enrolled 223 children between six months and 12 years old with a non-severe ear infection. Half were randomly assigned to amoxicillin right away. Half got medicine for pain and fever plus a plan to watch, with antibiotics held in reserve.
Two-thirds of the watch-and-wait children, 66%, finished the 30-day study without ever taking an antibiotic. Antibiotic prescriptions in that group dropped by 73% compared with the immediate-treatment group.
Now the other side, because it exists. Children who got antibiotics immediately had their symptoms settle faster over the first ten days, had 16% fewer treatment failures, and needed fewer doses of pain medicine. Parents in both groups reported the same level of satisfaction with the care. The treated children also had more drug side effects, and at the 12-day check the pneumococcal bacteria carried in their noses were more likely to be resistant to multiple antibiotics.
That is what a real trade-off looks like. Neither column is empty. The researchers were also clear that waiting only works with a safety net under it: a way to judge how sick the child is, real control of the symptoms in the meantime, a parent who knows what to watch for, and a fast route back to the office with an effective antibiotic ready.
Which gives you the shape of the question. “What happens if we wait” is not one question, it’s four:
- How long is it reasonable to wait, and what are we waiting for?
- What would tell us to stop waiting and act?
- What do I watch for at home, specifically?
- Who do I call, and how fast can I be seen?
If a clinician can answer those four, waiting is a plan. If nobody can answer them, waiting is just drift, and drift is the version that hurts people.
Six questions that change the answer you get
These are built to be asked in sequence, out loud, in a normal appointment. Pick three if you only have time for three.
1. “What are my options here, including waiting?” The phrase “including waiting” does the work. Without it, you get the list of treatments. With it, you get the full menu.
2. “Out of 100 people in my situation, how many are helped by this, and how many are harmed?” This forces the numbers into a form you can picture. More on why in the next section.
3. “Over what period of time?” A drug that prevents four heart attacks per 100 people over ten years is a different proposition at 52 than at 84. Benefits almost always have a clock on them, and the clock usually goes unmentioned.
4. “How will we know if it’s working, and by when?” This sets a review date. It also surfaces whether anyone is planning to check, which is how people end up on a medication for nine years that stopped making sense in year two.
5. “Is there anything here that isn’t a prescription?” Open-ended on purpose. It covers physical therapy, a sleep schedule, a brace, a procedure, a referral, a change in an existing drug.
6. “If this were your mother, what would you pick, and why the ‘why’ more than the ‘what’?” The reasoning is more useful than the recommendation, because the reasoning tells you which trade-off your clinician is weighting most heavily. You may weight a different one.
Write them on the back of an envelope. MedlinePlus, the consumer health site run by the National Library of Medicine, treats communication with your provider as one of the most important parts of getting good care, and its advice for making that happen is unglamorous: write your questions down before you go, and keep asking until the answer makes sense to you. The federal Agency for Healthcare Research and Quality runs a free tool called QuestionBuilder, available as an app for iOS and Android and on the web, that lets you assemble and email yourself a question list before a visit; AHRQ’s materials note the app keeps everything on your own device. Its companion brochure, “Be More Engaged in Your Healthcare,” makes the point that patients who ask and then confirm they understood the answer tend to get more timely and accurate diagnoses.
How to read “cuts your risk in half”
Here is the single most useful piece of arithmetic in a medical conversation.
Suppose your risk of some bad event over the next decade is 2 in 100. A treatment cuts it to 1 in 100. That is a 50% relative risk reduction, and it is also a 1 percentage point absolute reduction. Out of 100 people who take the drug, one avoids the event and 99 get no benefit from it at all.
Now suppose your risk was 40 in 100 and the same treatment brings it to 20 in 100. Still “half.” But now 20 people out of 100 are spared. Same headline, twenty times the real-world payoff.
This is why “cuts your risk in half” tells you almost nothing on its own. You need the starting number.
