Calculator of biological age and risk of imminent death
Swedish scientists from Uppsala University, working under the guidance of Professor Erik Ingelsson, have created a unique online calculator that allows you to estimate a person's biological age and risk of death over the next five years. When developing this tool, information stored in the largest research data bank, the UK Biobank, was used. The authors believe that their development will find wide application both in the clinic and public health, as well as in science.
The UK biobank stores the values of 655 indicators for about half a million Britons aged 40-70 years. Studies of associations between one or more risk factors and future mortality have been conducted before, but none of them used such a large amount of data. An extensive sample allowed the authors to identify several interesting patterns:
- It turned out to be quite unexpected that the variables that best predict the risk of death are not biologically measurable parameters, but independently assessed factors such as health status or walking speed. For example, the risk of death over the next five years for people with low walking speed was 3.7 times higher than for people with normal walking speed. Also, the risk of death was definitely increased for people who do not have a car, which is explained by socio-economic reasons.
- The variables that most effectively predict the risk of death differ for men and women, while the reliability of forecasts is generally higher for men.
- When analyzing data from only healthy people, smoking is the strongest predictor of death over the next five years.
- The risk of death from external causes, such as suicide or accident (data on the causes of death were also collected), is most effectively predicted using socio-economic and psychological factors.
- Forecasts are generally more reliable for the elderly than for young people.
The researchers used a computer algorithm to identify the combination of parameters that provides the most accurate prediction. As a result, 13 questions were identified for men and 11 for women, based on the answers to which the biological "Ubble–age" is calculated (Ubble, UK Longevity Explorer, is the name assigned to the calculator). If the value obtained is less than the actual age of a person, it means that the risk of his death over the next five years is below average, and vice versa.
The developers believe that their brainchild can be useful to a lot of people. It can be used in conducting prospective studies, identifying patients at high risk, making decisions about investing in health insurance, and in many other situations.
However, they note that the study demonstrates only statistical associations and emphasize that its purpose is not to identify risk factors. Users should refrain from hasty conclusions, since the results obtained show only average values and are not a personal forecast. For example, the probability of death within the next 5 years is 5%, which means only that out of 100 people with the same indicators (and even then – with a probability of about 80%), five will die, but who exactly, it is impossible to say.
Simultaneously with an article dedicated to the Ubble calculator, a commentary by two scientists was published in the same issue of The Lancet magazine: Simon Thompson and Peter Willeit from the University of Cambridge (UK Biobank comes of age). They note that the research done by the authors has both significant advantages and a number of disadvantages.
As the final result of testing, the authors use the C-criterion – the probability that a person who dies earlier has a higher prognostic indicator from a randomly selected pair. A C-criterion equal to 0.5 can be obtained by tossing a coin, whereas a C-criterion equal to 1.0 corresponds to a 100 percent probability. The C-criteria stated by the authors (0.80 for men and 0.76 for women) are very impressive.
At the same time, firstly, due to the very large number of analyzed variables, some of the strongest associations may be exaggerated and their influence may be partly due to a game of chance. The authors have done everything possible to mitigate these effects using internal cross-validation, but some exaggeration still persists.
Secondly, the laboratory data used in the analysis were presented only by the results of hematological studies, but there was no assessment of parameters such as the level of lipids in the blood, as well as the results of tests of liver and kidney functions. Therefore, the authors' statement that independently provided information in general makes it possible to predict mortality more effectively in comparison with biological measurable parameters may not stand the test of time.
Thirdly, the identified associations do not allow us to distinguish causal relationships from purely empirical (potentially erroneous) ones.
And most importantly, using only one parameter for calculating the C-criterion – age – allows you to get indicators of 0.68 for men and 0.67 for women, which greatly reduces the initial impression of a fairly high accuracy of the results of the questionnaire.
Commentators also note the possibility that the use of the Ubble calculator, in addition to improving users' awareness of their health status, is associated with the risk of developing so-called cyberchondria in some people. Moreover, 5-year mortality is much easier to predict than long-term morbidity or quality of life and its expected duration, whereas the latter two parameters are much more important both for people and for society as a whole.
Your risk of death over the next five years can be assessed on the Ubble website (those who do not know English can open it in the Chrome browser with a convenient machine translation function). But we remind you once again: the results obtained are not yours personally, but the average for the population. And keep in mind that all calculations are made for England, not for Russia.
The article by Andrea Ganna and Erik Ingelsson 5 year mortality predictors in 498 103 UK Biobank participants: a prospective population-based study is published in The Lancet.
Evgeniya Ryabtseva
Portal "Eternal youth" http://vechnayamolodost.ru based on the materials of Uppsala University:
New tool predicts your risk of death.
05.06.2015