FESTIVAL DELLA SCIENZA MEDICA

INTELLIGENCE OF HEALTH
BOLOGNA, 9-12 MAY 2019

Health depends on intelligence just as intelligence depends on health. Studies in a discipline known as “cognitive epidemiology” – a topic to be discussed at the Festival by psychologist Ian Deary – have shown that higher intelligence, measured as IQ, at the age of 11 lowers the risk of vascular disease, obesity, various chronic diseases, and some mental disorders. More intelligent people fall ill less both as adults and in old age. This effect is not caused by socio-economic status, which is known to inversely correlate with morbidity and mortality. It is directly linked with IQ, which is however an important factor in attaining a socio-economic status that protects health.

Cognitive epidemiology has shown evidence that once people reach the age of 80, health and longevity are mostly impacted by the cognitive decline taking place from the age of 11 to 79. In particular, it is a person’s fluid or critical cognitive abilities that count rather than initial crystallized intelligence. These observations apply equally to men and women.

Poor health penalize intelligence. In countries where children grow up undernourished, or contracting infections – some of which, like malaria, severely affect the brain – suffer cognitive impairments. This has negative knock-on effects not only for the individual but also for the economic, health, and social development prospects of a country. For when too many people remain of low intelligence, it will be more difficult to create a society founded on respect for the individual, economic freedom, justice, the right to health, and so on. Some scholars believe that the infectious disease load as well as undernutrition, which varies from country to country, explains the differences in IQ found around the world. In their opinion, this could also explain the so-called Flynn effect, term that refers to the increase in intelligence observed from the 1930s in Western countries, which could be caused in part by the drop in the prevalence of those infectious diseases, that sap the metabolic energy required in childhood to develop a cognitively efficient brain.

The challenge, facing increasingly personalized healthcare, is to understand how to enable patients to intelligently control their behavior, and therefore avoid or manage risks to health. This includes the threat to health posed by pseudo-medicine. in other words, how can intelligence be improved, in order to avoid the spread of beliefs like homeopathy, acupuncture, and the so-called complementary medicines.

Seen against the backdrop of the data provided by cognitive epidemiology, prevention becomes a question of understanding how education levels and educating women in developing countries can contribute to improving health. To what extent can we, in today’s affluent complex societies, prevent our cognitive biases and physiological constraints from giving rise to behavioral patterns that put health at risk?

There is a further forward-looking aspect to consider. The human species has developed machines with artificial intelligence, whose use in medicine is exploding. Indeed, AI will have an enormous impact on the treatment and prevention of disease, and, by the same token, on health promotion.

Some believe that AI will gradually make the physician obsolete. It is unlikely, however, that the physician of the future will resemble the robots of sci-fi movies, uttering diagnoses – almost always in female voices – and doing everything by themselves, with minimal assistance from man. AI will nonetheless lead to a new kind of physician, who must be able to work with machines, helping them to learn with increasing versatility and “intelligence” in order to implement practices and decisions taken on the basis of indicators, and study, i.e. verify, how to apply baseline knowledge to develop new therapies or prevention methods. Something similar happened in the past, when diagnostic techniques made some kinds of semiological expertise obsolete in the clinic.

AI has speeded up the diagnostic process and lowered error rates. Machines are more accurate and faster at checking large amounts of data, a task that would take months for a team of physicians. This means more lives saved. AI will also remove the dramatic uncertainties that are today part of many clinical practices, notably surgery, where robots will soon become smarter and more independent than humans. The robot’s ability to handle metadata, meta-models, augmented reality, deep learning, and machine learning etc. will increase its ability to work with increasing precision – which in turn means increasing patient safety and confidence. AI may also reduce defensive medicine. Physicians’ mistakes and erroneous diagnoses could be wiped out, as more accurate, efficient algorithms allow physicians to adopt indisputable standards.

The main problems surrounding the planning and use of AI today, arise from the fact that our psychological flaws and human prejudices are embedded in the algorithms we use, as data collection is based on our biased perspectives. These errors are then carried over to the machine’s decision-making processes and in turn feed into the clinical recommendations generated. As a result, some algorithms have been seen to be racist, discriminate against women and children, etc. Data sources must be critically assessed to enable understanding of how the statistical models generated actually work, and how their flaws can be eliminated. Perhaps AI is not a threat because it is likely to become too independent of man, rather because it is still too dependent on man.

 

Gilberto Corbellini
Scientific Director, Festival della Scienza Medica