June 4, 2015

Will you die in the next 5 years?

By Paul Laughlin

Feet on a morgue tableOnce in a while, a piece of public research captures the imagination, resulting in  positive interest and social media buzz.

Such is the case with the UK Longevity Explorer created by a Swedish team at Ubble.

Knowing how to the engage the press, the results of their research are also nicknamed the “death test“. Having previously shared on the dangers of public research, it is pleasing to have a positive & fascinating example to review.

Whilst not the first test into UK life expectancy, or likelihood of death in next 5 years, these researchers claim it is the most scientifically rigorous. It is encouraging to see that their approach echoes much of the best practice recommended by this site:

  1. Clear understanding of business/customer problem (Am I likely to die in the next 5 years & is there anything I can do about it?)
  2. Thorough secondary research to establish what is known already (they trawled through the data for almost half a million UK adults)
  3. Preparing quality data (they analysed Biobank data with 655 potentially predictive variables covering 948,000 people aged between 40-69). This data was collected between 2006-2010, so 5 years mortality data has be used to build, test & validate models since then.
  4. Analysing data & building predictive models (the factors with the highest correlations were reduced down to 13 for men & 11 for women). This model was also sense checked by considering which variables are actually proxies for known factors that are harder to capture via an online survey.
  5. Development of an easy to use online presentation tool for people to engage with the model (UK men & women between 40-70 can just answer those few questions online and be presented with both a percentage risk of dying in the next 5 years and an Ubble age).
  6. Driving action – the intention of the results presented is to help people be more self aware as to factors they can control. The construct of “Ubble Age” lets you see if yours is higher than your actual age & thus benchmarks you against the population. Similar to the idea of Wii age on the Ninentendo games platform, this will hopefully drive competitive improvement to achieve a better result.

How fast is your walking pace? Men aged 40-52 reporting a slower walking pace have 3.7 times greater risk of dying in the next five years than those reporting an average walking pace.

How fast is your walking pace? Men aged 40-52 who walk slowly are more likely to die in next 5 years. Click To Tweet

Much to praise and good to see such robust research & analysis raising the public awareness of good practice & how interesting insights can be.

You can find out more about the test via coverage in The Guardian:



You can also take the test yourself here, using Risk Calculator (or explore a data visualisation of predictive factors and mortality causes for your gender):

Please do also share your thoughts on this work & the coverage its receiving in the media.