Why Most Herbal Studies Are Small: Causes and Fixes 2026

If you have read a herbal study and counted the participants, the number usually looks thin. That is not an accident of a single bad lab. Herbal research is small because grant cycles are short, plants are hard to standardise, recruitment fails more often than anyone plans for, and many programmes never get past the exploratory stage. Small does not mean useless, but it does mean the study cannot carry the weight people put on it.

Reading across the phytotherapy literature for this site, the same handful of explanations keep showing up in review after review. They are structural rather than glamorous: money, logistics, and the basic difficulty of running a controlled test on a plant. Getting those straight helps more than any single herb does.

This guide covers where the small numbers come from, what small studies can and cannot tell us, and what a reader can check in a paper before repeating a claim as fact. It is research literacy, not medical advice. Nothing here is a recommendation to start, stop or change any treatment.

Why Most Herbal Studies Are Small

Most herbal studies stay small because herbal research is chronically underfunded, hard to recruit into, and constrained by the fact that a plant cannot be standardised as precisely as a single-molecule drug. Researchers also often start small on purpose, with pilot work, before committing to a full trial. Small studies still contribute useful evidence, but they cannot answer every question asked of them.

Why Most Herbal Studies Stay Small

The reasons stack on top of each other. Here are the ones that come up most often.

1. Short grant cycles cap the number. A typical university grant funds three years. Enrolling participants, screening them, following them and analysing the results eats most of that window. Trials are routinely designed around the grant length rather than the sample size needed, so the sample is whatever fits before the money runs out.

2. Recruitment fails more often than anyone expects. This is the big one, and it is not specific to herbs. An analysis of more than 700 trials run through the US National Cancer Institute’s CTEP network found that 81.5% missed their projected accrual goals and 37.2% never reached minimum enrolment. Herbal trials inherit the same problem, and they add a layer: participants are being asked to swallow something unfamiliar, sometimes alongside cancer treatment, and many decline or withdraw.

3. The eligible group is narrower than it looks. Once you exclude people with drug interactions, liver or kidney concerns, pregnancy, or therapies that cannot safely run alongside a herb, the pool of possible participants shrinks fast. In a lung cancer cohort, a 2021 study in Annals of Palliative Medicine found 46.2% of patients were willing to join a herbal trial. Willingness is not enrolment, and the gap between them is where sample sizes go to die.

4. Standardising botanical material is expensive. A single-molecule drug is the same batch every time. A plant is not. Chemical content shifts with soil, rainfall, harvest month, storage and plant part used. Building a preparation that holds its composition across every bottle in a multi-site trial requires botanical expertise, vouchered specimens and batch testing, and that work sits on the same tight budget.

5. Most programmes never leave the bench. Research on medicinal plants moves from ethnobotanical surveys to laboratory work on extracts and mechanisms, then to animal studies, and only sometimes to human trials. The human stage is last and most expensive, so many plants never get there at all. What reaches publication is the early exploratory work, which is small by stage rather than by quality.

How Funding and Research Priorities Affect Sample Size

The funding question is worth sitting with, because it explains more than any methodological complaint. A pharmaceutical company testing a new molecule expects to sell it, so it can fund a trial with thousands of participants across dozens of sites. There is no equivalent commercial return on a plant that grows in the wild and cannot be patented cleanly, so the money has to come from somewhere else.

That means public and philanthropic grants, which run on fixed cycles and reward investigator effort more than team scale. One principal investigator with a small grant runs a modest single-site trial. A multicentre study needs coordinating centres, data management, monitoring, statistician time and training across sites, which is exactly the part grants cut first.

The fix for any individual study matters less than the pattern it sits in. Researchers are rewarded for publishing, and publishing favours a completed study over an ambitious one that never recruits. Studies that finish, however small, go to press. Studies that recruit badly get quietly abandoned, which is why the published literature over-represents the trials that worked out.

One distinction is worth holding onto. An underpowered study claims more than its sample can support: it looks for a benefit, fails to find one clearly, and cannot rule out a real effect. A pilot or feasibility study claims only what it can: does this recruit, does the preparation hold up, does the outcome measure make sense. Small samples are legitimate in the second case. Reviewers and readers should be checking which kind they are looking at before anything else.

Study type	What it is for	Typical participants
Pilot study	Checking feasibility, recruitment and measurement	Often 10 to 30
Feasibility study	Testing procedures and estimating recruitment rates	Usually several dozen
Single-centre efficacy trial	Testing whether a preparation works better than placebo	Commonly 50 to 200
Multicentre confirmatory trial	Confirming an effect across sites and populations	Often several hundred upward

Small Herbal Trials Compared With Large Drug Trials

Seeing the gap side by side makes the scale of it obvious. Read it as typical rather than universal; there are exceptions in both columns.

Feature	Small herbal trial	Large pharmaceutical trial
Participants	Tens, occasionally a few hundred	Thousands
Duration	Weeks to a few months	Months to years
Preparation	A plant or plant extract, often not fully characterised	A single defined molecule at a fixed dose
Funding source	Public grants, universities, occasional industry	Manufacturer revenue
Multi-site	Uncommon	Standard
Detectable harm	Poor, because side effects are rare events	Good, because the numbers allow it

The last row is the one people skip. A trial of 60 people can detect a meaningful change in a common symptom such as fatigue. It cannot tell you whether a preparation raises liver enzymes in one person in two hundred. Small herbal trials are built to answer whether something does any good, not whether it does harm, and any summary that blurs those two questions is misleading.

