The H beside a result is the most-read character on a lab report and the least explained. It
is worth being precise about what it claims, because it claims less than it appears to.
What the flag is
A flag is a comparison between two numbers on the same page: your result, and the reference
interval the laboratory printed beside it. If the result sits above the upper bound, the lab
prints H. Below the lower bound, L. Some labs use an asterisk, bold type, or a separate
"abnormal" column instead. Some add HH or a panic-value marker for results far outside the
range, which is an internal signal for urgent notification.
That is the whole mechanism. It is arithmetic performed against that lab's own interval.
What it does not assert
It does not say a result is bad, and it does not grade how far outside the range a value sits —
a result one unit over the bound and one that is double it both print the same H.
It also does not mean something is wrong. MedlinePlus puts this plainly: a result outside the reference range "may or may not be a sign of a health problem. It's common for healthy people to sometimes have results outside the reference range." The reverse holds too — "a test result that is within a reference range isn't always a guarantee of good health."
Why flags accumulate on a big panel
Reference intervals are built from a reference group, and MedlinePlus describes them as based on "the test results from large groups of healthy people". A range that covers most of a healthy group necessarily excludes some of it. Run enough tests on one person and a flag or two becomes unremarkable — a comprehensive panel reports dozens of results, each with its own interval.
This is why a page with two flags on it is not twice as concerning as a page with one, and why the count of flags is not a score.
Why the same value can be flagged at one lab and not another
Because the flag is computed against the interval on that report, and intervals are not shared between laboratories. MedlinePlus notes that "because labs often use different testing methods, they often use different reference ranges to describe normal results." The same number can therefore be flagged on one report and unflagged on another. Neither lab has made a mistake.
Reading a flag usefully
Treat it as a pointer to the row, not a verdict on it. The useful questions are what the reference interval on that report actually was, what the value was the last time it was measured, and whether the same laboratory produced both — which is the only comparison that holds the interval constant.