Feeling better after trying something isn’t a test, but you can run one

Key takeaways

• In a 2021 UK trial, 200 people who blamed statins for muscle aches unknowingly took turns on the drug and on look-alike dummy pills, and on average their symptom scores came out nearly identical.
• In a small 2003 study using the same one-person-at-a-time method in Australian family practices, valerian was not clearly better than a placebo for chronic insomnia.
• NIH’s Office of Dietary Supplements suggests four questions before any supplement: what it might do for you, what the risks are, what dose to take and for how long.
• St. John’s wort can weaken birth control pills, the transplant drug cyclosporine, the blood thinner warfarin and many other medicines, so check interactions before any self-test.
• Check with a clinician first, or skip the experiment entirely, if you’re pregnant, choosing for a child, facing surgery, being treated for cancer or HIV, or tempted to replace proven care.

Between late 2016 and spring 2018, 200 people at 50 primary care sites in the United Kingdom signed up for an unusual test¹. All of them had stopped taking statins, the cholesterol-lowering drugs, or were thinking about stopping, because of muscle symptoms¹.

Over a year, each person took six two-month courses in a random order. Some courses held atorvastatin, a common statin. Others held placebo pills that looked the same. Neither the patients nor the study team knew which was which¹.

At the end of each course, people rated their muscle symptoms from 0 to 10. Among the 151 people with enough ratings to compare, symptoms were almost exactly the same on the drug as on the dummy pills, a gap of about one-tenth of a point¹. Dropouts over intolerable muscle symptoms were similar too: 18 people during statin courses and 13 during placebo courses¹. About two-thirds of those who finished said they were going back on statins for the long term¹.

This isn’t an article about statins. The trial, called StatinWISE, tested one dose of one drug, and it doesn’t mean no one ever has muscle side effects. It was a formal study run through doctors’ offices, not a home experiment. If you suspect a prescription is causing side effects, talk with your prescriber before changing anything.

The point is what a fair test can show that everyday experience can’t.

Researchers call this kind of study an N-of-1 trial: a trial in a single person (“N” is the number of people) that switches between treatments several times³. It has been used on herbs, too. In a study published in 2003, family doctors in Queensland, Australia, ran a series of N-of-1 trials comparing valerian with a placebo for chronic insomnia². Valerian was not clearly better than placebo, either for individual patients or for the group².

You can borrow the logic of these trials for any remedy you’re thinking about. First, though, come the questions.

Start with the problem, not the product

The most useful question comes before any product: what exactly am I trying to fix, and do I know what’s causing it?

Fatigue, gut trouble, aches and poor sleep have many possible causes. Some are easy to treat once they’re found, and a few are serious. NIH’s National Center for Complementary and Integrative Health (NCCIH) advises against letting an unproven approach become a reason to put off seeing a health care provider⁴.

Picture a hypothetical reader who spends three months on an energy supplement, then learns at a checkup that her fatigue came from low iron. The supplement didn’t fail. It was aimed at the wrong target.

Next, name your goal in a way you can measure. Not “feel better,” but “fall asleep within 30 minutes on most nights” or “keep my knee pain below 4 out of 10.” That target becomes the backbone of your self-test.

The questions health agencies tell you to ask

Federal health agencies publish short, overlapping lists of questions.

The NIH Office of Dietary Supplements (ODS) suggests asking your provider four things before you start a supplement: what it might do for you, whether it carries safety risks, what dose is right, and how, when and for how long to take it⁵.

NCCIH adds two habits. Find out what studies show about both safety and effectiveness, and tell every provider you see, conventional and complementary, about everything you use⁴.

The federal Agency for Healthcare Research and Quality (AHRQ) runs a campaign called Questions Are the Answer. One of its suggested questions fits here exactly: “Will this medicine interact with medicines that I’m already taking?” Another asks whether there are alternatives⁶.

Put together, they make five questions for any remedy:

1. What am I treating, and has the cause been checked?
2. What do independent summaries say about it, for this problem?
3. What dose, form and length of use did the studies test?
4. Could it interact with anything I take or any condition I have?
5. Is this product what its label claims?

