Periodic Reporting for period 1 - eCREAM (enabling Clinical Research in Emergency and Acute care Medicine through automated data extraction)
Berichtszeitraum: 2022-09-01 bis 2024-02-29
High-quality emergency departments (EDs) are crucial to ensuring the best patient care. In order to improve the clinical and organisational performance of EDs, their work needs to be continuously assessed based on validated criteria. Doing so requires the collection of trustworthy data as part of clinical and quality-of-care research, but this poses a significant challenge in the emergency medicine setting. Carrying out research in this setting is almost impossible because the large number of patients who visit the ED and the chronic staff shortages that characterise most EDs make ad-hoc data collection unachievable.
The only way to fill the gap between the need for clinical research and the availability of robust data is to automatically extract the necessary data from ED electronic health records (EHRs), avoiding dedicated, time-consuming data collection. This is a complex task because, while part of the data in EHRs is structured (e.g. laboratory test results), most useful information is contained in free text notes written by the nurses and physicians (e.g. the presence of signs and symptoms or the anamnesis).
The eCREAM project was launched to address this impasse. eCREAM is an international project involving partners from 7 countries (Greece, Italy, Greece, Slovakia, Slovenia, Switzerland, UK). More than 30 EDs from these countries will participate.
One of eCREAM’s main aims is to develop new solutions, including artificial intelligence-based natural language processing (NLP), to extract clinical information from different medical sources and create high-quality databases on patients visiting EDs. eCREAM will also develop a new EHR for the centres that wish to adopt it. It will prioritise the collection of solid, trustworthy data without increasing the time staff dedicate to inputting patient data.
eCREAM closely adheres to the Open Science paradigm and will take into account the diverse data access needs of different end users (clinicians, researchers, health policymakers, and citizens). It will also work to make the data easily accessible and reusable.
To test the practical validity of its solutions in real-life scenarios, eCREAM will address two use cases that represent major issues in emergency care. The first use case involves the propensity of an ED to hospitalise patients; the second involves exploiting ED data to develop real-time dashboards to inform clinicians, citizens, and healthcare policymakers about the current situation in EDs to help them make informed decisions.
Scale and significance of expected impacts
Sharing the unique data collected by eCREAM will profoundly impact the scientific, clinical, and public health communities, contributing to the European Health Data Space and giving the EU higher visibility in terms of health technology development. eCREAM will create databases of enormous value for the scientific community. Researchers and other end-users will have access to structured ED data for additional research, and the NLP community will benefit from the availability of eCREAM resources to advance the state-of-the-art in information extraction in the medical domain.
Once the eCREAM system has been set up, we intend to extend it to other groups of patients and EDs and enlarge the eCREAM network, even after the project’s official end, increasing the scale and significance of the action.
eCREAM’s creation of dashboards to empower users, permitting them to make appropriate decisions based on their needs, will have positive effects on several levels, from allowing patients with non-urgent conditions to choose less crowded EDs, thus reducing waiting times and overcrowding at other facilities, to allowing local emergency medical services to better organise emergency transfers, thus reducing the pressure on high-volume EDs and, ultimately, increasing the quality of care.
Lastly, eCREAM’s aim to create the conditions for making research in emergency medicine sustainable over the long term involves the development of a tool to extract data from EHRs and, even more importantly, the development of a new EHR. These tools will allow data to be collected directly in the field, freeing research from the limits imposed by emergency medicine and revealing many questions that have yet to be answered. In doing so, the barriers hindering research in emergency medicine will be broken, representing the game-changer in the field.
The only way to fill the gap between the need for clinical research and the availability of robust data is to automatically extract the necessary data from ED electronic health records (EHRs), avoiding dedicated, time-consuming data collection. This is a complex task because, while part of the data in EHRs is structured (e.g. laboratory test results), most useful information is contained in free text notes written by the nurses and physicians (e.g. the presence of signs and symptoms or the anamnesis).
The eCREAM project was launched to address this impasse. eCREAM is an international project involving partners from 7 countries (Greece, Italy, Greece, Slovakia, Slovenia, Switzerland, UK). More than 30 EDs from these countries will participate.
One of eCREAM’s main aims is to develop new solutions, including artificial intelligence-based natural language processing (NLP), to extract clinical information from different medical sources and create high-quality databases on patients visiting EDs. eCREAM will also develop a new EHR for the centres that wish to adopt it. It will prioritise the collection of solid, trustworthy data without increasing the time staff dedicate to inputting patient data.
eCREAM closely adheres to the Open Science paradigm and will take into account the diverse data access needs of different end users (clinicians, researchers, health policymakers, and citizens). It will also work to make the data easily accessible and reusable.
To test the practical validity of its solutions in real-life scenarios, eCREAM will address two use cases that represent major issues in emergency care. The first use case involves the propensity of an ED to hospitalise patients; the second involves exploiting ED data to develop real-time dashboards to inform clinicians, citizens, and healthcare policymakers about the current situation in EDs to help them make informed decisions.
