Mortality in Spain in the Context of the Economic Crisis and Austerity Policies

Abstract

Objectives. To analyze the mortality trend in Spain before, during, and after the economic crisis and austerity policies.

Methods. We calculated age-standardized annual mortality rates in 2001 through 2016 and estimated linear trends in mortality rates during 4 periods (2001–2007, 2008–2010, 2011–2013, and 2014–2016) using the annual percentage change (APC).

Results. All-cause mortality rate decreased during the period 2001 to 2016, although we found increases over the previous year. After adjusting for increased influenza activity (P = .743) and heat waves (P = .473), we found the greatest declines during the economic crisis (2008–2010) and the smallest in the period 2014 to 2016, in which the APC in mortality rates was −2.9 and −0.6, respectively. The APC before the crisis (2001–2007) and during austerity (2011–2013) was −2.0 and −2.2, respectively. We observed similar results in mortality from cardiovascular, respiratory, and digestive diseases and motor vehicle accidents. Mortality from cancer showed the smallest decline during the crisis and the austerity period, whereas suicide increased in the period 2011 to 2013.

Conclusions. Lifestyle changes could explain the decline in mortality during the economic crisis. Increased influenza activity and the 2015 heat wave may prevent identifying a possible delayed effect of austerity policies in the slowing down of mortality decline in the period 2014 to 2016.

A recent article in AJPH claimed to show that the downward trend in mortality in Spain observed in the first decade of the 21st century changed abruptly in 2011, resulting in a 29% higher mortality than expected on the basis of the previous trend.¹ Likewise, in 2015, the observed mortality was 41% higher than expected. According to the authors’ calculations, between 2011 and 2015 there were 500 000 more deaths than expected in Spain on the basis of the previous trend. This figure represents one fourth of the 2 million deaths recorded in those 5 years. The authors attributed this excess mortality to austerity policies implemented by the Spanish national and regional governments during the second half of 2010 to face an economic crisis that began in 2008.

Such an increase in mortality is necessarily a methodological error. Spanish and international civil society would have noted such a catastrophe at the time, since in the last 100 years there have been only 3 similar mortality increases in Spain: in the 1918 flu epidemic, during the Spanish Civil War (1936–1939), and in the first postwar years because of the subsistence crisis.

In any case, previous empirical evidence refutes these findings. Such a mortality increase contradicts various statistical analyses and investigations showing that mortality in 2011 continued to decline.^2,3 The findings also contradict estimates on life expectancy in Spain conducted by international organizations such as Eurostat and the Organisation for Economic Co-Operation and Development, which show an increase in 2011 compared with 2010.^4,5 Likewise, the mortality indicators developed by the Ministry of Health and the World Health Organization (WHO) reflect a clear downward trend in mortality in Spain since the beginning of this century, although in some years mortality increased with respect to the previous year.^6,7 Increased influenza virus activity during the winter or heat waves in summer were probably responsible for these increases.^7–11

The authors of the aforementioned study¹ concluded that the austerity policies implemented in Spain beginning in 2010 exacerbated the consequences of the 2008 economic crisis, as evidenced by the important increase in mortality observed between 2011 and 2015. In the present study, we have used the same data employed by those authors to show the evolution of mortality rates during the 21st century in Spain and to assess the extent to which this trend changed during and after the economic crisis and austerity policies.

METHODS

We obtained the data on deaths by 5-year age groups and cause of death between 2001 and 2016 (the last year with available data) from the mortality register developed by the National Statistics Institute. The institute was also our source of data on the resident population as of July 1 for each calendar year and by 5-year age groups.

We first calculated the age-adjusted annual mortality rates per 100 000 inhabitants for all causes. Given that the influenza virus and heat waves affect older persons in particular, we also calculated the age-adjusted mortality rate in people younger than 65 years and in those aged 65 years or older. We then calculated the age-adjusted annual mortality rates by cause of death. The causes of death in the mortality register are those reported on the medical death certificate and are coded using the International Classification of Diseases, 10th Revision (Geneva, Switzerland: World Health Organization; 1992). We analyzed the following specific causes of death: cardiovascular diseases, respiratory diseases, digestive diseases, cancer, motor vehicle accidents, and suicide. Weights for age standardization came from the 2013 European Standard Population.

