Socio-ecological perspective of older age life expectancy: income, gender inequality, and financial crisis in Europe

The proposed conceptual framework depicting the effect of the socio-ecological indicators on OLE is shown in Additional file 1. OLE can be affected by socio-ecological indicators. The term ‘OLE’ refers to being physically active with preservation of functional capacity and socioeconomic wellbeing as exemplified through an active life in society without diseases [3, 19, 21, 22, 26]. Consequently, OLE inequality may be affected or controlled by socio-ecological environment and hereditary factors, although the latter was excluded from this study. Thus, macro-level socio-ecological inequality factors were focused on in this study [3, 19‐21, 33].

Health promotion is determined by influences at multiple levels, including personal, community, and public policy factors from socio-ecological perspective [34]. Socio-ecological perspective in OLE may identify influences of personal, social environmental factors, and public policy on older adult’s health and individual behaviour. These factors might be human aggregate or characteristics of people and/or characteristics of the surrounding community and country [35]. Thus, this article proposed a socio-ecological inequality model for OLE focusing on the following: (1) personal perspective, (2) social environment such as human relations, and (3) public policy perspective as targets for health promotion [34, 36].

To implement socio-ecological changes, the model assumed that changes in national income per capita from personal perspective, changes in gender inequality of human relations from social environment perspective, and changes in the depth of credit information caused by financial crisis from public policy perspective would produce appropriate changes in country-level strategies of successful aging (Additional file 1). Thus, this study determined the association of elderly life expectancy (OLE) with gross national level income per capita (GNI) at personal level, gender inequality (GII) at socio-cultural level, and depth of credit information (DCI) on national finance in public policy at socio-ecological level to identify successful strategy for an aging society.

This study was a socio-ecological research that designed model affecting OLE according to personal level, social environment, and public policy. In other words, personal level reflects economic condition while social environment and public policy reflect gender discrimination level and national financial status, respectively. Personal level was set as gross national income per capita (GNI) because it had the most significant impact on medical expenditure of the elderly. Social environment was set as gender inequality (GII) considering that inequality between men and women could be an important stress in elderly of a modern society at socio-cultural level. Public policy was set as depth of credit information (DCI) in national finance condition because it might affect the health and welfare of the elderly. These selected indicators could be used in a model to predict factors that affect life expectancy of the elderly in each country in the long term. In other words, with a national socio-ecological analysis model, causal association and influences of factors on OLE could be predicted. Based on these results, strategies could be suggested to promote health policies and improve life expectancy of the elderly in an aging society (Additional file 1).

To examine associations between OLE levels and socio-ecological perspective indicators, models were developed to estimate income, gender, and financial crisis disparities in relation to each variable. Three socio-ecological models were developed considering the following: (1) GNI per capital from personal perspective, (2) GII from social environment perspective, and (3) DCI as financial crisis from public policy perspective. These models were used to depict the proposed framework for socio-ecological indicators on the basis of selected variables. Specifically, predictors of national income, gender inequality, and financial crisis disparities (i.e., GNI, GII, and DCI respectively) were used to create a combination model or a model comprising all three models. These variables were reflective of all selected socio-ecological inequality indicators. Thus, the relationship between OLE may differ according to national income, gender inequality, and financial crisis. From this model, it was hypothesized that increases in GII would result in corresponding decreases or increases in OLE. However, decreases in GNI and DCI would result in corresponding decreases or increase in OLE. Associations between these factors and OLE in three models were assessed using Pearson correlation coefficients and regression models [19, 21]. A regression analysis was run to determine whether each socio-ecological indicator was independently and significantly correlated with OLE. The following multivariate regression models were used to analyze there indicators: Model (1) for OLE (women & men), Y₁ = a + bx₁ + bx₂ + bx₃···· +ε, Model (2) for OLE (men), Y₂ = a + bx₁ + bx₂ + bx₃···· +ε, Model (3) for OLE (women): Y₃ = a + bx₁ + bx₂ + bx₃···· +ε, Where a was intercept term or constant, b was unknown parameter, ε was a random error term, x₁ was GNI, x₂ was GII, and x₃ was DCI. Hierarchal linear regression was used for final analysis.

Data for OLE analysis were obtained from life expectancy survey conducted by the World Health Organization [37]. DCI and GNI data used for this study were adopted from World Development Indicators of the World Bank [8, 38]. Data for GII indicators were obtained from United Nations [39] datasets. Necessary permissions were obtained to publish these data since no personal data were presented.

The following factors were used from socio-ecological perspective: (1) OLE (Older age life expectancy, years) for 2000 ~ 2012. It was defined as the average number of years a person expected to live in ‘full health’, including years lived with less than full health owing to diseases and/or injuries [37]; (2) GNI (Gross national income per capita) for 2005 ~ 2015. It was defined as PPP (current international $) for 2005 ~ 2012 converted to international dollars using purchasing power parity rates [40]; (3) GII (gender inequality index, value: 0 = women and men equally, to 1 = women poorly) for 2000–2010. It was a composite measure reflecting inequality in achievements between women and men in three dimensions: reproductive health, empowerment, and labour market [39]; and (4) DCI (0 = low, to 6 = high) for 2004–2012. A score of 1 was assigned for each of eight features of credit bureau or credit registry or both [5, 38].

