The story we dreamed between artificial intelligence (AI) and health has been surpassed. There are ideas that can be tried and improved, others that can be imagined, but the applications of them Big data and the possible uses of AI are still close to promising research. Opportunities include drug design through better identification of molecules, assessment of treatment adherence, diagnostics with image analysis, and the way control groups are recruited for clinical trials.
This is the prediction that Alberto Hegewisch, AstraZeneca’s medical director for Latin America, gave to technology journalist Rosa Jiménez Cano in the talk “Artificial and human intelligence: Changing the frontiers of health” at the WIRED Summit 2023.
Hegewisch studied medicine but has worked in the pharmaceutical industry for years. He pointed out that AstraZeneca has evolved from an industry that sells medicines to one that is able to support the medical ecosystem and create collaborations that improve patients’ lives; from studying medications to diagnosis and follow-up care.
Some very important events are taking place in Latin America, where four of the company’s 20 innovation centers are located. The project in Brazil focuses on oncological problems and timely detection, the project in Colombia on telemedicine and tools that allow to reduce the interaction or transfer of patients to their clinics, and the project in Mexico develops solutions for first contact medicine. Hegewisch saw the beauty of these centers in the fact that they integrate both the public parts as well as the private initiative and the academy.
The Startup The Ecuadorian uses artificial intelligence to hack people’s lives and improve their metabolic health through information from their glucose and consumption habits.
AI as a tool for health
One of the areas highlighted by the pharmaceutical company’s CEO was drug development, which used to take between 10 and 15 years. The time spent obtaining the necessary evidence has now been significantly reduced through the use of AI.
Regarding treatment adherence, the pharmaceutical director pointed out that data needs to be generated in order to make better decisions, but also to reduce hurdles for patients. For example, during the pandemic, they found that 70% of what is evaluated about a patient in a clinical trial can be done remotely via devices.
An application of these observations is taking place at the National Institute of Cardiology, where technologies are being evaluated in heart failure so that the patient is aware of some health variants and his doctor knows their adherence and parameters in real time.
Elena González-Blanco: “AI is losing a huge deal because it doesn’t speak Spanish”
The big challenge is to enrich artificial intelligence with data from the cultural heritage of Spanish speakers to expand their business opportunities.
Compliance with the treatment is in the company’s interest, says Hegewisch. “The vast majority of health expenditure, not only in Mexico but worldwide, is not due to the disease, but to the complications of the disease or late diagnosis.”
Among the pharmaceutical company’s decisions, Hegewisch highlighted that his area Data science provides everything that is learned, “whether positive or negative, learning from mistakes, this allows us to make the following developments faster and with a higher success rate.” The doctor points out that the collection and treatment information must be handled with respect and care. “We do not want this data to be used for commercial purposes.”
Other applications being tested involve shrinking the study groups that receive a placebo and using artificial intelligence to help identify and make comparisons with patient groups with the same characteristics as the group receiving the treatment . “This would help recruit fewer patients and enable more studies.”
They are also exploring the possibilities of diagnosis through image analysis: “Millions of tomography scans have been carried out worldwide during the pandemic. We are implementing a tool at the hospital level so that every person who undergoes a chest medical scan for any reason goes through intelligence machines. artificially. This will help the radiologist.” The same applies to mammograms. Specialists are expected to optimize their evaluations.