Authors: Mohamed Abdelhakim (Egypt), Shasta Addessi (USA), Martin Padi Amoatey (Ghana), Alison Girouard (Canada), Joyce Leido (USA) and Alisson Verissimo (Brazil)

Reviewers: Yosra Al Salami (YEL 2022, Emirates Health Services, UAE), Logandran Naidoo (YEL 2021, AIDS Healthcare Foundation, South Africa) and João Goldschmidt Oliveira (YEL 2023, Luz Saúde, Portugal)

1. Introduction

Healthcare organizations across the world are exploring how Artificial Intelligence (AI) and Big Data can be integrated to support both the efficiency and effectiveness of how care is delivered.  Whether applied to routine administrative tasks, resource triage, or more complex functions like identifying potential medical conditions in a patient’s Electronic Medical Record (EMR), AI is transforming how we operate in the healthcare sector. For the purposes of this commentary, we will use the Organisation for Economic Co-Operation and Development (OECD) AI Principle definition of AI, which is “a machine-based system that, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments. Different AI systems vary in their levels of autonomy and adaptiveness after deployment.” (Organisation for Economic Co-Operation and Development, 2019).

Leveraging these capabilities can support improved clinical decision-making, personalize treatment strategies, and streamline resource allocation, ultimately improving patient outcomes and increasing health system efficiencies.  While the potential advantages of AI in healthcare are vast, the associated risks and challenges are equally significant. As the technology rapidly evolves, healthcare leaders must address critical issues related to strategy, culture, biases, and capabilities. This article aims to motivate and equip healthcare leaders to tackle key aspects of AI implementation, ensuring they are prepared with practical strategies for deploying AI in both organizational and clinical practice while mitigating potential risks.

2. Emerging opportunities with AI in healthcare

AI and related technologies are increasingly being applied to the healthcare sector to improve efficiencies and patient care outcomes, and experts predict the global AI in healthcare market value will reach $272.91 billion by 2030 (Data Bridge Market Research, 2023).

Figure 1: Global artificial intelligence in healthcare market is expected by 2030. Adapted from Data Bridge Market Research, 2023.

Administrative applications

The rising administrative demands placed on healthcare providers is well understood and a key contributor to provider burnout (Herd, Moynihan, 2021). There are numerous AI applications that can reduce this growing administrative burden and increase satisfaction for clinical providers, including claims processing, automated scheduling, clinical and ambient documentation, scheduling, billing, and EMR queries. As an example, machine learning can be used for probabilistic matching of data across different databases, helping insurers verify the validity of millions of complaints and avoiding the costs and burden associated with incorrect claims (Davenport & Kalakota, 2019). Additionally, AI tools aiming to streamline clinical documentation and alleviate cognitive strain on healthcare providers can contribute to decrease workloads for clinicians (Buchanan et al, 2020). Ambient artificial intelligence (AI) scribes, which use machine learning applied to conversations to facilitate scribe-like capabilities in real time, have shown potential to reduce documentation burden, enhance physician–patient encounters, and augment clinicians’ capabilities (Tierney et al, 2024). In the US, the Permanente Medical Group’s rollout of ambient AI scribes to reduce documentation burdens has been deemed a success, saving most of the physicians using it an average of one hour a day, allowing them to spend more time directly with patients. (Robeznieks, 2024).

Clinical applications

In clinical settings, AI demonstrates significant potential in:

• Disease Detection and Diagnosis: AI-powered systems have shown high accuracy in interpreting medical imaging, such as detecting tuberculosis in X-rays (Qin et al., 2019).
• Robotic Assisted Surgery: AI is significantly advancing robotic surgery by providing enhanced intraoperative metrics and enabling automation of surgical tasks. It is also transforming surgical education through automated skills assessments and specialized feedback. (Knudsen et al., 2024)
• Predictive Analytics: Machine Learning (ML) models have been applied to predict patients at risk of ventilator-associated pneumonia, enabling early preventative interventions (Frondelius et al., 2024).
• Rare Disease Identification: AI algorithms have been developed to detect potential rare genetic conditions in children and adolescents, streamlining the diagnostic process (CHEO Research Institute, 2024).
• Treatment Optimization: AI is being used to predict the efficacy of pre-surgical chemotherapy in breast cancer patients, potentially improving treatment outcomes (Lavery & Ward, 2023).
• Clinical Decision Support: AI systems are being developed to aid in decision-making, tailoring treatment strategies to individual patient needs and preferences (Batra & Manishkumar, 2024).

