5 Keys to AI Success for GCC Health Organizations

While COVID-19 continues to unfold across the globe, the pandemic has catalyzed numerous AI-enabled healthcare development activities. For instance, after scientists decoded the genetic sequence of SARS-CoV-2—the virus causing COVID-19—and publicly posted the results, firms began using AI-enabled methods to rapidly develop potential vaccines, some of which are already proceeding to clinical trials. By comparison, traditional non-AI drug development processes take many months, if not years, to proceed to human clinical trials.

Increased Interest in Health AI in Gulf Healthcare Sector 

In the Gulf Cooperation Council (GCC), prominent health organizations are no strangers to the benefits of AI and have already expressed a growing interest in integrating AI into their operating model. Taking a closer look at their strategies and objectives shows that more than 50% of their strategic initiatives are centered around AI and data analytics solutions.

In the United Arab Emirates (UAE), the Ministry of Health and Prevention (MoHAP) and the Dubai Health Authority (DHA) aim to fully incorporate AI into medical services across providers and promote innovation across the healthcare industry. The DHA has already signed memorandums of understanding with international companies to introduce inventive healthcare innovations in AI. This includes the implementation of virtual health through an app that uses AI technology to provide remote general practitioner consultations round-the-clock and free-to-use AI pods across Dubai that will do quick health scans for the public and generate immediate results. In Abu Dhabi, the Department of Health (DoH) pioneered the development of an AI policy to regulate future usages of the technology in the healthcare sector and launched an AI laboratory that aims to develop groundbreaking solutions in wellness and prevention, chronic disease management, clinical care, and regulatory management.

Just recently, Saudi Arabia approved an AI strategy that is expected to contribute $133 billion in U.S. dollars to Saudi’s gross domestic product by 2030. In line with this focus, the Ministry of Health (MoH) in the Kingdom of Saudi Arabia devised its Digital Health Strategy as part of Vision 2030 that promotes reinvention across the Kingdom’s digital ecosystem and enables new models of care through connectivity and new technologies, with AI as a key pillar. In addition, the General Directorate for Research and Studies (GDRS) is actively promoting AI-related research projects as part of its Health System Research Fund Program 2020–2023.

In Qatar, Qatar University is already relying on AI capabilities in its research efforts to find a cure for COVID-19 infections. Moreover, Qatar is hosting the AI in Healthcare Qatar 2020 Forum that aims to connect global and local industry leaders and share experiences on how AI is transforming the health sector globally and position the country as a regional center of excellence for health AI.

5 Keys to AI Success for GCC Health Organizations

Building on these regional trends, GCC health players have an opportunity to achieve greater AI maturity as an enabler to predict and diagnose diseases early, inform clinical decision making for more effective treatments and therapies, and complement clinical skills and knowledge of the local workforce. Yet, achieving favorable returns of investment from AI initiatives requires careful planning and an experimental mindset that focuses on value extraction along the adoption journey. 

In an effort to tackle some of these challenges, Booz Allen has developed five key recommendations (Exhibit 1) that GCC health organizations can consider to better their chances of succeeding in their AI journey.

1. Strategize and build focus

• As an initial step, GCC health organizations will need to define the purpose and focus of initial AI efforts with care and clearly outline business challenges warranting AI adoption. That means, in part, identifying focus areas for AI applications and defining specific use cases within the identified focus areas that make a clear business case for AI investments. 
• Our research and experience in health point to several key areas where AI can yield significant value in the short to medium term. These areas include predictive analytics to improve clinical resource management and access to care, AI-assisted diagnoses to improve quality of care, and task automation to improve efficiency and cost of care.
• As a start, GCC health players can identify use cases within one or two of the above areas. They could then prioritize the identified use cases by assessing their potential impact on the organization and the wider healthcare system (e.g., optimization of service lines, impact on workforce efficiency, impact on quality of care), and ease of implementation (e.g., data availability, algorithm complexity, number of stakeholders involved).
• For example, RPA is a relatively easy AI solution to implement that can complete time-consuming and repetitive tasks such as data entry, data capture, and data transfer from one source to another. That way, RPA can serve as an easy gateway for the organization to tackle more advanced automation leveraging AI. 

