The passenger is still on the track.
They have been on the track since the first piece in this series and they are still waiting to cross the finish line. They are not counting laps. They are counting months, some weeks.
The car runs on a different clock. That clock counts database locks, interim analyses, and amendment cycles, and it is a reasonable clock for an institution to keep. But the two clocks do not run at the same speed, and the gap between them is not free.
Somebody pays for it, and they pay for it out of a life.
I call it the calendar tax.
The series has been a time story the whole way
Every piece I've written in this series has been about time wearing a different coat. A car rebuilt from scratch each study costs time. A crew that starts from zero costs time. A track re-learned every race costs time. Time has been the denominator under all of it.
This piece makes it the subject. Every tax gets levied by someone. Somebody proposes it, somebody votes, somebody signs. And nobody signed this one.
The Abilene Paradox
Jerry Harvey's family was on a porch in Coleman, Texas. One hundred and four degrees. A fan, cold lemonade, dominoes. His father-in-law suggested driving fifty-three miles to Abilene for dinner at a cafeteria.
They went. In an unairconditioned 1958 Buick. Through a dust storm. The food was terrible. Four hours and 106 miles later they were back on the porch, filthy and exhausted, and one by one they discovered that not one of them had wanted to go.
Harvey published this in 1974 and called it the Abilene Paradox. His finding was not that the family failed to manage its conflict. It was that the family failed to manage its agreement. Everyone wanted the same thing. Everyone did the opposite. And afterward, everyone went looking for someone to blame.
His line for it: "It takes a real team effort to go to Abilene."
The industry has a signed map showing Abilene is the wrong destination
In September 2025, fourteen biopharmaceutical companies published, through TransCelerate and the Tufts Center for the Study of Drug Development, what they found when they pooled 105 protocols and measured how much of what they collect gets used.
More than a third of it comes from procedures that are non-core or non-essential.
They ran the study because earlier Tufts work had already shown where this was heading. Over the past decade, procedures in Phase III pivotal trials are up around 40 percent and data points collected are up 283 percent.
Read that with the paradox in mind. The companies doing the collecting funded the study proving a third of the collection is unnecessary. They agreed. Publicly, in writing, with numbers. Ten months ago.
Yet, almost a year later the collection continues unabated.
Nobody signs a check that says Abilene
That is what makes it hold.
No line item in any budget reads "gas to Abilene." They read: additional monitoring visit. Reconciliation support. Audit readiness. Inspection preparation. Data quality review.
Not one of those is indefensible. Not one of them was requested by the science.
The trip is funded, entirely, in items that each look like care.
I won every lap and lost the race
I spent a recent weekend at a go-kart track with my closest friends. Two races. My friend started ahead of me both times and I finished ahead of him both times.
In the second race I got aggressive, cut inside a slower kart, opened a gap, then spun the hairpin on the second to last lap and gave back three seconds. I still crossed the line first. I walked to the printout certain I had won.
He won. I had the fastest lap of the day. I also had the slowest. He had neither. He never spun.
I was not lying to myself about the overtakes. They happened. I was measuring the thing I could feel and the scoreboard was summing the thing I could not.
The track was full of traffic. Slower karts, kids, no room to move. The traffic did not spin me.
There are two variances
Scientific variance is the race. Does the drug work, in whom, by how much? It is irreducible and it is the entire reason the trial exists. You never want it gone. You want it measured cleanly.
Operational variance is the spin. The reconciliation cycle. The query loop. The audit scramble. The recurring meeting where five people stare at a spreadsheet and try to work out what happened, and when, and why.
The science never asked for a second of it.
In the second piece in this series (F1 economics) I argued that the money in clinical research flows to a race the science does not require. So does the time. Same root, different currency.
Humpty Dumpty
Chris Decker, President and CEO of CDISC, published an article about exactly this, and he reached for the nursery rhyme.
Data is captured in rigid, siloed structures. Adverse events collected with no real connection to medications or vital signs. Even inside a single domain the relationships are not kept: the MedDRA body system and the preferred term sit next to each other in a dataset, and the only reason anyone knows they are related is that a human said so.
Context is lost at the moment of capture. And, like Humpty Dumpty, it cannot be put back together.
That is the spin, named by the person who runs the standards body.
Dave Iberson-Hurst wrote something this year, coming back from the CDISC Interchange in Milan, that locates where the break happens. He has said for years that a small part of him dies when he hears the word "mapping". The point is not the joke. The point, as he put it, is that "mapping exists because we cut the relationships in the first place."
It was CUT.
Not lost, not degraded but cut, at capture, by us, on purpose, and then reconstructed later at enormous cost by people who were not there.
What can be recovered, and what cannot
Kerstin Forsberg published something this month that sharpens this to a point.
In USDM, the key "text" appears on twenty-three different classes. A study title has a text. So does an address, an endpoint, an objective, a person's name. Twenty-three meanings, one key, and the file cannot tell you which one you are holding.