What Makes African Herbal Research Difficult to Scale

African ethnomedicine research runs into the same money problem as everywhere, plus conditions that make a controlled trial genuinely harder to run.

Preparation is not one thing. What a community calls a remedy may be a decoction of one species in one region and a mixture of three in another, prepared from roots rather than leaves. A trial has to pick one, and every community that prepared it differently has to be treated as a separate intervention rather than folded into the same result.

Seasons and plant parts matter. The right harvest window can be weeks long. Harvesting outside it changes the chemistry, and a study that cannot collect material on schedule either waits or compromises.

Species identity is often unresolved. Related species look alike, trade names overlap, and vernacular names shift between languages and districts. Misidentified material invalidates a trial no matter how many people enrolled, which is why herbarium vouchers and taxonomic verification are a cost line and not a formality.

Consent takes longer than a protocol allows. Community elders, traditional practitioners and individual participants may all need to agree, and some knowledge is meant to stay with the people who hold it. Research that treats consultation as a checkbox gets poor cooperation and bad data in equal measure.

Language access cuts both ways. Consent forms and result summaries have to exist in the languages people actually speak, and results published only in a national journal in one language can sit unread for years. Reviews from two decades ago already flagged herbal studies that were effectively invisible because they were not available in English, and the problem has not disappeared.

Does a Small Study Mean a Herbal Remedy Is Unproven or Unsafe?

No. Sample size is one dimension of evidence quality, not the whole of it, and treating it as the whole leads people to condemn remedies on weak grounds and endorse them on equally weak ones.

A better read asks what kind of study it was. Was it randomised, and was allocation concealed? Was the control a matched placebo, or an unequal comparison? Were outcome measures defined before data collection or invented afterwards? Were outcome assessors blinded? Were participants who dropped out counted in the analysis?

Then look at the numbers themselves. A confidence interval that runs from a large harm to a large benefit is not a result, whatever the headline says. Ask whether the effect size was specified in advance, whether the study was registered before it began, whether the funding source is disclosed, and whether anyone has tried to repeat it. The people who quote a confidence interval rather than saying studies show are usually reasoning better than the press release they read.

What small studies cannot do is establish safety. If a preparation shows a plausible interaction risk in laboratory work, a 60-person trial designed to measure symptom improvement will almost never surface it. Absence of reported harm in a small herbal trial means absence of data, not absence of harm. Take questions about interactions and side effects to a doctor or pharmacist rather than to a study abstract.

What Can Readers Check Before Trusting an Herbal Study

Here is the appraisal I run through any herbal paper, in about five minutes.

1. What was the question? A clear question names the preparation, the population, the comparison and the outcome. Vague framing in the abstract usually predicts vague design underneath.

2. How many people, and how many finished? Count both. A trial that enrolled 90 and analysed 54 tells you something about retention that the headline will not.

3. What was it compared against? Placebo, standard care or nothing at all produce very different results, and only the first two let you attribute an effect to the preparation.

4. Was it randomised and blinded, and was it preregistered? Registration is checkable. If the outcome measures changed after enrolment started, that is worth knowing before you repeat the result.

5. Was the plant actually pinned down? Species name with an authority, plant part, extraction method and a voucher specimen. If the paper says only a common name or a brand, the thing tested may not be the thing on any shelf.

6. What were the limits, who paid, and does it agree with the rest of the field? Authors who list their own limitations are usually the ones whose conclusions you can trust. One small positive result that conflicts with everything else is a reason to wait, not a reason to start.

The pattern that worries me most is a single small trial with a dramatic headline and no follow-up. Real communities notice this, and calling a cinnamon bark trial a pilot rather than a breakthrough is a healthier habit than most press coverage manages.

How Larger and More Reliable Herbal Studies Can Be Built

Scale is not the only fix, and most herbs will never see a 5,000-person trial. That is fine. The improvements below are achievable within ordinary grant budgets.

Register the study before it starts. Preregistration fixes the outcomes in advance and gives later reviewers something concrete to check against.

Run pilots deliberately. A pilot that measures recruitment rates, retention and preparation consistency answers real questions, and its samples then inform a realistic power calculation instead of a hopeful one.

Document the plant properly. Voucher specimens, taxonomic verification, plant part, batch chemistry and a stated preparation method. Without this, a negative result is uninterpretable because nobody can tell what was tested.

Build community partnership into the protocol. In African settings this means funding local researchers and practitioners as collaborators rather than data sources, translating materials properly, and respecting limits on traditional knowledge. It also improves recruitment, which is the practical bottleneck.

Report completely. Follow CONSORT for randomised trials and register botanical details alongside the clinical ones. A well-reported small study is far more useful than a large one that reports only its headline.

Power trials for the question asked. If the real interest is a symptom that only 35% of patients have, the eligible pool and the required sample are both much larger than intuition suggests. Where a definitive trial is impossible, systematic review of several small studies can still give a defensible answer, which is why trial registries and honest negative-result publication matter so much.

Frequently Asked Questions

What to Check First in Any Herbal Study

Start with the design, not the headline. Ask what kind of study it was, what it was compared against, how many people finished, whether the plant was properly identified, who funded it and whether anyone has repeated it. A small sample is one signal among several, and reading the rest of the signals tells you far more than the participant count alone.

And if the question is whether something is safe for you specifically, take that to a doctor or pharmacist. No paper, small or large, can account for your other medicines, your health history or your liver and kidneys.