Skip to content

Search documentation

Find a documentation page by the words on it.

Documentation

How Serelora Reads an Uploaded Document

When you upload a file, Serelora does more than store it. It reads the file, labels it, and, if your clinic has document extraction turned on, turns the facts it finds into entries in the chart. This page explains each step so you know what to expect and what to check.

Step 1: The file is stored

Uploading puts the file in the patient's documents. It appears on the Documents tab straight away, and you can open it even if nothing else happens. See Documents.

Step 2: The text is read

Serelora reads the text of PDFs and images automatically, including scanned pages and photos, so you do not need to run a scan program first. A file that Serelora cannot read shows Extraction failed, and a file with no readable text shows No text found. See Why a Document Was Not Read Into the Chart.

A lab report in a PDF or image gets extra care when extraction is on. Serelora reads its tables, and each row of a results table becomes its own result, with its value and unit. If the report prints a reference range or a flag, they are kept with the result and shown in its claim detail. See Source Details on a Claim.

Files that already carry structured data, such as C-CDA and FHIR files, are not guessed at. Their entries are mapped directly. See Importing a C-CDA Record.

Step 3: The document is labeled

When you do not supply them, Serelora fills in the document's Category, Type and Author from what the file says, and from its file name when that is needed. You can change the category if it is wrong. See Changing a Document's Category.

Step 4: Facts become entries

If your clinic has document extraction turned on, Serelora finds the facts in the text, such as diagnoses, medications, allergies and results, and adds each one to the chart. See Feature Flags. What to know about the entries:

  • They usually start unconfirmed. An entry shows Unconfirmed until a person confirms it. An entry from a health record file can arrive with the status that file stated, such as Provisional.
  • They keep their evidence. Each entry links to the words it came from. See Document Highlights and the Original View.
  • A fact that says the patient does not have something is kept apart from one that says they do. Serelora does not combine opposite statements.
  • The same fact from several documents is one entry. Serelora recognizes entries that share a code and adds the new document as another statement of the existing entry. Some duplicates are not recognized, and you can merge those in Manage Claims.
  • A fact with no date gets the document's date. The chart then shows Date not stated, and hovering shows which date Serelora used.
  • Codes may come with the entry, and you can add or generate more. See Code Vocabularies in the Chart.

Narrative text is analyzed only when extraction is on. Without it, uploads are stored and readable, and structured files are still imported. A clinic can have separate switches for uploaded documents and for signed notes and addenda.

Step 5: Check the result

The chart updates by itself when the reading finishes. See Live Chart Updates. To check a document, open it and look at the highlights, and open any entry to see the words it came from. See Claim Detail and Where a Fact Came From.

What Serelora does not do

  • It does not confirm entries for you. Reading a document is a starting point, and a clinician decides what is true.
  • It may miss a fact or read one wrongly, especially on a poor scan. A Partially read badge tells you the scanner could not read every page. Compare important entries with the document.
  • It does not replace your judgment. To fix an entry that was read wrongly, see Correcting a Claim.