In 2001, the New Jersey Supreme Court took up a malpractice case that turned on a document nobody was supposed to still have. A chiropractor named Zimmerman had recopied, altered, and destroyed a patient's chart after receiving a complaint. The rewritten chart described a patient who had improved and was satisfied with her treatment. The original, which the patient had copied before litigation began, described continuing complaints and dissatisfaction.
The trial court kept the alteration evidence out, and the jury found for the chiropractor. The state supreme court reversed. A jury, it held, could reasonably infer that a doctor who rewrote the chart believed the original would hurt him — so the rewriting was itself admissible evidence about what the original had said.
What the court was protecting goes beyond one doctor's credibility. A clinical record fixes what was known and believed at a particular moment. Other clinicians consult it. So do pharmacists, and the patient herself, and whoever is deciding what to do next week. Any of them may have acted on what the chart said when they read it. Change the entry afterward and you have quietly cut the reasoning out from under decisions that were already made on the strength of it.
Florida had reached a related conclusion fourteen years earlier. In Public Health Trust v. Valcin, a surgeon's operative note was missing or too thin to be useful after a patient suffered complications. The Florida Supreme Court held that where an essential record is unavailable through the defendant's negligence, and its absence substantially hinders the plaintiff, the burden of proof can shift: the hospital has to show it wasn't negligent rather than the patient having to show it was. The gap in the file proved nothing about the surgery. But the party responsible for keeping the file shouldn't profit from the fact that its contents can no longer be reconstructed.
The rules that grew out of this line of cases are mechanical rather than moral. CMS's documentation integrity requirements for Medicare review say that amendments, corrections, and late entries must be permanently identified, must carry a date and an author, and must leave all original content legible. An electronic record system has to be able to show the original content, the modified content, the date of each change, and who made it. Every system that updates a patient record runs into the same question, whether it is correcting a medication list, revising a diagnosis code, or migrating a chart between platforms: does the update preserve the previous state, or overwrite it?
Medicine was taught this principle rather than deducing it — case by case and at considerable expense — by courts drawing adverse inferences from rewritten charts, by licensing boards treating undisclosed additions as misconduct, and finally by legislatures specifying how a correction must be made. The accumulated lesson is narrower than it first sounds. A record's integrity depends on whether the history of how it reached its current state remains readable, because the original entry, even when wrong, was for some period the best information anyone had, and other people acted accordingly.