POSTDOC POSITION in Using machine learning techniques to identify anomalous data in electronic health records used for international stroke care quality monitoring
Masaryk University, Brno, Czech Republic invites excellent scientists to apply for
in Using machine learning techniques to identify anomalous data in electronic health records used for international stroke care quality monitoring
The Stroke research group at FNUSA-ICRC is one of the most active participants in the European Stroke Organisation’s ESO EAST programme. ESO EAST (Enhancing and Accelerating Stroke Treatment) is focused on improving the quality of care in participating countries, and raising the overall level of stroke treatment throughout Europe.
In 2016, as an initiative of the ESO EAST programme, the FNUSA-ICRC Stroke group developed RES-Q, the Registry of Stroke Care Quality (https://qualityregistry.eu/). RES-Q is an international registry designed to provide a free-to-use platform for hospitals to monitor the quality of their stroke care based on key performance metrics. RES-Q provides an online eCRF tool for collecting data from hospitals, as well as a real time analytics engine for delivering updates and reports to participating hospitals and national stroke societies. In the last 3 years, RES-Q has grown to include many countries within Europe, as well as globally, and now has more than 1000 participating hospitals from 61 countries. Data from more than 130,000 patients is currently housed within RES-Q.
One of the key challenges in operating a clinical registry with the scope of RES-Q is in ensuring the quality and validity of data which is captured by participating hospitals. However, traditional approaches to clinical trial management, especially on-site auditing, are not possible in this context as most participants are from low-resource areas, and the funding required would be unsustainable.
To address these issues, the Stroke group will be launching a research project focused on implementing machine learning algorithms for the detection of anomalous data in electronic health records (EHR). The project will focus on the implementation of unsupervised anomaly detection algorithms, primarily utilizing clustering-based methods, to identify outliers in the collected data with greater sensitivity and specificity than possible with conventional anomaly detection approaches such as standard deviation analysis.
The primary project output will be a viable automated detection process, implemented in software, which can identify anomalous data entered in the RES-Q EHR database. The software implementation will allow participants to review anomalous data for correction or explanation. Secondary outputs will be research papers regarding the development and implementation of the necessary machine learning algorithms, and potentially analysis of the existing RES-Q data.
The successful candidate should:
- be a researcher who has received a PhD or its equivalent within the last 7 years
- be a researcher who has worked at least two whole years in the last three outside the territory of the Czech Republic in the field of research with a working time of at least 0.5 full-time equivalent, or who has been PhD student (or equivalent) abroad
- have a publishing record – in the last three years at least two publication outputs registered in the Thomson Reuters Web of Science, Scopus or ERIH PLUS databases and at the same time publications such as “articles”, “books”, “book chapters”, “letters” and “reviews”.
Specific criteria can be filled, i.e.:
- have experience in applying machine learning algorithms to healthcare research data
- the primary educational background should be from the fields of computer science, applied mathematics, health informatics, medicine, or other related fields
- a cross-functional understanding of both the medical and IT fields, however as the project outputs are primarily algorithms and software applications, experience in these areas is strongly encouraged for successful completion of the project
- experience with programming languages and frameworks common to machine learning tools
- a background in neurology or cardiovascular treatment would be an asset
- have excellent communication skills and the ability to collaborate in teams
The application should include:
- a CV including a summary of education and research experience, publication activity, involvement in research grants, etc.
- a scanned copy of the PhD diploma or an official letter certifying submission of a doctoral thesis for thesis defence and the planned defence date
- a motivation letter
- at least two reference letters
MU offers the opportunity to get:
- an interesting job in a dynamically-expanding university area
- diverse and challenging work in an excellent research environment
- tenure track with an initial appointment for 2 years
- a professional team and pleasant working conditions
- interaction with leading scientists in an inspiring, internationalised environment
- a welcoming service for the successful candidate and his/her family
Anticipated start date: The position is available from March/April 2020 or upon agreement but no later than by 30th November 2020.
The submission deadline is 29th February 2020.
Please submit your application by e-mail to email@example.com
A review of applications will commence immediately after the deadline. Short-listed candidates will be invited for interview within one month of the deadline.
Further information about:
- Postdoc@MUNI is available at http://postdoc.muni.cz
- Masaryk University is available at https://www.muni.cz/en
- Brno is available at https://www.gotobrno.cz/en/
We will send you a link to the e-application