The Wrong Man on the Train

A worker was out on a neck and shoulder claim. His restrictions were specific: limited turning of the head, restricted use of the injured arm and hand, no lifting away from the body. Weeks into the claim, the file had the usual friction. Treatment was moving slowly, the restrictions were not improving, and his employer was asking fair questions. Was he really as limited as the chart said, or was the file paying for a recovery that had already happened?

Surveillance was ordered to answer that question. It was a normal step, taken for normal reasons. Nobody involved set out to cut a corner. That is exactly why what happened next is worth telling.

The tape that seemed to settle it

Working from the description and the photograph in the file, the investigator found a man who matched both and followed him onto a commuter train. For the length of the ride, the camera had a clear view. The man typed on a laptop with both hands. He turned his head freely, watching the doors, looking out the window, following conversations around him. At his stop he lifted a bag down from the overhead rack with no hesitation and walked off into the crowd.

Read against the restrictions, the footage was devastating. Both hands in full use. A neck rotating well past the treated range. Lifting overhead. The investigator's report used the word "inconsistent" several times, and within days the mood around the file had changed completely. This was no longer a slow medical claim. People were talking about what the tape would do at a hearing, and who should see it first.

The word that got skipped

The report contained one qualifier that everyone, in the momentum, read straight past. The identification of the man on the train was probable. Not confirmed. Probable.

Here is what that word actually meant. The subject matched the general description and resembled the photograph on file, across a crowded platform and in uneven light. He had not been followed from the claimant's verified home. He had not been matched to a confirmed appointment, a known vehicle, or a single person who could say, without hesitation, that is him. The tape showed a man doing things the claimant was not supposed to be able to do. Whether the man was the claimant was an assumption, and the file was quietly promoting that assumption into a fact.

That promotion is where good files get hurt. Once a report like this circulates, it stops being treated as a lead and starts being treated as evidence. Readers anchor on it. Someone drafts the confrontation. Someone else suggests that the independent medical examiner should watch the footage before the upcoming exam, so the doctor knows what to look for. Every one of those moves is reasonable if the man is the claimant, and corrosive if he is not.

How it unraveled

Before the footage went anywhere that mattered, a reviewer asked the boring question out loud: how do we know this is him? The answer was not good enough, so the identification was checked properly, against images of the claimant that carried no doubt at all. The kind of images a file accumulates without thinking about them: identification shown at appointments, photographs taken through the ordinary course of the claim. The conclusion was uncomfortable and clear. The man on the train was somebody else. Similar build, similar age, similar hair. A different person.

The catch had to be withdrawn, and withdrawn thoroughly. It was pulled out of the working narrative of the file. Every summary that had repeated it was corrected. Everyone who had received the report was told, plainly, that the subject had been misidentified and the footage showed a stranger. That is a painful exercise, and it is exactly the right one. A bad catch that gets quietly shelved has a way of resurfacing later, in a deposition or a hearing, quoted back by the other side. A bad catch that gets formally retracted is an embarrassment for a week. A bad catch that gets used is a disaster for the whole defense.

Rebuilding the record

The claim still had the open questions it had before the tape, so surveillance was done a second time, the slow way. This time, the subject's identity was established before the camera rolled, and it was anchored to the one event that belonged to the claimant beyond argument: his independent medical examination. The IME put the claimant at a known office at a known time. The man who arrived for that appointment could be identified with certainty and followed continuously from there.

The second session produced a smaller, duller, and far more valuable record: the right man, identified the whole way, doing whatever he actually does. It did not matter, for the integrity of the file, whether that footage helped or hurt. What mattered was that it could be used with confidence, and that no one on the other side could take it apart by asking a single question.

What this file teaches about surveillance

None of the discipline in this story is exotic. It is a short list, and most of it costs nothing but patience.

• Confirm the subject before the camera rolls. Identity is the assignment; the footage is secondary. A verified starting point, a confirmed photograph, and continuous observation are what turn a tape into evidence instead of a rumor with a timestamp.
• Treat "probable" as a warning label. If an investigator grades an identification as anything less than confirmed, that grade should travel with the footage everywhere it is mentioned. Qualifiers that fall off in summaries are how probable quietly becomes proven.
• Show no one until the identification is locked. Do not send unconfirmed footage to a treating doctor, an examiner, or anyone preparing for a hearing. An opinion formed by watching the wrong man does not simply get corrected later. It damages the credibility of every opinion that follows it.
• Retract loudly when a catch goes bad. Notify everyone who received the report, in writing, and mark the footage in the file as a misidentification. The goal is that nobody, including your own side, ever relies on it again by accident.
• Anchor the redo to a fixed event. Examinations, hearings, and scheduled appointments give surveillance a time and place where the subject is certainly the claimant. Build the session around that anchor instead of around a description and a hope.

Where it stands

The claim continues on its medical merits, which is where it should have stayed all along. The file now holds one retracted tape that proves nothing and one properly identified record that both sides can test fairly. The money spent on the first session bought an expensive education. The discipline shown in the second one is what should have been purchased the first time.

The lesson for employers

Surveillance feels like the most objective evidence in a claim. It is moving pictures. It looks like the truth. But a tape is only ever as good as the identification underneath it, and identification is a process, not an impression. When a report lands on your desk announcing that your injured worker has been caught, the first question is never what he was doing. It is how we know it was him. Ask that question every time, out loud, before the footage is shown to anyone whose opinion you will need later. The employers who get hurt by surveillance are rarely the ones whose claimants were dishonest. They are the ones who acted on the wrong man.

CompShield is a workers' compensation claims-defense and cost-control firm. This case study is anonymized; identifying details have been changed to protect the privacy of all parties. It describes investigation practices in general terms and is not legal advice. Investigation and privacy laws vary by state, so consult counsel for your jurisdiction.