Spain registered a continuous decrease in gross domestic product (GDP) during the last 6 months of 2008, causing the country to enter an economic recession for the first time in 15 years.¹² Spain began recovering from the recession in 2014 with a 1.4% increase in GDP. During the last years of recession, the Spanish population was also exposed to austerity measures—reduced public expenditure and increased taxes—implemented by the government. These measures were put into effect from the second 6 months of 2010 until 2014, although the most profound measures ended in 2013.¹² Therefore, we estimated the annual percentage change (APC) in the death rate for the periods 2001 to 2007, 2008 to 2010, 2011 to 2013, and 2014 to 2016. We selected these periods on the basis of the years the economic crisis began (2008) and ended (2013), and the date that austerity measures were implemented (2010).

For this purpose, we estimated segmented Poisson regression models.¹³ The number of deaths was the outcome variable, and we included person-years as an offset variable. Each model included 4 independent variables: time and 3 interaction terms with time. We defined time as a continuous variable from year 2001 to 2016. If Β₁, Β₂, Β₃, and Β₄ are coefficients of the models for time and for each of the interaction terms, Β₁ reflects the trend in the period 2001 to 2007, Β₁+Β₂ reflects the 2008–2010 trend, Β₁+ Β₂+Β₃ the 2011–2013 trend, and Β₁+Β₂+Β₃+Β₄ the 2014–2016 trend. In a first model we included, as an adjustment variable, the age of individuals. In a second model—for deaths from all causes, from cardiovascular, respiratory, and digestive diseases and from cancer—we added 2 variables because of their probable confounding effect: the presence or absence of increased influenza activity and of heat waves. We assigned the presence of influenza to those years in which the Epidemiological Surveillance Area of the Carlos III Health Institute detected increased influenza activity in Spain or a greater than expected number of deaths in the weeks of greatest flu activity¹⁴; these years were 2002, 2005, 2007, 2012, and 2015. Likewise, we assigned the presence of heat waves to the years 2003 and 2015 because, according to the State Meteorology Agency, these were the years with the most important heat waves since the beginning of the century.¹⁵

The APC is 100 × [exp (Β1) −1] for the time period 2001 to 2007, 100 × [exp (Β1 + Β2) −1] for 2008 to 2010, 100 × [exp (Β1 + Β2 + Β3) −1] for 2010 to 2013, and 100 × [exp (Β1 + Β2 + Β3 + Β4) −1] for 2014 to 2016. To simplify the presentation of results, we did not calculate the APC difference of each time period with respect to the APC of other periods. However, such differences can be estimated from the results presented. The estimates of the model parameters are shown in Table A (available as a supplement to the online version of this article at http://www.ajph.org), and the information on the goodness-of-fit of models compared with hierarchical models with fewer parameters is shown in Table B (also available online).

RESULTS

The all-cause annual mortality rate for the whole population between 2001 and 2016 is shown in Table 1 and Figure 1. Overall, we observed a decrease in all-cause mortality rate throughout the period, although it increased with respect to the previous year in 2002, 2003, 2005, 2007, 2012, and 2015. We also observed these fluctuations in the mortality rate for individuals aged 65 years and older, whereas among those younger than 65 years the all-cause mortality rate declined each year with respect to the previous one, although the magnitude in 2014 and 2015 was quite similar (Table 1).

TABLE 1—

Age-Adjusted Mortality Rate per 100 000 Population From All Causes: Spain, 2001–2016

Year	All-Cause Mortality per 100 000 Population
	All Ages	< 65 Years	≥ 65 Years
2001	1091.9	219.9	4691.7
2002	1092.2	215.9	4709.7
2003	1110.6	214.4	4810.7
2004	1046.1	204.7	4519.6
2005	1064.0	200.9	4627.1
2006	983.2	194.4	4239.6
2007	986.1	192.3	4263.4
2008	956.5	186.6	4134.7
2009	925.3	178.7	4007.6
2010	889.8	173.3	3847.5
2011	875.0	170.4	3783.7
2012	887.2	166.6	3862.2
2013	837.8	161.5	3629.5
2014	829.0	158.4	3597.2
2015	864.1	158.1	3778.6
2016	821.4	154.7	3573.6

FIGURE 1—

Age-Adjusted, All-Cause Mortality per 100 000 Population: Spain, 2001–2016

Note. Annual mortality rates for all of the analyzed causes of death are shown in Figure A (available as a supplement to the online version of this article at http://www.ajph.org).