Scale and significance of expected impacts
Sharing the unique data collected by eCREAM will profoundly impact the scientific, clinical, and public health communities, contributing to the European Health Data Space and giving the EU higher visibility in terms of health technology development. eCREAM will create databases of enormous value for the scientific community. Researchers and other end-users will have access to structured ED data for additional research, and the NLP community will benefit from the availability of eCREAM resources to advance the state-of-the-art in information extraction in the medical domain.
Once the eCREAM system has been set up, we intend to extend it to other groups of patients and EDs and enlarge the eCREAM network, even after the project’s official end, increasing the scale and significance of the action.
eCREAM’s creation of dashboards to empower users, permitting them to make appropriate decisions based on their needs, will have positive effects on several levels, from allowing patients with non-urgent conditions to choose less crowded EDs, thus reducing waiting times and overcrowding at other facilities, to allowing local emergency medical services to better organise emergency transfers, thus reducing the pressure on high-volume EDs and, ultimately, increasing the quality of care.
Lastly, eCREAM’s aim to create the conditions for making research in emergency medicine sustainable over the long term involves the development of a tool to extract data from EHRs and, even more importantly, the development of a new EHR. These tools will allow data to be collected directly in the field, freeing research from the limits imposed by emergency medicine and revealing many questions that have yet to be answered. In doing so, the barriers hindering research in emergency medicine will be broken, representing the game-changer in the field.
Concerning the development of a solution to extract EHR data, the partners have been collecting general medical domain data for training the NLP models to recognize the necessary information from the EHRs' free text notes and have been assessing the models’ performance to choose the most suitable option. The capacity of the participating EDs’ computer systems has been assessed to ensure the data extraction system can run smoothly.
Concerning the new EHR and the first use case, the partners have met with clinicians in the participating countries and agreed on a set of data to collect. A mock-up of the future EHR, mimicking what the EHR will look like and how it will work (displaying the variables to be collected, with clickable buttons activating available functions and options), is well underway and is being used to discuss with nurses and physicians their needs and preferences. For the second use case, work has begun between the partnering countries to define the data to be collected and shown on the dashboards to satisfy the different end users.
Work has also been carried out to guarantee the project’s aim to enhance the findability, accessibility, interoperability and reusability of the services and databases created. A strategy for maximizing the exploitation of the collected data for secondary research and ensuring compliance with ethical, legal and societal issues (ELSI) requirements, especially the GDPR and national data protection regulations, has also been set up.
Communication and dissemination activities to raise awareness of the project, including the creation of the eCREAM website, a LinkedIn account, a professional video, posters, and participation in scientific meetings, have been carried out.
Concerning the new EHR and the first use case, the partners have met with clinicians in the participating countries and agreed on a set of data to collect. A mock-up of the future EHR, mimicking what the EHR will look like and how it will work (displaying the variables to be collected, with clickable buttons activating available functions and options), is well underway and is being used to discuss with nurses and physicians their needs and preferences. For the second use case, work has begun between the partnering countries to define the data to be collected and shown on the dashboards to satisfy the different end users.
Work has also been carried out to guarantee the project’s aim to enhance the findability, accessibility, interoperability and reusability of the services and databases created. A strategy for maximizing the exploitation of the collected data for secondary research and ensuring compliance with ethical, legal and societal issues (ELSI) requirements, especially the GDPR and national data protection regulations, has also been set up.
Communication and dissemination activities to raise awareness of the project, including the creation of the eCREAM website, a LinkedIn account, a professional video, posters, and participation in scientific meetings, have been carried out.
Although the project is in its first phase, we have already achieved some results, involving the design of the new EHR, that can be disseminated.
We promoted qualitative research in two EDs to study the processes, relational dynamics, human-computer interactions, and the organisational cultures present. A sociologist conducted an in-depth analysis based on organisational ethnography using participant observation, interviews, documentation collection, and shadowing. The aim was to gain a thorough understanding of the context in which the new EHR will be installed to provide valuable insights to its developers to maximise its functionality.
The second result of note involves the systematisation of the data flow in the ED and was based on the qualitative research described above, on numerous focus groups with ED professionals, several site visits to the wards, and a literature review. This is a relevant result that may help define a standard format of ED EHRs, in line with what has already been done by the European Commission under the European Electronic Health Record exchange format (EEHRxF) recommendations (https://digital-strategy.ec.europa.eu/en/library/recommendation-european-electronic-health-record-exchange-format(öffnet in neuem Fenster)).
We promoted qualitative research in two EDs to study the processes, relational dynamics, human-computer interactions, and the organisational cultures present. A sociologist conducted an in-depth analysis based on organisational ethnography using participant observation, interviews, documentation collection, and shadowing. The aim was to gain a thorough understanding of the context in which the new EHR will be installed to provide valuable insights to its developers to maximise its functionality.
The second result of note involves the systematisation of the data flow in the ED and was based on the qualitative research described above, on numerous focus groups with ED professionals, several site visits to the wards, and a literature review. This is a relevant result that may help define a standard format of ED EHRs, in line with what has already been done by the European Commission under the European Electronic Health Record exchange format (EEHRxF) recommendations (https://digital-strategy.ec.europa.eu/en/library/recommendation-european-electronic-health-record-exchange-format(öffnet in neuem Fenster)).