The Promise of AI-Augmented Healthcare
Healthcare systems around the world face significant challenges in achieving the ‘quadruple aim’ for healthcare: improve population health, enhance the patient’s experience of care, boost caregiver experience, and reduce the rising cost of care. Aging populations, the growing burden of chronic diseases, and soaring healthcare costs are pushing governments, payers, regulators, and providers to innovate and transform models of healthcare delivery.
The COVID-19 pandemic has further highlighted the shortages in the healthcare workforce and inequities in access to care. Against this backdrop, the application of technology and artificial intelligence (AI) in healthcare has the potential to address some of these supply-and-demand challenges.
AI is a powerful and disruptive technology with the capability to fundamentally transform the practice of medicine and the delivery of healthcare. The increasing availability of multimodal data (genomics, economic, demographic, clinical, and phenotypic) coupled with innovations in mobile, internet of things (IoT), computing power, and data security herald a moment of convergence between healthcare and technology.
This convergence is enabling the transition of effective and safe AI systems into mainstream healthcare delivery, powered by the scalability and computing capacity of cloud computing. Many technology providers are increasingly seeking to partner with healthcare organizations to drive AI-driven medical innovation.
Understanding AI in Healthcare
Simply put, AI refers to the science and engineering of making intelligent machines that can mimic human cognitive functions, such as learning and problem-solving. AI systems have the potential to anticipate problems, deal with issues as they come up, and operate in an intentional, intelligent, and adaptive manner.
AI’s strength lies in its ability to learn and recognize patterns and relationships from large, multidimensional, and multimodal datasets. For example, an AI system could translate a patient’s entire medical record into a single number that represents a likely diagnosis.
AI is not a single, ubiquitous technology, but rather a collection of subfields, including machine learning (ML) and deep learning (DL). ML refers to the study of algorithms that allow computer programs to automatically improve through experience, while DL is a class of ML algorithms that learn by using a large, many-layered collection of connected processes and exposing these processors to a vast set of examples.
Despite the significant focus on AI in healthcare over the past decade, the use and adoption of AI in clinical practice remains limited. Many AI products for healthcare are still in the design and development stage, as there have been attempts to force AI solutions into healthcare problems without due consideration for local context, such as clinical workflows, user needs, trust, safety, and ethical implications.
A Human-Centered Approach to Effective AI-Augmented Healthcare
We believe that AI should amplify and augment, rather than replace, human intelligence in healthcare. When building AI systems for healthcare, it is crucial to not replace the important elements of human interaction in medicine, but to focus it and improve the efficiency and effectiveness of that interaction.
To build effective and reliable AI-augmented healthcare systems, we recommend a problem-driven, human-centered approach:
-
Design and Develop AI Solutions for the Right Problems: Start by understanding the key problems, needs, constraints, and workflows in healthcare organizations through user-designed research. Identify which problems are appropriate for AI to solve and ensure the availability of applicable datasets to build and evaluate the AI.
-
Contextualize Algorithms in Existing Workflows: By contextualizing algorithms in an existing workflow, AI systems can operate within existing norms and practices, providing appropriate solutions to existing problems for the end-user.
-
Pilot New Experiments with Tight Feedback Loops: Focus on piloting new, stepwise experiments to build AI tools, using tight feedback loops from stakeholders to facilitate rapid experiential learning and incremental changes.
-
Iteratively Evaluate and Validate AI Predictions: Evaluate the AI tool’s performance based on three dimensions: statistical validity, clinical utility, and economic utility. High model performance on retrospective, in-silico settings is not sufficient – the algorithm must demonstrate clinical effectiveness and generalizability in real-time environments.
-
Scale Up with Deployment Modalities and Regulatory Considerations: Even after an AI system has been deployed clinically, it must be continuously monitored and maintained to monitor for risks and adverse events. Healthcare organizations, regulatory bodies, and AI developers should cooperate to collate and analyze relevant datasets for AI performance, clinical and safety-related risks, and adverse events.
By following this human-centered approach, healthcare organizations can build AI systems that are effective, reliable, and truly transformative for patient care.
Realizing the Potential of AI in Healthcare
AI has the potential to enable healthcare systems to achieve their ‘quadruple aim’ by democratizing and standardizing a future of connected and AI-augmented care, precision diagnostics, precision therapeutics, and ultimately, precision medicine.
We can expect to see a non-exhaustive suite of AI applications in healthcare in the near, medium, and longer term:
Near-Term AI Applications
Currently, AI systems are beginning to be adopted by healthcare organizations to automate time-consuming, high-volume, repetitive tasks. There is also considerable progress in demonstrating the use of AI in precision diagnostics, such as diabetic retinopathy screening and radiotherapy planning.
For example, AI algorithms for diabetic retinopathy screening have demonstrated robust diagnostic performance and cost-effectiveness, leading to their approval for Medicare reimbursement in the United States. Similarly, AI-based tools can significantly reduce the time required for image segmentation and planning tasks in radiotherapy, dramatically reducing waiting times for potentially life-saving cancer treatment.
Medium-Term AI Applications
In the medium term, we expect to see significant progress in the development of powerful AI algorithms that are more efficient, able to use unlabeled data, and can combine disparate structured and unstructured data (e.g., imaging, electronic health records, multiomics, behavioral, and pharmacological data).
