Toward an AI-augmented future in health care

AI in health care is leading to efficient care and lower costs.


A Markets and Markets research study states that artificial intelligence (AI) in the health care sector is estimated to be valued at $45.2 billion by 2026, with its expansion at a compound annual growth rate (CAGR) of 44.9% from 2020 to 2026. 

The health care industry has been witnessing the application of AI in increasingly complex tasks, leading to faster turnarounds at much lower costs. 

Hence, it is no surprise that the global medical community has accelerated the adoption of AI-based technologies in the battle against the coronavirus. 

On one hand, AI is being deployed to review lung scans for signs of COVID-19-related pneumonia, on the other it is being used to screen billions of molecules to counter the infection — all toward delivering faster and more accurate diagnosis and treatment. 

Therefore, considering the ongoing pandemic, it has become amply clear that intelligent technologies will be more and more crucial to help people live healthier lives and help make the health system work better for everyone. 

From improving efficiencies in day-to-day operations to population health management and even getting new drugs to market faster, the implementation of AI-based technologies can augment human capabilities to enhance and improve the care delivery ecosystem. 

Let’s take a closer look at the application areas ahead.

AI in the care delivery ecosystem

AI and other advanced technologies are transforming the modern health care ecosystem in many profound ways. 

With WHO projecting a shortage of 12.9 million health care professionals globally by 2035, the implementation of intelligent technologies can help in bridging this gap between demand and availability, leading to a paradigm shift in the treatment and care of customers/patients.

The application of AI in care delivery impacts how tasks are performed and how they are transformed by improving the processes or changing care requirements. 

Machine learning (ML) — a subset of AI — can help care organizations to lay the foundation for innovative AI tools through the creation of a robust and scalable ML lifecycle. It defines each step that an organization should follow to derive practical business value from their AI investments. 

This includes developing a data pipeline, creating a relevant data model to understand its relevance, interpreting it, and, finally, implementing the interpretation to derive maximum value.

Health care analytics can then comb through vast volumes of data leading to better-informed decisions and improved outcomes for both clinicians and patients. 

A subset, descriptive analytics uses data to find answers to “What has happened?”; predictive analytics uses statistical models and forecasting for “What could happen?”; and prescriptive analytics can advise on possible outcomes and find answers to “What should we do?” 

AI can create significant impact across each tenant of the care delivery value chain:

Providers: AI is increasingly being leveraged by health care providers in diagnostics, treatments, patient monitoring, patient risk prediction, readmission reduction, as well as streamlining administrative workflows, including admissions, patient monitoring, discharges, post-discharge care. 

Case in point, AI can reduce 30-day hospital readmissions by an estimated 12%, leading to lower annual health care costs and improved patient care. AI capabilities empower providers to design their treatments to reduce the cost and care burden of undertaking unnecessary treatments and improve patient outcomes.

Member (patient): Today, patients are looking at innovative, patient-centric care delivery services that are cost-effective and easily accessible. 

Remote monitoring allows for patients to be under care in the comfort of their home while being continuously monitored across several parameters though wearables and (in the near future) embedded biosensors. 

Take, for instance, AI-enabled pregnancy management that allows for monitoring an expecting mother and her fetus for term-wise fetal growth and identifying fluctuations in her existing and potential health issues, such as hypertension, high blood sugar and urinary infections. 

Additionally, AI-based solutions aid in predicting the most effective methods for utilizing available resources at vital care sites, such as emergency departments and urgent care centers. 

This ensures optimal staffing levels to account for fluctuations in patient flow, thereby translating into lesser waiting time for patients and resulting in an enhanced patient experience.

Payers (insurance companies): Health care insurers are looking at AI and other technological solutions for the accurate identification of at-risk individuals and to better manage costs. 

By streamlining processes, such as the review of medical records prior authorization, pre-payment review, and post-payment auditing, AI can aid in framing customized health insurance plans with tailored options for various customer segments. 

It can also facilitate faster detection of frauds by providing insights, such as unusual claims patterns against existing data.

Through automated discovery and analysis of patterns, AI can improve claims denial management by differentiating between claims that have a low chance of acceptance and guide insurance agents to close those claims that have a higher chance of being approved.

Clinical: AI solutions can be applied across a vast number of areas, including drug development and application, drug formularies, development of new and advanced medical devices for augmenting health care providers, as well as chronic diseases management.

For example, AI-based algorithms can reduce the time required to scan potential test subjects based on their medical history to select the most suitable candidates for the human trials phase of drug discovery and development.

But will AI replace humans in health care?

While the role of AI in improving efficiencies across the health care delivery value chain is widely acknowledged, it cannot replace the comfort of human touch, or even decision-making in entirety. That said, it is necessary to address the concerns related to the adoption and integration of AI within the health care ecosystem. 

Ethics, data privacy and security are key areas that lead to questions related to who has the final onus of applying AI in diagnostics or treatment decision-making processes. This also includes considerations such as, is the technology in line with public values, does it address the patient’s emotional needs in care situations, etc. 

Another concern relates to the accuracy of machine-based outcomes. Since the outcomes are based on the quality of raw data, and organizational silos often make data sharing difficult, this creates a bigger challenge of solely depending on the automated outcomes. 

Lastly, there are questions related to the need to reskill and upskill the health care stakeholders — such as clinicians, and ancillary health care staff — to harness the best of AI. 

As intelligent technologies transform procedures and workflows in care delivery organizations, providing training opportunities to the workforce will be key for ensuring operational efficiency and retaining a competitive edge. 

Even considering the ML lifecycle discussed earlier, there will be need for data scientists, statisticians, analysts and data engineers to design AI-based solutions; implement processes; run operations; and monitor efficiencies at its various stages.

The road ahead...

AI augments human capacities to create process efficiencies, effective care delivery mechanisms and, finally, a high-performing health care workforce.

The successful application of AI across the care delivery chain needs a cohesive, multi-stakeholder approach that will bring together the efforts of clinicians, IT operations, finance, health insurers, etc. This is important in order to fully integrate the potential of AI into managing the health care system.

The need of the hour is to come out of the hype that surrounds technology; introspect to define your needs; understand the fundamentals of data science; focus on business value creation; and enable people to live healthier lives.