The annual mortality rates for the causes of death analyzed from 2001 to 2016 are shown in Table 2 and Figure A (available as a supplement to the online version of this article at http://www.ajph.org). The mortality rate exhibited a decrease for all causes of death except suicide in the most recent calendar years. The mortality rate from respiratory diseases showed the same fluctuations as mortality from all causes. The mortality rates from cardiovascular and digestive diseases declined each year with respect to the previous one, except in 2003 and 2015 for cardiovascular diseases and 2002, 2003, and 2015 for digestive diseases; the mortality rate from cancer showed a continuous decline. The mortality rate from motor vehicle accidents showed a major reduction in the period analyzed, although it increased in 2014 and 2015. The magnitude of the mortality rate from suicide was similar in 2006 and 2016, although it experienced several fluctuations between those years.

TABLE 2—

Age-Adjusted Mortality Rate per 100 000 Population From Several Causes of Death: Spain, 2001–2016

Year	Mortality per 100 000 Population
	Cardiovascular Disease	Respiratory Disease	Digestive Disease	Cancer	Motor Vehicle Accidents	Suicide
2001	388.8	116.9	55.4	275.2	13.9	8.2
2002	383.5	126.3	55.6	269.7	13.1	8.5
2003	384.5	129.9	56.0	268.8	12.9	8.5
2004	357.1	113.0	54.2	264.8	11.4	8.4
2005	356.3	132.4	53.5	258.1	10.4	8.0
2006	325.4	106.6	51.1	255.4	9.4	7.5
2007	322.1	114.1	50.1	253.4	8.6	7.4
2008	306.9	110.4	48.2	249.0	6.7	7.7
2009	289.9	104.2	47.0	245.9	5.7	7.5
2010	277.3	93.5	44.5	244.7	5.2	6.8
2011	265.4	94.4	44.3	243.9	4.7	6.8
2012	266.4	103.0	43.7	242.9	4.3	7.5
2013	248.6	89.8	41.7	239.0	4.0	8.2
2014	241.6	90.0	40.8	232.8	4.1	8.3
2015	249.3	103.6	42.0	230.8	4.2	7.6
2016	234.5	91.2	40.4	230.1	4.1	7.5

Note. The cause-of-death codes are from the International Classification of Diseases, 10th Revision, as follows: cardiovascular disease, I00–I99; respiratory disease: J00–J99; digestive disease: K00–K93; cancer: C00–C97; motor vehicle accidents: V02–V04, V09.0+V09.2, V12–V14, V19.0–V19.2, V19.4–V19.6, V20–V79, V80.3–V80.5, V81.0–V81.1, V82.0–V82.1, V83–V86, V87.0–V87.8, V88.8–V88.9, V89.0–V89.2; suicide: X60–X84, Y87.0.

The results of the time-period analysis are shown in Table 3. After adjustment for the presence of increased activity from influenza and heat waves, it can be seen that the all-cause mortality rate in the overall population showed the greatest decline in the period 2008 to 2010 and the smallest decline in 2014 to 2016. The decline in the second part of the economic crisis and during the austerity policies (2011–2013) was slightly greater than before the crisis (2001–2007). The APCs in the periods 2001 to 2007, 2008 to 2010, 2011 to 2013, and 2014 to 2016 were −2.0, −2.9, −2.2, and −0.6, respectively. We observed similar findings in the time-period analysis of all-cause mortality rates for people younger and older than 65 years.

