The road ahead for healthcare leaders continues to be fraught with challenges. According to a recent McKinsey survey, health system executives see digital and AI transformation as crucial to overcoming ongoing challenges such as rising costs, workforce shortages, aging populations, and increasing competition from nontraditional players. While AI, traditional machine learning, and deep learning are projected to result in net savings of $200 billion to $360 billion in healthcare spending, 75% of execs fear their investments may fall short.
One growing complication is the sheer volume of data generated in healthcare. By 2020, experts estimated that healthcare would generate 2.3 zettabytes of data globally. Recent estimates say that a typical hospital produces 50 petabytes per year (and there are a million petabytes in a zettabyte). The adage “Garbage in, garbage” out is truer than ever.
Quality data is critical for healthcare organizations to gain measurable returns using AI such as productivity improvement, cost reduction, and revenue capture. Consider the radical concepts of Dr. W. Edwards and their impact on the automotive and other industries. He asserted that “Quality is everyone’s responsibility” and that organizations focused on improving quality would automatically reduce costs, while those focused on reducing cost would automatically reduce quality – and simultaneously increase costs.
This principle is particularly applicable in the data- and labor-intensive area of healthcare data abstraction an essential driver for research and the shift to value-based payment. According to the Agency for Healthcare Research and Quality (AHRQ), the use of patient registries is growing rapidly. To illustrate, the AHRQ’s review of cancer registries showed over 650 entries for patient registries (four of which have over 100,000 participants) with a wide range of purposes. Data abstraction is ripe for AI advancement without introducing new risks, uncertainties, and liabilities. Here are five considerations to help healthcare leaders balance the fear of embracing new technologies with the fear of missing out (FOMO) on the potential gains in quality and cost reduction.
Drive operational efficiency and cost savings
Time is money, and one of the biggest draws for AI solutions is the promise of time savings. As Deming’s philosophy advocates, however, efforts to reduce costs shouldn’t come at the expense of quality. AI-enabled clinical data abstraction using large language models (LLMs) – that are well-trained on clinical data – can accelerate the gathering of information required by registries for human review. This relieves staff of the tedious aspects of the job and allows them to focus their clinical expertise on the quality of the information. Technology can take the lead in improving accuracy and speed to meet complex data input requirements.
Improve financial performance
For healthcare executives responsible for their organization’s overall financial health, efficient and cost-effective data abstraction directly impacts profitability. When evaluating AI solutions, it’s essential to understand the upfront investment required, the implementation time, and ongoing costs related to the expected profit contribution. Organizations need assurance that the solution will free up time for staff and enable them to work at the top of their license, resulting in higher quality, job satisfaction, and overall value.
Ensure accurate financial reporting
In addition to its importance for registries, abstracted data feeds into performance metrics, cost analysis, and revenue projections. AI solutions that ensure the timeliness and quality of that data will have a direct, positive impact on financial reporting and strategic planning.
Enhance patient care and outcomes
Accurate data is fundamental to better decision-making, leading to better patient outcomes. With AI-enabled technology that facilitates clinical registries, abstracted data can improve ongoing care processes. High-quality patient outcomes have far-reaching impacts, including higher patient satisfaction, higher quality ratings, increased revenue, and an enhanced reputation that attracts and retains patients and clinical staff.
Manage compliance and mitigate risk
Abstracted data supports compliance with regulatory guidelines and quality reporting requirements for accurate reimbursement. Significant concerns about AI have come to the forefront because some applications, such as ChatGPT, are known to make mistakes or even fabricate information. There are AI and LLM solutions, however, that are more reliably trained using clinical data. More importantly, AI solutions in healthcare should always adhere to the “human in the loop” philosophy, which states that humans review, fine-tune, and continuously improve and maintain information gathered for them using AI. This combination of clinical expertise and AI technology ensures high-quality data while lowering costs.
Balancing fear and FOMO results in the right strategy
Despite the buzz and confusion in the market about AI, healthcare leaders have much to gain with a reasoned approach to evaluating and adopting AI solutions. With the broad need for abstracted data, focusing on improving quality has historically been proven to reduce costs. Measurable returns are ready to be gained using AI to enhance productivity, reduce costs, and advance the financial performance of healthcare organizations.
About Lucas Tanner
Lucas Tanner is the Chief Financial Officer at Carta Healthcare, a leader in clinical data abstraction, applying both artificial intelligence and expert clinical data abstractors to serve a wide range of health systems nationwide, from standalone community hospitals to large academic medical centers and multi-state health systems.
link