Home › Know Your Body

How to Research a Treatment Your Doctor Hasn't Offered You

By the Strength from the Well Editorial Team · Evidence last reviewed September 2026 · Not reviewed by a licensed clinician (why we say that)

The short answer

Start by finding out whether the treatment has been studied in humans at all, using free public databases: MedlinePlus for plain-language background, PubMed for published studies, ClinicalTrials.gov for trials in progress, and the Cochrane Library for summaries that pool many studies together. Then sort what you find by study type, because a randomized trial in people and a promising result in mice are not the same kind of fact. Finally, bring one page of specifics to your doctor and ask a question rather than making a case.

Key points

Almost every treatment worth asking about has a public paper trail, and the databases that hold it are free to anyone.
Study type matters more than study count. Ten mouse studies do not outweigh one good randomized trial in humans.
An abstract answers five questions: who was studied, how many, for how long, compared to what, and what actually improved.
Bring one page, not a folder. Specific questions open conversations; stacks of printouts tend to close them.
If the only enthusiastic sources are also selling the thing, you have left research and entered marketing.
'We don't know yet' is a real finding, and it is often the honest answer for newer treatments.

Why this skill matters

A doctor has roughly fifteen minutes with you and a few hundred other patients to think about. You have one body and all the time you are willing to give it. That asymmetry is not a scandal and it is not anyone's fault. It is just the shape of the system, and it means that the person most likely to notice that a treatment exists for your specific situation is you.

Being that person well is a learnable skill. It is not the same as self-diagnosing, and it is not the same as deciding you know more than your clinician. It means arriving at an appointment having already done the reading, able to ask a precise question, and able to tell the difference between something with real evidence behind it and something with a good website.

Step one: turn your hope into a searchable question

Vague searches return junk. Before you open a database, write your question in a specific form: this treatment, for this condition, in people like me, compared to what I am doing now.

So instead of 'does creatine help the brain', you write 'creatine supplementation for cognitive function in healthy adults'. Instead of 'peptide for gut healing', you write the actual name of the substance plus the actual condition. Instead of 'natural thyroid', you write the specific preparation and the specific diagnosis.

This forces two useful things. It forces you to name what you actually want to happen, which is usually a symptom you want reduced or a number you want moved. And it gives you search terms that will match the language researchers use, which is how you find the studies instead of finding the ads.

Step two: get the plain-language background first

Start with MedlinePlus, which is run by the National Library of Medicine. It is written for patients, it covers drugs, supplements, conditions, and lab tests, and it will not try to sell you anything. If a treatment is well established, MedlinePlus will say so in ordinary English. If it barely appears there, that itself is information.

For dietary supplements specifically, the NIH Office of Dietary Supplements publishes fact sheets in two versions, one for consumers and one for health professionals. The professional version is more useful than it looks, even if some of it goes over your head. Skim it for the phrases that tell you where things stand: 'evidence is limited', 'results are mixed', 'several randomized controlled trials have found'.

The point of this step is orientation. You want to know, before you go deeper, whether you are researching something mainstream that your doctor simply did not think to mention, or something genuinely experimental.

Step three: find out whether it has actually been studied

Now go to PubMed, the free index of biomedical literature maintained by the National Library of Medicine. Type your specific question in as plain terms. You do not need special syntax to get started. Two habits make a large difference right away.

First, use the filters in the sidebar. Filter to 'Clinical Trial', 'Randomized Controlled Trial', 'Meta-Analysis', or 'Systematic Review' and see what survives. If nothing survives, that is a finding: nobody has run a proper human trial on this yet. Second, sort by date and look at the last five years, then sort by relevance and look at the most-cited work. You want both the current state and the foundational studies.

Next, check ClinicalTrials.gov, the registry of clinical studies run by the National Institutes of Health. It tells you what is being tested right now, by whom, on which conditions, and whether any of it is recruiting near you. A treatment with several active trials for your condition is in a very different position from one with none. Trial listings also tell you honestly what researchers are still unsure about, because the outcome they chose to measure is the thing they do not yet know.

Finally, look for a systematic review in the Cochrane Library. A systematic review gathers every decent study on a question and weighs them together, which is much more trustworthy than any single result. Cochrane reviews include plain-language summaries near the top. If one exists for your question, read that summary before anything else. It may save you a week.

The evidence ladder, from weakest to strongest

This is the part that separates an informed patient from a worried one with a browser. Every result you find sits somewhere on a ladder, and where it sits determines how much weight it can carry.

Most exciting health headlines come from the bottom three rungs. That does not make them worthless. Early work is how everything starts. It makes them preliminary, and preliminary is a fair thing to say out loud.

Test tube or cell studies: a substance did something to cells in a dish. Interesting, and very far from you. Cells in a dish are not a person with a liver.
Animal studies: it worked in mice or rats. Most things that work in mice do not go on to work in humans. Look for the word 'mice' in the abstract before you get excited.
Case reports and case series: one patient, or a handful, got better. No comparison group, so you cannot know whether the treatment did it.
Observational studies: researchers watched people who happened to take the thing and compared them to people who did not. Useful for generating questions, weak for proving cause, because the two groups usually differ in other ways too.
Randomized controlled trials: people were assigned by chance to the treatment or to a comparison, which is the only reliable way to isolate what the treatment itself did. This is the rung that matters most.
Systematic reviews and meta-analyses: all the decent trials, gathered and weighed together. The strongest single thing you can cite.

