By Justin Yamashita, MSc. Benchtop, site, CRO: three levels of basic and clinical research, explained without spin.

Plain Talk is Root to Rx's plain-language teardown: the myth, the evidence, and the one question that beats it.

In 1981, when the FDA wrote the rules for the boards that approve clinical trials, it said out loud that those boards have no obligation to check whether a study's data are any good, and that they "are generally not in a position to follow human subjects."

That’s the regulator describing the limits of its own safeguard, in the document that created it. Most people arguing about clinical trial oversight, in both directions, have never read that sentence.

Root to Rx is free, three times a week, written by a clinical researcher. Subscribe at RootToRx.com, or help us grow and send it to someone you know. This is a small business run by one person, so every share and comment goes a long way.

Monday, Coordinator Nani told a worried husband that she builds the safety record and never gets to read the ending. That’s not modesty and it’s not a dodge. Here’s the machinery underneath it.

Two Inspectors, Not One

Before anyone breaks ground on a building, a permit office reviews the plans. It looks at the design, the safety provisions, and who’s allowed on site, and it can approve, demand changes, or refuse outright. Nothing starts without that.

Once construction is underway, a second inspector shows up. This one doesn’t work for the builder. They can look at things the crew isn’t looking at, and if they see a real hazard, they can call it in and get the job halted.

In the US, a clinical trial runs on the same 2 inspector shape. Both are separate from the company running the study. Other countries use their own version of the first inspector, usually called an ethics committee, structured differently place to place. Everything below is the US system specifically.

Now here’s where the analogy breaks, and it breaks in 2 places worth knowing.

A building inspector checks a public code. Anyone with the right training could stand in that same spot and check the same things. The second inspector in a clinical trial is different in kind, because they’re the only people permitted to see the one thing that matters: how the treatment group compares to the placebo group while the study is still running. Not the doctors, not the coordinator, and not the sponsor, meaning whoever is paying for and running the study. Their power isn’t that they know the rules better than you. It’s that they’re the only ones allowed to look.

And a building inspector can red tag a site on their own authority. A monitoring board recommends, and then the sponsor acts on the recommendation. That gap is small in practice, but it isn’t nothing, and anybody who tells you a safety board can slam a trial shut by itself is overselling it.

Inspector One: The Board That Clears the Study for Launch

The first inspector is the Institutional Review Board, the IRB. Before a single person enrolls, it reviews the study's design, the informed consent form, patient facing materials like recruitment materials, and under federal regulation it has authority to approve the study, require changes before approving it, or disapprove it outright.

The composition is set in regulation too. An IRB must have at least 5 members, including a scientist, a nonscientist, and at least one person with no other affiliation to the institution running the research. Read that last one again. Federal regulation requires that somebody in the room does not work there. That’s a deliberate structural answer to the obvious failure mode, which is a hospital reviewing a hospital.

The IRB also reviews the study again while it runs, at intervals appropriate to the risk and not less than once a year.

Now the honest part, and it’s the FDA's own. An IRB is not a data monitor. It doesn’t watch the safety numbers accumulate. The FDA said so plainly when it wrote the rule: IRBs have no obligation to assure the validity and reliability of data, and they are generally not in a position to follow human subjects through a study. An IRB approval is a judgment that a study is ethically acceptable to start, not a promise about how it turns out.

Which is exactly why a second inspector exists.

Why the First Inspector Exists at All

For 40 years, nothing like an IRB was required.

From 1932 to 1972, the US Public Health Service ran the Tuskegee Study, following the progression of untreated syphilis in Black men in Macon County, Alabama, without informed consent. Issue 001 covered that one. In 1943, penicillin became the treatment of choice for syphilis and became widely available. The study didn’t offer it, and it kept going for another 29 years.

It ended because a journalist told the country. Associated Press reporter Jean Heller broke the story on 25 July 1972. The study was terminated that November, after a federally appointed advisory panel reviewed it.

Congress passed the National Research Act on 12 July 1974. That law created the requirement for IRB review of federally funded human research, and it established a national commission to work out what ethical research actually required. That commission published the Belmont Report on 18 April 1979: respect for persons, beneficence, justice. Those 3 principles are still the frame every IRB in the country works inside.

