The lab's liquid-biopsy panel served 60 partner clinics, and its AI-assisted analysis pipeline produced results clinicians could act on quickly, once those results actually reached them. In practice they did not: every result left the lab as a PDF report, sent onward by fax or portal download, with no connection to the EHR either side of the transfer relied on.

That gap pushed the work onto clinic staff. Someone at each clinic had to open the PDF and re-key the findings into the patient chart by hand, a process that introduced both delay and the chance of transcription error. Results that the lab's pipeline had already produced still took days to reach the physician actually treating the patient.

Ordering suffered from the same disconnect. Test ordering still ran on paper requisitions rather than through the EHR, and once a sample was sent there was no shared visibility into its status, so clinics and lab staff alike were left guessing whether a given liquid-biopsy panel was still in transit, in analysis or already reported.

We built an HL7 FHIR integration layer connecting the lab's existing AI-assisted analysis pipeline directly to partner clinic EHRs, replacing the fax-and-portal handoff with a structured data connection both sides could rely on, rather than a document either side had to chase down, open and re-key by hand each time.

Structured results now post directly to the patient chart as discrete fields rather than arriving as a PDF attachment, so the finding lands where the physician already looks instead of in a separate document that has to be opened, read and manually transcribed by clinic staff before anyone can act on it.

We extended the same integration to ordering and tracking. Physicians now order tests from inside the EHR rather than on paper, and a shared status tracker follows every sample from accession through to released report, giving both the lab and the clinic the same visibility into where a sample stands.