The impact of an unemployment insurance reform on the incidence rates of hospitalizations for alcohol-related disorders: a quasi-experimental study of the heterogeneous effects according to ethnic origin, level of education, employment status and gender in Sweden | BMC Public Health

0

Descriptions

Descriptive characteristics of the study population are presented in Table 1. The estimates show higher incidence rates of hospitalizations due to alcohol-related disorders in men compared to women, relatively high numbers in the foreign-born European population, but low rates among non-European foreign-born. Large differences in incidence rates by educational level and employment status are shown. Among people who were unemployed but did not receive unemployment benefits, the incidence rates are more than twice as high as those of the unemployed who received such benefits. The unemployed and the less educated also show a notable increase over the two specified periods (2001-2006 and 2007-2012).

Table 1 Descriptive characteristics of the study population, 2001-2012

Supplementary files 2, 3 and 4 show the percentage of unemployed with and without unemployment benefit according to ethnic origin, level of education and sex. It can be noted that the percentage of unemployed was generally stable or slightly decreasing in most groups before the implementation of the reform. A more obvious decrease is then observed in 2007-2008, followed by an increase from 2009. Regarding the results stratified by ethnicity (Supplementary File 2; Supplementary Figure S2), the orange lines reveal that the share of unemployed born in Sweden with Swedish-born parents who received benefits decreased between the years 2005 and 2009 (two years before and after the implementation of the reform). This trend is consistent across all subgroups. The proportion of unemployed without unemployment benefits is broadly stable for both groups of Swedish-born people, with some tendency to increase over time. For Europeans born abroad and non-Europeans born abroad, there is a clear increase from around 2008, especially in the latter group. The blue bars illustrate the gap between beneficiaries and non-beneficiaries compared to the overall average among the unemployed. The mirror-reversed chronological pattern with a breakpoint in 2007 in the group of foreign-born non-Europeans is particularly evident. Graphs that stratify by education (Supplementary File 3; Supplementary Figure S3) and gender (Supplementary File 4; Supplementary Figure S4) show broadly similar patterns. Among the less educated, the increase in non-recipients after 2007 is somewhat more pronounced than among the highly educated.

Results of regression discontinuity analyzes

Figures 1, 2, 3, 4 and 5 show RD plots with incidence rates of hospitalization due to alcohol-related disorders over the period 2001-2012 in the total study population, as well as by ethnicity, level of education, employment status, and gender. Again, it should be noted that the bins represent the local averages of group-specific hospitalization incidence rates due to alcohol-related disorders at each month during the study period. Point estimates, p-values ​​and confidence intervals are available in Supplementary File 5 (Supplementary Table S1).

Regarding the results for the total population studied (Fig. 1), the results show a slight reduction in the incidence rates around the time when the reform was implemented (-0.49, p

Fig. 1

Regression discontinuity plots with alcohol use disorder incidence rates. Total study population (30 to 60 years old, 2001 to 2012)

Then the results are stratified by ethnicity (Fig. 2). People born in Sweden whose parents were born in Sweden show a slight decrease in incidence rates (-0.42, pppp

Figure 2
Figure 2

Regression discontinuity plots with alcohol use disorder incidence rates. Total study population, stratified by ethnicity (30–60 years, 2001–2012)

Stratification by level of education (Fig. 3) reveals substantial differences in absolute levels between low and high educated individuals. However, both groups show a slight decrease at the time of the reform (low level of education: -0.47, pp

Figure 3
picture 3

Regression discontinuity plots with alcohol use disorder incidence rates. Total study population, stratified by level of education (30-60 years, 2001-2012)

For results stratified by employment status (Fig. 4), the slight decrease in incidence rates for employed individuals (-0.56, pp

Figure 4
number 4

Regression discontinuity plots with alcohol use disorder incidence rates. Total study population, stratified by employment status (30-60 years, 2001-2012)

Large absolute differences in incidence rates can be seen between males and females (Fig. 5). The relatively high incidence rates in men follow the trend observed for the general population. For both groups, however, there are decreases in the incidence rate of hospitalizations for alcohol-related disorders (men: -0.77, pp

Figure 5
number 5

Regression discontinuity plots with alcohol use disorder incidence rates. Total study population, stratified by gender (30-60 years, 2001-2012)

The next set of figures (Figures 6, 7 and 8) show results specifically for the unemployed population, stratified by ethnicity, education level and gender. As mentioned earlier, colored bins were used to illustrate changes in the absolute difference between unemployed people who received unemployment benefits and unemployed people who did not receive such benefits during the study period.

Figure 6
number 6

Regression discontinuity plots with alcohol use disorder incidence rates. Unemployed population, stratified by ethnic origin (30 to 60 years, 2001 to 2012)

Focusing on the ethnic origin of the unemployed (Fig. 6), people born in Sweden to parents born in Sweden show increasing trends after January 2007, when the unemployment insurance reform was implementation (2.25, pppp

In terms of education (Fig. 7), incidence rates are comparatively higher in people with little education. For this group, there is also a notable increase at the time of the reform (2.50, pppp

Picture 7
number 7

Regression discontinuity plots with alcohol use disorder incidence rates. Unemployed population, stratified by level of education (30-60 years, 2001-2012)

Figure 8
figure 8

Regression discontinuity plots with alcohol use disorder incidence rates. Unemployed population, stratified by sex (30-60 years, 2001-2012)

Robustness checks

The first set of sensitivity analyzes (see Supplementary Files 6, 7 and 8; Supplementary Figures S5-S7) examines the robustness of the selected threshold. In general, the January 2004 and January 2010 alternative cut-off points do not show larger increases in alcohol-related disorders compared to January 2007, when Unemployment Insurance reform was implemented. This is particularly clear among some of the groups in the unemployed population that previously had the largest effects, including people born in Sweden to Swedish-born parents, foreign-born Europeans, people with low education, and men. . Since the differences between the cut-offs are less pronounced in the Swedish-born to foreign-born parents and in women, the effects previously identified in these groups should be interpreted with caution.

The next set of robustness checks (see Supplementary Files 9, 10 and 11, Supplementary Figures S8-S10) examines the effect of reform using a narrower bandwidth (2005-2008). Whether the regression lines are based on quartic or linear polynomials, the results largely reflect analyzes using the entire tracking period (2001-2012).

Share.

About Author

Comments are closed.