Here is where to find answers fast:

• ODS fact sheets and NCCIH herb pages, which cover many vitamins, minerals and herbs and include safety sections.
• MedlinePlus, from the National Library of Medicine. Since July 29, 2025, its herbs and supplements index no longer carries the commercial database it once hosted; entries now link to NIH pages such as NCCIH’s⁷.
• The NIH Dietary Supplement Label Database and the FDA’s Dietary Supplement Ingredient Directory, for checking a product’s label and what the FDA has said about an ingredient⁸.
• Your pharmacist, who can check your whole list at once.

On the fifth question, NCCIH warns that a label may not match what’s inside. Products sold for weight loss, sexual improvement or bodybuilding are a special concern because some contain hidden drugs⁸. Words such as “standardized,” “verified” or “certified” don’t guarantee quality⁸.

Interactions are the question people skip

A supplement can make a medicine stronger, make it weaker, or combine with it in harmful ways, as NCCIH’s free “Know the Science” lesson on interactions explains⁹.

The danger is greatest with drugs that work only within a narrow range. NCCIH lists carbamazepine and phenytoin for seizures, cyclosporine for preventing organ transplant rejection, digoxin for heart problems, and the blood thinner warfarin¹⁰. With these, even a little too much or too little can cause serious, possibly life-threatening reactions¹⁰.

St. John’s wort shows how far one herb can reach. NCCIH says it can weaken birth control pills, cyclosporine, warfarin, digoxin, some antidepressants, some HIV and cancer drugs, and the statin simvastatin¹¹. Taken with certain antidepressants, it may raise the risk of serious serotonin-related side effects¹¹.

ODS adds two more examples. Vitamin K can blunt warfarin’s effect, and antioxidant pills such as vitamins C and E might make some chemotherapy less effective⁵.

Surgery is another flash point. Supplements can affect bleeding, anesthesia, heart rate or blood pressure, so some clinicians want patients off herbal products for several weeks before a planned operation¹⁰.

The fix is simple. Bring a written list of everything you take, with doses and how often, or bring the bottles themselves, and update it at every visit¹².

Why “it worked for me” is so easy to get wrong

Most people test a remedy the casual way: start it, wait and see how they feel. AHRQ’s guide to N-of-1 trials explains why that often misleads. Casual trials have no comparison, no blinding and no steady way of measuring, so they’re open to bias³.

Symptoms also rise and fall on their own. Without careful tracking, patients and clinicians alike can be fooled about what a treatment did³.

Other things shift at the same time. People often start a remedy at a low point, when symptoms are likely to ease anyway, and they may sleep more or worry less once they feel they’re doing something. Expectations shape how we feel, too, as our article on placebo effects explains.

None of this means your experience is wrong. It means one unstructured experience can’t separate the remedy from everything else going on.

Many natural-health traditions teach close attention to your own body. A self-trial takes that attention seriously. It turns “I think it helps” into a record you can trust, share and act on.

How to run a fair test on yourself

This routine borrows from formal N-of-1 trials. It won’t match a blinded study, but it’s a fairer comparison than memory.

1. Pick one or two outcomes and a scale. The AHRQ guide calls careful, consistent measurement perhaps the most important part of an N-of-1 trial, and it advises choosing one or two main outcomes³. Rate the same thing, the same way, at the same time each day.
2. Record a baseline first. Rate it daily for at least a week or two before you start, so you know your normal ups and downs.
3. Change one thing at a time. Keep your food, exercise, sleep habits and other products steady. Write down anything that changes anyway, such as a cold, travel or a stressful week.
4. Match the trial to the evidence. Use the dose and length that studies used. For St. John’s wort, the evidence covers about 12 weeks of use¹¹. The right length depends on how quickly a remedy starts and stops working³.
5. Write your success and stop rules before you start. Decide what change would make it worth continuing, and what would make you stop at once.
6. Add an off period if you can. For a supplement or herb, never a prescription drug, stop for a week or two when the trial ends and keep rating. Formal N-of-1 trials switch on and off several times in a random order, often with look-alike placebos, as StatinWISE did¹³.
7. Review the log on decision day, and share it. Compare your baseline average with your trial average, and bring the numbers to your clinician.

Expect it to take discipline. In the Australian valerian study, only 24 of the 42 people who enrolled produced enough data to analyze².

