Evidence and uncertainty
Is functional medicine evidence based?Judge the model and each treatment separately.
There is published research on functional medicine care, but it does not prove that every test, supplement or theory used under the label is effective. Evidence needs to be separated by question, treatment and study design.
There is research on functional medicine, but the evidence is not one simple yes-or-no result.
Some studies have examined functional medicine care models and reported improvements in patient-reported outcomes. However, much of the published whole-model evidence is observational, comes from a small number of centres, or tests one component such as health coaching rather than proving that every part of “functional medicine” works.
The most accurate way to judge the evidence is to separate three questions: does the overall care model improve outcomes, does the specific treatment being proposed work for the condition in question, and is the test or treatment safe and useful for this individual?
Studies of a functional medicine clinic or programme can tell us something about the package of care, but they do not prove that every test, supplement or theory used within that package is effective.
Diet, exercise, medicines, psychological care and supplements each need their own evidence.
Testimonials can describe how care felt. They cannot establish treatment effectiveness.
What studies have tested the functional medicine model?
Several frequently cited studies are useful, but each answers a narrower question than “does functional medicine work?”
| Study | What it found | What it cannot prove |
|---|---|---|
| JAMA Network Open, 2019 | A retrospective propensity-matched comparison found greater improvement in a patient-reported physical-health score at 6 months for patients seen at a Cleveland Clinic functional medicine centre than matched family health centre patients. [1] | The study was observational. Differences in patients, care intensity or other unmeasured factors can affect results. |
| BMJ Open, 2021 | A retrospective single-centre study compared shared and individual functional medicine appointments and found small differences in patient-reported outcomes at 3 months. [2] | It did not compare functional medicine with usual NHS care and does not establish which components caused the changes. |
| Medicine, 2024 | A randomised trial compared an elimination diet with or without functional medicine health coaching. The main overall comparison did not show a between-group difference, although some subgroup results favoured coaching. [3] | It tested coaching added to one dietary programme, not the entire functional medicine model or every population. |
Why does study design matter?
A randomised controlled trial can reduce some forms of bias when comparing treatments. An observational study can show what happened in real clinical practice, but it cannot fully separate the effect of the treatment from differences between the people who chose or received it.
Patient-reported outcomes are important, but they are not the same as showing a lower risk of hospital admission, fewer complications or longer life. A study can also be statistically significant while the real-world size of the benefit is small.
Funding, author affiliations, setting and whether results have been reproduced by independent teams also matter. Evidence becomes stronger when different methods and groups point in the same direction.
Evidence for a component is not evidence for the whole package.
Functional medicine consultations often include familiar parts of healthcare: medical history, medication review, diet, exercise, sleep, stress, standard diagnostic testing and referral when needed. Many of those components have their own established evidence in specific conditions.
Other proposals may be much less certain. A specialist laboratory panel, supplement combination or restrictive diet should not inherit credibility simply because it sits inside a longer medical consultation. The evidence for that exact test or treatment still needs to be assessed.
That is why the definition and limits of functional medicine on this site separates the care model from individual treatment claims. The functional medicine service page describes the doctor-led practice scope without claiming that the label proves an outcome.
How strong is a patient testimonial as evidence?
A review can tell you whether a patient felt listened to, found the process clear or liked the way care was organised. Those are legitimate experience signals. A testimonial cannot tell you how many people improve, whether the improvement was caused by a treatment, or whether the same result would occur in another patient.
This distinction is especially important in healthcare. People may improve naturally over time, receive several treatments together, or have expectations that affect how they report symptoms. That is why clinical effectiveness needs controlled research, not only positive reviews.
Reviews answer a different question
Use reviews to judge the consultation experience and communication. Use clinical research and authoritative guidance to judge treatment claims.
What should an evidence-based functional medicine decision look like?
A doctor should be able to explain the aim of a proposed test or treatment, the evidence that supports it, the important uncertainties, likely harms or burdens, alternatives and what would happen if you chose not to proceed. GMC and NICE guidance both place this kind of balanced discussion at the centre of shared decision making. [4] [5]
Evidence also needs to match the patient. A finding from a small study in one population may not apply to someone with a different diagnosis, medicine list or risk profile. The safest approach is to use the strongest evidence available for each decision while being explicit where evidence is limited.
How to read a study without getting lost
Start with one question: what did the study actually test? A study may test a clinic, a diet, a coaching plan or one group of patients. It does not test every idea that may sit under the same label.
