Public Health Informatics

We have stopped using paper requisitions in favor of online ordering, but when we did we used a software called Teleform that is supported by a company called Optiform.

Forms can be designed in the software (or imported), and each field on the form is made into a data element.  The forms are scanned in batches using the form template.  Then the user is able to verify the data in the software and correct any mistakes the software made in interpreting the writing on the form (the software pulls up an image of the form with the data overlaid on it and the user tabs through the fields correcting any mistakes they see).  Once the batch is verified, the software converts the data into HL7 order messages that are transmitted to our LIS.  There is also scripting as part of the form evaluation that can do various things with the data.

The software has other uses as well.  There are different modules that can be added to the license.  One is called automerge publisher.  This allows you to create forms and use a data file to fill them in.  Then the filled-in forms can be emailed to recipients.  Or you can create forms and have it autonumber them for you.  We used this when we had to have a unique requisition number on our forms for a specific reference lab.

Additionally, you can set it to monitor a folder and have it pick up files from there to process and send either the data or an image of the form somewhere.  We use this for lab reports we download as a PDF file from an outside reference lab.  It is able to separate the batch into the individual reports, find the sample ID on the form and use it as the file name, then put that file in a folder where our LIS can pick it up and attach it to the order as an image.

Feel free to reach out with questions or for a demo of the process.

Are these typed forms or hand-written?  Hand-written are far harder to convert. 

Are they in PDF files?  How many pieces of paper are you scanning?

We use a Fujitsu scanner and it comes with Paper Stream Capture.  It can do some of this.  I also handles bar codes.

Thanks,

Hi!

Apologies for the delay – ColLaborate and my new email address didn't like each other for a moment there...

James, they are hand-written and I want to say mostly associated with outbreaks (which means numbers vary seasonally and with disease strains). I'm curious where your data land when you scan them? (Do you import directly into your LIMS or a middleman application?)

And, thank you all for sharing your ideas and expertise! ��

Neelima, I also had a steep learning curve for posting! (And responding!)

Let's see – in the interest of transparency I'll share back with you what I have learned (some folks responded directly):

• Several LIMS can be extended to support paper form workflows, including labeling and attaching paper forms into the LIMS. I'm particularly loving the chain of custody capabilities many of you highlight.
• Many LIMS have moved away from using paper entirely. (I have a million questions here about what you use when the internet is down?)
• Options for text recognition of handwritten info on paper req forms include:

• Telform by Optimize
• https://www.extractsystems.com/
• Rumors and intrigue about AI (including maybe these options for Python?):

• EasyOCR
• Handprint
• TrOCR
• TensorFlor

Thank you all!

Laurel

Laurel Boyd, MPH

Laboratory Data Systems Informatician

OREGON HEALTH AUTHORITY
Public Health Division

Oregon State Public Health Laboratory

*New* email: Laurel.Boyd@oha.oregon.gov 

Cell: 971-470-5539 | Fax: 503-693-5602

Pronouns: She, Her, Hers