This is the Clincial Reasoning Platform central repository. It is designed to be a launch point for content and all the specific components and additional artifacts that enable and make Clinical Reasoning an extensible platform can be found. For general awareness, any Clinical Reasoning components ARE ONLY intended to be run within resources that leverage the needed upstream software or equivalent licensed software as defined below within the content. In addition, this effort through its open source licensing selection ensures appropriate protection for all parties. Red Hat Healthcare is doing this to help the broader healthcare community and while a market within Red Hat this is an independant effort.
Healthcare customers are under constant pressure to comply with Quadruple Aim guidelines(1) in their organizations. In addition, these organizations are under scrutiny to operate securely and to comply with mandates such as those from CMS(2) to provide secure access to information. If you layer on the typical business operating requirements to keep costs down and maximize profits you have the formula for a very complex IT environment, which represents a rich set of potential opportunities for Red Hat to not only help our customers but to disrupt and transform Healthcare for the better.
Healthcare organizations are looking for a common, interoperable platform of AI tools with a consistent app development and operations experience. Our goal is to make Red Hat the preferred platform for open AI computing solutions in Healthcare. We will do this by working with our ecosystem of partners and open source communities to help our customers capitalize on AI solutions based on open source building blocks. Just as importantly, we introduce our Open way of working to our customers to enable the type of cultural transformation that is a necessary precursor to any kind of digital transformation. Healthcare tends to be an average of 10 years behind other industries in modernization efforts, so it becomes even more poignant to take a holistic approach to the transformation of the classic “People, Process, Technology” (for us it’s People-Process-Platform) to make them experience successful adoption.
(1) The Triple Aim is an approach developed by the Institute for Healthcare Improvement (IHI) in order to optimize health system performance. According to IHI, the goal of the Triple Aim is to “improve the patient care experience, improve the health of a population, and reduce per capita health care costs.” IHI stresses that the strategy is a single aim with three dimensions. While this model has worked great in guiding the optimization of health systems since its inception, recently, an additional aspect has been adopted by many healthcare professionals — improved clinical experience — leading to the creation of the Quadruple Aim.
(2) CMS Final Rule mandates cover a range of areas including access to patient data, notification of significant health events, and EHR certification. For a good summary and calendar, refer to this link.
"Between 12 to 18 million Americans every year will experience some sort of diagnostic error," said Paul Cerrato, a journalist and researcher.
This is the best description we have found yet: Clinical Reasoning: Defining It, Teaching It, Assessing It, Studying It
Excerpt from the above article: “[...]we can describe the clinical reasoning process as including the physician’s integration of her own (biomedical and clinical) knowledge with initial patient information to form a case representation of the problem. The physician uses this problem representation to guide the acquisition of additional information and then, on the basis of this information, revises the problem representation. She repeats the information gathering – representation revision cycle until she reaches a threshold of confidence in that representation to support a final diagnosis and/or management actions.”
In addition, there are two main parts of CR: diagnosis and management. Most of the time the focus is on diagnosis because it can more easily be “scored” as right or wrong while management (ongoing therapy) often involves opinion, best practices, standards of care, etc. and can vary quite a bit.
As the author notes, there is no universally accepted definition of clinical reasoning, but I think this gives a good working definition.
To understand where Red Hat can play in this space and offer differentiated solutions, let’s first define a few terms:
- Artificial intelligence is intelligence demonstrated by machines, unlike the natural intelligence displayed by humans and animals, which involves consciousness and emotionality.
- Machine learning is the study of computer algorithms that improve automatically through experience and by the use of data. It is seen as a part of artificial intelligence.
- Data science is an interdisciplinary field that uses scientific methods, processes, algorithms and systems to extract knowledge and insights from many structural and unstructured data.
Clinical Reasoning solutions encompass the AI domains of Machine Learning, Decision Management (business rules), Business Optimization (business process automation, complex event processing, constraint optimization), and IT automation. This translates to the middleware products Decision Manager, Process Automation Manager and OptaPlanner, IT automation with Ansible.
Clinical Reasoning also includes running Machine Learning workloads on OpenShift with or without Open Data Hub, and key partner solutions, particularly in the areas of Data Science and algorithm development.
All of the AI tools mentioned above run on OpenShift and have a common architecture (i.e. run as Operators) and common set of APIs. This gives Red Hat a unique advantage in the marketplace, giving our customers one place to run all of our AI tools and to choose the right tool for the job, without sacrificing interoperability or experiencing vendor lock-in, since all of the tools have corresponding upstream communities.
Any of these tools can give answers that are predictive or prescriptive and they are all data-driven and can be knowledge-driven, meaning they can incorporate conclusions (example: scores) from prior iterations or other tools. Any good toolset should also allow you to simulate models before moving them into production.