Healthcare organizations and medical practices will also evolve from being adopters of AI platforms to becoming co-innovators with technology partners in the development of novel AI systems for precision therapeutics.
Longer-Term AI Applications
In the long term, AI systems will become more intelligent, enabling AI healthcare systems to achieve a state of precision medicine through AI-augmented healthcare and connected care. Healthcare will shift from a one-size-fits-all approach to a preventative, personalized, data-driven disease management model that improves patient outcomes and cost-effectiveness.
AI could significantly reduce inefficiency in healthcare, improve patient flow and experience, and enhance caregiver experience and patient safety through the care pathway. This could include the application of AI to remote patient monitoring, intelligent telehealth, and ambient intelligence, where patients, caregivers, and healthcare providers are all connected to a single, interoperable digital infrastructure.
Transforming Healthcare with AI-Powered Automation and Diagnostics
AI is already making significant strides in healthcare, particularly in the automation of administrative tasks and the enhancement of diagnostic capabilities.
Natural language processing (NLP) technology has the potential to automate tasks such as documenting patient visits in electronic health records, optimizing clinical workflows, and enabling clinicians to focus more time on patient care. Solutions like Nuance’s Dragon Ambient eXperience are already demonstrating the power of AI-powered clinical documentation.
The automated classification of medical images is another leading AI application in healthcare today. Studies have shown that AI can meet or exceed the performance of human experts in image-based diagnoses across specialties like radiology, dermatology, pathology, and cardiology.
For example, AI algorithms have demonstrated robust diagnostic performance and cost-effectiveness in screening for diabetic retinopathy, a leading cause of preventable vision loss worldwide. The Centers for Medicare & Medicaid Services in the US has already approved reimbursement for the use of an FDA-approved AI algorithm for this application.
Another impactful AI application is in assisting clinicians with image preparation and planning tasks for radiotherapy cancer treatment. AI-based tools can cut the time-consuming, manual segmentation of images by up to 90%, dramatically reducing waiting times for potentially life-saving radiotherapy.
Powering Precision Therapeutics and Drug Discovery
To make progress towards precision therapeutics, we need to considerably improve our understanding of disease. Researchers are exploring the cellular and molecular basis of disease, collecting multimodal datasets that can lead to digital and biological biomarkers for diagnosis, severity, and progression.
Two important future AI applications include immunomics/synthetic biology and drug discovery.
Through the application of AI tools on multimodal datasets, we may be able to better understand the cellular basis of disease and the clustering of diseases and patient populations to provide more targeted preventive strategies. This could have a revolutionary impact in the cancer, neurological, and rare disease spaces, personalizing the experience of care for the individual.
AI will also drive significant improvements in clinical trial design, the optimization of drug manufacturing processes, and any combinatorial optimization process in healthcare. We’ve already seen breakthroughs like DeepMind’s AlphaFold, which sets the stage for better understanding disease processes, predicting protein structures, and developing more targeted therapeutics.
In the longer term, healthcare professionals will leverage AI in augmenting the care they provide, allowing them to deliver safer, standardized, and more effective care. Clinicians could use an ‘AI digital consult’ to examine ‘digital twin’ models of their patients, testing the effectiveness, safety, and experience of an intervention in the digital environment before delivering it to the patient in the real world.
Overcoming Challenges to Widespread Adoption
While the advances in AI have tremendous potential to transform healthcare, there are significant challenges related to the wider adoption and deployment of AI into healthcare systems. These include:
- Data quality and access: Healthcare data is highly sensitive, inconsistent, siloed, and not optimized for machine learning development, evaluation, implementation, and adoption.
- Technical infrastructure: Sufficient computing power is required to generate real-time decisions, which is being transformed by the advent of cloud computing.
- Organizational capacity: Healthcare organizations need to develop the processes, skills, and culture to effectively integrate AI into clinical workflows and decision-making.
- Ethical and responsible practices: There are critical issues around data privacy, security, equity, and safety that must be addressed.
Healthcare leaders must consider these challenges and invest in the necessary infrastructure, talent, and governance to leverage AI for the benefit of patients and the healthcare system as a whole.
Conclusion: Realizing the Promise of AI-Augmented Healthcare
Advances in AI have the potential to transform many aspects of healthcare, enabling a future that is more personalized, precise, predictive, and portable. It is unclear whether we will see an incremental or radical adoption of these technological innovations, but the impact they can have requires health systems to adapt to the changing landscape.
For the NHS and healthcare systems globally, the application of AI technologies has the potential to release time for care back to healthcare professionals, enabling them to focus on what matters most to their patients. In the future, healthcare providers could leverage a globally democratized set of data assets and the “highest levels of human knowledge” to deliver a common high standard of care, wherever and whenever it is needed.
Ultimately, AI could become a key tool for improving health equity around the world. As much as the last 10 years have been about digitizing health records for efficiency, the next 10 years will be about the insights and value that can be gained from these digital assets, and how they can be translated into driving better clinical outcomes with the assistance of AI.
We are at a pivotal moment in the convergence of medicine and technology. While there are significant challenges to overcome, the potential of AI-augmented healthcare is undeniable. By taking a human-centered approach and addressing the critical issues of data, infrastructure, skills, and governance, healthcare systems can harness the power of AI to deliver transformative improvements in patient care.