Figure 2: Global artificial intelligence in healthcare market by application in 2023 in USD million. Adapted from Data Bridge Market Research, 2023.

Patient engagement

Equally, there are numerous documented examples of AI technologies supporting increased patient engagement. When patients become active participants in their care, rather than passive recipients, individuals take an active role in managing their health and health behaviors (Batra & Manishkumar, 2024). If deeper involvement of patients in their care results in better health outcomes, then health education is very important in improving the health condition of patients and healthcare in general. Following a study by Davenport & Kalakota (2019), the more patients proactively participate in their own well-being and care, the better the outcomes. AI has the ability to customize patient education, ensuring it is both relevant and comprehensible for each individual. AI can also be used to forecast which patients might struggle with adherence and customize interventions to address these risks. Additionally, AI-powered chatbots and virtual health assistants offer ongoing support and motivation, helping patients adhere to their treatment plans.

3. Challenges and risks associated with AI in healthcare

Although there are several benefits of AI in healthcare, leaders need to take pause and consider the inherent risks associated with its use. These include inaccurate, incomplete, biased or false responses; data quality and data bias; automation bias; skills degradation; and undermining of informed consent (World Health Organization, 2024). One of the primary challenges for leaders is developing a cohesive organizational strategy and approach to AI implementation that addresses these concerns. According to a study by Petersson et al. (2022) healthcare leaders stress the importance of having “an approach and shared plans and strategies at the regional level” rather than relying on individual driven initiatives. Without a clear commitment from executive leaders, the board of directors, and other key stakeholders, an innovative and transformative AI effort can be met with skepticism and may start under-resourced. It is crucial for healthcare institutions to establish a framework for change management and a governance structure led by individuals with the authority and who are empowered to drive, not just ideas for AI use, but also AI adoption across the organization (Petersson et al., 2022). A commitment to actualizing an organizational strategy, coupled with the leadership ability to navigate teams through the ebbs and flow of a change management journey is paramount.

Securing resources, both financial and human capital, is another challenge healthcare organizations may face as they begin this endeavor.  AI initiatives will require investments in the technological infrastructure, data aggregation, and recruiting talent with skill sets to support the transformation. Many healthcare institutions lack personnel with the skills required for AI development and implementation. Batko and Ślęzak (2022) observed differences in the readiness of healthcare institutions, surfacing that many are still at early stages. Developing the capabilities is vital. Organizations need to invest in training existing staff while also attracting and hiring talent proficient in AI. It may be necessary to establish roles that bridge technical and operational domains.  The costs of integrating data are commonly high and doing so “is expensive because of the lack of data standardization, the volume of data and the shortcomings of system connectivity” (Peng-Ting et al., 2020). With the growing amount of sensor data and remote monitoring, the opportunity to enhance patient care, particularly through AI-driven predictions, will expand, and the volume of data will continue to rise.

Managing data and information, with a defined governance that represent various interests across all stakeholder groups is another hurdle to overcome. Healthcare institutions often struggle with scattered data spread across systems. According to Batko and Ślęzak (2022) healthcare facilities encounter challenges in integrating and analyzing unstructured data. Fanelli et al. (2022) point out that the legal challenges and complexities of big data storage and sharing among organizations may limit collaboration across organizations and health systems. Improved data governance structures and updated policies are essential for integration of data sources needed for advanced analytics.