The rules kept moving after that, which is the part people miss. The modern federal human subjects regulation was revised as recently as 2019, and multi site studies now use a single board rather than one per site. This isn’t a system that was fixed once in 1979 and left alone.

Inspector Two: The Board That Watches It Run

The second inspector is the Data Safety Monitoring Board, also called a Data and Safety Monitoring Board (DSMB) or an Independent Data Monitoring Committee (IDMC). Statisticians and physicians with relevant expertise and no financial stake in the result.

3 things define it.

It sees what nobody else sees. At intervals set in advance, it reviews the unblinded data while everyone running the trial stays blinded. That’s the whole engine.

It works from a charter written in advance. The charter says how often the board meets, what it looks at, and where the stopping lines are, and those lines are set before the first person enrolls. Setting the threshold before you see the data is how you keep from setting it somewhere convenient later.

It recommends. Continue, change, or stop early, and the sponsor acts on it.

What That Looked Like in 2002

The Women's Health Initiative was testing whether combined hormone therapy, estrogen plus progestin, protected postmenopausal women against heart disease and other conditions. It enrolled 16,608 women aged 50 to 79 at 40 clinical centers, and it was designed to run 8.5 years.

At a regularly scheduled meeting on 31 May 2002, the monitoring board saw invasive breast cancer cross a boundary that had been set before the trial began, and judged that overall harms now outweighed benefits. The arm ended after a mean of 5.2 years, more than 3 years early. Participants started receiving letters on 8 July 2002 telling them to stop their study pills, and the results went public the next day.

The findings were genuinely mixed, and this is where most retellings go wrong in one direction or the other. Breast cancer, heart disease, stroke and blood clots all went up in the treatment group. Hip fractures and colorectal cancer went down. It wasn’t a poison and it wasn’t a miracle, and no single person running that study was in a position to weigh those against each other in real time. A group that had no stake in the answer was.

The Part That Is Still Open

4 honest limits, because an issue about oversight that ends on reassurance isn’t doing its job.

The second inspector is mostly not required. IRB review is required by regulation. A data monitoring committee is not, for most trials. FDA strongly recommends one when participants face serious harm or death, and there’s exactly one situation where federal regulation flatly requires an independent committee, which is emergency research conducted under an exception from informed consent. Everywhere else it’s a strong professional norm, and norms aren’t law.

The rulebook for it is out of date and the update is unfinished. The FDA's guidance on these committees was finalized in March 2006. A revised draft came out in February 2024 and is still marked draft. So the current final word on how these boards should operate is 20 years old, and the replacement has been sitting in draft for over 2 years. If that pattern feels familiar, it’s the same shape as last issue's unfinished diversity guidance. Put the date on the claim, every time.

And the first inspector’s backstop is thin right now. The IRB requirement is enforced by 2 different agencies depending on who pays for the study. FDA handles research on regulated products. For federally funded research it is a small HHS office called the Office for Human Research Protections, which covers more than 13,000 institutions. Reporting published this June put that office at 10 employees, 3 investigators, and no director in any of its 3 divisions for close to a year. It was set up in 2000 on a recommendation of at least 42 people. A requirement on the books is only as strong as the office that checks it, and Thursday has those numbers with their sourcing.

And oversight can fail. Both rooms only work if the people in them are actually independent and actually paying attention. There are documented cases where that didn’t hold. What the regulation guarantees is the structure, not the performance. That’s a real limit, and it’s also precisely why the claim that no such structure exists at all is the one that falls apart first.

What You Can Actually Go Check

Look up the study on ClinicalTrials.gov. Its registration record includes an oversight element that says whether a data monitoring committee has been appointed, yes or no, and a field recording whether a human subjects review board has approved it. What the record does not give you is the name of the reviewing board. Approval status is public, the identity of the IRB generally is not.

The rest is a conversation, and it’s a fair one to have. A study coordinator can tell you whether the study has a monitoring board, how often it meets (though they may need to check and get back to you), and who to contact with a safety concern that you do not want to raise with the treating doctor. Those are ordinary questions. A coordinator who answers them easily is telling you something.

