Results were printed by analyzers, retyped at a desk and checked a second time.
01
What the floor actually looked like
The first morning we spent in the lab, the bench ran fine. Samples came in, tubes got spun, the analyzers hummed through their racks. The bottleneck was not the science. It was the fifteen metres between the analyzer’s thermal printer and the desk where somebody typed those values into a reporting template.
Every result was produced twice: once by an instrument accurate to three decimal places, and once by a person at the end of a shift. Between the two sat a paper tray, a ruler to keep your place on the printout, and a rule that any transcription had to be double-checked by a second technician, which meant the error rate was low and the cost of that low error rate was two people’s afternoons.
Then the phone calls. Patients ringing to ask whether the report was ready. The receptionist walking to the bench to ask. The technician stopping mid-rack to answer.
02
What we changed
LabCare speaks to the instruments directly. We implemented the ASTM E1381/E1394 and HL7 v2 dialects the analyzers already emit, over serial and TCP, so the result lands in the record the moment the run completes. No retyping, no second reader, no paper tray.
Specimens get a barcode at reception, so the tube that goes into the rack is already bound to a patient, a requesting doctor and a panel. The analyzer reports against that barcode. The match is made by the machine, not by someone reading a handwritten number off a cap.
Validation stayed human. A pathologist still reviews and signs, because that is the part that requires judgement. What changed is that signing is now one screen and one action, and the signature releases the report automatically.
03
What happened after
Turnaround, measured from run completion to the patient holding the PDF, dropped by about three quarters. The receptionist stopped walking to the bench, because patients stopped calling: they get an SMS or WhatsApp with a secure link the moment the report is signed.
The number the lab director cared about was not speed, though. It was that transcription errors went to zero, because there is no longer a transcription step to get wrong.
