AEGIS High Level Scenarios – Scenario #3 – Data-Driven Monitoring and Alert Services for Impaired or High Risk Groups Individuals

posted in: Blog | 0

This is the third blog post out of the 5 blog posts that will provide a view of the 5 AEGIS high level scenarios.

To find more about the AEGIS scenarios, you can visit the first blog post of this series.


Scenario 3: Data-Driven Monitoring and Alert Services for Impaired or High Risk Groups Individuals


The background Amanda is the head data analyst of SameHealthForAll Inc, a private-public social care service provider that is co-funded by the government and some private funds in order to provide health services to individuals and to other relevant organisations as well, acting both as an intermediary expert and service provider to primary and secondary health institutions and professionals, but also as an organisation that has direct interface with individuals.

Amongst the key strategic SameHealthForAll Inc goals, is the design and roll-out of specific health related services addressed to individuals that have certain diseases or conditions that classify them to high risk groups (like for example Alzheimer or heart condition patients, disabled people and pregnant women), to which also elderly people belong. The services offered to this group by the organisation vary from formal recommendation and notes that are issued towards the health care providers, to informal or simplified guidelines and notifications issued to the general public via the press, flyers and media channels. Those are mostly based on previous experience, and one of Amanda’s daily tasks is to go through the news coming from the press, other health agencies and related organisations (such as environmental agencies) and indicate incidents above certain thresholds that are connected with past experiences. Once such an incident is identified, Amanda gathers her team and, after consultation, the risk factor of an incident is determined. If this is characterised as “high”, then the appropriate procedures for raising awareness are triggered, with announcements being issued, updates of protection guidelines being drafted and communication to the relevant stakeholders being performed.

The identification of such incidents has been made lately more straightforward thanks to the introduction of novel ICTs that are able to grasp data from external resources, whether these are of public and open nature, or private, belonging to organisations with which SameHealthForAll Inc. collaborates and has access to notification and data subscription services. However, the organisation is still constrained to the provision of recommendations and services of generic nature, which although valuable, are not that intelligent and their potential is limited by various constraints. The main constraints of those have to do with the inability to tackle needs which are way more complicated and highly dependent on the context of each individual, as well as locality, as most data acquired refer to large (often nation-wide) geographical areas.

The driving need In the quest of finding ways to offer advanced services, which could be directed both to other organisations, but also directly to high risk groups individuals, Amanda has forwarded a proposal to the board of directors to establish a strategic partnership called “Health2home2020” with specific healthcare providers, smart home and automation service providers and other organisations, who are able to provide the missing pieces to complete the envisaged services offering, by utilising their resources and infrastructure to serve SameHealthForAll Inc. with data coming from individuals and from their environment, as well as data coming from other known clinical data sources that are essential for conducting the analyses. Of course, apart from the willingness of organisations to share data and value, other issues have to be resolved, ranging from the absence of infrastructure for analytics, to common data schemas and data confidentiality.

During the latest discussions in a cross-sector meeting of the “Health2home2020” alliance, Amanda suggested to the partnership to base their operations on the offerings of the AEGIS platform, as it promises many solutions that solve most of the challenges that have been identified.

The analysis Phases Visiting the AEGIS platform, Amanda creates an organisation profile for the company and then creates a project called “Assisted ALL Analytics” marking it as a “private” project. In the project, Amanda invites some of her colleagues to have access to the same project, giving them full access to the data. At the same time, she also sends invitations to external collaborators with specific rights to upload data to the project’s repository, indicating also the expected data structures and schemas.

The other members of the “Health2home2020” alliance, having received the invitation from the platform, join the project and are presented with the list of data that they have been asked to upload. At their disposal are a number of tools that could anonymise, clean and transform data, in order to meet the expectations of Amanda. After specifying what kind of data they will upload, specific dialogues guide them through the process, prompting them to take actions (possibly leveraging AEGIS tools) towards ensuring compliance with data privacy terms, in case the nature of the data to be uploaded imposes such requirements. Once they do this, they are presented with their data, while Amanda is able to cross-check that these data comply with the necessary requirements, and if not transform them accordingly. This step is done multiple times, as some data are initially uploaded only for experimentation purposes (e.g. some samples of datasets), while there is also the option to connect data to the platform through a project specific API endpoint.