TABLE 3—

Time Trends in Mortality Rate, From All Causes and From Several Specific Causes of Death, as Annual Percentage Change (APC): Spain 2001–2016

Time Period	All Causes, APC			Individual Causes, All Ages, APC
	All Ages	< 65 y	≥ 65 y	Cardiovascular Disease	Respiratory Disease	Digestive Disease	Cancer	Motor Vehicle Accidents	Suicide
Model 1a
2001–2007	−2.1	−2.6	−2.0	−3.5	−1.4	−2.0	−1.5	−8.2	−1.8
2008–2010	−3.4	−3.4	−3.3	−5.0	−5.7	−3.9	−1.0	−17.3	−3.9
2011–2013	−1.8	−2.5	−1.6	−3.2	−1.0	−2.3	−0.9	−5.9	6.0
2014–2016	−0.3	−1.2	−0.1	−1.5	0.8	−0.7	−1.2	0.6	−2.7
Model 2b
2001–2007	−2.0	−2.5	−2.0	−3.4	−1.5	−2.0	−1.5	. . .	. . .
2008–2010	−2.9	−3.3	−2.9	−4.6	−4.0	−3.7	−1.0	. . .	. . .
2011–2013	−2.2	−2.5	−2.1	−3.6	−2.6	−2.5	−0.9	. . .	. . .
2014–2016	−0.6	−1.4	−0.4	−1.8	0.2	−0.9	−1.2	. . .	. . .

^aAPC adjusted for age.

^bAPC adjusted for age, increased influence activity, and heat waves.

The time-period analysis of mortality rates for cardiovascular, respiratory, and digestive diseases showed results similar to those observed for all-cause mortality, although the APC in the period 2014 to 2016 was positive for respiratory diseases. After adjustment for the presence of increased activity from influenza and heat waves, the APCs in the periods 2001 to 2007, 2008 to 2010, 2011 to 2013, and 2014 to 2016 were, respectively, −3.4, −4.6, −3.6, and −1.8 in cardiovascular mortality, −1.5, −4.0, −2.6, and 0.2 in respiratory mortality, and −2.0, −3.7, −2.5, and −0.9 in digestive mortality. The APC for cancer was highest in the period 2001 to 2007 and lowest during the crisis periods. The decreases in motor vehicle fatalities were intense in the first 3 time periods, especially in the first part of the economic crisis (2008–2010), with an APC of −17.3, whereas in 2014 to 2016 the APC was positive (0.6). The greatest decrease in the suicide mortality rate was in the first part of the economic crisis, with an APC of −3.9, whereas it was positive in 2011 to 2013 (6.0).

DISCUSSION

After adjustment for the presence of increased influenza activity and heat waves, the downward trend in all-cause mortality in Spain showed the greatest annual decline in the period 2008 to 2010, the first part of the economic crisis, and the smallest in 2014 to 2016, after the crisis. The APC in the mortality rate in the second part of the economic crisis, during the implementation of austerity measures (2011–2013), was similar to the APC before the crisis (2001–2007). The trend in mortality from cardiovascular, respiratory, and digestive diseases was similar to the trend in all-cause mortality. By contrast, the upward trend in mortality from cancer showed the smallest decline in the 2 periods of the economic crisis. With regard to external causes of death, the important reduction in mortality from motor vehicle accidents, especially in the first part of the economic crisis, was inverted in the period 2014 to 2016, whereas suicide mortality, which showed the greatest decline in the first part of the economic crisis, showed an important rise in the second part of the economic crisis, during the implementation of the austerity measures.

Comparison With Other Studies and Possible Explanations

Our findings confirm that the results of the study mentioned in the introduction are wrong.¹ All-cause mortality in Spain did not increase beginning in 2011, but rather continued to decline, although the magnitude of the decline in the period 2014 to 2016 was smaller than in the previous time periods. The authors of that work used different weightings to calculate the age-adjusted annual mortality rate before 2011 and the subsequent period, which strongly distorted the between-period comparison, as noted in an editorial comment accompanying the article.¹⁶