What the research says about how to ask
A systematic review in the Annals of Internal Medicine in 2014, led by Daniella Zipkin, pulled together 91 studies comparing ways of presenting probability to patients. Visual aids, especially icon arrays and bar graphs, improved both understanding and satisfaction. Absolute framing produced more accurate comprehension than relative framing, and it was also less likely to nudge people toward accepting treatment. Presenting the number needed to treat made understanding worse.
That last finding is worth pausing on, because the number needed to treat is a beloved shorthand among clinicians. It is simple enough to state: if 100 people take a drug and two avoid an event, the number needed to treat is 50, meaning 50 people take it for one to benefit. People reliably misread it anyway.
The evidence is not perfectly tidy. In a randomized study published in the Journal of General Internal Medicine in 2003, Stacey Sheridan and colleagues gave 357 patients aged 50 to 80 at a university internal medicine clinic one of four formats and asked them to compare two treatments. The relative risk format produced the most correct answers at 60%, ahead of a combined presentation at 43%, absolute risk at 42%, and number needed to treat at 30%. When the same people were asked to apply the information to a stated baseline risk, almost everyone struggled: 21% correct with relative risk, 17% with absolute, 7% with the combination, 6% with number needed to treat.
Two studies, two emphases. What they agree on is the practical part. The number needed to treat is a poor thing to hear by itself, and no format works unless you are also told the baseline risk you are starting from. So ask for both numbers, in people out of 100, over a stated number of years.
Why the same benefit can be graded differently
It helps to know how national recommendations get made, because the answer explains a lot of apparent inconsistency. The US Preventive Services Task Force, created in 1984, is a volunteer panel of experts in prevention and primary care. The Agency for Healthcare Research and Quality has been authorized by Congress since 1998 to convene it. The Task Force describes itself as a “scientifically independent, volunteer panel” of national experts.
It assigns each recommendation a letter grade from A to D, or an I statement when the evidence is insufficient. The grade reflects both the strength of the evidence and the balance between benefit and harm. Cost is deliberately excluded from that judgment.
One detail matters more than any individual grade: Task Force recommendations are written for people who have no signs or symptoms of the condition being screened for. If you already have a lump, a bleed, or a symptom, a screening recommendation was not written about you, and arguing with your clinician about its age cutoffs is beside the point.
Asking about the options that are not drugs
Plenty of good options are not prescriptions. Physical therapy for a bad knee or back. A supervised exercise program after a cardiac event. Cognitive behavioral therapy for insomnia. A splint, a hearing aid, a different mattress, a change in the timing of a medication you already take.
The National Cancer Institute’s patient question list, updated in 2023, is a decent model for how to phrase this without friction: it tells patients to ask what the treatments are for their specific situation, what the benefits and risks of each one are, and why the doctor recommends the one she recommends. The structure works outside cancer care too.
The non-drug path has honest costs, and a clinician who skips past it may be reading your calendar rather than dismissing the idea. Physical therapy means eight to twelve appointments you have to physically attend. Behavioral treatment for insomnia takes weeks before it pays off. Coverage is less certain. A prescription takes four minutes.
So ask the question in a way that acknowledges that: “Is there a non-drug option that’s genuinely worth the extra effort here, or is this one of those cases where it isn’t?” That phrasing gives your clinician room to say no without feeling accused of laziness.
One note on scope. Herbal products and dietary supplements belong in this conversation too, but they carry their own interaction risks and deserve their own discussion with your clinician and pharmacist. And when a decision is genuinely close, formal decision aids exist, which is a separate subject.
When the reaction is defensive
Most clinicians welcome questions. A minority bristle, and it is worth knowing what that means and what it doesn’t.
It often means nothing about you. It can mean the schedule is eleven patients behind, or that the question landed as a challenge to competence rather than a request for information. Rephrasing sometimes fixes it on the spot. “I’m not second-guessing you, I just want to be able to explain this to my husband tonight” lowers the temperature remarkably often.
