LynxCare had the opportunity to attend the exciting AIMed North America Conference that was held in Dana Point, CA from the 13th until the 15th of December 2018: Three-days dedicated to the transformative impact that AI-inspired technology is having on healthcare; a  platform designed by clinicians to showcase latest thinking and facilitate new ideas and partnerships. The Conference agglomerated 150+ speakers such as Clinicians, Physicians, CIOs, CMOs, CTOs & Senior Technology Leaders, Healthcare & Medical App Designers & Developers, Data & Computer Scientists, Academics such as PhD researchers, engineers & practitioners and finally Investors and Entrepreneurs, Industry Leaders in Healthcare, Biotech & Pharma industries.

In this blog post we would like to share the main highlights of the Conference, particularly 3 very impactful discussions initiated by Lynda Chin, Associate Vice Chancellor and Chief Information Officer at the University of Texas, Dennis Wall, Associate Professors of Pediatrics at Stanford University and Leo Anthony Celi, Principal Research Scientist at MIT.

Data conundrum & security in healthcare

Dr. Lynda Chin – Executive director of Real-World Education, Detection & Intervention at University of Texas – highlighted the importance of security, transparency and trust in the data conundrum. To start her story with, she stated it is important to see the patient as a person and not just as a diagnosis. There is an important need to change the infrastructure in a secure way in order to connect data TO analytics & insights TO value for patients.” It is indeed true that analytic insights alone don’t help the patients, we have to consequently create and implement a workflow (based on the insights) in order to get the value to the patients and improve quality of life of patients. Dr. Chin praised for a connected ecosystem in order to go from connected data to analytics to insights to action to value (see pic below). She compared it to a ‘digital highway‘ meaning that it’s not because we have highways that we could decide that once per week only yellow cars can drive on the highway. It’s about infrastructure and organization; all cars can access the highway and it should be the same for healthcare, no patient should feel left out!

In order for us to overcome the data challenge, we need to start thinking about infrastructure (where the data is, how to use it and when), only then we will be able to aggregate the correct data. Dr. Chin also mentioned the fact that data science is a tool that clinicians can use and so the clinician needs to be in the driver seat (meaning INVOLVED when developing clinical AI tools). However, when it comes to delivering healthcare, clinicians are not ready and don’t have the knowledge on how to use these tools. It is very important for healthcare in general that they learn how to make use of them and that’s why clinicians and data scientists need to closely work together.

Medical imaging: Cardiology/ Pathology/ Radiology

Dennis Wall, founder of Cognoa, shared his thoughts on AI in medical imaging.

Cognoa is developing AI-based digital diagnostics and personalized therapies that are designed to make earlier diagnoses and effective treatments available to more children to improve outcomes and lower behavioral healthcare costs. When more physicians are empowered to identify and begin treating behavioral conditions and developmental delays earlier, more children have the opportunity to benefit from treatments at a younger age when there is the greatest potential for improved lifelong outcomes. Cognoa currently provides the Cognoa Child Development app via partnerships with employers, payers, and ABA therapy centers to empower parents and caregivers to better support their children’s unique behavioral health and growth.

During this panel, Dennis and the other panel members: John Mattison (Moderator of the discussion and Chief Health Information Officer at Kaiser Permanente), Kevin Seals (Resident Physician and Diagnostic Radiology at  UCLA Health), Dr. Kathy Jenkins (Senior Associate in Cardiology and Professor of Pediatrics at Harvard Medical School) and Sina Bari (Solutions Architect at iMerit Technology Services) addressed the current challenges AI is facing in Medicine:

• The approach: We are currently using AI either for very broad solutions vs. individual applications.
• The execution: How are we implementing AI and who is responsible for it?
• The implementation: Either via medical devices or Picture Archiving and Communication Systems (PACS).
• The needs of enterprise: AI needs interoperability and the right infrastructure to be successfully implemented. Medical IT start-ups are therefore a huge support for a successful implementation of AI in a healthcare system.
• The regulatory restrictions: Requirements are very broad but are restricted by very narrow approvals.
• The acceptance: Administrators, providers and patients have to fully accept the fact that AI is a tool that is there to help them improve the healthcare system, otherwise implementations are doomed to fail.
• And finally, the legal aspect: As new laws are emerging surrounding AI, it is important to always be fully aware of them.

AI in medicine: around the world

Leo Anthony Celi has practiced medicine in three continents, giving him broad perspectives in healthcare delivery. As clinical research director and principal research scientist at the MIT, he brings together clinicians and data scientists to support research using data routinely collected in the intensive care unit (ICU). His group built and maintains the Medical Information Mart for Intensive Care (MIMIC) database. This public-access database has been meticulously de-identified and is freely shared online with the research community. The ultimate goal is to scale the database globally and build an international collaborative research community around health data analytics. Leo founded and co-directs Sana, whose objective is to leverage information technology to improve health outcomes in low- and middle-income countries. It is an open-source mobile tele-health platform that allows for capture, transmission and archiving of complex medical data (e.g. images, videos, physiologic signals such as ECG, EEG …), in addition to patient demographic and clinical information.

In the panel discussions they talked about the differences worldwide. We’ll present 2 cases specific to Canada and Europa.
Given that rural and northern Canadians generally have poorer access to health services, telemedicine tools collecting data and AI tools  analyzing these data can have a great impact. At Mimosa e.g. they are building a simple handheld device for imaging the foot in order to early recognize and treat deep open wounds of diabetic patients that can lead to amputation if not detected early.

Starting from May 25th 2018 European companies need to comply with The General Data Protection Regulations (GDPR), which ensures the protection of EU citizens’ personal data. Individuals will need to be given clear and specific information about what is being done with their personal information. Providers of connected devices and apps will need to fully disclose what the service provider intends to do with collected data. Broad statements like “we may use your personal data for research purposes” buried within lengthy terms and conditions are no longer acceptable. GDPR has reinforced key complexities of data sharing in a shared care record, most notably the concept and mechanisms surrounding data sharing consent and opt-outs (e.g. the need for clear affirmative action -opt-in rather than opt-out). In this context, the panel members expressed their fear that sharing data will become more complex and might as such slow down data sharing platform initiatives using AI. However, it is possible to comply with GDPR and still share data as we at LynxCare are doing that in Europe. It is just a matter of taking time to understand what applies to your business and making sure you take care of it – in the end it is in the interest of the patient.

To conclude with 3 questions

These 3 speakers were only the tip of the iceberg as the Conference had so much to offer!

In general, the AIMed panel discussions all agreed that better cooperation and close engagement between researchers/clinicians/executives and data scientist is needed to be able to design and implement AI tools in a successful way. This AIMed conference definitely fostered these cooperation and engagement between all stakeholders!

We’d like to end with 3 questions that should be asked when designing AI projects:

• 1. What is going to be the operational impact of the AI tool?
• 2. How easy will it be to use the AI tool in the clinic and to implement/integrate in the current workflow?
• 3. Does the AI project align with the strategic goals of the department/hospital in general?

‍