AI for Wellness: The New Frontier in Healthcare – Part 1
Welcome to a brand new episode of The Zista Podcast where we are joined by Sumeet Maniar, an expert with over 25 years of experience across healthcare and tech.
Join us as we explore how AI is revolutionizing healthcare, focusing on its most significant applications, the ethical considerations of its integration, and the challenges healthcare providers face.
We also discuss the use of synthetic data and how it is reshaping pharmaceutical research.
Tune in for an insightful conversation that highlights the transformative impact of AI on healthcare, offering practical insights for industry professionals and enthusiasts.
Welcome to the latest episode of The Zista Podcast, where we explore the transformative power of AI in healthcare. In this episode, we sit down with Sumeet Maniar, an expert with over 25 years of experience in product development and leadership across multiple industries, including healthcare and gaming. Sumeet, currently the fractional CPO with Cipra.ai and formerly the CEO of Wellbrain, shares his extensive knowledge on the real-world applications of AI in healthcare.
Join us as we explore the most significant applications of AI in healthcare today. We discuss how companies are managing ethical considerations when integrating AI into wellness and healthcare, the biggest challenges healthcare providers face with AI technology, and upcoming game-changing developments in healthcare infrastructure. We also touch upon the use of synthetic data in pharma and healthcare.
He also explores the potential uses of synthetic data in the pharmaceutical and healthcare sectors, offering concrete examples and actionable insights.
Tune in for an insightful conversation on how AI is transforming the wellness and healthcare landscape.
KEY TAKEAWAYS
- Combining knowledge in AI, data science, and healthcare is essential.
- AI can assist in solving complex healthcare problems, but human critical thinking is essential for interpreting AI recommendations.
- Understand the ethical implications of AI in healthcare to maintain patient trust and compliance.
- Successful AI solutions must optimize workflows and reduce workloads for healthcare providers.
- AI is accelerating drug development and precision medicine, opening new career paths.
- Synthetic data can improve the efficiency and effectiveness of AI models in healthcare. Learn how to work with synthetic data and its applications in pharmaceutical and healthcare solutions.
Questions
Q1. What is the most significant application of AI in healthcare today?
A: Sumeet shares that the most significant application of AI in healthcare today is its generative aspects, particularly in nursing care. The vast amount of knowledge required to deliver care, especially in hospital settings, benefits greatly from AI assistance. He believes AI’s impact on delivering better healthcare, minimizing mistakes, and making better recommendations using different types of data will be immense.
Historically, AI applications involved working with large datasets and supervised learning, where models are trained with properly labeled data. Sumeet has experience in this area, where AI algorithms used supervised learning to provide accurate insights. However, generative AI takes it to a whole different level by handling both structured and unstructured data to deliver insights and recommendations.
Sumeet provides an example with Cipra.ai, a company he’s working with. One component of their solution uses semi-supervised learning based on population health data to make precise, personalized recommendations. This approach is particularly powerful for reducing hypertension. The AI incorporates generative elements to fine-tune recommendations for diet and other health factors, making them more impactful.
The AI solution offers numerous advantages, such as reducing staff time needed for tasks like patient outreach, addressing staffing shortages, and enhancing efficiency. This leads to better results and improved health outcomes without the complexity of relying heavily on human staff. Sumeet believes the opportunities in healthcare, from patient care to back-end systems management, will be immense with AI integration.
Q2. How are companies managing the ethical considerations when integrating AI into wellness and healthcare?
A: Sumeet believes the focus or emphasis is on privacy first. All digital health entities or any health entity must follow various U.S. standards for health privacy laws, and AI must be used in a way that ensures data is doubly encrypted. For example, at Sumeet’s company, they ensure that everything is processed on their own premises. Data is not pushed out externally, and if it is ever used externally, it is de-identified. This means that no one can attribute the data to a specific person, ensuring privacy.
In some cases, to conduct certain types of studies, they might need to merge their data with external data. When doing so, they work with third-party companies to ensure that all steps adhere to privacy regulations. Personal Identification Information (PII) is protected, and they comply with the Health Insurance Portability and Accountability Act (HIPAA) in the United States, which mandates how health records should be stored and processed.