Be realistic about what self-trials can do. In a 2018 study in JAMA Internal Medicine, 215 adults with chronic muscle and joint pain got either usual care or app-guided N-of-1 trials designed with their clinicians¹³. After six months, pain’s interference with daily life was not significantly different between the groups¹³.

The N-of-1 group did report more shared decision-making about medicines, and 88% agreed the app could help people like them¹³. A self-trial is a decision tool, not a treatment.

When not to experiment on yourself

Some situations call for a clinician first, or for no experiment at all.

• Pregnant or breastfeeding. Many supplements haven’t been tested in pregnant or nursing women⁸. If you’re trying to conceive, ask first, too.
• Choosing for a child. ODS urges caution about giving a child supplements unless the child’s provider recommends them⁵.
• Surgery ahead. Tell your surgeon and anesthesia team about everything you take, as early as you can⁵¹⁰.
• Taking prescription drugs, especially narrow-range medicines such as warfarin, or being treated for cancer or HIV¹⁰¹². Never stop or change a prescription to make room for a trial. Talk with your prescriber.
• Sudden, severe or fast-worsening problems. The AHRQ guide says N-of-1 methods don’t suit acute illnesses or conditions that can turn catastrophic without warning³.
• Instead of proven treatment. In a 2018 study in JNCI, a team led by Skyler Johnson compared 281 people who chose alternative medicine alone for curable breast, prostate, lung or colorectal cancer with matched patients who got standard treatment¹⁴. The alternative-medicine group had 2.5 times the risk of death¹⁴. The study was observational, so it can’t prove cause. The gap was widest for breast cancer¹⁴.

A one-page worksheet to fill in before you start

Print this or copy it into a notebook. Finish Part 1 before you buy anything.

Part 1: Before you start

Question	Where to look	My answer
What exactly am I trying to change, and has a clinician checked the cause?	Your clinician	________
What do independent summaries say about it for this problem?	ODS fact sheets, NCCIH herb pages, Cochrane summaries	________
What dose, form and length of use did the studies test?	The same summaries and the studies they cite	________
Could it interact with anything I take, including other supplements?	A pharmacist, with my full list or the bottles	________
Is it riskier for me because of pregnancy, surgery, age or a health condition?	Clinician, pharmacist, fact-sheet safety sections	________
How do I know the product contains what the label says?	A third-party seal; FDA warnings and recalls	________
What will the whole trial cost?	Price per day times the number of days	________
Which of my providers will I tell?	Everyone who prescribes or treats	________

Part 2: My trial plan

• The symptom I’ll rate each day, from 0 (none) to 10 (worst): ________
• The time of day I’ll rate it: ________
• Baseline: I’ll rate it daily for ______ days before starting (at least 7).
• Product, dose and time of day: ________
• Trial length: ______ weeks (no shorter than the studies used).
• Success means: ________ (decide now; for example, “my weekly average improves by 2 points or more”).
• I’ll stop at once and call ____ if: ______
• Changes I’ll avoid during the trial: ________
• Decision day: ________

Part 3: Daily log

Add one row per day, starting with your baseline days.

Date	Took it?	Score (0-10)	Side effects	Anything else different?
__	Yes / No	______	______	Sleep, stress, food, illness, other new products
__	Yes / No	______	______	________
__	Yes / No	______	______	________

Part 4: Decision day

• My average score during the baseline: __. My average in the last two weeks of the trial: ____.
• Did I reach my success mark? Yes / No
• Optional check (supplements and herbs only): I stopped for one to two weeks and kept rating. Did the symptom come back? Yes / No
• My decision: keep it / stop it / talk with my clinician first.
• If I keep it, my next review date: ________

