Skip to content
Prof. Dr. Mustafa Özdoğan · Medical Oncologist

Transforming Health Completely with Artificial Intelligence

Transforming Health Completely with Artificial Intelligence

Why Is There an Urgent Need for AI-Powered Clinical Decision Support Systems?

People living in low- and middle-income countries struggle to access health services and face unfavorable treatment conditions because of staff shortages and financial constraints. The toll of these shortcomings is thought to be very heavy, causing approximately 8.4 million deaths and around 1.6 trillion dollars in lost productivity each year. The difficulty of training competent health professionals, providing systems that enable these professionals to attain adequate clinical knowledge and management standards, and shortcomings in reliable, data-driven clinical decision-making systems are important elements of the complex causes of the problems faced in these countries. Artificial intelligence is thought to have the potential to revolutionize the improvement of health services that are inadequate because of gaps in knowledge and management in these countries. Artificial intelligence uses machine learning or deep learning networks to learn through trial and error and adapt to changing circumstances. Today, advanced technologies and the exponential growth in data generation are increasing the applicability of artificial intelligence technologies while reducing their costs. In addition, many countries have announced national artificial intelligence policies for use in health services. In the figure below, you can see the definitions and differences between 3 widely used, closely related new information technology terms: artificial intelligence, machine learning and deep learning.

what artificial intelligence, machine learning and deep learning are, and the differences between them

The Miracle of Artificial Intelligence – Or Another Disappointment in the Digitalization of Health?

The beneficial use of artificial intelligence in global health systems faces 3 main challenges:

  • The first is the availability and reliability of suitable data. Artificial intelligence systems are trained using Big Data (large datasets), so obtaining high-quality outputs requires training these systems with data that are also high-quality and reliable. In low- and middle-income countries, the main problem is limited access to data, and even the available data may not be of sufficient quality for clinical application. Although there has been a substantial increase in data obtained from health systems, when we look at their sources, we see that they mostly come from developed countries. Using artificial intelligence systems trained and developed solely on these data in underdeveloped and developing countries may cause serious problems. Criticism of these systems also mostly concerns their applicability, reliability and ethical use. In low- and middle-income countries, the introduction of new technologies may also create problems in terms of sociological and biological applicability. However, artificial intelligence systems may cause more critical problems. As systems that constantly change and develop, they can support human decision-making mechanisms or, in some cases, even make decisions in their place. In particular, as an example of systems that can make decisions instead of well-trained pilots, the Boeing MAX aircraft crashes that we have recently followed in the news show that such systems can also cause fatal and irreversible errors. Artificial intelligence technologies with high potential may have more serious consequences than other technologies. For this reason, oversight of these technologies is also vital. However, because overseeing these systems is costly, much of the technology in this field is implemented without sufficient oversight. Machine-assisted decision support systems must never be put into practice without being properly overseen.
  • Second, there are significant challenges in the implementation of artificial intelligence tools in health systems. These tools must be used by the right people, and making this match is often difficult. However, even the best tools can be ineffective or even harmful in the wrong hands. This raises the following questions: How do health systems motivate untrained or poorly trained professionals to use these tools? Similarly, how do they assure providers that they will not be replaced by tools that may possess far more knowledge than even the most skilled professionals? Despite the obvious enthusiasm for artificial intelligence, there has been little focus on ensuring that the workforce can use these tools effectively and is motivated to do so.
  • Third, most health systems lack the regulatory capacity to oversee and manage rapidly changing technology. Most low- and middle-income countries have inadequate regulations even for traditional healthcare systems and technologies. Introducing AI-powered systems into healthcare brings complex regulatory requirements. Examples include responsibilities such as obtaining user consent, ensuring ethics and privacy, and maintaining data security for the processes that take place within the stages of artificial intelligence models that cannot be fully explained and are described as a black box.

In many low- and middle-income countries, the absence of strong regulatory structures makes it difficult to scale artificial intelligence technologies effectively at the system level. The World Health Organization and others have begun to establish guiding principles to regulate the use of artificial intelligence, but each country will need to adapt these principles to its own political circumstances and health services. Building the capacity to understand this rapidly changing environment and implement regulatory principles will be difficult even in developed countries; it will be a major obstacle almost everywhere.

How Can Failures of Artificial Intelligence Solutions Be Overcome?

One approach could involve substantial forward-looking investment in data collection, the harmonization of standards and the improvement of tools with insufficient resources. Although adapting health systems in low- and middle-income countries to artificial intelligence technologies initially requires high expenditure, the potential future gains are thought to be worth it. This feasibility is becoming increasingly apparent as universal health coverage programs are implemented and countries seek low-cost solutions to strengthen their healthcare delivery systems. To obtain the full benefit of artificial intelligence technologies, it will probably also be necessary to invest in less expensive technologies such as mobile applications and sensors, through which different types of high-quality data can also be obtained via various channels. - Related topic: Closing the distance between doctors and patients with a patient-reported outcomes application. Then, with economic incentives and regulatory measures, companies that produce artificial intelligence solutions could ideally use these new types of data to create, or at least validate, their own solutions for underdeveloped countries. Another part of the solution is to train health workers to use these technologies well. Since governments often lack the capacity to deliver this training, it is a task that can be undertaken with suppliers. The private sector can also invest more in user-friendly interfaces and ensure that the right applications reach the right types of providers.

Conclusion

Finally, we believe that a key part of the solution is for governments to invest in regulatory agencies. Artificial intelligence is advancing too rapidly for most governments to regulate it using traditional methods. Identifying new governance models to oversee these technologies, ensuring that their effects are monitored, and training regulators to manage them adequately are critical. Although regulation should initially be light, ideally there will be sufficient regulatory capacity to allow real-world applications without putting public health at serious risk. The balance here will necessarily vary according to specific needs, but now is the time for each country to begin developing its own formula. Artificial intelligence is emerging as a powerful and disruptive technological innovation with the potential to transform universal health systems. Nevertheless, we cannot ignore the significant challenges ahead. We believe that if we take them into account and take appropriate steps, we can ensure that artificial intelligence becomes a key tool for improving health equity around the world.

Sources

Mehul C. Mehta, M.D., Ingrid T. Katz, M.D., M.P.H., and Ashish K. Jha, M.D., M.P.H.
Article.
Transforming Global Health with AI.
NEJM, 27 feb 2020.

Appointment request

DROZDOGAN AI

Explore this article together

Ask about the article and read answers alongside published sources from our site.

DROZDOGAN AI

Drozdogan Assistant

Transforming Health Completely with Artificial Intelligence

Ask about the article and read answers alongside published sources from our site.

Conversation and privacy information

Drozdogan Assistant answers this conversation. Authorized site administrators may review it for quality. Do not share personal information.

Replies use OpenAI infrastructure. Questions and replies are processed by that service; conversation records are stored on our site. Do not share identifying, contact or personal health information.

0 / 1000

For information only; no personal diagnosis or treatment advice.

Related articles

Explore other articles on this topic.

Stay informed

Receive new articles and scientific updates by email.

Google Analytics runs on public pages independently of this window and may use analytics cookies. Your choice here controls external services only.

Cookie policy