LEVEL Real-World Evidence
Real-world evidence is what happens when people use a product in their own lives, tracked day by day over weeks. It matters to us because that is where a cannabinoid's real effect shows up.
LEVEL has run four human studies across more than a thousand participants, published one in a peer-reviewed journal, and partnered with university researchers on another. Some confirmed what we expected. One did not find what we set out to measure, but pointed us somewhere new. Every one of them taught us something about how cannabinoids actually work.
An industry built on stories, not data
Cannabis spent fifty-one years as a Schedule I substance, so the research infrastructure other wellness categories take for granted was never built. What circulated instead was anecdote, and anecdote scaled faster than science ever could. The legal hemp market inherited that gap. Hemp products are regulated as supplements, so nothing requires a company to demonstrate that a formulation does what its label implies before it goes on sale.
Real-world evidence is the practical way out, and regulators have spent a decade building the framework for it, from the 21st Century Cures Act to the FDA's 2024 guidance on real-world data and on trials run remotely. It also suits how cannabinoids behave. They modulate how the whole system regulates itself, so effects build across days rather than arriving at once.
LEVEL's real-world studies
Every study below was run with an independent research partner, and each one states its design, its sample size, its findings, and its limitations exactly as they were published.
Real-world observational study
Sleep Protab
Baseline
Day 29
Nightly awakenings2.84 → 2.14
Woke feeling refreshed19% → 63%
+41minutes of sleep per night
Findings and limitation
Read the study▾
Participants averaged forty-one additional minutes of sleep per night by day twenty-nine. Nightly awakenings fell roughly twenty-five percent, from 2.84 to 2.14. The probability of waking up feeling refreshed rose from nineteen percent to sixty-three percent.
Consistency of use was the only significant moderator of outcomes. Dose was not. Taking the tablet every night mattered more than taking more of it.
Observational, no placebo group. It describes what participants experienced rather than proving the formulation caused it.
Real-world observational study
Relief Protab
Baseline
Day 29
Reported pain, ten-point scale5.26 → 3.56
Days pain derailed the day34% → 22%
80%said Relief Protab reduced their daily pain
Findings and limitation
Read the study▾
Reported pain fell about thirty-two percent within hours of each dose, from 5.26 to 3.56 on a ten-point scale. The share of days when pain derailed the day dropped from thirty-four percent to twenty-two percent. Validated well-being scores moved with it, WHO-5 rising from 46.41 to 50.16.
Observational, no placebo group, and participants knew what they were taking.
Randomized, triple-blind, placebo-controlled
CBG Veterans Trial
CBG group, primary sleep measureImproved
Placebo groupImproved
Schematic. Both groups improved and the gap between them was not statistically significant, which is the finding.
Findings and limitation
Read the study▾
Schematic. Both groups improved and the gap between them was not statistically significant, which is the finding.
CBG did not separate from placebo on the primary sleep measure. Both groups improved, and the difference between them was not statistically significant. CBG was well tolerated at both doses. One exploratory signal held: participants who dosed in the afternoon showed a lower resting heart rate two hours after taking it than placebo.
A null result is still a result, and this field has too few of them on the record. Publishing only the studies that work is how a category ends up with a literature nobody can trust. This one is peer-reviewed, and it is the first placebo-controlled human CBG trial anywhere.
Small completed sample, and the trial was designed before we understood how much dose timing might matter. The afternoon heart-rate finding is exploratory and needs its own study.
Randomized head-to-head, no placebo arm
University of Michigan Chronic Pain Study
Formulation one
Formulation two
Acidic cannabinoid formulation
Schematic. Symptom improvement was close to identical across the three, in a market that sells on formulation differences.
Findings and limitation
Read the study▾
Schematic. Symptom improvement was close to identical across the three, in a market that sells on formulation differences.
Across fibromyalgia, rheumatoid arthritis, and knee or hip osteoarthritis, participants improved in nearly every symptom category measured, with cognitive function the exception. The acidic cannabinoid formulation showed specific reductions in neuropathic pain intensity. Overall the three formulations performed remarkably similarly, which is itself useful in a market that sells on formulation differences.