What the evidence shows

Claim	Evidence rating	What it rests on
In people who quit, or wanted to quit, statins over muscle symptoms, atorvastatin 20 mg caused more symptoms than placebo on average	Evidence says no, or unsafe	200 people in randomized, blinded N-of-1 trials; one dose of one statin; doesn’t rule out side effects in some individuals1
Valerian beat placebo for chronic insomnia in the 2003 Australian N-of-1 series	Evidence says no, or unsafe	Not clearly better for individuals or the group; side effects similar to placebo; only 24 of 42 patients gave usable data. One small study, not a review of all valerian research2
St. John’s wort weakens many prescription drugs	Strong evidence	NCCIH lists interactions with birth control, transplant, heart, HIV and cancer drugs11
Casual “try it and see” trials reliably show whether a remedy works	Evidence says no, or unsafe	No comparison, blinding or steady measurement; symptoms vary on their own. In StatinWISE, people who blamed a statin for their aches rated them about the same on placebo13
App-guided self-trials improve chronic pain more than usual care	Not enough evidence	One randomized trial of 215 adults found no significant difference at six months13
Alternative medicine alone is a safe substitute for standard treatment of curable cancer	Evidence says no, or unsafe	National cancer database study; 2.5 times the risk of death; observational, so not proof of cause14

Questions readers ask

The bottom line

Asking good questions before trying a remedy takes minutes. Check what you’re treating, what independent summaries say, what dose and how long, what it might interact with, and whether the product can be trusted.

Then test it the way a researcher would test it for one person: a baseline, one change at a time, a daily score, a set end date and a stop rule written in advance. One careful month can teach you more than a year of guessing, and it leaves a record your clinician can use.

Related reading

• Six questions to ask before you buy anything for your health
• Most health products aren’t scams, but many just haven’t been proven
• What happens when you don’t tell your doctor about supplements
• Third-party tested says what’s in the bottle, not that it works
• Honey “probably” eases a child’s cough, and that word is a precise rating

Sources

1. Herrett E, Williamson E, Brack K, et al. “Statin Treatment and Muscle Symptoms: Series of Randomised, Placebo Controlled N-of-1 Trials.” BMJ, 2021;372:n135. discovery.dundee.ac.uk
2. Coxeter PD, Schluter PJ, Eastwood HL, Nikles CJ, Glasziou PP. “Valerian Does Not Appear to Reduce Symptoms for Patients With Chronic Insomnia in General Practice Using a Series of Randomised N-of-1 Trials.” Complementary Therapies in Medicine, 2003;11(4):215-222. research.bond.edu.au
3. Kravitz RL, Duan N, eds. “Design and Implementation of N-of-1 Trials: A User’s Guide.” Agency for Healthcare Research and Quality, Publication No. 13(14)-EHC122-EF, February 2014.
4. National Center for Complementary and Integrative Health. “Are You Considering a Complementary Health Approach?” NCCIH, updated September 2016. nccih.nih.gov
5. NIH Office of Dietary Supplements. “Dietary Supplements: What You Need to Know.” ODS, updated January 4, 2023. ods.od.nih.gov
6. Agency for Healthcare Research and Quality. “The 10 Questions You Should Know.” AHRQ, Questions Are the Answer, reviewed November 2020. ahrq.gov
7. MedlinePlus, National Library of Medicine. “Herbs and Supplements” (index and July 29, 2025 notice). medlineplus.gov
8. National Center for Complementary and Integrative Health. “Using Dietary Supplements Wisely.” NCCIH, updated January 2019. nccih.nih.gov
9. National Center for Complementary and Integrative Health. “How Medications and Supplements Can Interact” (Know the Science). NCCIH. nccih.nih.gov
10. National Center for Complementary and Integrative Health. “Some Medication-Supplement Interactions Can Be Serious” (Know the Science). NCCIH, accessed October 2026. nccih.nih.gov
11. National Center for Complementary and Integrative Health. “St. John’s Wort.” NCCIH, updated May 2025. nccih.nih.gov
12. National Center for Complementary and Integrative Health. “Talk With Your Health Care Providers” (Know the Science). NCCIH. nccih.nih.gov
13. Kravitz RL, Schmid CH, Marois M, et al. “Effect of Mobile Device-Supported Single-Patient Multi-crossover Trials on Treatment of Chronic Musculoskeletal Pain: A Randomized Clinical Trial.” JAMA Internal Medicine, 2018. pubmed.ncbi.nlm.nih.gov
14. Johnson SB, Park HS, Gross CP, Yu JB. “Use of Alternative Medicine for Cancer and Its Impact on Survival.” Journal of the National Cancer Institute, 2018;110(1):121-124. academic.oup.com