2. Experiment, iterate, then scale 

• GCC health organizations need to adopt a radically different approach to digital transformations, one that fosters acceptance of uncertainty and continuous value improvement through rapid experimentation, iterations, and incremental development of minimal viable products and services. By adopting an Agile approach that fosters experimentation and learning, GCC health organizations could promptly sort out successes, quickly deliver business value, maintain adaptability, and keep a rein on AI spending.  
• Furthermore, organizations should develop their capacity—technical and otherwise—to scale successful AI models quickly and efficiently and to integrate them seamlessly with existing business services and processes. Setting up the right AI ecosystem (platforms, partnerships) can be an effective and cost-efficient approach to agility and scale. 

3. Curate core data assets

• GCC health organizations need to keep in mind that the foundation of a successful enterprise-wide AI deployment is effective data management and integration. This typically entails identifying, acquiring, transforming, and integrating the relevant data sets to provide a coherent, multidimensional information model on the problem at hand. Integrating diverse data sets thus increases the odds of success of data-hungry AI algorithms. 
  
• When possible, the data management approach should ensure that both structured and unstructured data from diverse sources can be used. This can include detailed genomic data, clinical data from electronic health records, claims data from payers, radiology images and reports, among many other sources. Health data relied on for analysis also needs to be well documented in terms of its composition, collection process, pre-processing, distribution, and maintenance processes. It is also critically important to identify the broad array of data from public sources, published papers, visualizations, maps, and application programming interfaces, especially given the rapidly unfolding pace of analysis that is required during the COVID-19 outbreak.
  
• To that end, health players can learn from the longer standing experiences of the other industries that have relatively matured in the data management field, such as the security sector, to improve their capabilities and lay the ground for a robust AI platform. Key lessons that can be extracted include:
  
  • Focusing on what matters, through the identification of defined and relevant sets of inputs (e.g., patient demographics and comorbidities) along with the corresponding correct output (e.g., expected number of visits, re-admission risk), enabling algorithms to learn and adjust by comparing correct output to actual output
    
  • Identifying and documenting core internal (e.g., patient history) and external data (e.g., health utilization information) assets required for the identified use cases 
  • Automating the extraction, validation, and transformation of data to prepare data for AI and machine learning algorithms
  • Institutionalizing data quality management practices within organizations to ensure a consistent level of data quality

4. Manage potential risks

• With the increase of AI-enabled activities, organizations can face an increase in risks related to ethics, cybersecurity attacks, poor data quality, and non-compliance with regulations and standards.
  
• With the adoption of AI in mission-critical activities, organizations need to be deliberate about identifying and managing risks emanating from AI deployments. Such risks include but are not limited to data privacy breaches, data and algorithm bias, and cyber attacks.
• Risk mitigation measures should be developed accordingly and comprise:
  • Ensuring compliance with applicable AI policies, through a periodic and systematic review of the national regulatory landscape and international standards
  • Increasing data completeness and quality, through the development of data quality management capabilities, the review of government and industry best practices of applying data quality criteria to ensure a consistent level of data quality
  • Monitoring and continuously improving AI outcomes, through the establishment of dynamic, flexible strategic targets over static ones
  • Ensuring risks are clearly understood and risk management guidelines for AI-enabled activities work are widely practiced through education and awareness 

5. Stand up digital capabilities and pivot talent

• To properly leverage AI, it is critical to develop the Agile and digital competencies that are required to building the dynamic organic and inorganic capabilities needed to manage AI initiatives and thrive in today’s and tomorrow’s environment.
  
• A health organization should therefore set up a digital ecosystem that supports the efficient development and scaling of AI solutions. The ecosystem should comprise at least three complementary components:
  • Technology platforms, including advanced data management; machine learning and analytics platforms; data modeling and visualization tools; and scalable, cloud-based compute and storage capacity
  • AI service providers with expertise in the health domain
  • AI talent pipeline, covering data management, AI modeling, and domain experience
• Finally, to support this digital ecosystem, it is important to promote a diverse digitally savvy culture—one that focuses on the customer, favors experimentation, and supports employee boldness. This mindset should emanate from the leadership of the organization that fosters innovation and truly accepts failures. Instill the Agile mentality among staff to promote collaboration, communication, adaptation, and innovation.
• It is also highly beneficial to establish partnerships with research and academic institutions to enable recruiting of rising talent with the necessary science, technology, engineering, or mathematics background for skills such as innovation and critical thinking that support the successful implementation of AI initiatives.

Although there are many factors that contribute to success, these considerations can be key as GCC health organizations aim to improve their odds in their AI journey and strive to better harness the benefits of this capability while side-stepping its potential caveats.