No placebo arm, so the comparison is between formulations rather than against no treatment.
The scientific record, in the order it happened
Almost everything we understand about how these compounds work was learned inside a single human lifetime, and most of it inside the last thirty-five years.
The compounds are pulled apart and named, long before anyone knows what they act on.
Wood, Spivey, and Easterfield separate cannabinol from hemp resin at Cambridge. Their molecular formula is wrong, and it takes until 1899 to purify it.
J Chem Soc 1896;69:539-546
Roger Adams pulls cannabidiol from wild hemp at Illinois. In separate work the same year, the structure of CBN is finally settled.
J Am Chem Soc 1940;62(1):196-200
Mechoulam and Shvo at the Weizmann Institute publish the correct structure, opening modern cannabinoid chemistry.
Tetrahedron 1963;19:2073-2078
Gaoni and Mechoulam isolate delta-9-THC and establish its structure. The same team characterizes CBG that year.
J Am Chem Soc 1964;86(8):1646-1647
Paton and Pertwee at Oxford show cannabis constituents prolong barbiturate sleeping time in mice. Most later CBN sedation folklore traces here, to a study that was never about human sleep.
Br J Pharmacol 1972;44(2):250-261
Receptors, the body's own cannabinoids, and the enzymes that clear them.
Lisa Matsuda's team at NIH clones the first cannabinoid receptor, turning a pharmacological inference into a defined molecular target.
Nature 1990;346:561-564
Devane, Hanus, and Mechoulam isolate a brain lipid that binds the receptor. The body makes its own cannabinoids.
Science 1992;258:1946-1949
Munro's group in Cambridge finds a second receptor, mostly in immune tissue, extending the system well beyond the brain.
Nature 1993;365:61-65
Taura, Morimoto, and Shoyama establish CBGa as the precursor from which THCa, CBDa, and CBCa are each built. The mother cannabinoid framing belongs here, to CBGa.
J Biol Chem 1996;271(29):17411-17416
Cravatt's team at Scripps characterizes the enzyme that breaks anandamide down, creating the off switch that CBD is later shown to inhibit.
Nature 1996;384:83-87
Bisogno and Di Marzo demonstrate CBD activates TRPV1 and blocks anandamide breakdown. It does not bind the classical receptors.
Br J Pharmacol 2001;134(4):845-852
Russo's group at Montana shows CBD acts at a serotonin receptor, a route to mood and stress effects with no cannabinoid receptor involved.
Neurochem Res 2005;30(8):1037-1043
Takeda at Kyushu shows CBDa selectively inhibits COX-2, a mechanism CBD does not share. CBDa is not weaker CBD.
Drug Metab Dispos 2008;36(9):1917-1921
O'Sullivan at Nottingham finds CBD's slower effects depend on a nuclear receptor, adding a third distinct pathway.
Eur J Pharmacol 2009;612(1-3):61-68
Leweke's team in Cologne finds CBD raises serum anandamide, and the rise tracks with improvement. CBD works by enhancing the body's own tone.
Transl Psychiatry 2012;2:e94
Regulation, legalization, and the first human trials worth citing.
The 21st Century Cures Act directs the FDA to evaluate registries and patient-reported outcomes as regulatory evidence.
Public Law 114-255
A NASEM committee reviews over ten thousand abstracts, finding substantial support in a few areas and large gaps nearly everywhere else.
NASEM, doi 10.17226/24625
The FDA approves plant-derived CBD as a prescription drug in June. In December, hemp comes off the Controlled Substances Act and the FDA publishes its real-world evidence framework.
FDA; Public Law 115-334
Chris Emerson, who founded LEVEL in 2016, files the priority application for a controlled-release cannabinoid tablet. It grants as two US patents, in 2023 and 2024.
US 11,596,606 B2; US 12,076,442 B1
In 293 adults, seven nights of 20 mg CBN reduced awakenings and sleep disturbance versus placebo. Sleep onset did not improve, and adding CBD did not help.
