While waiting for the wave of Covid patients, Sebastian recognized that admitting symptomatic patients to the hospitals is inefficient and fills up capacities quickly, and unnecessarily. There had to be a better way to keep the not-so-heavy cases out of the hospital, but would you know they're ok? Welcome to the idea of DigitalHospital (formerly hospital@home).
What it does
Vision statement: We slow the transmission and save the critical hospital resources by monitoring patients at home through hospital@home, a decentralized medical monitoring system.
During a pandemic crisis, such as the Covid-19 crisis it is vital to prevent a healthcare-system-meltdown.
- If not managed properly, the exponential growth of case numbers will overhelm the healthcare system, leading into the worst-case scenario of a “melt-down”.
- The first goal in the pandemic is to protect and and maximize the use of core healthcare assets, staff, beds and machines. The key strategy to implement this goal therefore has to be to keep Covid-19 cases out of hospitals for as long as possible, while making sure they`re safe.
Target Groups: The system is relevant for different target markets and user roles, such as public hospitals, or healthcare-system-responsible, and it enables them to securely monitor and support infected patients in homecare.
Needs: These are the most relevant needs that the system and the proof-of-concept is addressing:
- Robust system that can securely take care of patients with Covid-19 pneumonia for days, isolated and safe in their home environments, while being monitored for changes in vital signs and health status.
- Ability to securely monitor signs of deterioration in the group of patients at home by wearable monitoring, such as declining blood oxygen saturation (SpO2).
- Enable a central operator at hospital/ regional level to detect and react to these signs early enough, so that the patient at home can immediately be hospitalized, when in need.
- Enablement of data support for fact based management of healthcare infrastructure
- Full legal compliance for processing sensitive personnel data as by region of deployment and use (i.e. DSGVO)
Product: Key parts of the product concept are
- Open interface to wearable devices used locally including abilities to securely forward vital-sign-data into a central cloud infrastructure (i.e mobile app)
- Fault-tolerant stream processing of inbound datastreams within proven cloud infrastructure
- Features and processes that support the operator-in-duty to take educated decisions for hospitalization events for patients in need including implementation of these decisions.
- Features that support patient and operator communication.
- Features that support medical personnel to define and fine-tune the hospitalization decision.
- Analytics and features that support the management of the healthcare system.
Business Goals: Our product's proof of concept enables the following business goals:
- Support the public good "preservation of human lives" during a pandemic
- Manage relevant parameters for over-utilization of hospitals
- Reduction of overall infection risk (decentralization of infected persons)
Enable a revenue perspective that supports ongoing implementation, operation and maintenance:
- Open source platform that is available globally
- Enablement of local or regional managed-cloud business cases
- Establish common and open platform for suppliers that work on interfacing solutions (i.e. sensors, emergency intervention services,….) including the ability to create business benefits.
- Cost optimization: Optimization of the scare, system-critical resource “hospital bed”, including related infrastructure of all kind.
Architecture and Design: The following characteristics describe the software architecture and the software design:
- Collection of data by means of a locally-deployed wearable monitor.
- Transmission of this health-related data by means of local desktop or mobile devices into a stream processing engine within a cloud based data-lake infrastructure. We base our proof-of-concept on the Google cloud platform, but from the technical perspective it is possible to use offerings from different cloud infrastructure providers as well.
- Stream processing by means of a pub-sub logic, including data-parsing and message processing, with data storage in SQL/NOSQL data warehouse
- The ingested and cleaned data is visualized in dashboards that are exposed as https services to the outside world (at current bokeh server)
- Users securely access data and dashboards by means of a identity-aware proxy
UI/UX: Using figma as a tool, we implemented mockups that demonstrate the interface design for these major building blocks and use cases. During the hackathon our design team added some further use cases to the initial concept.
Scroll through the assets above in the video section (graphical product, presentations, etc) and see the links further below to find more detailed information about the solution.
How we built it
For the initial hackathon, we started as a newly-founded global team of diverse qualifications and skills (healthcare, product management, backend-, and mobile-development, design, etc). We mostly did not know ourselves beforehand.
Starting from a common product vision, in an team effort, and using standard agile development methods, we
- Collected an initial set of user stories involving different roles (including users/ medical personnel).
- Ranked and evaluated the user stories including scoping of a minimum viable product.
- Mapped the user perspective to requirements and by this defined the feature scope of our product.
- Used the product concept to support downstream product development phases for UI/UX or software architecture.
- Implemented an initial set of software modules and created assets that cover identified and central requirements. These modules demonstrate the technical feasibility of important parts of the concept. Examples are features that demonstrate the data ingestion processes using big query.
- Implementated mockups for mobile patient apps and for the operator dashboards by UI/UX experts.
Progress Highlights in the EUvsVirusHackathon
We continued to add features and functions to our initial concept, and we started to implement the different building blocks. In particular we enhanced the product in the following direction:
- Implemented the Mobile App for patients
- Started to develop the Desktop Frontend of the Operators' Dashboards
- Collected feedback from potential users in hospitals or from ICU - Teams, and enhanced the original pure home-care usecase to cover scenarios for make-shift hospitals or nursing homes.
- Ranked the requirements and created/ enhanced the project development plan
- Implemented chatbot functionality in the patient app
- Following user feedback the initial version will concentrate on patient-operator dialog, and will add then sensor data process in a subsequent step.
See the links further below to find more detailed information about our process and the deliverables.
Challenges we ran into
- We did not face any substantial content-related challenges.
- As a global team we face time-zone issues
Accomplishments that we are proud of
- As a newly established diverse and global team we almost instantly hit the road running.
- By means of different building blocks that integrate one into the other, we are able to show that it is realistically possible to implement the intended approach to monitor patients at home.
- Therefore it is secure to assume that a decentralized approach can be employed to help prevent meltdowns of health systems during an epidemic crisis.
What we learned
- Experience and diversity is key
- Hackathons help to create momentum
- User feedback is important to create a meaningful solution
- Implementation with volunteers
What's next for DigitalHospital (formerly hospital@home)
- Due to time restrictions of a hackathon, it is not possible to give each work package the depth of attention that a fully professional product concept would deserve. However it demonstrates the feasibility of the concept. And it is a perfect starting point for continuous work.
- Along the given lines, all deliverables need to be extended and further detailled-out (i.e user interviews with different roles and in a larger sample), once development activities on the scoped product reach the go-milestone, and decision.
- We are aware that the system processes highly sensitive personal information. Therefore the entire aspects of data security need to be detailed out and implemented (i.e. DGSVO) for example using thread modeling techniques.
- Due to the size of the solution, support by experienced technology providers might be of advantage in the realization phase.
The sourcecode of the different modules is stored in our GitHub. You find further information about the working-results in the video section above and by following these links (all deliverables were created or enhanced in the EUvsVirus Hackathon):
- Project presentation - Project Overview + Business Plan
- Concept, use cases, requirements and project plan - Detailed descriptions of the project
- Mockup for the Operator's Dashboard - Mock on Figma
- Mockup of the Patient App - Mock on Figma
- Mockup for Logo and Onboarding - Mock on Figma
Sources and Downloads