Then look at who took part. Were they like you? Were they people with one clear condition or a mixed group with many health problems? A result in one group may not apply to all patients.
Look at what the study compared. A trial that compares two active plans answers a different question from a study that compares care with no added treatment. An observational study can be useful, but the people who choose one type of care may differ from those who choose another.
Next, look at the outcome. Did people report that they felt better? Did a lab value change? Did the study show fewer hospital stays or less disease? These outcomes are not the same. Each can matter, but the claim should match what was measured.
Finally, look for limits. Was the group small? Was the study done at one centre? Was it short? Has another team found the same result? A fair evidence summary should make those limits easy to see.
What does this mean for a patient?
You do not need to grade every paper yourself. Ask the clinician to explain the strength of the evidence in plain language. Ask what is known, what is not known and what choice you have.
If a treatment has good evidence for your condition, that should be clear. If evidence is weak or early, that should be clear too. The same rule applies to tests. A new or specialist test is not more useful just because it is more complex or costs more.
It is also reasonable to ask what happens if you choose not to proceed. NICE shared decision-making guidance treats “no change” as an option that may need to be discussed alongside treatment choices. [5]
The aim is not to demand perfect proof for every decision. It is to match the confidence of the claim to the quality of the evidence, the risks of the choice and your own goals.
What patients say about the consultation experience.
Patient reviews can help you understand how a consultation feels. They do not prove that a particular diagnosis will be found or that a treatment will work.
Individual reviews reflect personal experiences and do not predict diagnosis, treatment response or outcome.
“takes incredible care to get to know her patients”
Google reviewer
“patient, understanding and kind”
Google reviewer
“warm and compassionate bedside manner”
Google reviewer
Common questions
Is functional medicine evidence based?
Some parts of functional medicine use well-established medical and lifestyle methods, and several studies have examined functional medicine care models. However, whole-model evidence remains limited and much of it is observational. Each proposed test, treatment, diet or supplement should therefore be judged on its own evidence rather than being assumed effective because it is labelled functional medicine.
What studies test the whole functional medicine model?
Published studies include retrospective Cleveland Clinic comparisons of patient-reported outcomes and a smaller number of trials testing individual components. Observational studies can suggest associations but cannot prove causation. Trials that test coaching or a specific programme do not automatically validate every element used across functional medicine practice. [1]
Did the 2024 functional medicine coaching trial prove functional medicine works?
No. The 2024 randomised trial tested an elimination diet with or without functional medicine health coaching. The main overall between-group analysis did not show a difference, although some subgroup findings favoured coaching. It did not test the whole functional medicine model or all patient groups. [3]
Are patient testimonials clinical evidence?
No. Testimonials are useful for understanding communication, trust and patient experience, but they cannot show treatment effectiveness or how often an outcome occurs. Clinical claims need research designed to reduce bias and compare outcomes, not only reports from people who chose to leave a review.
How should I judge evidence for a specialist test or supplement?
Ask what clinical question it answers, whether the test or treatment has been validated for that use, what benefit is expected, what the important harms or uncertainties are, and what would change after the result. Do not assume evidence for the overall consultation proves every specialist test or supplement is useful.
What if the evidence is uncertain?
Uncertainty should be stated openly. Shared decision-making guidance supports discussing benefits, harms, alternatives and the option of no change. A reasonable plan may be to use a better-established option, monitor, seek specialist advice, or decide that the uncertain intervention is not worth its cost or burden. [5]
Medical sources and further reading
- Beidelschies et al., JAMA Network Open 2019. A retrospective propensity-matched cohort compared patient-reported quality-of-life outcomes in a functional medicine centre with family health centre care; it was observational and cannot establish causation.
- Beidelschies et al., BMJ Open 2021. A retrospective single-centre study compared shared and individual functional medicine appointments and reported small differences in patient-reported outcomes and delivery cost.
- D'Adamo et al., Medicine 2024. A randomised trial of elimination diet with or without functional medicine health coaching found no between-group difference in the primary overall analysis, with some subgroup findings favouring coaching.
- GMC: Decision making and consent. GMC guidance requires balanced discussion of options, risks, benefits and uncertainty for medical decisions.
- NICE: Shared decision making. NICE recommends discussing the aims, risks, benefits and consequences of tests and treatments, including the option of no change.
- NICE NG197: Shared decision making. Recommends discussing benefits, risks, alternatives and the option of making no change where relevant.
Evidence is strongest when the claim is specific.
Ask what was studied, how it was studied, what changed, and whether that evidence applies to the decision in front of you.