Exp Clin Psychopharmacol 2024;32(3):277-284
Across six trials and roughly 1,077 adults, THC or CBN improved subjective sleep versus placebo. CBD alone did not. Different cannabinoids are doing different jobs.
Sleep Med Rev 2025, PMID 40929927
LEVEL's decentralized triple-blind trial in US veterans publishes. CBG did not separate from placebo on the primary sleep measure, but did lower resting heart rate for those taking it in the afternoon.
Med Cannabis Cannabinoids 2026;9(1):1-14
The first polysomnography trial of CBN in diagnosed insomnia misses its primary endpoint. The higher dose shortens sleep onset and improves subjective quality. The picture stays complicated.
J Sleep Res 2026, doi 10.1111/jsr.70284
Most of the entries above are younger than the people reading them. That is not a reason to dismiss the field. It is a reason to keep measuring.
Join the next study
The next round of real-world evidence needs participants. Add your name and we will contact you when a study opens that you may be eligible for. Studies typically run a few weeks, happen entirely from home, and involve short daily questions. Nothing you are invited to is an obligation.
What each kind of study can and cannot tell you
Those four studies use three different designs, and the differences matter. Here is what each one does well, and where it stops.
| Study design | What it does well | What it cannot tell you | Where LEVEL uses it |
|---|---|---|---|
| Randomized, placebo-controlled trial | Isolates cause. If the treatment group separates from placebo, the formulation did it. | Little about real life. Small samples, tight eligibility, short windows. | CBG veterans trial |
| Real-world observational study | Captures ordinary use, at scale, over weeks, with people managing their own routines. | Cannot prove cause. Without a placebo group, expectation is in the result. | Sleep and Relief Protab studies |
| Head-to-head comparison | Shows how formulations differ from each other in the same population. | Nothing about whether either beats no treatment. | University of Michigan pain study |
| Patient registry | Follows outcomes over long periods and diverse populations, generating hypotheses worth testing properly. | Not built to answer a single question definitively. | Ongoing evidence registry |
Where to look next
Every study above has a full write-up with the methods, the measures, and the numbers we did not lead with. Beyond our own work, the blog holds more than thirty pieces on what the outside literature says. Every batch we make is third-party tested and the certificates are posted publicly, so you can check the specific bottle in your hand. The work above was designed and run by Chris Emerson, PhD, a small molecule chemist and the inventor on both LEVEL patents.
Run your own experiment.
Conduct your own experiment by trying our Discovery Kit featuring six unique formulations. Try one, keep a journal, and identify the perfect formulations for you to maximize your wellness.
What is real-world evidence, and why does it matter for cannabinoids?▾
Data collected during ordinary use rather than in a clinic. Participants take a product at home, on their own schedule, and report how things go over weeks. It matters here because cannabis research was restricted for decades, so the category has very little data of any kind, and because cannabinoids build effects over time rather than producing one dramatic dose response. A short lab visit measures that poorly. A month in someone's own bedroom measures it well.
Why publish a study that found no difference from placebo?▾
Because a result pointing the wrong way is still information, and a field where only successful studies get published has a literature nobody can rely on. Our CBG trial in veterans did not separate from placebo on its primary sleep measure. It is peer-reviewed, it is the first placebo-controlled human CBG trial anywhere, and it is on this page at the same weight as the studies that went well.
Are LEVEL's studies peer-reviewed?▾
Two of them. The CBG veterans trial is published in Medical Cannabis and Cannabinoids, and the University of Michigan chronic pain study in Clinical Therapeutics. The Sleep and Relief Protab studies were run by MoreBetter as real-world observational studies. Those two are not peer-reviewed publications, and we do not describe them as clinical trials.
How can I take part in a LEVEL study?▾
Add your name to the research list on this page. When a study opens we contact people on that list with the details, the time commitment, and the eligibility requirements. Studies are usually a few weeks and run entirely from home. Being on the list commits you to nothing.
Where can I see the test results for what I bought?▾
Every batch is third-party tested and the certificates are posted at /pages/test-results, organized by formulation. Match the batch number on your bottle to the certificate.