CereCore organized a cohort of healthcare CIOs representing different healthcare organizations to delve into the realm of AI, analytics and automation. As they explored the endless possibilities, they acknowledged the tangible risks and valid concerns. The event, led by Peyman Zand, chief strategy officer, and Josh Dunaway, senior director of advanced data solutions, featured the CIOs sharing their expertise on existing and potential use cases, the level of acceptance among healthcare leaders, and crucial factors for health leaders to bear in mind amid rapid advancements.
Artificial Intelligence
A report by MarketsandMarkets estimates that the global market for healthcare AI is expected to reach $102.7 billion by 2028. Recent and ongoing advancements could significantly impact strategies to improve operational efficiencies, alleviate burnout, and boost patient satisfaction.
Along with the excitement around AI capabilities, many health leaders are justifiably hesitant about the safety and accuracy of AI tools such as ChatGPT. As AI continues to advance, information technology leaders must think strategically about some considerations pointed out by CereCore’s CIO Cohort members including:
Low-risk ways that AI solutions can be integrated seamlessly into a physician’s workflow.
This strategy has the potential to alleviate physician strain and reduce burnout. “It's all about taking a cautious, reasonable approach [to AI] —not blocking it, but making sure we don't move too fast,” said Alan Ariel, Interim CIO for Delta Dental of Missouri.
Some possible lower-risk healthcare use cases for AI solutions include:
- Discharge instructions: Tools such as ChatGPT can expedite the process for writing discharge instructions by using data and predictive technology to draft aftercare steps. Though a physician will proofread and revise instructions as needed, the availability of a first draft can save time and effort.
- Training videos: AI video generation platforms can be used to create hospital staff training videos. These tools require no hired actors or lengthy production time, which saves time and money to generate educational training videos for onboarding or other processes.
Predictive analysis based on readings collected by devices and alerts issued to care teams.
According to Dunaway, one lifesaving use case for AI solutions is sepsis prediction. “Organizations like HCA Healthcare have gained a better understanding of sepsis risk with the ability to continuously monitor patient vitals and other patient information and quickly alert their health team during the patient’s continuum of care,” said Dunaway.
Cybersecurity and privacy measures for complexities introduced by AI.
Cybersecurity and privacy concerns are always top of mind. Two concerns voiced by the cohort were:
- Chatbot security: “AI adds another complex layer to an already complicated cybersecurity environment,” said Denis Tanguay, vice president and CIO for Sturdy Memorial Hospital. “Specifically, with AI chatbots, it is essential to have the answers to questions such as, where is our data going and is it secure? It is important to understand all aspects of AI, especially with the prevalent threat of ransomware and malware attacks.”
- Data integrity: “We must be mindful of how we use AI tools. People underappreciate the significance of integrity and quality data issues. Deploying AI for straightforward tasks like creating a first draft of discharge instructions is a low-risk way to alleviate strain on clinicians, but we have to be very careful when using AI for more sophisticated purposes,” said Rich Pollack, former CIO for Sturdy Memorial Hospital. “It is important to tread carefully and remember our overly inflated expectations about analytics 10 years ago that failed to fully materialize.”
Analytics
Predictive analytics technology is currently used in various aspects of healthcare to improve patient outcomes. “The technology supports the ability to better assess the overall health and risk of the patient during a visit,” said Dunaway. “Documentation of patient encounters provides physicians with a contemporary chart, which improves clinical outcomes and supports coding and reimbursements.”
Internally, health systems use these tools for readmission risk determination and for service desk optimization. Use cases can drive better business outcomes by improving patient satisfaction and preventing the need for readmission. Analytics technology is also being leveraged to expedite insurance claim submission.
Automation
“Transitioning to new platforms of any kind requires a significant amount of build. Often whenever these conversions are occurring within a health system, clinicians have to take time out of their day to do manual data entry which adds additional strain on the clinical staff. Automation can eliminate roughly 40% of the data entry, which helps mitigate burnout and drive efficiency,” said Dunaway. This allows them to focus more of their time and energy on higher priority tasks. However, some CIOs are hesitant to leverage automation for quick fixes. “Simplification should be the predecessor to automating tasks that are not being well managed,“ said Pollack.
In Summary: Navigate AI, Analytics and Automation Carefully
Before implementing these tools, consider the following points to protect your organization and the patient:
- Quality and Integrity Issues: AI, analytics and automation are only as accurate as the data provided, which can be problematic if healthcare data is not clean.
- Physician Understanding: Be transparent with physicians and make sure they understand how these operations work and where the data is coming from so they can be confident in the results.
- Human Touch: All artificial content should be reviewed for accuracy by a clinician.
- HIPAA Compliance: AI and analytics have the potential to include sensitive and private patient data. It is critical to keep compliance top of mind when determining how your health system intends to implement these solutions.
- Cybersecurity: Having these technology solutions in place adds a complex layer to cybersecurity concerns as they could all pose new threats.
AI, analytics and automation offer unprecedented potential to solve current and future problems in healthcare. Responsible implementation is required to ensure positive outcomes for your organization and the patients you serve.
