A radiologist spends his life turning images into decisions other people can act on. So when his own neuropsychological report came back, he read it the way he reads everything, looking for the line that tells him what to do.

There wasn’t one.

He was forty-five. A year earlier his car had slid on ice at night, sober, and struck a fixed structure. The airbag did not deploy. His head did. The mechanism was high-energy. The MRI, read weeks later, was equivocal, with subtle white matter findings described as nonspecific and of uncertain clinical significance. The words were accurate. They were also the problem. He had spent his career telling other physicians that a clean scan does not mean an intact patient. Now the same truth was turned back on him, and it left him with nothing to hold.

He came on his own, through his neurologist. No employer sent him. He had one question, and it was not abstract: could he keep reading scans, where a missed finding can be someone’s missed cancer, or did he need to step away? He wanted the truth more than he wanted to be reassured.

Everyone he saw was careful. The neurology workup was careful. The testing, when it came to me, did not hand over a clean answer either. And at every step the same error repeated, quietly enough that no one noticed they were making it.

Uncertainty about cause was treated as if it were uncertainty about consequence.

No one could say for certain why his processing speed had dropped, so no one would say whether the drop mattered. Those are different questions. The first is about etiology, and it was genuinely hard. He carried a head injury, a long-controlled seizure disorder, a year of sobriety after heavy use, and medication, any of which can press on cognition. None of that made the injury the only explanation. It made it one explanation that could not responsibly be dismissed. The second question was not about cause at all. It was about a man whose work depends on speed and vigilance he might no longer reliably sustain. You can be honestly unsure of the first and still owe a clear answer to the second.

That distinction is the whole case, so the data underneath it is worth stating exactly.

Against the general population, his scores were mostly average. A few fell lower: processing speed, sustained attention, error monitoring. Nothing dramatic. Something harder. Average is the wrong reference point for this man. He did not get through a radiology residency with average attention, so whatever his exact premorbid baseline was, the estimate had to be anchored well above the population mean, because the work selects for it. A score at the population average, in someone who started well above it, is not reassurance. It is evidence consistent with decline that a population norm, on its own, will not reveal. The point is not that every average score in a high-functioning person proves injury. It is that average scores in the wrong domains, in the wrong person, under the wrong occupational demands, cannot be waved away as normal. Testing does not reproduce a full radiology shift, but it can show whether the systems that shift requires are straining under structured demand. His were.

The honest reading was not “within normal limits.” It was that the findings were most consistent with a meaningful reduction from his own baseline, in precisely the domains his work depended on most.

I could have written the other sentence. Every cautious report keeps it ready: performance broadly within normal limits, with areas of relative weakness that may warrant monitoring over time. That sentence would have been defensible. It would also have been useless, and worse than useless, because it would have sent a radiologist back to a reading room holding a document that technically said nothing was wrong.

Two things made the soft sentence tempting, and neither was the diagnosis.

The first was self-protection. A clear opinion can be wrong, can be read back to you in a deposition, can threaten a career, and if you are wrong, the person whose career you threatened is entitled to ask why you were so sure. Hedging is a survival behavior. It keeps the clinician safe. But the patient did not come for the clinician’s safety. When the clinician refuses to commit, the risk does not disappear. It transfers, whole, onto him.

The second was the question of where a finding would go. He was a physician; his patients depended on his accuracy. As a clinical matter, the law did not appear to force anyone’s hand. The mandatory reporting machinery for physician impairment in California practice settings generally runs through hospitals, peer-review bodies, and employment-based oversight. None of that was in play. A subtle reduction in processing speed is not a communicated, serious threat of physical violence toward a reasonably identifiable victim. No mandatory rule compelled a report, no rule erased the concern, and no form resolved the responsibility. That gap is exactly where hedging lives. The safest document is the one soft enough that the larger question never has to be asked.

The vagueness was not only protecting against being wrong. It was protecting against the consequences of being right.

Here is what that softness does to a man like him.

He went back and forth for months. The cognitive evaluation was supposed to break the tie and instead handed him language that could be read either way. With no one willing to commit, the decision fell entirely to him, which sounds like autonomy and is actually abandonment. He has a mortgage, a family, an identity built on the work, and a recovery the work helps hold together, and every human reason to read an ambiguous report in the direction he needs. So the default won. He kept working.

Hold the two errors side by side. If he was impaired and kept reading, the cost was not his alone; it fell on patients whose imaging crossed his desk on a day his attention slipped the way the data hinted it might. If he was fine and left out of fear, he dismantled a career and a recovery for nothing. A clear opinion exists to prevent both. The hedge prevented neither. It only ensured that whichever error occurred, no clinician’s name was on it.

This is the specific cruelty of it. The people who most need a clear opinion, the surgeon, the pilot, the physician, are the ones for whom a clear opinion is hardest to give. Their stakes are what make clinicians flinch, and the flinch leaves them more exposed, not less.

The patients who can most afford ambiguity get clean answers. The ones who can least afford it get hedged.

None of which required a verdict. The honest opinion was not “he can never read a scan again.” At twelve months, with the acute phase over and the major confounds either resolved or stabilized, his data showed enough reduction in the systems his work depends on that unsupervised reads could not be responsibly endorsed without safeguards. Not a global judgment about competence. A work-specific opinion about a high-demand task. Restricted duties, a second reader, re-evaluation at a set interval. That commits, and it leaves him a future. Hedging offered neither.

The alternative to hedging was never false confidence. It is accountable judgment: here is what the evidence supports, here is what it does not, here is the limit of what I can know, and here, within that limit, is what I think.

You can be uncertain why a man is slower and still owe an opinion on whether the slowing matters.

He did not need certainty. He needed someone willing to turn a year of frightening ambiguity into a position he could stand on. The opinion, in a case like this, is the intervention. Not the testing. Not the imaging. The willingness to look at a frightened man whose whole life is balanced on one question, and answer it.

What he needed was the one thing the system teaches clinicians to withhold: someone who would rather be useful than be safe.

Daniel Hai, Psy.D., is a clinical neuropsychologist and the founder of Neuro Assessment Center in Encino, Los Angeles. He provides comprehensive neuropsychological assessment for adolescents and adults with complex, high-impact, and difficult-to-clarify presentations.

This is a composite case; no individual patient is depicted. The discussion of reporting obligations reflects the author’s clinical reading of the regulatory landscape and is not legal advice.