Ai and Robotics in Healthcare
2 likes590 views
The document discusses the integration of artificial intelligence (AI) in healthcare through various use cases, focusing on how AI can empower personas like virtual nursing assistants and health advisors in India. It highlights the challenges faced by the Indian healthcare system, such as a shortage of doctors and the need for better data governance and digitization of health records. The text also points to the importance of establishing clear policies on health records and adopting standardized information systems to leverage AI effectively.
Ai and Robotics in Healthcare
- 1. AI in Healthcare SUBHENDU DEY (subhendu.dey@in.ibm.com) Sr. Architect, Global Cognitive Center of Competence, IBM GBS
- 2. Content Introduction to some personas Top value-added Healthcare use cases AI can bring in Solution Architecture to support such use cases Indian context Next step for India from the solutions perspective Q&A 2
- 3. Meet some of our very known personas 3 Monalisa, a student 2nd grade is suffering from child depression. Her mother uses a healthcare buddy on her mobile which not only measures the quality time spent with her daughter but also notifies her on social events where she can take her daughter to. The virtual avatar creates an augmented reality based environment that has high empathy quotient. Laltu, a 10 years old boy hails from a village near Kaliaganj, Uttar Dinajpur, is suffering from high fever for last 3-4 days. There are various kinds of unknown fevers nowadays in the locality, causing huge queue in nearest health center which is 3 km away. Safiul, Laltu’s father calls a virtual healthcare assistant who speaks to him in Bengali and advises on therapeutic treatment options based on symptoms. Dr. Sharma, one of the leading radiologist of the town travels through ~10 diagnostic centers every weekday to read through brain scan images and write reports. 10 years back he had unbearable workload causing him not to do full justice to his work always. Now he uses a virtual assistant which can “read” through the images and can generate report for standard cases. Dr. Ganguly, one of the leading oncologist nowadays uses a of medical insight advisor which reads through the medical journals published world wide related to melanoma to distill the information for his effective consumption. This has been helping him a lot to keep up-to-date on latest R&D, effectiveness of recent medications on other and accordingly advise treatment.
- 4. 4 Yes, you guessed it right. These personas exist today but their activities are futuristic in Indian context! However that is not fully true in other parts of the world though. AI is making the personas empowered.
- 5. Top value-added Healthcare use cases AI can bring in* 5Engagement Virtual Nursing Assistant • Avatar based augmented reality to respond to medical symptoms • Conversation service • Speech to text and text to speech • Information retrieval • Tone analyzer • Integration through IoT • Social and other public data integration • Cognitive API integration Healthcare bots • Standard information on medicines • Appointment and other operations management Personal health advisor • Medication alerts and adherence check • Empathy score and subsequent advise • Encourage Healthy habits DataExplorationandInsight Mining of EMR as well as other medical records • Evidence based treatment options • Targeted treatment (e.g. focus on actionable genes in case of tumor analysis) • Precision medicine • Pharma insights generations • Medical language understanding • Hypothesis generation and inference derivation • Ranked information retrieval • Explorative visualizations • Advanced analytics models • Deep learning (e.g. CNN) Improved clinical trials • Easily find potential list of clinical trials for patient Drug discovery • Discovery of hidden connections • Evidence based predictions Radiology Image analysis • STP for standard cases, human intervention to critical cases • Detection of anomalies and alert * There are other robotics use cases which are not included here.
- 6. Cognitive elements in Healthcare solutions demands Data, Analytics + AI focused architecture Extract Understand Relate Reason Learn React 6 Application AI + Analytics Data Cloud • DATA-DRIVEN ANALYTICS Medical Image analytics, patient cohorts, on real world evidence Discovery and exploration, statistical modeling, machine learning, reporting and visualization • KNOWLEDGE-DRIVEN ANALYTICS Annotator on clinical data, medical insight, patient insight NLU and classifier, Cognitive pipeline, Knowledge base and query, concept detection • Infrastructure • Dependability • Security / Compliance • DevOps • Data pipeline – ingest/transform/persist in data lake • Core Health Services – FHIR, Master patient index, medication DB etc. • Data Governance – consent, de-identification, audit etc. • Healthcare solution for end-users • Developer experience including the based services / boiler-plates • Client and application services
- 7. Indian context – there is a clear demand* 69.2 million Indians suffer from Diabetes as per data from World Health Organization. Prevalence of diabetic retinopathy (DR) in patients with diabetes was recently estimated to be 34.6%. For every 1681 citizens in India, there is only one doctor available as per medical council of India report published in May 2016. 68.84% of population reside in rural India as per recent census. Last mile delivery of healthcare amenities is still a major challenge in India, especially to places that lack the connectivity of routes. India is known to suffer from a dual disease burden. While the country has witnessed and cured chronic diseases, it still hasn’t overcome challenges posed by infectious diseases and malnutrition. Decentralization of diagnostic testing needs to be inculcated through AI based diagnostic technology. 7 * Sourced from ET Healthworld, December 26 2016
- 8. Next steps for India – from solutions perspective A large part of healthcare specific NLP services are English based. Any AI based solution depends on large amount of data. India at a country level needs to come up with clear policy on citizens health record. Even a large amount of urban health record is not effectively digitized yet. A common information standard like FHIR should be adhered to. As evident from the architecture, cloud based solution is almost inevitable. A clear governance policy needs to be established on data security and privacy. 8
- 9. Q&A 9
- 10. References Frost & Sullivan analysis: From $600 M to $6 Billion, Artificial Intelligence Systems Poised for Dramatic Market Expansion in Healthcare. Accenture report: Artificial Intelligence: Healthcare’s new Nervous System IBM Watson health offering Sense.ly – the virtual nurse 10