HIMSS - Health Analytics Current data situation and …...Health Analytics Current data situation...

Post on 22-May-2020

8 views 0 download

Transcript of HIMSS - Health Analytics Current data situation and …...Health Analytics Current data situation...

@ ehealthNORWAY

Health AnalyticsCurrent data situation and use in Norway

Anne Torill Nordsletta, Director Health Analytics, Norwegian Centre for E-health Research

2

NORWEGIAN CENTRE FOR E-HEALTH RESEARCH

2016 – Established

3

• F

NORWEGIAN HEALTH AND CARE SERVICES

@ ehealthNORWAY

4

REAL-WORLD DATA/EVIDENCE

@ ehealthNORWAY

Environmental data

Electronic health records and registries

Claims databases

Medical imaging

Genomics

Patient monitoring devices and social media data

Digital phenotyping data

Health Data Program -Health Registries

One Health Record –One Patient

NATIONAL ACTIVITIES

5

RESEARCH INFRASTURCTURES IN NORWAY

@ ehealthNORWAY

• The Norwegian Primary Care Research Network

– Real time data from EHR

• Health Registries for Norway

• ELIXIR Norway

• Services for sensitive data

• SAFE

6

@ ehealthNORWAY

40%Clinical data is unstructured

60%Clinical data is structured

7

• Advanced statistical and machine

learning methods: – Provide solutions that turns data into

actionable insights

– Prediction

– Extract useful information for generating

advanced decision-making

• Health analytics (HA) methods– Machine learning

– Natural language processing

– Data mining

– Process mining

HEALTH ANALYTICS

@ ehealthNorway

8

Important in the new health paradigm

– Precision medicine will reshape the

healthcare

– Provide clinicians with decision tools for

decision-making

– From descriptive to predictive and

prescriptive analytics

HEALTH ANALYTICS

@ ehealthNorway

Source: BIGMED report

9

• Medical imaging

• Treatment queries and suggestions

• Drug discovery and development

• Improved care - multiple diagnosis

• Clinical pathways

• Population risk management

• Robotic surgery

• Precision medicine

• Automatic treatment

• Performance improvement

MACHINE LEARNING IN HEALTHCARE

@ ehealthNorway

10

HEALTH ANALYTICS RESEARCH ACTORS IN NORWAY AND

SWEDEN

@ ehealthNorway

• Use of deep learning networks for medical image analysis– Identify fractures in orthopedic radiographs – Karolinska Institutet, Sweden

– More targeted cancer treatment, use image analysis and deep learning, quantification of

DNA- DOMORE! project at Radiumhospialet, Norway

– Designed neural network for automatic detection of lood vessels in real-time from ultrasound

images – NTNU, Norway

• Genomics– Rare diseases, sudden cardiac death, metastatic colorectal cancer – BIGMED, Norway

• Mental health– Predict mental states such as depression in bipolar patients – INTROMAT, Norway

11

• Predict anastomosis leakage

• Early detection in pre-operativ

planning

• Early warning and decision support

• Previous study had a sensitivity of

100% and specificity was 72% with

use of bag-of-words model

Source: Ferris, Robert. Retrieved from https://www.slideshare.net/RobertFerris5/anastomotic-leak-following-colorectal-resection

STRUCTURED AND UNSTRUCTURED DATA

12

NORKLINTEKST PROJECT

Data available

NLP, statisticsand machinelearning

Prediction algorithm

Predict and identifyrisk patients

• Pre-operative planning, early warning and decision support.

• With improved specificityless expensive false alarms

Improve specificity

13

REPORTS

• BIGMED– Legal and regulatory

– Organisational

– Competence and knowledge

– Technological

– Financial and political

• The Norwegian Data Protection Authority– Privacy

• Norwegian Centre for E-health Research– Health analytics

14

HEALTH ANALYTICS IN THE FUTURE

• More health analytics research

• Raise awareness and knowledge among decision-makers and other

stakeholders

• Innovations for equitable healthcare ecosystem

@ ehealthNorway

Contact

Anne Torill Nordsletta, Director Health Analytics, Norwegian Centre for E-health Researchhttps://ehealthresearch.no/

@ ehealthNORWAY

16

REFERENCES

• Soguero-Ruiz, C., Hindberg, K., Rojo-Alvarez, J. L., Skrovseth, S. O.,

Godtliebsen, F., Mortensen, K., … Jenssen, R. (2016). Support Vector

Feature Selection for Early Detection of Anastomosis Leakage From Bag-of-

Words in Electronic Health Records. IEEE Journal of Biomedical and Health

Informatics, 20(5), 1404–1415. https://doi.org/10.1109/JBHI.2014.2361688

@ ehealthNorway