Special Session on Applications of AI Methods

Written by Ameen Abu-Hanna   



Much of the literature on AI in Medicine pertains to the question “Does the system work?” while the questions “Does it help?” and “Why it succeeds or fails?” receive far less attention. This session is dedicated to two kinds of papers addressing these latter two questions. In particular the following types are sought:

1. Papers on substantial fielded applications. The paper should attempt to demonstrate that the application actually improves clinical care. The applications should meet one or (preferably) more of these criteria:

* the application addresses a non-trivial problem which affects a substantial number of people.

* the application’s impact has been subjected to adequate validation (ultimately by interventional multicentre controlled trials).

* the impact affects a relevant process (e.g. behavior of care professionals) or outcome of care (e.g. adverse events).

* the application has been integrated into the clinical workflow and IT infrastructure and adopted in daily clinical practice.

2. Papers on important lessons learned by research teams developing and implementing applications or series of applications in clinical practice. Application papers should not (only) present concrete application results; they are primarily meant to convey insight obtained during research on real applications for non-trivial problems. One should describe general lessons learned e.g. about the nature of working with clinicians; integrating an application with an existing information infrastructure; barriers experienced such as resistance to the application or reasons for success or failure; the difference in experience between implementing a prototype and an implementation meant for a longer term, etc.


Regardless of the paper type please address the relevant points from the following list:

General: Clinical/medical setting, patient/client population, the perceived need for the application, the problem owner, the initiator of the project (research group? Clinicians?), the sponsor, the initial intended scope (prototype for research or fielded application), the progress of implementation, the publication strategy.

System: Requirement elicitation, approach for developing the application, the interface with the clinical/medical counterpart, motivation for the chosen development environment, arrangements to accessing the information infrastructure, application maintenance, plans for collecting data for future or ongoing research.

Impact/lessons learned (the most important part of the paper):  Design and results of the (clinical) trials, factors of success and experienced barriers, lessons learned, do’s and don’ts, what would you do differently in the future?


1. Authors should specifically submit to this track (the submission system will provide a way to indicate this).

2. Only long papers should be submitted in order to adequately assess their merit. The length and layout of the papers in the track is exactly the same as any other AIME paper.

3. The papers are initially reviewed for this track by the session organizers.

4. Dates for notification are the same as for the other papers.

5. Papers in this session, like the other AIME papers are eligible for selection in the planned journal special issue.

Last Updated on Tuesday, 11 May 2010 07:00