The form gave you three choices. Field 7b — Relationship to Product — offered a dropdown: caused by, associated with, coincident with. Each triggered a different pipeline. Elena had learned this the way you learn which doors in a building are locked: by trying them all and remembering.
Caused by went to Safety. Safety had twelve people and a ninety-day review clock. They opened an investigation, assigned a case number, and produced a report that traveled upward until someone with enough authority decided whether the product changed or the report got a footnote.
Associated with went to Pharmacovigilance. Pharmacovigilance had forty people and no clock. They added the event to a signal-detection database, where it became a data point in a statistical model that would flag the product for review if enough data points accumulated. No single report mattered. Every report mattered.
Coincident with went to a file server in Maryland. Quarterly, someone ran a script that counted the filings by category and produced a summary no one was required to read.
The patient was a fifty-eight-year-old man who’d taken the medication for eleven days and then couldn’t feel his left hand. The hand still worked. He could grip, release, turn a doorknob. He just couldn’t feel himself doing it. He described it to his neurologist as watching someone else’s hand obey my instructions.
The neurologist referred him back to his prescriber. The prescriber ordered tests. The tests came back normal — nerve conduction, MRI, bloodwork, all within range. The hand worked. The hand was fine. The patient said it was not fine. The prescriber filed the adverse event report and reached Field 7b.
She could have called Elena. Most prescribers did, eventually. Elena was the medical affairs liaison, which meant her job was to answer the question the prescribers were actually asking, which was never the question they said they were asking.
They’d say: Is this a known side effect?
They meant: Which verb do I pick?
Because the three verbs weren’t three levels of confidence. They were three theories of what had happened. Caused by meant the drug did this — the molecule entered the body and the body changed in this specific way as a result. Associated with meant something happened and the drug was present and the two facts belonged in the same sentence but not necessarily the same clause. Coincident with meant two things occurred in the same window and the window was the only link.
Same patient. Same hand. Same eleven days. Three verbs, and each one constructed a different event.
Elena had a folder she kept to herself. Not hidden — just not shared, because sharing it would have required explaining why she’d made it, and the explanation would have sounded like a complaint, and she was not in the business of complaining.
The folder held fourteen cases where she’d seen the same event filed under all three verbs by different prescribers. Same drug, same symptom, same temporal profile. One prescriber wrote caused by. Another wrote associated with. A third wrote coincident with. The fourteen cases weren’t errors. No one had gotten the verb wrong, because there was no right verb. The dropdown offered a choice and the choice constructed the reality the system would respond to.
In seven of the fourteen, the caused by filing led to a safety review that produced a label change. The associated with filing entered the signal-detection database, where it became part of a trend that was detected eighteen months later. The coincident with filing went to Maryland.
Same molecule. Same hand. Same patient, in some cases — the same person, filed three times by three doctors who each chose a different verb and therefore created three different events in three different systems, each of which responded to a different thing that had happened, none of which was wrong.
The prescriber called Elena.
“Is this a known side effect?”
“There have been reports.”
“Reports under which category?”
She paused. This was the question behind the question. Not has this happened before but what kind of thing is this, and which door does it go through?
“All three,” she said.
The prescriber was quiet for a moment. “So which one is it?”
“It’s Field 7b.”
“That’s not an answer.”
“It’s the most honest answer I have. The field asks you to commit to a relationship the data doesn’t resolve. The verb you choose will determine what the system does with the report. The system will respond appropriately to the verb. It won’t respond to the patient.”
He chose associated with. Elena knew he would. It was the middle option, the one that felt like neither commitment nor dismissal. It acknowledged the event without claiming to understand it. It was the verb you chose when you wanted the system to know something had happened without telling it what.
The report entered the signal-detection database. It became a row in a table. Eighteen months later, when the signal was detected, this report would be one of four hundred and twelve that contributed to the statistical threshold. The label would change. The word associated would appear in section 6.1 of the prescribing information, in a sentence that read: Sensory dissociation has been associated with use of this product.
The verb survived. It made it all the way from the dropdown to the label. It arrived at the end meaning exactly what it had meant at the beginning: something happened, and we are not prepared to say what kind of thing it was.