Several previous studies have shown that during economic crises, short-term cancer mortality does not vary, whereas mortality from cardiovascular, respiratory, and digestive diseases and from motor vehicle accidents decreases and suicide mortality increases.^17–24 Our time-period analysis reveals similar findings in Spain, except for cancer and suicide. Given the long induction period required for development of cardiovascular, respiratory, and digestive diseases, a more intense decrease in mortality during the economic crisis compared with other periods could be more the result of improvements in prognosis than of decreases in incidence. A hypothetical improvement in treatment is implausible. The adoption of healthy lifestyles—decreased smoking and alcohol consumption^25,26—likely contributed to the important reduction in mortality from these diseases. It is known that smoking is a risk factor for mortality in patients with cardiovascular or respiratory diseases or other chronic diseases.^27–29 Alcohol consumption has also been associated with decreased survival in patients with chronic diseases.^29–30 In Spain, per capita cigarette and alcohol consumption decreased more during the economic crisis than in the previous period.^25,26 The decline in commercial and recreational transportation contributed to reduced mortality from motor vehicle accidents, and to lower mortality from respiratory diseases due to reduced air pollution. Furthermore, it has been observed that in oncology patients, concurrent cardiovascular and other chronic diseases affect the mortality risk from cancer.³¹ The smaller decline in cancer mortality during the economic crisis could be a result of cancer deaths during that time among individuals who, in the absence of the crisis, would have died from concurrent cardiovascular or respiratory diseases.

Because of its low magnitude, suicide mortality did not contribute to an all-cause mortality trend, but the findings observed were unexpected. In Spain, there was an increase in suicides in 2008 with respect to 2007—especially among men, as some authors have reported.³² This finding is consistent with different studies that show increased short-term mortality from suicide with economic crises.^19,22–24 In 2010, however, suicide mortality decreased greatly. The definitive implementation in that year of a new medical death certificate, which included changes in the reporting procedure of violent deaths with respect to the previous medical certificate, could have contributed to this decrease. Beginning in 2012, there was a significant increase in the number of suicides. This increase was probably due to a more precise assignment of the cause of death in violent deaths with judicial intervention—that is, it was possible to assign specific causes of death to deaths previously coded as poorly defined causes. This would explain the extraordinary increase observed in the period 2011 to 2013, as other authors have also noted.³³ In fact, the National Statistics Institute warns of the difficulties of comparing mortality rates from suicides in those years.³⁴ Beginning in 2014, suicide mortality again decreased intensely, perhaps because of an attenuation of the administrative practice aimed at recovering poorly defined causes of death. This poor classification of deaths from suicide explains the heterogeneous results in studies of the economic crisis and suicide in Spain, since the findings depend on the calendar years analyzed in each study.^2,32,33,35

A previous study showed little variation in the mortality rates for older people in Spain during the economic crisis compared with the decrease in the previous period.³⁶ The authors attributed this finding to the effect of the austerity measures. Our findings show that during the implementation of austerity measures, in the period 2011 to 2013, the decline in all-cause mortality slowed, although when we adjusted for increased influenza activity and heat waves, the decrease was similar to that observed before the crisis. The same thing occurred with mortality from cardiovascular, respiratory, and digestive diseases. By contrast, cancer mortality declined very little in the period 2011 to 2013. The concurrent cardiovascular and other chronic diseases in oncology patients previously mentioned is the most reasonable explanation, because there is no reason to think that the austerity measures would have specifically affected the care of cancer patients and not others. In addition, the low decrease in cancer mortality was also observed during the economic crisis of 2008 to 2010, a period in which the health care expenditure in Spain grew by 5.1%, according to Eurostat.

Our findings differ from those of a previous study on the effects on mortality of austerity policies implemented since 2010 in England.³⁷ The authors of that study attributed an immediate effect of the austerity policies on mortality in the English population, since they found a stabilization of the mortality trend in the period 2011 to 2014 with respect to the decline observed in the previous period. It is possible that the effect of the austerity measures in Spain was not immediate, as some authors have suggested.³⁸ However, a large part of this slowing down of the mortality decline in the period 2014 to 2016 was due to the high mortality in 2015. That year, there was an increase in winter deaths in Spain as well as many European and non-European countries, probable due to influenza virus, which affected not only the elderly but also younger persons.¹⁰ In fact, mortality in Spain among people younger than 65 years varied little between 2014 and 2015. Furthermore, mortality—especially from respiratory diseases—may have increased in 2015 because of a summer heat wave. The summer of 2015 was the second hottest in several decades in Spain and other places worldwide.³⁹ Some authors noted that in the United Kingdom, the percentage increase in deaths in 2015 with respect to 2014 had not been seen in the last 50 years.⁴⁰ Something similar happened in other high-income countries.⁴¹ In fact, in carrying out a similar analysis using the information available in several countries, we observed that there was a slowdown in the decline in mortality in the period 2014 to 2016, both in countries with important austerity policies, such as Greece and Croatia, and in countries where austerity policies were less drastic, like Germany (Table C, available as a supplement to the online version of this article at http://www.ajph.org). This may explain why adjusting for the presence of increased influenza activity in winter and heat waves in summer did not completely attenuate the increased mortality of 2015 in the 2014 to 2016 trend in Spain.

