Why AI is becoming the disease detective

Tufts Center for the Study of Drug Development estimates that it costs, on average, $2.8 billion to develop a prescription drug that gains market approval. This figure could conceivably be reduced through AI by providing better patient recruitment for clinical trials, leading to quicker discoveries, for example. The greater the insight the medical profession can have into diseases, the better it can diagnose more quickly and provide personalized treatments.

AI also raises the prospect of affordable healthcare for all. According to the World Health Organization (WHO), 400 million people do not have access to one or more essential health services, and 6% of those in low and middle-income countries are pushed further into extreme poverty because of health spending.

In many cases, AI can diagnose diseases more accurately than doctors. In the future, we will see physicians working in partnership with AI – enabling technology to free up their time to concentrate on treatment of the disease as opposed to the diagnosis.

Here we look at areas where AI promises to have a real impact on chronic and infectious diseases, from diagnosis and treatment plans to containing the global outbreaks of the likes of SARs and Ebola.

Cardiovascular disease

Nearly 18 million people die each year from cardiovascular disease, according to WHO. Current approaches to predicting cardiovascular disease, which include a set of standard guidelines from the American College of Cardiology (ACC), fail to identify people who would benefit from preventative treatment.

Researchers at the University of Nottingham in the UK, however, believe they may have an answer. In a recent study, they used AI to predict disease rates by exploiting complex interactions between risk factors harvested from routine data in patient’s electronic records. The study showed that AI is better at predicting the absolute number of cardiovascular diseases correctly, while successfully excluding non-cases. Present guidelines rely on only eight criteria including age, blood pressure and cholesterol levels, which are too simplistic to account for factors such as medication or multiple disease conditions.

Zika and Ebola

With our interconnected world, contagious diseases such as Zika and Ebola can spread incredibly fast. Quickly identifying the cause of an incoming hospital patient’s illness could be crucial to overall public health.

Researchers at Rush University Medical Center in Chicago are using AI to spot contagious diseases. Geographic Utilization of AI in Real-Time for Disease Identification and Alert Notification suite, dubbed Guardian, can predict if a patient is suffering from ordinary flu-like symptoms or an emerging tropical disease such as Zika or Ebola. It can also detect biological or chemical agents such as anthrax and sarin.

As physicians enter data on a patient into Rush’s electronic medical records, Guardian analyzes them in real time. The system uses computerized algorithms that sift through and add up various clinical variables to detect a contagious disease.

In addition to predicting disease, a new AI application in Guardian can forecast if a patient needs to be hospitalized within 15 minutes of their arrival to 90% accuracy, expediting bed allocation and patient flow.


New cancer diagnoses are expected to rise by 70% in the next decade, according to WHO. AI is already being explored extensively as a tool for spotting various types of cancers, leading to early detection and treatment. IBM's Watson for Oncology, an AI platform trained by researchers at the Memorial Sloan Kettering Cancer Center in New York, is already being used in hospitals to identify and evaluate cancer treatment options. Other examples include start-up Curemetrix, which is working on an algorithm for breast cancer detection in mammograms to provide personalized treatments. Researchers at Stanford University have used a database of nearly 130,000 skin disease images to train an algorithm to visually diagnose potential skin cancer. The research team hopes to make the algorithm available in a smartphone app.

Alzheimer’s disease

Researchers at the University of Bari in Italy have developed AI that can understand the changes in the brain related to Alzheimer's disease, making a diagnose possible ten years earlier than doctors.

Using MRI scans, the AI can detect early signs with 84 percent accuracy by spotting changes in how parts of the brain connect. Doctors are currently not able to identify the disease until a decade past this mild cognitive impairment (MCI) stage. This ten-year head start in treatment could be critical in slowing its progress. The research team is now planning to look at how it can apply AI to early diagnose other neurodegenerative diseases, such as Parkinson’s disease.

Globally healthcare services are under growing pressure to deal with a growing and aging population. How can telemedicine and e-health meet these challenges, while keeping the patient central to the system. Find out more at Orange Healthcare.