HSE Undertakes the Implementation of the DCEAF Project
DCEAF (pronounced: /’disi:f/) stands for Dual Clinical Evaluation of the ASCAPE Framework and is an Open Call project in the context of ASCAPE EU-funded* project. HSE is running this cascade funding project to evaluate the ASCAPE Open AI (Artificial Intelligence) Framework in the context of a study to be conducted at the General Oncology Hospital of Kifissia “Agioi Anargyroi“, collaborating with the hospital’s clinicians and key scientific personnel. DCEAF commenced in March and is expected to deliver its results in November 2022.
ASCAPE is an ambitious research project, but is it relevant to real-world clinical practice? DCEAF will aim to install, integrate, and even extend the ASCAPE Open AI Framework at a busy general oncology hospital and have it evaluated by its clinicians all within a very short time frame. The technical aspects of this exercise will aim to demonstrate ASCAPE can be deployed at hospitals around the world, even extended, with minimal effort and that patient reporting of Quality-of-Life issues can be integrated in clinical practice in an effective and efficient manner. The clinical evaluation will provide a clinical perspective on whether the ASCAPE Open AI Framework is a useful addition to clinical practice or not, based on a study that aims to simulate (or otherwise factor into the evaluation) key aspects of a realistic vision for ASCAPE’s integration into clinical practice in busy oncology clinics.
Bringing AI into the hospital setting
DCEAF researchers will conduct a non-interventional prospective study during which patients will be asked to provide information on their Quality-of-Life issues. The design of the study is meant to minimise patient and clinicians’ effort in this process, in order to emulate a real-world scenario of ASCAPE being integrated into clinical practice. The study will involve breast cancer patients and colon cancer patients. As ASCAPE did not support colon cancer prior to DCEAF, the project will add support, demonstrating also the abilities of ASCAPE and identifying any challenges in extending ASCAPE to other cancer types. Finally, diverging from ASCAPE’s doctor-focused approach, DCEAF will aim to demonstrate how ASCAPE facilities can be useful in the context of clinical practice where nurses have a significant role also.
What’s DCEAF approach?
Researchers from HSE and hospital’s clinicians will evaluate the ASCAPE AI framework based on the DCEAF study’s data. As DCEAF values data privacy, a dedicated ASCAPE server will be deployed on the Hospital premises so the study participants’ data will remain on premises. The DCEAF study will record Quality-of-Life data at baseline and three months later (follow up). Doctors and nurses will be able to evaluate ASCAPE’s early warning/proactive monitoring capabilities, predictions, and intervention suggestions at both these time points while they will be also asked to provide Quality-of-Life predictions and recommendations of their own, as well as their opinions on the usefulness of the ASCAPE Open AI Framework. Cross-validation based evaluation of the ASCAPE’s AI predictive capabilities will also be conducted. Evaluation of results provided by ASCAPE AI will focus on the breast cancer data, since ASCAPE has already collected enough relevant data for AI models training, whereas for colon cancer the evaluation will focus primarily on the ease of adding support for a new cancer type.
Does it matter?
It sure does! Since DCEAF will aim to evaluate ASCAPE’s utility in scenarios informed by the challenges of clinical practice, acknowledging the role of nurses in it and the importance of minimising clinicians’ workload by providing all the automations necessary to achieve this aim. At the same time, it will add a new cancer to ASCAPE’s list of supported cancers. In a broader perspective, DCEAF aims to facilitate the uptake of ASCAPE’s AI solution, streamlining it and making it an attractive option for more hospitals.
*The ASCAPE project, and subsequently the DCEAF project, have received funding from the European Union’s Horizon 2020 research and innovation programme under grant agreement No 875351