Exploring the Evolution and Future of AI in Healthcare

AI in Healthcare Exploring Its Evolution and Future

Artificial Intelligence holds the potential to transform the healthcare industry. In this introductory article of a series, we examine the dynamic landscape, offering an up-to-date analysis of the latest breakthroughs and their ramifications. We envision AI as a catalyst for change in healthcare; this transformation will be progressive and far-reaching. Ultimately, it could streamline not only diagnostics but also a substantial portion of administrative tasks, profoundly altering the doctor-patient dynamic.

This inaugural article, the first in a series exploring AI in healthcare, offers a contemporary snapshot of the present landscape, examining current developments and contemplating the future possibilities.

AI in Healthcare

Thus far, AI’s influence on healthcare has been comparatively mild, showcasing intriguing advancements in areas such as:

Primarily, these products serve as point solutions, catering to the requirements of existing stakeholders within their established workflows.

In the coming years, we anticipate AI to:

  1. Automate a significant portion of healthcare administration, freeing up time and resources, and ultimately enhancing clinical outcomes
  2. Integrate various components of the healthcare system, such as payers and providers, more cohesively
  3. Additionally, AI is predicted to have a profound impact on the BioPharma industry, as transformer technology that understands natural languages, like ChatGPT, can also analyze large molecules, such as DNA and proteins.

Over the long term, we expected AI to bring about fundamental changes in healthcare, including:

  1. Increased automation of diagnoses
  2. A shift in the role of family doctors towards health coaching
  3. Enhanced productivity of hospital specialists, leading to more personalized and ambitious treatments

Consequently, AI is poised to usher in a significant transformation in healthcare delivery. In the previous year, the top two academic papers frequently cited in the realm of AI concentrated on the application of AI to protein structures, rather than on generative AI, as one might assume given the surrounding hype.

The Transformative Impact of Transformers on AI and Healthcare

The Rise of Transformers

Transformers, introduced in 2017, are considered the second great breakthrough in AI, after Neural Networks. They’re poised to have a much more significant impact on healthcare than image recognition. Large Language Models analyze sequential information like words in a text or amino acids in a protein by considering the relationships between elements and their importance.

Transformers: The Key to Self-Supervised Learning

Unlike traditional models, transformers don’t need large labeled datasets for training. They excel at adjusting their predictions based on context. As Nvidia’s CEO Jensen Huang mentioned recently, transformers make self-supervised learning possible, taking AI to “warp speed.”

Enhancing Human-Computer Interaction

Transformers have revolutionized natural language understanding. Transformers grasp the context and relationships between words, enabling Large Language Models (LLMs) to understand languages like English, Chinese, and Spanish more effectively. AI can now extract key points from unstructured text and write in a convincingly human-like manner, as demonstrated by ChatGPT and GPT-4.

Advancements in Speech Transcription

AI has significantly improved its ability to transcribe human speech, as evidenced by Nuance, a company acquired by Microsoft for $20 billion. Nuance recently launched a system that uses GPT-4 to instantly draft clinical notes based on doctor-patient conversations.

Challenges: Truthfulness in LLMs

Despite their impressive capabilities, LLMs like ChatGPT and GPT-4 have limitations in medicine due to their lack of understanding of the concept of truth. They may produce plausible yet misleading statements, which can be dangerous in healthcare. We believe that AI will eventually overcome this issue, but if not, the use of LLMs in healthcare must be cautiously monitored.

Unlocking the Language of Biology

Originally designed for natural language analysis, transformers have proven effective in analyzing biological structures like DNA and RNA. AlphaFold can predict 3D protein shapes, and Nvidia’s AI-as-a-service aims to help researchers predict biomolecule functions. These breakthroughs hold profound implications for the BioPharma industry.

The Future of AI in Healthcare: Consumer Push, Industry Adoption, and Medical Devices

Consumer Push: The Demand for Health-Focused Products

Disruption occurs when there’s an unmet need and a technology to address it. The growth of “quantified self” technology, such as wearables, demonstrates the increasing consumer demand for health-focused products. We anticipate that these technologies will deepen (e.g., smartwatches advising on blood sugar levels) and reach a larger percentage of the population. AI is expected to generate health insights from this data, with consumers wanting their doctors to assess these AI-driven suggestions.