The One Question: Skeptic’s Toolkit Q3

Q3. Who benefits if this claim is true?

Run it on the claim that oversight is theater. Whoever is selling you something that has never been through either room benefits enormously from you believing that the rooms are fake. If the referee is corrupt, the unlicensed fight looks reasonable.

Now run it the other way, because a tool you only point in one direction isn’t a tool. Who benefits if you believe oversight is airtight? Anyone who would rather you not ask whether this particular trial has a monitoring board, given that most trials aren’t required to have one.

Both questions have answers. Ask both. Run it on the supplement shelf too. Most of what sits there never went through either room, and nothing requires it to. That isn’t an argument about whether anyone should take it. It’s that when you actually need something to work, one option comes with a record you can check and an independent group that read the safety data, and the other comes with neither.

Next issue: the manufacturing floor, and why the same molecule made in 2 different plants is not automatically the same product.

Send this to someone who needed it

Send this to the person in your life who says nobody is checking, and to the person who says it is all handled. Same issue, both answers.

Thursday's Informed has the full Women's Health Initiative numbers with confidence intervals, the regulation by regulation breakdown, and every citation, so you can check my work. It comes automatically.

Run it on yourself first. Got Skepticism is our 2 minute game on catching the version of a claim that flatters you. Free, no signup, at openlabelmedia.com/games.

Refer a friend - every share helps

You just learned what 2 separate boards actually check. Send this to someone who needs the tools, and collect a Series 1 card while you're at it.

References

1. US Food and Drug Administration. Institutional Review Boards, 21 CFR Part 56, sections 56.107 and 56.109. ecfr.gov

2. US Food and Drug Administration. "Protection of Human Subjects; Standards for Institutional Review Boards for Clinical Investigations." Final rule, 46 Federal Register 8958, 27 January 1981. fda.gov

3. Centers for Disease Control and Prevention. "The Tuskegee Timeline." cdc.gov

4. National Research Act of 1974. Public Law 93-348, signed 12 July 1974. govinfo.gov

5. The Belmont Report. 44 Federal Register 23192, 18 April 1979. hhs.gov/ohrp

6. US Department of Health and Human Services, Office for Human Research Protections. "Revised Common Rule." Compliance dates 21 January 2019 and 20 January 2020. hhs.gov/ohrp

7. Rossouw JE, Anderson GL, Prentice RL, et al. "Risks and Benefits of Estrogen Plus Progestin in Healthy Postmenopausal Women: Principal Results From the Women's Health Initiative Randomized Controlled Trial." JAMA. 2002;288(3):321-333.

8. US Food and Drug Administration. "Establishment and Operation of Clinical Trial Data Monitoring Committees." Guidance for clinical trial sponsors, final, March 2006. fda.gov

9. US Food and Drug Administration. "Use of Data Monitoring Committees in Clinical Trials." Draft guidance for industry, Revision 1, February 2024, still draft. fda.gov

10. US Food and Drug Administration. Exception from informed consent requirements for emergency research, 21 CFR 50.24. ecfr.gov

11. ClinicalTrials.gov protocol registration data element definitions, oversight information. clinicaltrials.gov

New here? Root to Rx runs three pieces a week. Monday you watch a real health claim try to work on someone. Wednesday you get the tools to run it yourself. Thursday you get the full evidence so you can check my work. Written by a clinical researcher, free, at RootToRx.com.

For educational purposes only. Nothing in this newsletter is medical advice. Talk to your doctor before making any health decision.

Instagram: instagram.com/roottorx | LinkedIn: linkedin.com/company/roottorx | Facebook Lab: facebook.com/groups/theroottorxlab | Subscribe: RootToRx.com

You are receiving this because you subscribed at roottorx.com.

Disclosure: Root to Rx is published independently by Open Label Media LLC. Views expressed are personal views of Justin Yamashita and do not represent his employer or any affiliated organization. No employer resources or proprietary information are used. Every claim is sourced from publicly available materials.

Root to Rx | 228 Park Ave S, #29976, New York, New York 10003, United States