Resistance from the internal stakeholders is an important barrier that the organizational governance structure must address from the start.  With the expertise in healthcare and the high stakes of affecting patients, skepticism from various stakeholders is to be expected. Lack of explainability and transparency of the algorithms, potential biases, and difficulty to improve or validate the AI models by clinicians generates distrust among the users (Esmaeilzadeh, 2024).  Physicians and staff can also perceive additional workload and mistrust in the system (Peng-Ting et al., 2020) and therefore should be included in the governance structure to ensure their concerns are heard and integrated into the planning.  Additionally, the increasing dependence on AI in clinical practice may lead to skill degradation, whereby clinician critical thinking skills are eroded, leaving them unprepared to question or correct AI decisions in critical situations (WHO, 2024). This may be exacerbated by automation bias, where clinicians place excessive trust in AI outputs without critical evaluation. While AI can offer accurate and efficient results, it can operate as a “black box”, without clear understanding of its logic and rationality (Nguyen, 2024). Furthermore, AI has the potential to perpetuate biases in clinical care and the outputs need to be critically examined from a health equity lens (Kelly et al., 2018). As such, healthcare professionals, as a critical component of human capital, must be equipped with AI foundational competencies to interpret and apply insights AI tools generate in order to maintain trust and credibility with patients.

4. Discussion

The findings highlight the complex landscape of AI implementation in healthcare. While AI offers substantial opportunities to address longstanding challenges in healthcare delivery, it also presents significant risks and implementation hurdles. The potential for AI to alleviate administrative burdens and enhance clinical decision-making is particularly promising. However, the challenges of data quality, bias, and integration cannot be overstated. Healthcare organizations must carefully navigate these issues to realize the benefits of AI while maintaining the highest standards of patient care and ethical practice. The resistance to change among healthcare professionals underscores the need for a comprehensive change management approach. Engaging stakeholders early and often in the AI implementation process is crucial for success.

Based on our analysis, we propose the following recommendations for healthcare leaders:

1. Develop a cohesive organizational strategy for AI implementation, ensuring clear commitment from all levels of leadership.
2. Invest in necessary resources, including technology infrastructure and specialized talent.
3. Improve data management and governance structures to support AI initiatives.
4. Address stakeholder concerns through transparent communication and involvement in the implementation process.
5. Ensure ethical implementation by critically examining AI outputs for biases and maintaining robust patient privacy protections.
6. Foster collaboration and co-design of AI initiatives with patients, their families, and the healthcare team.
7. Stay at the forefront of emerging AI technologies, and actively engage in and integrate healthcare innovations to avoid being left behind.

5. Conclusion

As we start to see AI in healthcare move from experimental pilots to enterprise-wide initiatives, leaders face critical decisions around investment, prioritization, and change management. The call to action for healthcare leaders is clear: embrace AI as a strategic priority, invest in the necessary foundations, and foster a culture of innovation and continuous learning.  Collaborating to create guiding principles for AI innovations is crucial for ensuring that health leaders are well-equipped to advocate for and facilitate the broader adoption and integration of AI solutions (Healthcare Canada Now, 2021). This includes ensuring that initiatives are co-designed with patients, families, caregivers and healthcare providers to ensure there is buy-in and trust of these systems.

Although the hurdles may seem daunting, they are not impossible to overcome. Leaders should prioritize establishing the groundwork. To fully harness the transformative power of these technologies, leaders must consider strategy, culture, talent, data management and processes. Through navigating change management and developing a strategic approach, leaders can unlock the potential for AI to enhance patient care and outcomes.

