DASH

News and Events

Maria Skeppstedt’s Licentiate Dissertation Defense

Health Informatics, Licentiate Seminar, SYSLAB

Maria Skeppstedt defended her Licentiate thesis convincingly and successfully on May 15, 2012! The dissertation, entitled From Disorder to Order – Extracting clinical findings from unstructured text, investigates the use of rule-based techniques to extract mentions of findings (e.g. breathing problems) and disorders (e.g. asthma) from electronic health records written in Swedish, while also being able to determine whether or not they occur in a negated context.

Associate Professor Beáta Megyesi from Uppsala University conducted the opposition, which yielded a fruitful discussion with the respondent. The examiner from DSV was Professor Harald Kjellin. Maria’s supervisors are Professor Hercules Dalianis and Professor Gunnar Nilsson.

Congratulations!

Comparing certainty level annotations in clinical NLP: a case study on Swedish and English within the Interlock project

Health Informatics, Presentation, SYSLAB

On the recent return from her research visit to the University of California, San Diego (UCSD), Sumithra Velupillai presented the results of a comparative study of certainty-level annotations in clinical NLP written in Swedish and English respectively. The research visit took place within the Interlock project – a collaboration between the IT for Health group at DSV and the School of Medicine’s Division of Biomedical Informatics at UCSD.

The aim of the research visit was to create a taxonomy for a characterization of the expression of (un)certainty in clinical text and to compare these expressions in Swedish and English emergency ward documentation . The overall goal of this research is to improve automatic information extraction from clinical narratives.

Paper accepted to the 13th Conference on Artificial Intelligence in Medicine (AIME)

Health Informatics, Information Systems, Publication, Research Paper, SYSLAB

We’ve had a paper accepted to the 13th Conference on Artificial Intelligence in Medicine (AIME’11), which this year will be held in Bled, Slovenia (July 2-6). The paper is written by myself, Martin Hassel and Maria Kvist and is titled Diagnosis Code Assignment Support Using Random Indexing of Patient Records – A Qualitative Feasibility Study. We have explored the possibility of using statistical data on co-occurrences of words and assigned diagnosis codes (ICD-10) in coded patient records and applying these (word space) models as a means to provide computer-aided diagnostic coding support.

Third HEXAnord Meeting, Utö 22-24 September 2010

Health Informatics, Networking, Photo, SYSLAB
The third HEXAnord meeting took place on Utö in the Stockholm archipelago last week and was hosted by the IT for Health group at DSV.

HEXAnord is a network with participants from the Nordic (Sweden, Finland, Denmark, Norway) and Baltic countries (Estonia and Lithuania). In addition to SU/DSV, there were representatives from Göteborgs universitet and Linköpings universitet. The participants from DSV were Hercules Dalianis, Martin Hassel, Gunnar Nilsson, Sumithra Velupillai, Maria Skeppstedt, Helen Allvin, Mia Kvist and I.


This time around the meeting was designed as a PhD training course (3 ECTS) for the 14 PhD students, with seminar talks given on SNOMED CT (Systematized Nomenclature of Medicine – Clinical Terms), how to build essential tools and resources for the clinical domain, the nature and problems of nursing documentation, as well as on applying machine learning on clinical text to provide decision support. An assignment was given to the students by each of the lecturers, which are to be completed in groups across the universities.

The next meeting will be held in Trondheim, Norway at the beginning of next year.

Aron Henriksson, new employee

Employee, Health Informatics, SYSLAB

Hi!

I have been working full-time at DSV for a month now and would like to take the opportunity to introduce myself on the Syslab blog. I’m currently working on adapting the Web Mining course to a more distance-friendly form. I will also be involved in the research work of the “IT for Health” team, performing natural language processing on patient health records.

I have a BSc in computer science from RMIT in Melbourne, Australia and I just recently completed my Master’s degree from KTH and the EMIS program. In my thesis I explored the possibility of applying the automatic classification methods (corpus annotation and machine learning) used in the biomedical and clinical domain to that of knowledge management. More specifically, I wanted to find out whether it is possible to assess automatically the level of certainty in knowledge-intensive documents. Together with my supervisor, Sumithra Velupillai, the thesis was turned into a paper that was published and presented at a workshop (Negation and Speculation in Natural Language Processing) in Uppsala in July.


I look forward to getting to know you all more in the coming months!

Aron