How to read an abstract in five questions

You do not need to understand the statistics to get real value from an abstract. You need to interrogate it. Ask these five questions in order and write down the answers.

Who was studied? Look for age, sex, health status, and diagnosis. A trial in 25-year-old athletes tells you little about a 62-year-old with kidney disease. How many people? Twelve is a pilot. Two hundred is a study. Two thousand is a serious one. How long did it run? Six weeks tells you almost nothing about a chronic condition you will live with for decades. Compared to what? If there was no comparison group, or the comparison was nothing at all, the result is much softer than it sounds. What actually improved? This is the one people skip. There is a large difference between 'fewer heart attacks' and 'a lab value moved in a direction we think is good'. The second is called a surrogate marker, and surrogate markers have a long history of improving while patients do not.

Turn it into one page

Now compress everything into a single page you can hand across a desk. This is the step most people skip, and it is the step that determines whether the conversation goes well.

Your page should have five things on it: the exact name of the treatment, what you are hoping it would do for you specifically, the two or three strongest studies you found with the journal and year, what those studies actually showed in one sentence each, and the question you want answered. Nothing else. No printouts of full papers, no forum screenshots, no videos.

Keeping it to a page does something useful to you as well as to your doctor. If you cannot summarize the evidence in a paragraph, you probably have not understood it yet, and the summarizing is where you find that out.

How to bring it up without closing the conversation

Tone decides the outcome here more than content does. A doctor who feels quizzed gets defensive. A doctor who feels consulted usually engages, because most of them went into this work for exactly that.

Say the appointment out loud at the start: 'I read about something and I want your honest read on whether it makes sense for me. I am not asking you to prescribe it today.' That sentence lowers the stakes for both of you.

Then ask a real question rather than presenting a case. Good ones sound like: 'Is this something you have seen used for my situation?' 'What would make you consider it, and what would rule it out for me?' 'Is there a reason it would be a bad idea with the medications I am already on?' 'If not this, is there something in the same family you would consider first?'

And ask for the reasoning behind a no, not just the no. 'Can you help me understand what makes it a bad fit?' is not a challenge. It is the question that turns a dead end into information you can use.

How to tell you have wandered out of research

There is a line between informed patient and mark, and it is crossed quietly. Watch for these signals in the sources you are relying on.

The strongest signal is simple: are the people who are most enthusiastic about this also the people selling it? If every positive account traces back to a clinic, a supplement company, or an affiliate link, you are reading advertising with citations attached.

Notice, too, what happens to your own thinking. Research makes you more precise and usually more uncertain. Marketing makes you more certain and more urgent. If a week of reading has left you completely convinced and in a hurry, step back and look at what you actually read.

Testimonials stand in for trials, and the stories are always dramatic and never detailed.
The claim covers many unrelated conditions at once. Real treatments have narrow indications.
The seller explains away the absence of evidence with a story about suppression rather than pointing to the trials.
Urgency is manufactured: limited supply, prices about to rise, a window that is closing.
The named studies, when you look them up, turn out to be in animals, tiny, or on a different question entirely.
You are told to avoid telling your doctor, or that your doctor will not understand.

Knowing when to stop

Research can become its own anxious hobby. You are done when you can answer three things: whether this has been tested in people like you, what the best available evidence actually showed, and what the specific risks and interactions are for your situation. Once you have those, more reading tends to add worry rather than knowledge.

Some of the time, the honest answer you arrive at is that nobody knows yet. That is not a failure of your search. It is the current state of the field, and knowing it is worth far more than a confident answer you assembled from the wrong kind of study.

Bodies are worth the trouble of paying attention to. Doing this carefully is a form of stewardship, not suspicion, and it works best when it stays on that side of the line.

Common questions

I cannot get past the paywall on a study. What now?

The abstract is free on PubMed and usually contains the five things you need. Many papers are also free in full through PubMed Central, which PubMed links to directly when a free version exists. Public and university libraries can often get you the rest, and authors will frequently email you a copy if you ask politely.

Is it wrong to look things up before an appointment?

No. The problem is never that a patient read something. The problem is arriving with a conclusion instead of a question. Doctors generally welcome a prepared patient and dread a patient who has already decided what they are going to be prescribed.

What if I find a study that contradicts my doctor?

Bring it, and bring it as a question. One study rarely overturns clinical practice, and there may be later or larger work you have not seen, or a reason specific to your history. Ask what they make of it. If the reasoning does not hold together, that is a legitimate reason to seek a second opinion.

How much weight should I give a preprint?

Less than a published study. A preprint is a manuscript posted publicly before peer review, which means no independent reviewer has looked for its errors yet. Preprints are useful for seeing where a field is heading. They are not a basis for a treatment decision.

Does a trial being listed on ClinicalTrials.gov mean the treatment works?

No. Registration means a study is planned or running, not that it succeeded. Many registered trials find nothing. Look for whether results have been posted, and whether the trial finished or was terminated early, which is itself worth noticing.

Sources

This is information, not medical advice. Strength from the Well is an independent publisher. Nothing here is a diagnosis, a prescription, or a recommendation to start or stop any treatment. Talk to a clinician who knows your history before you act on anything you read here — including anything you read here that contradicts them.