Healthcare AI Technology Guide Part II: Deploying a Realistic Approach to AI Utilization

By MacKenzie Gonnelly

The proliferation of AI has been a massive topic in healthcare over the past year, serving as a major discussion point among healthcare leaders at the recent Becker's AI + Digital Health Virtual Event. As a continuation from part I in this series, we've highlighted key conference takeaways for leveraging healthcare AI technology within a hospital or health system, starting with steering committee development.

1. Start an AI steering committee early
   Hospitals and health systems have an opportunity to take an enterprise-wide approach to AI implementation. One key strategy for this is to begin an AI steering committee early on in the journey, bringing together experts from across the healthcare organization to jointly make AI decisions. Stakeholders should include leaders representing — but not limited to — operations, patient experience, ethics and health equity, compliance, data analytics, legal, security, and nursing/clinical champions. Once an expert internal team is aligned, consider inviting in several AI experts from external markets, who better understand the complexities of AI models and algorithms. Collectively, a strong steering committee can help establish clear AI policies within a health system, addressing concerns such as privacy, bias, and safety.
   
   
2. Emphasize organization-wide AI education
   While 87% of healthcare professionals do not know the difference between machine learning and deep learning — according to research from the National Library of Medicine — 79% of respondents still believe that AI could be useful in their field of work. These results show that building technical awareness and emphasizing AI education is highly important in a health system's AI journey. HIT executives may choose to develop training sessions to teach both clinical and non-clinical staff about the differing types of AI, how these technologies work, and ways they can reduce administrative burden or cognitive overload. This can be especially critical education for physicians who often get pitched by various AI vendors. Providing valuable education may also cultivate clinical AI champions, who then advocate for additional AI initiatives in the future growth of the organization.
   
   When thinking about one's own healthcare, most individuals tend to view AI more critically. In fact — according to the Pew Research Center — 60% of Americans would feel uncomfortable if their healthcare provider relied on AI for medical care. This distrust indicates that AI education must extend to the patient population as well. Healthcare organizations can offer educational materials and FAQs to emphasize the value of various AI solutions, potential health outcomes, as well as safety of new technologies to get patients on board. Teaching patients that tools like chatbots or virtual assistants can bolster a more engaging, accessible healthcare experience may reduce technology uncertainties. However, this education should not be the sole responsibility of clinicians; it needs to be implemented at an organizational level.
   
   
3. Empower healthcare staff to find areas of opportunity for AI solutions
   Since frontline staff understand day-to-day operational workflow better than anyone, it is pivotal to consider their opinion during the AI decision-making process. HIT executives might try rounding the hospital or health system, inquiring about various departmental pain points and experiences. Ask targeted questions about administrative assignments or repetitive tasks that staff encounter to develop a clear understanding of opportunities for AI use. Seek staff input on areas where they believe AI solutions could reduce their workload. This will empower employees to deliver creative solutions that they will likely champion down the line. Additionally, consider hosting focus groups to allow additional time for meaningful pain point discussions. Creating a collaborative environment between IT and non-IT staff can help generate excitement behind these often-stressful technology advancements.
   
   
4. Re-negotiate contracts to include AI-specific language
   HIT leaders must still focus on vendor risk management with their solution partners. Amid enthusiasm for new technology, HIT leaders may overlook the need to re-negotiate or update contracts to include specific AI-related language when tools are introduced. However, this contract review must become a priority, discussing topics such as data usage or data protection with their current vendors to safeguard the hospital or health system. While AI governance and procedures will vary across organizations, it is important to assess a vendor's AI policies to ensure they align with your health system's framework. Establishing these guardrails early on, even with established partners, will enhance organizational and patient safety long term.
   
   
5. Prepare for eventual AI vendor consolidation
   According to research by Fortune Business Insights, the global market value for AI in healthcare is expected to grow from $19.54 billion in 2023 to $490.96 billion in 2032. Along with the projected growth, the AI vendor market in healthcare will likely become increasingly saturated. Thus, we can expect vendor consolidation to ultimately occur. Although an exact timeline is uncertain, HIT executives should prepare for this, being ready to adapt and respond to any vendor shift. HIT leaders should always know what third parties are using their health system's data. In regard to these vendors, ensure a thorough understanding of the full scope of organizational data they access. Flexibility and adaptability are crucial in any eventual shift, so being prepared with as much information as possible — and taking corrective action if needed — will ease any transition concerns.
   
   

By combining these five best practices for the deployment of AI initiatives within a health system, HIT leaders can strategically advance hospital technology to align with patient and staff needs, while reducing vendor risks and rollout hurdles along the way.

To read additional guidance for hospital IT directors and CIOs, stay tuned to Stoltenberg blog.