Hello Tom,
sorry for my delay with answering to your mail.
>To your questions:
>>-The typical solution in a hospital information system and archiving system in an Austrian hospital nowadays would be:
>>- make a (paper) copy of the medical report that the patient provides for the doctor and put it into the records
>Thank you for that information!
>So "put into" means that they have in case of doubt: hand work.?!
Yes, because it depends on the workflow.
>So I it is to say, I am working (searching) on an ocr-solution.There are some free ocr tools around.
>And I have read a post from 2010 where they say xsane free software had already some ocr integrated.
>The question marks above indicates that I hardly can believe that yet. In all these bureaus of 2015's world they, they do not use OCR, yet?
I do not say “in all bureaus of 2015´s world”,
but in the typical Austrian hospital in the typical workflow of the physician,
If there is still a paper record in use (which is still in the majority of hospitals).
There are some hospitals which already use the electronic record,
but the workflow does not make sense with using the ocr-program if the ocr result
will not be checked very intensive for correctness.
And if a doctor has to cite the information of the medical (paper) report it is easier to read
and dictate it as there are medical secretaries who have to write the medical reports.
>I wanted to build some hello-World script with it.
>So the report could be an arbitrary sentence "
><Owner of the place read from database of gnuhospital> greets: Hello World", located in the middle of a page.
>So if you or else someone knows of a preferably non-interactive way
>of such an hello world report, I would be glad, for a few lines of script, anyway.
Yes, that would be great. I tried it some months ago and failed at the end,
mainly because I had a problem with accessing LibreOffice from GNUHealth.
>So what might one believe: are there any sense and_ chance that such
>an OCR-Tool-Chain will be integrated in (especially free) future development?
Yes, I believe so.
>Are there maybe risks?
>Such as waste of effort?
No, but it is very important to integrate as a next step into this OCR-Tool-Chain the “benefit” for the doctors,
e.g. an automatic plausibility check, automatic coding or a kind of automatic knowledge extraction.
>Would it be a big step forward against (non-free) competitors?
>Other relevant aspects?
Always keep in mind support for the relevant medical workflow.
And maybe think of Natural Language Toolkit (http://www.nltk.org/).
Opinions?
In my opinion the most important thing is that GNUHealth will be used now in real life in hospitals and in doctor´s offices.
Therefore the first three steps have to be done:
1. provide program versions (server and client) which are very easy to install (let us say “like an app”),
2. meet the legal requirements (e.g. ICD-10 encoding, export functions for reimbursements),
3. provide functionality for importing data which is already in use in the hospitals or offices from their Electronic Health Records
(or at least from their official data sets, e.g. MBDS – minimum basic data sets).