References

1. Batra, P., & Dave, D. M. (2024). Revolutionizing healthcare platforms: The impact of AI on patient engagement and treatment efficacy. International Journal of Science and Research, 13.
2. Batko, K., & Ślęzak, A. (2022). The use of Big Data Analytics in healthcare. Journal of Big Data, 9(1), 3.
3. Bill C-27, An Act to Enact the Consumer Privacy Protection Act, the Personal Information and Data Protection Tribunal Act and the Artificial Intelligence and Data Act and To Make Consequential and Related Amendments to Other Acts, 1st session, 44th Parliament, 2022.
4. Buchanan, C., M. L. Howitt, R. Wilson, R. G. Booth, T. Risling and M. Bamford. (2020). Predicte Influences of Artificial Intelligence on the Domains of Nursing: Scoping Review. Journal of Medical Internet Research, 3(1).
5. CHEO Research Institute. (2024). World-first AI algorithm developed at CHEO leads to rare disease diagnosis for families. https://www.cheoresearch.ca/about-us/media/news/world-first-ai-algorithm-developed-at-cheo-leads-to-rare-disease-diagnosis-for-families/
6. Data Bridge Market Research. (2023). [Global Artificial Intelligence in Healthcare Market]. Data Bridge Market Research. AI in Healthcare – Statistics and Trends – FreeAgent CRM Resources
7. Davenport, T. and R. Kalakota. (2019). The Potential for Artificial Intelligence in Healthcare. Future Healthcare Journal, 6(2): 94-98.
8. Esmaeilzadeh, P. (2024). Challenges and strategies for wide-scale artificial intelligence (AI) deployment in healthcare practices: A perspective for healthcare organizations. Artificial Intelligence in Medicine, 151, 102861. https://doi.org/10.1016/j.artmed.2024.102861
9. Fanelli, S., Pratici, L., Salvatore, F. P., Donelli, C. C., & Zangrandi, A. (2022). Big data analysis for decision-making processes: challenges and opportunities for the management of health-care organizations. Management Research Review.
10.  Frondelius, T., Atkova, I., Miettunen, J.,Rello, J., Vesty, G., Chew, JHS., and Jansson, M. (2024). Early prediction of ventilator-associated pneumonia with machine learning models: A systematic review and meta-analysis of prediction model performance. European Journal of Internal Medicine. 121 (76-87).
11. Healthcare Excellence Canada. (2021). Implementing Artificial Intelligence in Canadian Healthcare: A Kit for Getting Started. https://www.healthcareexcellence.ca/media/qdmn0p3b/20211208_aireport_en.pdf
12.  Herd P, Moynihan D. (2021) Health care administrative burdens: Centering patient experiences. Health Serv Res. 2021 Oct;56(5):751-754.
13.  Kelly, C. J., Kathikesalingam, A., Suleyman, M., Corrado, G., & King, D. (2018). Key challenges for delivering clinical impact with artificial intelligence. BMC Medicine, 17(195).
14.  Lavery, I., Ward, K. (2023). AI and Breast Cancer: How a Canadian Lab Plans to Use New Tech to Treat Patients. https://globalnews.ca/news/9496451/artificial-intelligence-breast-cancer-research-canada/
15.  Maximize Market Research. (2023). Global Artificial Intelligence (AI) Healthcare Market. Retrieved from https://www.maximizemarketresearch.com/market-report/global-artificial-intelligence-ai-healthcare-market/21261/
16.  Nguyen, T. 2024. ChatGPT in Medical Education: A Precursor for Automation Bias? JMIR Medical Education, (10)
17.  Organization for Economic Co-Operation and Development. 2019. OECD AI Principles Overview. https://oecd.ai/en/ai-principles
18.  Peng-Ting Chen, Chaia-Li Lin, Wan-Ning Wu (2020) Big data management in healthcare: Adoption challenges and implications.  International Journal of Information Management
19.  Petersson, L., Larsson, I., Nygren, J. M., Nilsen, P., Neher, M., Reed, J. E., … & Svedberg, P. (2022). Challenges to implementing artificial intelligence in healthcare: a qualitative interview study with healthcare leaders in Sweden. BMC Health Services Research, 22(1), 850.
20.  Robert, N. (2019) How artificial intelligence is changing nursing. Nursing Management. Sep;50(9):30-39.
21.  Robeznieks, A. (2024). AI scribe saves doctors an hour at the keyboard every day. Found at: https://www.ama-assn.org/practice-management/digital/ai-scribe-saves-doctors-hour-keyboard-every-day
22.  Tierney, A., Gayre, G., Hoberman, B., Mattern, B., Ballesca, M., Kipnis, P., Lui, V., and Lee, K. (2024). Ambient Artificial Intelligence Scribes to Alleviate the Burden of Clinical Documentation. NEJM Catalyst. 5 (3).
23.  Wang, Y., Kung, L., & Byrd, T. (2016). Big data analytics: Understanding its capabilities and potential benefits for healthcare organizations. Technological Forecasting and Social Change.
24.  World Health Organization. 2024. Ethics and Governance of Artificial Intelligence for Health Guidance on Large Multi-Modal Models.(whohttps://www.who.int/publications/i/item/9789240084759.int)