Present: Rune, Arezoo, Soudabeh, Ingrid, Endre, Øystein and Laura

Status:

Everyone should have their introductions ready very soon.

Arezoo:

Changing from Patient Trajectories to Pathways (Why?)

Spending time on ICD10-codes, instead of writing the introduction and problem description (Make introduction and goals before next meeting)

Papers: Shahar is only about Streams/Records.... Trajectories... I need.... Pathways? to make queries about specific times in the patients life.

--Øystein: You have to find a way to separate (instant) events/actions from (diagnostic codes for) chronic situations aka states. Points/events/intervals/states (for reasoning)

...For the length of stay: How can

What is left: ICD-10, DRG, Operational codes... (What else?). Can Kjartan give us 0-day, 1-day or many-day length-of-stay information

Ingrid:

Working on NER for Clinicle notes. Collaboration with Maria and Tanushi-san in KTH (Stockholm)

-Contact Thomas Brox Røst to get in touch with Roar and Jaspreet about POS tagging. (Who is Lars Erik Brus, Mendelay)

-Hans is indexing the same text.

-Real data has to be on a computer which is not connected to Internet!

--But the system itself (without data) is not sensitive.

--Ingrid needs REK-approval to get full access to the clinical notes to test her system.

Endre:

I think my part has changed from inventing algorithms to reusing algorithms in a domain I don't know.

I was anonymizing data by making it more general by merging similar rows in a table.

The data: Starting with NPR data. Create data that can actually be shared with the public. We need this to reason about patient states (in general).

Øystein

Maybe we should add more students: Aleksander, Roar and Jaspreet

Later: Trond Elde (from DIPS, lead architect) and Haldor Husby (Master at HØK/AHus, data contact)

TODO

Keep working on the outlines and goals

Get started with the literature review.