OLE at age of 60 years is derived from life tables. It is based on sex and age-specific death rates [37]. The estimated number of deaths in a life table and population by age group are aggregated in a given region in order to compute regional life tables [37]. The WHO uses a standard method as described above to estimate and project life tables for all member states using comparable data. This may lead to minor differences compared to official life tables prepared by member states themselves [19, 21]. With regard to year, OLE in this retrospective study reflected total women and men. OLE of women or men from 2000 to 2012 [OLE = (OLE in 2000 + OLE in 2012) / 2] [19, 21] was also analyzed.

DCI measures rules that might affect the scope, accessibility, and quality of credit information available through public and private credit registries. DCI is an index that ranges from 0 to 8, with higher values indicating availability of more credit information [8]. The index shows how DCI (0 = low to 6 = high) varies by country. In 2012, countries with the highest DCI value in Europe were the United Kingdom, Germany, and Lithuania with an average value of 6.00. However, Malta reported the lowest value of 0.00 [8].

GII measures gender inequalities in three important aspects of human development [41]. In other words, GII reflects inequality in achievements between women and men in three dimensions: reproductive health, empowerment, and labor market [18, 41, 42]. Thus, GII reflects women’s disadvantage in three dimensions for as many countries as possible with data of reasonable quality. It ranges from 0 (women and men fare equally) to 1 (women fare as poorly as possible in all measured dimensions) [18, 41, 42].

GNI per capita is based on purchasing power parity (PPP). PPP GNI is a gross national income (GNI) converted to international dollars using PPP rates [40]. Data are in current international dollars based on 2011 ICP round [40]. Thus, GNI in this retrospective study reflected the GNI from 2005 to 2012.

To analyse socio-ecological inequalities, this study excluded countries with insufficient gender inequality or financial crisis information. Because OLE inequalities are based on cohorts with healthy lives, national income, gender inequality, and financial crisis indicators might have changed during the study period. Thus, examining the association between OLE and socio-ecological indicators would be appropriate. Time series data are available for GNI, GII, DCI, and OLE, allowing for more robust results. However, the present study has two comparability issues [3]: comparability between countries and comparability across time periods. These are accounted for in socio-ecological indicators. Because OLE level is a type of dynamic output indicator that represents morbidity improvement. It is closely related to country-level socio-ecological inequalities. Therefore, an equivalence indicator must be employed [3]. Thus, national income, gender inequality, and financial crisis levels were used as “mean rate of change index (MC)”. Because some countries experienced greater progress than others between 2000 and 2012, MC was used to supplement faults in such information. Specifically, MC considered both the present and changed value of indicators from 2000 to 2012. MC is the change in value of an indicator divided by time elapsed [3, 19, 43, 44]. Specifically, MC is the sum of the current value of socio-ecological indicator and its changed values from 2000 to 2012 divided by the number (N) of elapsed years: MC of national income indicators = [the value of indicators in 2005 + the value of indicators in 2012) / N]), MC of financial crisis indicators = [the value of indicators in 2004 + the value of indicators in 2012) / N], MC of gender inequality indicators = [the value of indicators in 2000 + the value indicators in 2010) / N] [3, 19, 21]. It was used as an estimate of time series data of socio-ecological inequality indicators. Indicators without sufficient time series data were excluded from this study.

Results of descriptive statistics for OLE ranges across countries and socio-ecological inequality indicators values are summarized in Table 1. The mean GNI was $26,938 with a between-country disparity of $50,060. Mean DCI generally ranged from 1.51 in Slovenia to 6 in the United Kingdom. Mean GII was 0.186 with a between-country disparity of 0.38. OLE (Women and men) ranged from 17 years in the Russian Federation to 24 years in Europe (Iceland, Switzerland, France, Italy, and Spain), with a mean of 21 and 7 years in disparity between countries, respectively.

Results of OLE, GII, and DCI indicators of socio-ecological in 34 European countries are shown in Table 2. Country-level OLE was correlated with GNI, GII, and DCI caused by financial crisis. Although OLE had positive correlations with GNI (r = 0.834, p = 0.001) and DCI (r = 0.704, p = 0.001), it showed significantly negative correlation with GII (r = −0.798, p = 0.001).

To investigate the direct relationship between OLE and GNI, GII, or DCI, a multiple regression analysis was conducted (Table 3). Regression analysis of socio-ecological indicators revealed the strongest predictors in hierarchal linear regression models (Table 4). Significant predictors of OLE in univariate analysis were used to build a model for multivariate analysis to predict OLE. Finally, higher GNI and DCI but lower GII were found to be predictors of OLE level (women and men) (R² = 0.804, p < 0.001).

Thus, the greater the country-level OLE value, the greater its effect on increasing GNI and DCI but decreasing GII in Europe. This result indicates that OLE has significant impact on corresponding country-level GNI, GII, and DCI. The country with higher OLE is expected to have higher GNI and DCI but lower GII.