How AI Adds Value To Healthcare Systems

Artificial intelligence (AI) is being increasingly incorporated into various healthcare domains. With this trend in mind, examining whether the technology adds tangible value to healthcare systems is critical. This article will examine three main ways through which AI most meaningfully results in positive healthcare outcomes.

How AI Affects Patients, Clinicians, Hospitals, and Pharma

AI's impact is not limited to niche uses. AI directly affects all involved parties in a healthcare setting:

• AI Impact on Patients: Access to improved and more customized care by combing through medical data to predict or diagnose disease faster 

• AI Impact on Physicians: Higher accuracy of image study interpretation, faster diagnoses, earlier detection of high-risk patients for select diseases, prioritization of emergency cases
  
  
• AI Impact on Hospitals: Fewer misdiagnoses, lower costs, reduced wait times, higher patient satisfaction
  
  
• AI Impact on Pharmaceutical Companies: Reduced time and cost for drug discovery, advanced study of drug interactions, faster identification of viable drug targets

Improved access to data is a central part of the above benefits, which include:  

• Data that is more accurate to improve decision-making (and better patient outcomes by extension)
• Data that can be rapidly accessed 

Leveraging AI and deep learning in medical imaging allows radiologists and physicians to leverage other systems and workflows. This is because AI thrives on data input and can often "see" what the human eye misses. 

One instance of this phenomenon is observed in a pre-print study presented at the "Building Bridges" conference hosted by the European Congress of Radiology earlier this year:

This study compared two methods for 2D knee imaging in healthy volunteers: A deep learning AI algorithm and traditional MRI compressed SENSE. 

Here are the main takeaways:

"The AI-based reconstruction allows a 64% reduction in scan time compared to the fully unaccelerated sequence (CS-AI3). Except for artifacts, the subjective rating was significantly higher for CS-AI than conventional CS for at least one acceleration factor.  

Sequences reconstructed using AI were rated better than the time-equivalent conventional CS for almost all acceleration factors. Signal-to-noise and contrast-to-noise were significantly better for all CS-AI reconstructions."

Value Add #1: Automation Of Repetitive Tasks

Radiologists often engage in mundane, repetitive tasks as part of their daily workflow process. A recent study found that during a typical 8-hour shift at a PACS workstation, a radiology resident logged in 10,778 keyboard inputs and 1.37 miles of travel on their mouse. 

This is concerning when paired with the finding that radiologists who use more mouse clicks to evaluate radiographs end up having longer turnaround times. 

Using AI to automate such tedious tasks dramatically reduces burnout due to complex workflows, decreasing radiologists' workloads that have quadrupled over the past 15 years and leading to significant savings in time and cost. 

An example of this is Arterys' AI-assisted cardio MRI software, which has been evaluated in credible peer-reviewed journals to have the following time-saving benefits:

• Up to 25 minutes per exam are saved
• A 30% reduction in total scan time
• Accuracy of quantifying biventricular cardiac size and function is identical to those of expert clinical readers

Not only does AI automate low-value tasks, but it can also help eliminate them in situations where they are unnecessary.

A Journal of Digital Imaging study published in 2019 found that CAD-based on AI reduced false positives per image in mammogram scans by 69% compared to traditional CAD. Based on prior literature, this could reduce radiologist reading time by 17%.

Value Add #2: Democratization Of Healthcare

In the context of AI, it is important to define precisely what the term "democratization of healthcare" means. 

A recent "Health Trends Report" published by Stanford Medicine identifies this term by making a distinction between patient empowerment and data access:

"...health care democratization is characterized by two major factors: the distribution of data and the ability to generate and apply insights at scale. It promises a world in which patients—armed with data, technology, and access to expertise—can take charge of their own well-being and manage their own health. 

Democratization will mean that providers focus less attention on routine tasks and more on the areas where they provide the most value and find the most satisfaction. And individuals managing their health care concerns will put less strain on the health care system, lower costs, and improve public health overall."

This allows for several sources of data to be drawn upon for patient care: Genetics, family history, biochemistry lab results, symptoms, data from "smart" wearable devices, and much more. 

The democratization of healthcare through AI gives doctors a "predictive" way to understand how a patient has changed over time and the specific elements of their lifestyle and/or medications that have led to their current condition.

Other notable aspects of democratizing healthcare include: 

Pooling of Resources

Consolidating resources into larger systems across numerous locations that each operate according to different standards (financial, clinical, etc.), reduces treatment time and improves patient outcomes. This is especially important during critical care situations where a delay of a few minutes can mean the difference between life and death for a patient. 

Leveraging Cloud Technology

Often coupled with AI is cloud technology, allowing clinicians to provide high-quality patient-centered solutions regardless of location. Even in the most remote areas, standard of care will not have to be compromised.

AI Diversification

AI has expanded far beyond diagnostic tools, expanding to easier integration into existing workflows and automating the operational needs of providers. 

Value Add #3: Better Return On Investment  

The ROI observed for AI can be divided into two subtypes: Hard ROI and soft ROI. 

• Soft ROI -- Better patient experience through personalization, higher provider satisfaction, skill retention and acquisition, more agile responses to new challenges
  
  
• Hard ROI -- Productivity gains via faster execution of tasks and better decision-making, automation of repetitive manual/cognitive tasks, revenue growth through increase of customers and/or money spent per customer, cost savings through fewer employees needed to complete the same volume of work

The hard ROI financial gains fall into two categories of motivation:

The first source of motivation is the declining reimbursement and the resulting financial pressures that ensue. Recently, Medicare reduced reimbursements for the radiology industry by 11%. This necessitates the efficient use of a cost-saving solution such as AI technology that allows for universal image sharing.

The second source of motivation comes from the growing trend of coverage for AI-based medical services, as observed in a pivotal reimbursement decision made by the Centers for Medicare & Medicaid Services (CMS) in Summer 2020:

"In a key step towards broader dissemination of AI, in August 2020, the US Centers for Medicare & Medicaid Services (CMS) announced its intention to provide coverage for the first AI-specific Common Procedural Terminology (CPT) code and the creation of the first New Technology Add-On Payment (NTAP) for an AI device—both historical precedents for reimbursement of AI devices."

 And to further elaborate on the creation of an NATP and what it encompasses:

"The NTAP is equal to the lesser of 65% of the amount by which the total covered costs of the case exceed the DRG payment or 65% of the costs of the new technology. This formula requires Medicare and hospitals to share the financial risk of providing costly new technologies. The NTAP for each patient is determined individually on the basis of CMS calculations, with a maximum reimbursement set at $1040."

To put matters simply, the hard ROI benefits of AI are downstream of its soft ROI benefits. Not only can more high-quality radiology scans be made in a shorter period of time, but more accurate—and fewer inaccurate—diagnoses can be made. 

This directly translates into significant savings when one considers how medical errors alone result in roughly 100,000 deaths per year and cost the healthcare system $20 billion annually in the United States alone. 

The Bottom Line

AI provides significant benefits to healthcare's bottom line, enables rapid and more accurate diagnosis and treatment of illnesses, and contributes to the overall productivity of physicians. This naturally extends to higher patient satisfaction, which in and of itself directly loops back to affect the three factors just mentioned positively. 

If you’d like to learn more about how a web-based AI imaging platform can integrate into your existing workflow through the automation of manual processes and the improvement of vital patient outcomes, click here to schedule a free demo today.