If that fails, try a narrower question. Broad questions are expensive in a short visit; specific ones are cheap. “What’s the one thing that would make you switch me off this?” is answerable in fifteen seconds.
You can also move the question out of the room. Send it through the patient portal, where the clinician answers between visits and the answer arrives in writing. Ask the nurse. Ask the pharmacist. The National Institutes of Health, in its own patient communication guidance updated in March 2025, points readers toward nurses and pharmacists as sources of information in their own right, not as consolation prizes.
What is not acceptable is a clinician who refuses to tell you what the alternatives were, discourages you from getting another view, or gets angry at being asked for numbers. That is a signal worth acting on, and getting a second opinion is a normal, expected part of medical care rather than a betrayal.
When to stop asking and start moving
This whole approach assumes you have time. Sometimes you don’t, and knowing the difference matters more than any question on the list.
Stop and call 911 for chest pressure or pain that spreads to the arm, neck or jaw, for sudden weakness or numbness on one side, a drooping face, or speech that comes out garbled, for serious trouble breathing, for heavy bleeding that won’t stop, or for a sudden headache that is the worst you have ever had. Stroke and heart attack care is measured in minutes, and nothing you ask in those minutes is worth what it costs.
Get seen promptly, within days rather than weeks, for a new lump, blood in your stool or urine, unexplained weight loss, a mole that is changing, or a fever with a stiff neck. These are not diagnoses. They are reasons to move the appointment up rather than research it first.
There is also a middle category worth naming. For most chronic conditions, taking a week to think, read, and talk it over costs nothing medically. Ask your clinician directly: “Is this a decision I can take a week on?” The answer is usually yes, and hearing it out loud makes the week feel less like avoidance.
The two sentences worth memorizing
If you keep nothing else from this, keep these.
“What are my options, including waiting?”
“Out of 100 people like me, how many does this help, and over how long?”
The first one surfaces the menu. The second one converts a persuasive-sounding percentage into something you can weigh against the side effects, the cost, and the hassle.
Say them in that order, write down the answers, and ask for the numbers in writing if they are complicated. A clinician who can answer both has told you nearly everything you need. A clinician who won’t has told you something too.
Sources
- Agency for Healthcare Research and Quality. “Questions Are the Answer.” ahrq.gov
- Agency for Healthcare Research and Quality. “QuestionBuilder App.” ahrq.gov
- Agency for Healthcare Research and Quality. “Be More Engaged in Your Healthcare.” ahrq.gov
- MedlinePlus, National Library of Medicine. “Talking With Your Doctor.” medlineplus.gov
- National Institutes of Health. “Talking With Your Doctor or Health Care Provider.” nih.gov
- National Cancer Institute. “Questions to Ask Your Doctor about Treatment.” cancer.gov
- U.S. Preventive Services Task Force. “About the USPSTF.” uspreventiveservicestaskforce.org
- MedlinePlus Medical Encyclopedia, National Library of Medicine. “Shared decision making.” medlineplus.gov
- Braddock CH et al., JAMA, 1999. PubMed. DOI: 10.1001/jama.282.24.2313. “Informed decision making in outpatient practice: time to get back to basics.” doi.org
- Zipkin DA et al., Annals of Internal Medicine, 2014. PubMed. DOI: 10.7326/M14-0295. “Evidence-based risk communication: a systematic review.” doi.org
- Sheridan SL et al., Journal of General Internal Medicine, 2003. PubMed. DOI: 10.1046/j.1525-1497.2003.21102.x. “A randomized comparison of patients' understanding of number needed to treat and other common risk reduction formats.” doi.org
- McCormick DP et al., Pediatrics, 2005. PubMed. DOI: 10.1542/peds.2004-1665. “Nonsevere acute otitis media: a clinical trial comparing outcomes of watchful waiting versus immediate antibiotic treatment.” doi.org