How AI Has Impacted Global Healthcare

ai
AI(Artificial Intelligence) concept. 3D rendering brain.

In this DocWire review, we analyze four use cases for artificial intelligence (AI) in the global healthcare space. These cases were recently included in a USAID publication regarding AI, and detail the power this technology can have in transforming the medical space.

AI-Enabled Population Health

This application includes AI-tools that assess and monitor population health, as well as select public health interventions based on analytics. Included among this category are data processing methods that characterize the spread of disease, as well as analytics that project future disease spreading and potential outbreaks.

This approach was used by Dr. Dhesi Baha Raja, who uses AI analytic tools in his work with the Ministry of Health (MoH) in Malaysia. Though his work focuses on controlling infectious diseases, Dhesi was without proper analytic tools to predict disease spread until last year. The MoH had a small pool of health data from regions across the country and no method of including non-health data into their work.

To resolve this issue, the MoH created integrated and comprehensive electronic medical records (EMRs) from healthcare providers throughout Malaysia. Applying AI tools to these EMRs allowed the MoH to map out different health risks and disease outbreaks throughout the country and integrate non-health information such as weather patterns into the data set to better predict outbreaks. In doing so, Dhesi’s team was able to view and analyze health data in real time, as well as generate appropriate interventions.

Assisting Frontline Health Workers (FHWs)

FHWs benefit from using AI in that it provides a convenient means of transporting healthcare to low-income regions that lack proper access to resources. These tools assist in diagnosing patients outside of healthcare facilities, clinical decision support, and tracking patient compliance. The speed and accuracy of these AI-driven tools offer a reliable means of real-time data collection. By offering this remote healthcare option, patient burden is reduced on smaller facilities that may already be overwhelmed with incoming patients.

USAID described one such application, in which Anita, a community health worker from a small village in Western Kenya uses AI-enabled apps on a smartphone to help local citizens. One day she encountered a 2-year-old patient with a high fever and a rash, whose mother was unsure what was wrong. Using AI-apps, Anita enters in the child’s symptoms as well as a photo of the rash to be analyzed by advanced algorithms. The rash is diagnosed as a spider bite, but the child is deemed a candidate for malaria and blood testing is conducted. Using the smartphone camera and a disposable blood test, another AI-powered app determines the child does not have malaria and needs no further care.

Patient Virtual Health Assistant

This application of AI puts the power in the patient’s hands, equipping them with tools for self-referral, diagnosis, medical record collection, self-care functions and more. It should be noted that these tools are not intended to replace the physician in healthcare, but rather to provide recommendations and insights to patients to alleviate their symptoms in time’s they cannot seek professional help.

These technologies can provide patients with medical information, lifestyle and behavioral recommendations, direction to proper care, and potential diagnoses without having to go to a professional. This holds similar utility in low-income areas as the AI-tools available for FHWs to use.

Offering Physicians Clinical Decision Support

Tools under this category include those that help physicians in tasks such as reading diagnostic images, strengthening decisions, and quality assurance regarding previous performance and where errors were potentially made. Like the virtual health assistant tools, the purpose of this AI-support is not meant to replace one’s physician. These support tools ultimately enhance the physician’s ability to serve their patients, offering enhanced speed and accuracy in some diagnoses.

Examples of this include IDx-DR, an AI-driven tool that can diagnose diabetic retinopathy from an image of a patient’s retina. FDA-approved technologies such as these offer great utility to low-income regions with limited access to specialists. Such a condition often requires a highly-trained optometrist for diagnosis, and with AI-driven diagnostic tools these conditions can be detected without the presence of a physician.

Source: USAID