Closing the last mile in healthcare AI implementation
This is the first of three articles, sponsored by Philips Healthcare, exploring how healthcare organizations can realize the powerful promise of AI to improve efficiencies and help improve patient outcomes by addressing AI implementation barriers strategically, balancing the best of AI and humans, and focusing on ethical AI and related governance.
Picture a hospital where a radiologist, aided by artificial intelligence (AI), swiftly identifies a tiny tumor hidden within a complex scan, potentially saving a patient’s life. Or a clinic where an AI-powered system alerts a physician to a patient’s early signs of sepsis, enabling prompt treatment and preventing a life-threatening situation. These scenarios are no longer futuristic fantasies; they are becoming a reality, thanks to the rapid advancements in artificial intelligence in healthcare. However, there remains a critical challenge: closing the “last mile” and ensuring that these powerful AI-driven insights translate into tangible actions that improve patient outcomes.
Barriers to closing the “last mile”
Consider the scenario where an AI system predicts a patient’s deteriorating condition hours before symptoms manifest, yet the insight fails to reach the clinical team in time to prevent a crisis. This scenario illustrates one of the key challenges in closing the ‘last mile’ in healthcare AI: ensuring that AI-generated insights are seamlessly integrated into clinical workflows and lead to timely interventions.
While healthcare organizations are heavily investing in AI, many struggle to operationalize these insights effectively.
1. Cultural and organizational barriers
One of the primary success factors is active engagement and collaboration with clinicians. When clinicians are actively involved in the AI implementation process, they can help ensure that the technology aligns with clinical workflows, addresses their concerns, and ultimately improves patient care. For instance, Lahey Hospital & Medical Center successfully integrated six AI algorithms into clinical workflows by actively engaging with clinicians and offering them comprehensive education on the tool.
2. Technological barriers
Data quality and accessibility remain critical issues. The effectiveness of AI depends on the availability of high-quality data, which is often hindered by fragmented technology infrastructures. Additionally, a lack of AI literacy among staff can impede the seamless integration of AI tools into existing systems.
3. Ethical and Trust barriers
Transparency and explainability in AI are crucial for building trust. Healthcare professionals need to understand how AI-derived insights are generated to feel confident in using them. This requires comprehensive upskilling initiatives across all levels of the organization, including clinical, executive, IT, and administrative staff. Training programs focused on AI literacy can empower staff to effectively use AI tools, understand their ethical implications, and contribute to the responsible development and deployment of AI in healthcare. Ensuring continuous feedback loops between AI systems and users can help maintain trust and improve system reliability.
Case Studies and Examples
Several organizations recognized as “Most Wired” have successfully integrated AI into clinical workflows, showcasing the potential benefits:
- Real-Time Alerts for Sepsis Risk: Some Electronic Health Record (EHR) systems now incorporate AI-driven alerts for sepsis risk, allowing for timely interventions. Johns Hopkins University developed an AI system that analyzes electronic medical records (EMRs) and clinical notes to identify patients at risk of sepsis. In a study involving 5 hospitals and 590,000 patients, the system reduced sepsis mortality by 20% and detected sepsis an average of 6 hours earlier than traditional methods.
- AI-Powered Diagnostic Support: The Mayo Clinic’ cardiology team uses AI applied to CT scans and ECG/EKG tests for early risk prediction and diagnosis of serious or complex heart problems, including stroke, weak heart pump, and atrial defibrillation. The provider is looking to grow its use of AI for prediction and diagnosis and is developing AI tools that are compatible with smartphones and high-tech stethoscopes.
- Personalized Treatment Recommendations: Predictive analytics models offer personalized treatment plans based on patient data, improving care outcomes. MD Anderson Cancer Center is using AI to analyze large datasets of patient information (genomics, medical history, etc.) to develop personalized cancer treatment plans. This helps oncologists determine the best therapies for individual patients based on their unique tumor profiles.
Strategies for overcoming AI implementation barriers
Change management
Addressing cultural and organizational resistance requires robust change management strategies. Engaging clinicians early in the AI integration process and providing continuous education can foster acceptance and enthusiasm. Cleveland Clinic actively involves clinicians in the selection, development, and implementation of AI solutions. This includes gathering feedback on the usability of the tools, addressing concerns, and ensuring that the AI aligns with clinical workflows. This participatory approach fosters a sense of ownership and buy-in among clinicians.
Data fluency and AI literacy
To fully leverage AI’s potential, healthcare organizations must prioritize continuous upskilling of their workforce, focusing on key areas such as data literacy, AI ethics awareness, and change management skills. These investments will empower staff at all levels to effectively utilize AI, ensuring its responsible implementation and maximizing its potential to improve patient care. By fostering a data-fluent and AI-literate workforce, healthcare organizations can bridge the “last mile” gap and transform healthcare delivery.
Data infrastructure investment
Healthcare organizations must prioritize investments in robust data infrastructures to ensure data quality and accessibility. Reliable data pipelines are essential for the effective functioning of AI systems. UCLA Health turned to cloud-based high-performance computing (HPC) and data lake storage to help it more quickly process large amounts of clinical and research data into AI-driven insights physicians and researchers can use to improve care and outcomes.
Transparency and explainable AI
Implementing explainable AI (XAI) can address ethical concerns and foster trust. By making AI processes transparent, clinicians can better understand and trust the insights generated, leading to higher adoption rates. For instance, research has shown that incorporating XAI into AI diagnostic tools for dermatology enhances clinicians’ confidence in their diagnoses and increases trust in the AI system.
Bridging the gap between AI insights and actionable outcomes
Closing the “last mile” is essential for realizing the full potential of AI in healthcare. By addressing cultural, technological, and ethical barriers, healthcare leaders can ensure that AI-driven insights lead to actionable outcomes, enhancing patient care and operational efficiency. It is imperative for healthcare organizations to prioritize strategies that bridge the gap between AI insights and real-world application, ensuring that the benefits of AI implementations are fully realized at the point of care.
Closing the ‘last mile’ is not merely a technological challenge; it is a call to action for healthcare leaders to embrace innovation, invest in robust data infrastructure, and foster a culture of collaboration and trust. By doing so, they can unlock the full potential of AI, transforming healthcare delivery and improving the lives of patients.
Healthcare leaders must address the “last mile” problem to fully realize AI’s potential. By tackling cultural, technological, and ethical barriers, they can ensure AI-driven insights lead to actionable outcomes, enhancing patient care and operational efficiency. Healthcare organizations should prioritize strategies that bridge the gap between AI insights and real-world applications, fostering a future where AI truly transforms patient care
About Philips
Royal Philips is a leading global health technology company focused on improving people’s health and well-being through meaningful innovation, employing about 74,000 employees in over 100 countries. Our mission is to provide or partner with others for meaningful innovation across all care settings for precision diagnosis, treatment, and recovery, supported by seamless data flow and with one consistent belief: there’s always a way to make life better. For more information, please visit https://www.philips.com/global.