Having all the data in one place, Amanda is now able to invoke several analyses, choosing which data to combine as well as the algorithms to utilise. Those come out of a predefined algorithms library, while it is also possible for Amanda to conduct an analysis using her own algorithms, which, however, requires an additional step in order to pass an approval process by the system’s administrator, to guarantee proper resources utilisation in AEGIS, following the request to include a new algorithm that has to trigger a consultation and testing process led by the AEGIS team. The overall results are then presented in a dashboard that visualises the outputs of the analyses, where access can be provided to any member of the project, while the results can be also exported in various formats. What is especially interesting for Amanda, is the option to export the data through an API. This can come either from an external stimulus, such as a weekly call from an external system residing in the “SameHealthForAll Inc” datacentre, or from triggers specified and enabled in the AEGIS platform, such as the updating of a dataset or the occurrence of an event.

At the moment, she has saved two analyses, which are called “IndoorConditionsRisks” and “TemperatureRiskPlots”. The first one is able to relate indoor air quality conditions, as retrieved from IAQ sensors installed in premises, further correlated with HVAC devices operational status, IAQ data from external weather stations and IAQ Exportconstrains/regulations as defined by national and international health organisations. This correlated analysis will further enable:

  1. home automation services to trigger the appropriate control strategies on HVAC devices towards addressing IAQ requirements
  2. social care service providers to trigger notifications about IAQ conditions.

The second one is correlating positioning and health information gathered from wearable devices, location and weather conditions along with Public Health Information Sources towards the identification of:

  1. possible environmental conditions that pose risk for the health and well-being of individuals (e.g. alert individuals suffering from COPD in case of increased humidity outdoors),
  2. pattern irregularity which could signify cognitive deterioration (e.g. wandering off without any apparent reason),
  3. physical wellbeing deterioration and frailty status (e.g. detection of falls or taking the individual more time to complete typical physical routines).

In all those analyses, specific user groups are formed based on various aspects such as their demographics, their conditions etc.

At the same time, all involved stakeholders of the alliance are also permitted to upload their data publicly, setting specific licence requirements for reuse, and can also explore data and services already uploaded in the platform that will help them to further improve the value of their in-house information. In the same manner, Amanda is also able to identify in the platform a social media mining service that sends notifications based on specified keywords specified. After she accepts the license agreement she starts experimenting with it by asking the service to return alerts coming out of Twitter stream analysis regarding ice incidents at the greater district of Cambridge. She also decides to utilise these notifications as a trigger to run again the analysis called “TemperatureRiskPlots” that has been part of the project she initiated. Furthermore, she also finds a free service about weather conditions that offers more detailed information than the one she is currently using, and decides to replace this data stream in her analysis.

The benefits Having conducted the analyses, Amanda presents to the alliance the benefits of AEGIS and persuades them to utilise the service at production level, as it fully covers the needs of the whole alliance, while it does not require high investments costs, or the presence of a data analyst in each organisation. By doing this, Amanda is able to gather data for the platform to build the services she has envisaged based on the outcomes of the analyses. These enable her organisation to develop application services that offer personalised information and recommendations to stakeholders. These applications utilise both the data streams of the AEGIS analytics, as well as the intelligence coming out of them, to define specific rules and triggers for sending out information. At the same time, some of these services (depending on data nature and strategic value) provide data back to the AEGIS platform, either directly, through API calls to the platform’s endpoints, or indirectly, as in the case of healthcare institutions who have stated that they prefer to first retrieve and review data and then re-upload selected parts of them to the platform.

At last, the services are deployed, and the organisation, together with all the value chain collaborators are able to offer added value services to their end-users, as well as to each other. Furthermore, new data and service exploitation models are introduced through the platform, as essential information such as indoor conditions from sensors, as well as activity tracking information of anonymised personas are offered online, and other interested stakeholders can experiment with them and then come to an agreement for their reuse with the data owners.


If you want to learn more about the AEGIS scenarios and the steps we take towards bringing them to life, stay tuned for our upcoming blogposts!

If you can’t wait, you can always start by checking out all our scenarios here.

Blog post authors: SUITE5