On the other hand, several studies have found that all-cause mortality would tend to decline faster in economic recessions than in expansions.^17,18,21 Mortality from motor vehicle accidents increased in the period 2014 to 2016 because of the economic recovery. Their contribution to the trend in all-cause mortality was marginal, though, since such deaths represent only 0.5% of the total. However, changes in lifestyle—cigarette and alcohol consumption stabilized beginning in 2013 after the previous decline^25,42—and their impact on cardiovascular, respiratory, and digestive mortality, could have contributed to the deceleration of the all-cause mortality trend.

Aside from the austerity measures, in 2012 the Spanish government enacted a regulation restricting the access of undocumented immigrants to the health system.^16,38 However, there is no evidence that this restriction was actually carried out—probably because the regional governments did not apply this national measure and because the measure did not restrict the access of undocumented immigrants to emergency services.³⁸ In any case, the possible impact of this measure on the trend in mortality in the period 2014 to 2016 would have been small, given that such immigrants represent a very small percentage of the population. Furthermore, most of them are young adults, who have low mortality.

Strengths and Limitations

Unlike the study referred to in the introduction,¹ our study’s use of the same standard population made it possible to observe the fluctuations in the annual mortality rate between 2001 and 2016. Furthermore, adjustment for the factors responsible for these fluctuations allowed us to estimate the trends in all-cause mortality and mortality from various causes of death in different time periods, including those of the economic crisis and austerity policies. The adjustment has modified the trend in mortality from cardiovascular diseases and, above all, respiratory diseases, since they are the causes of death where the fluctuations are more evident.

The goodness-of-fit indicators of the Poisson models (Pearson χ² and deviance statistics) reflect overdispersion, which is common for count data. Therefore, we also estimated negative binomial regression models. These models reflect a good fit of the data, and the Β parameters of the different time periods do not contradict the findings we obtained with the Poisson models.

The inclusion of age as an adjustment variable would be inadequate if, in each time period, the linear trend was ascending in some age groups but descending in others. Inspection of the evolution of mortality risk at different ages in the time periods analyzed suggested that this was not the case.⁴³

We did not assess the possible impact of the immigrant population on the findings. Among those younger than 65 years, the immigrant population represented 7% of the population residing in Spain in 2001. This percentage increased to 15% in 2008 and then fell to 14% in 2016.⁴⁴ Given that mortality for most immigrants is lower than for the native population,^45,46 the inclusion of immigrant status as an adjustment variable would have increased the downward trend in mortality in this age group after 2008. This potential impact does not affect the trend in mortality of the population older than 65 years, since the proportion of immigrants among the residents of this age group remained around 3% between 2001 and 2016.⁴⁴ Consequently, the impact on the mortality of the population as a whole is of little importance given that the majority of deaths occurred among people older than 65 years.

We cannot know what the mortality trend would have been in the period 2014 to 2016 if there had not been such high mortality in 2015 and if the economic recovery had not occurred. Thus, we cannot rule out a possible delayed effect of the austerity measures on mortality in this period. Nor would having definitive data on deaths in 2017 help to clarify this possible effect, given the intense heat wave—the third most severe in Spain in recent decades—and the elevated influenza activity in that year.^47,48

Conclusions

Mortality in Spain since the beginning of the present century has shown a downward trend, which was greater during the economic crisis. During the implementation of the austerity measures the decline in mortality slowed, although when we adjusted for increased influenza activity and heat waves, the decrease in the period 2011 to 2013 was similar to that observed before the economic crisis. The high mortality in 2015, probably caused by increased influenza activity and the summer heat wave that year, may prevent identifying a possible delayed effect of austerity policies in the slowing down of mortality decline in the period 2014 to 2016.