Sumeet emphasizes that liability is much larger in the United States if these protocols are not followed. Therefore, the mindset is always to keep patient records and data as intact and secure as possible, ensuring patient privacy and ethical management of data throughout the process.
Q3. What are some of the biggest challenges that healthcare providers face when deploying AI technology in the solutions they offer?
A: Sumeet identifies several key challenges healthcare providers face when deploying AI technology:
- Noise and Confusion: As with all new technologies, there is a lot of noise and many companies claiming they can implement AI solutions. For example, physicians might talk to several companies, each claiming to offer AI solutions, which creates confusion. It takes time to help them understand what true AI is and what it can do, slowing down the adoption of new technology.
- Behavioral Change and Trust: Implementing AI requires a behavioral change, meaning providers need to trust the AI solution. Currently, many AI concepts sound promising but are often just simple concepts that need validation. In generative AI, there’s a concept called Retrieval-Augmented Generation (RAG), which uses APIs with external data sources to validate solutions. Despite this, having a human in the loop to validate the data remains crucial, which can be a challenge.
- Human-Centered Design: Another challenge is ensuring that AI solutions consider the entire workflow of practitioners and their staff. Many digital solutions fail because they add features that increase the workload. For example, requiring staff to use an iPad or open specific apps can be burdensome. Successful solutions need to optimize workflows, even down to small details like eliminating the need for a shift key to type capital letters, to speed up processes.
Sumeet emphasizes that it’s important to think through these solutions from a human-centered design perspective to get adoption. Overcoming these challenges involves understanding the noise, helping providers understand it, taking the time to trust the solutions, and ensuring they reduce workloads and integrate seamlessly into existing workflows.
Q4. In terms of future developments in AI, are there any upcoming game changers in healthcare infrastructure or in how deeply AI is integrated into the entire service delivery process?
A: Sumeet says that in the service delivery process of healthcare, he anticipates significant changes with many more AI solutions emerging. One notable development is the increased use of voice-based processing. This technology allows users to make requests via speech, with the AI agent responding to resolve issues such as insurance claims. This will simplify processes for users and may also involve machine processing on the backend, leading to more efficient resolutions.
Sumeet also mentions initiatives in end-to-end healthcare services, such as precision medicine and prototyping how molecules interact. He highlights Sorcero, a company focused on medical affairs management. Pharmaceutical companies often have vast amounts of structured and unstructured data, making it challenging for analysts to retrieve necessary information quickly. Sorcero’s AI compresses this time, speeding up the delivery of drugs to market.
Another example Sumeet provides is Hippocratic AI, whose founder has developed multiple specialized AI agents in different verticals. These agents can provide live consultations, such as a nurse consulting an AI agent to get proper care recommendations while interacting with a patient. This represents a significant paradigm shift, making healthcare delivery faster, better, and more efficient.
Sumeet believes these advancements should be seen as complementary or augmentative to existing solutions, particularly in scenarios with staff shortages. The future of AI in healthcare promises to be even faster, more efficient, and capable of delivering better outcomes.
Q5. Do you see use cases of synthetic data in the pharma or healthcare space?
A: Sumeet says that he does see synthetic data being utilized in the pharma and healthcare space. While he hasn’t worked extensively with synthetic data due to having access to sufficient patient and outcome data in his historical cases, he acknowledges its potential. Synthetic data can make models more efficient and effective. For example, having a baseline of all pharmaceutical compounds and inferring synthetic data from them could significantly enhance the process of developing solutions.
Sumeet believes that synthetic data can lead to better solutions in specific use cases involving molecules and pharmaceuticals. He also sees other AI applications that could work in tandem with synthetic data. For instance, audio-based AI solutions for nurses could isolate and interpret sounds from crowded environments. AI technologies already exist that can analyze facial expressions and breathing patterns to determine health parameters.
Therefore, it’s not far-fetched to think that capturing and modulating voice data in crowded settings could lead to responsive alert systems. Sumeet sees a lot of scope for such AI solutions to be effective in various environments, making it highly possible for synthetic data to play a significant role in future healthcare and pharmaceutical applications.
