Review outlines methods to estimate life expectancy

Life expectancy measures, particularly years of life lost or gained, have captured a great interest in the last years by public health institutions and healthcare professionals given their immediate and actionable interpretation.

Academics from the University of Leicester investigated the common methods used to predict the ‘years of life lost’ by reviewing a range of methods: from basic methods – such as life tables – to most recent and advanced methods using statistical modelling. The research was supported by the ARC East Midlands.

They highlighted the important differences between the methods, current software used, and how they can be implemented in healthcare research and real-world settings using an example from a clinically-relevant topic of multimorbidity in the UK Biobank. 

According to the findings, the review found that using the same data and research question, the estimated years of life lost value differed among methods, as each method focused on estimating different quantities. 

Lead researcher Dr Yogini Chudasama, an Epidemiologist at the Leicester Real World Evidence Unit, said: “This project was initially from my PhD, where at the time there was no simple and comprehensive paper explaining or comparing the methods for the years of life lost calculation, or showing which method I should use for my data. This was a challenging task and required a lot of patience and learning to overcome the difficulties encountered. Collaborating with co-authors and supervisor Dr Francesco Zaccardi, I was able to explore the methods in detail and understand how they are applied to real-world data.”

“I am very proud to have this published in the Journal of Clinical Epidemiology, as a tutorial guide for researchers and healthcare professionals, as it will help promote a better and easier understanding of life expectancy metrics in healthcare research.” 

As a result, the researchers also developed a general life expectancy analytical programme, which has been applied to the following three NIHR ARC-funded published studies:

https://journals.plos.org/plosmedicine/article?id=10.1371/journal.pmed.1003332

https://onlinelibrary.wiley.com/doi/full/10.1111/joim.12987

https://bmcmedicine.biomedcentral.com/articles/10.1186/s12916-019-1339-0“

Dr Francesco Zaccardi, Clinical Epidemiologist at the University of Leicester, main supervisor for the study, said: “Findings from epidemiological studies are commonly reported in terms of relative risk – for example, type 2 diabetes increases the risk of death by two-fold. This statement is difficult to interpret because it does not convey information on the absolute risk of death.

“Years of life lost, conversely, are easily interpretable: in a middle-aged person, type 2 diabetes reduces the life expectancy by five years. As different methods exist to estimate life expectancy, in our study we highlighted that decisions should be based on the purpose of the research, the type of available data, and the required flexibility when using complex statistical modelling.”

Professor Kamlesh Khunti CBE, Director of NIHR ARC East Midlands and the Real World Evidence Unit and Professor of Primary Care Diabetes and Vascular Medicine at the University of Leicester, added: “This review is important as years of lost life metrics are simple summary measures that can enhance the interpretation of the findings from epidemiological studies for healthcare professionals and the public.”

To access the entire research findings, click here.