Monitoring sub-national longevity in a small country: comparative analysis of three approaches

Life expectancy is the key indicator describing social and health aspects of human development (UNDP, 2021). Despite numerous international and national health policy documents indicating that long and healthy life is a principal goal for socioeconomic change, the reality suggests about the persisting or even increasing longevity disparities between and within countries. Substantial longevity disparities between socioeconomic groups and areas exist even in countries with strong social policies and systems, including the Nordic countries (Mackenbach, 2017). International comparative studies indicate that there is no systematic evidence about narrowing mortality disparities (Mackenbach et al., 2019). Therefore, persisting disparities should be considered as important obstacles for future sustainable health progress at the national levels.
There are several approaches for measuring longevity and mortality disparities. The majority of the ongoing research focuses on mortality disparities by socio-economic status. Nevertheless, the spatial dimensions of mortality changes are considered as equally important because they are principal components of the health development at the national level. The small-area analysis builds the bridge between these two approaches and connect spatial variation with socio-economic environment. Moreover, in the absence of register-based or census-linked individual-level mortality data by socioeconomic group (in many cases due to data protection regulations), the analysis of small small-area data is the only option to study disparity by socio-economic characteristics. The biggest challenges in applying this approach is estimating life tables for small territorial units. The UK Office of National Statistics (ONS) made important methodological contributions proposing how to apply and adjust traditional life table estimation methods for small populations. (ONS, 2003; Silcocs, 2004). Due to recently observed slow-downs or even reversals in life expectancy, there is an “explosion” of new studies based on mortality and life expectancy estimation for small areas in the USA and UK (Arias et al., 2018; Rashid et al., 2021).
Prior research indicates that estimations of life tables for small areas might be based on standard demographic methods, statistical modelling, or combination of these two approaches. The main advantage of standard methods is related to the maximal reflection of reality, i.e. relying on official statistical data. The main disadvantage of this approach for producing small area
estimates is related to the sensitivity of the results to small numbers leading to large fluctuations and wide confidence limits. It has been shown that it is almost impossible to fully apply this approach in case of multiple zeros in death counts and/or population exposures. Therefore, the standard computation of life tables is usually complemented by some (often arbitrary) adjustments, smoothing or modelling parts of mortality curves. During the last decades the field of small area life table estimations has been increasingly relying on advanced statistical modelling approaches, including Bayes modelling. These approaches allow to obtain age-specific mortality estimates by using estimated parameters from a standard mortality schedule (e.g. national or higher rank regional unit) or by borrowing information for mortality estimation from either neighboring or similar (e.g. by socioeconomic characteristics) areas or from areas with better quality data. As a result, one gets much more stable and statistically robust results, allowing to assess the magnitude and directions of changes of inequalities. The limitations of modelling include arbitrary choice of parameters, risk of overlooking specifics of mortality patterns in some areas. This paper is focusing on the performance of a modified real data and TOPALS model approaches and their capabilities to produce consistent life expectancy estimates at various levels of subnational division in Lithuania.