Industry Adoption: The Expansion of AI in Healthcare

Currently, AI is deployed in pockets within the healthcare industry, but we expect this to expand as doctors recognize AI’s clinical benefits and its potential to reduce administrative burdens. As cost pressures increase, more clinicians and managers will likely adopt AI for various purposes, driving its growth in the industry.

The Convergence of Health, Tech, and Consumer Industries

We believe that the future of healthcare will be shaped by the convergence of health, tech, and consumer industries. AI is a major component of the healthcare data revolution,  which involves new ways of collecting, integrating, and analyzing data. This revolution is particularly crucial in healthcare due to the granular nature of data, its integration from multiple sources, and real-time communication with individuals, clinicians, and public health authorities. Over time, we think healthcare will become more preventative, convenient, and personalized, with AI driving much of this shift.

AI’s Growing Impact on Medical Devices

AI is already making noticeable strides in medical devices, particularly in radiology and medical imaging. Highlighting the growth of AI-enabled products in the field: the number of FDA-approved AI-enabled radiology devices has consistently grown, with about 40% of all Class II radiology devices and 5% of all Class II devices now AI-powered. We expect these trends to continue upward.

The Future of AI and Doctors: Changing Roles and Increased Efficiency

The Emergence of AI-Driven Diagnosis Tools

In China, over a third of doctors use an AI-driven tool called AskBob to help diagnose and treat patients. While there is no equivalent system in the United States yet, the success of AskBob implies a significant opportunity to improve primary care in areas with weak existing provisions. As AI becomes more user-friendly and accurate, its role in healthcare will increase globally.

AI’s Expanding Medical Knowledge and Access to Data

AI systems, like ChatGPT, are already approaching the ability to pass the USMLE exam, which is required for becoming a medical doctor in the U.S. With specific medical training, generative AI systems will likely surpass this benchmark soon. AI has the advantage of being able to analyze the entire corpus of medical literature, stay up-to-date on medical advances, track potential drug interactions, and digest more individual data than human doctors.

The Changing Role of Doctors

We expect the role of doctors to change as AI becomes more integrated into healthcare. More specifically, we see the following trends as inevitable:

  1. Family doctors: In the future, family doctors will likely focus more on being health coaches and addressing patients’ health holistically. With AI handling diagnosis and prescriptions, doctors can concentrate on preventative, proactive medicine.  Indeed, progressive systems such as Kaiser Permanente already explicitly espouse this preventative model of medicine.
  2. Hospital specialists: AI will improve clinical decision support devices, leading to faster and more accurate systems. As a result, specialists will spend less time on diagnosis and utilize more personalized, ambitious treatments, ultimately increasing their productivity.

The Path Forward: Medical and Consumer Systems

We expect two major classes of diagnosis products to evolve:

  1. Medical systems that work alongside doctors, assisting in clinical terms and administration.
  2. Consumer systems that work at home or on the go, which may or may not be FDA-approved. These systems will be linked to wearables and remote monitoring devices, providing lifestyle suggestions or acting as accessible AI doctors from smartphones.

In the long term, we see a convergence between medical and consumer systems , with continuous monitoring, analysis, and nudges replacing annual check-ups.

Conclusion: Embracing the AI Revolution in Healthcare

In this article, we have emphasized that AI will gradually transform healthcare in numerous important ways. Radiology has already undergone a significant change, and healthcare administration could be next. The drug discovery process is also due for disruption, and the role of doctors will eventually change. The integration of AI into healthcare promises a transformative shift in the way we approach medical care. From the development of AI-driven diagnostic tools to the changing roles of family doctors and hospital specialists, the industry is on the brink of a revolution that will lead to more efficient, personalized, and proactive care for patients.

As medical and consumer systems continue to evolve and converge, we can expect a future where healthcare is seamlessly integrated into our daily lives, with continuous monitoring and real-time insights to help prevent illnesses and improve overall well-being.  These AI monitoring systems will greatly extend and augment physicians in their ability to make timely interventions, proactively reach out to patients, and drive better health outcomes.

To fully harness the potential of AI in healthcare, it is crucial for stakeholders, including healthcare providers, researchers, and policymakers, to collaborate and address the challenges that come with the adoption of these technologies. By embracing the AI revolution and ensuring its responsible development and deployment, we can usher in a new era of healthcare that benefits society at large in unprecedented ways.