She generated a context file mechanically from CDISC's own published model and applied it. The JSON did not change. Byte for byte, the same file. What changed is that every key now resolves to precisely what it always meant. In her words, "the ambiguity did not survive the lift."
No transformation and no mapping specification. She did not recover the meaning, she declared it, because it had been sitting in a published file the whole time, undeclared. Which shows you exactly where the floor is.
The meaning of a field was never lost. It was undeclared, and a declaration brings it back.
But let's think about this carefully.
Why a value changed on a Tuesday in March, who decided, on what basis, and whether they would say the same thing under oath eighteen months later: that was never written into anything. No context file recovers it, no lift produces it. There is nothing to declare, because nobody wrote it down.
Humpty Dumpty at the schema layer can be reassembled. The pieces were in the box. At the instance layer, start opening boxes.
The lot number is in one system. The batch record is in a plant on another continent. The specification is in a file someone has to request. The reason for the deviation (partial auto-injector dose) is buried in a PDF that arrives by email, if it arrives at all. Some of those rooms are locked, and the key takes weeks to get.
So go and get the keys. All of them. Wait the weeks. Open every box in every room.
The shard you actually need is not in any of them. Nobody ever wrote down why!
We already pay for it, and then we bury it
Since 1997, 21 CFR Part 11 has required a secure, computer-generated, time-stamped audit trail. Who touched the value, what it was before, what it is now, and when. That is federal law, and the industry complies with it. Most EDCs also capture a reason for change, which Part 11 does not strictly require and every data integrity guidance expects anyway.
So the record already exists, ALL OF IT. Right now, in your system, but it exists in a log.
Part 11 says the audit trail must be available for agency review and copying, and the industry satisfied that sentence by making it retrievable, not necessarily usable.
Try it yourself. Ask your EDC rep whether they can attach the Part 11 audit trail to each CRF field as a property that travels with the value and can be queried alongside it. Not export it, not run a report. Attach it to the data.
Count the days until the answer comes back as a statement of work.
We are already paying for provenance. We dig the hole, we throw it in, and then we hire people to dig it back out.
What it would take?
Four things, and none of them are exotic.
1. Attached, not filed. The why sits with the value, not in a log on the far side of a wall from the thing it explains.
2. Surviving the boundary. Site to CRO to sponsor to agency. Provenance that stops at a company wall is NOT provenance, it is a log. Which means the format cannot belong to any single vendor, including the one you like.
3. Meaning the same thing everywhere. A reason for change written at a site and a reason for change read at an agency have to be the same object, not two strings that happen to resemble one another. Attaching a log to a field is engineering, and engineering gets done. This one is vocabulary, and it is the least glamorous and most load-bearing item on the list.
4. Somebody going first. Which is the exact move the paradox punishes.
The most honest room in this industry has no budget
Go and read the comments under Chris Decker's piece.
Kerstin Forsberg is in there, having built the identifiers and then gone and asked the governance group who guarantees that a name still means the same thing in fifteen years, which she correctly calls governance rather than technology. Johannes Ulander is in there, having independently rendered the same model a different way and taken it to Basel. Someone else turns up to say he is building something similar and they should collaborate. Anthony Chow is in there, pointing out that inferred context should not be mistaken for preserved context.
Senior people, named, in public, agreeing with each other, in writing, at length, for free.
Not one of them can stop a purchase order.
That is the whole thing. The confrontation Harvey said had to happen is happening. It is happening in a room with no pen in it.
Harvey was specific about this. The confrontation only works in the group setting, with the people who hold the decision, because the failure is collusive. A comment thread is not the car.
So the words are being shouted out of the window, at seventy miles an hour, on the way to Abilene.
The question
Chris ended his article with a question, and I am not going to answer it here.
"How do we get the right people at the table?"
I would add only one thing to it. The table has never had the person with the pen at it.
Which makes the next question where the money goes, and who decides. That is the next piece.
And the passenger, the one running out of time ...
They are still on the track. They did not attend the meeting. They did not fund the study. They are not on the working group, they have not read the guidance, and they have never heard of Abilene, Texas.
They are STILL waiting.
The calendar tax does not come out of a sponsor's budget. It comes out of a human life.
Let that sink in.
About the author
Bo Lora spent four years at AWS deeply embedded across more than thirty roles in the clinical research value chain, from sites to sponsors, CROs, tech vendors, start-ups, and regulators. He held Core Technical Field Community Leadership in Health and Life Sciences. Before that he was a consultant at BearingPoint, working across finance, telecommunications, manufacturing, and semiconductors. He started his career as a BBN Field Engineer in Stuttgart in 1989, where he once flew to Oslo to fix a modem that was unplugged.
He is active in CDISC, Vulcan, and TransCelerate, and serves as AI Faculty for ValueAdd Group.