1 Suicides in old and new EU Member States: 2007=1 1.35 1.3 1.25 - PDF document

The health effects of a financial crisis: epidemiology on a large scale Martin McKee Hong Kong, August 2016 Twitter: @martinmckee What happened? GDP per capita (2007=1) 1.10 1.05 1.00 0.95 Germany What happened to health? United States 0.90 United Kingdom Spain 0.85 Greece 0.80 0.75 0.70 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 1

Suicides in old and new EU Member States: 2007=1 1.35 1.3 1.25 1.2 1.15 “We estimate that the Great Recession is associated with at 1.1 least 10,000 additional economic suicides between 2008 and 1.05 2010.” 1 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 EU members before May 2004 EU members since May 2004 Reversal in long term decline in unmet need Deaths on the roads 30 25 20 deaths/100,000 15 10 5 0 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 Reeves A, McKee M, Stuckler D. The attack on Universal Health Coverage Hungary Lithuania Netherlands in Europe: recession, austerity, and unmet needs. Eur J Publ Health 2016 Suicides in England Increase in unmet medical need in Portugal, by economic status, 2010 to 2012 • Each 10% increase in the number of unemployed men was significantly associated with a 1.4% (0.5% to 2.3%) increase in male suicides. • About two fifths of the recent increase in suicides among men (increase of 329 suicides, 126 to 532) during the 2008-10 recession can be attributed to rising unemployment. Barr B, Taylor-Robinson D, Scott-Samuel A, McKee M, Stuckler D. Suicides associated with the 2008-2010 recession in the UK: a time-trend analysis. BMJ 2012: 345 doi: 10.1136/bmj.e5142 2

A question of priorities • Health of all Greeks worsened during austerity after 2010 • However, impact of job loss on chance of reporting worse health much greater during austerity Country Last year of mortality • Financial data data (June 2016) – Instantaneous Iceland 2009 Pre-austerity Austerity Ireland 2010 • Mortality data France 2011 – Within the decade Estonia 2011 Belgium 2012 Epidemiological issues Another source? • Data availability & accuracy • Understanding effect modifiers • Cyclical trends in search activity for suicide and depression- • Defining exposure – What makes things worse related terms, with peaks in – What makes things better – What is a financial crisis or a autumn and winter months, recession? • Natural experiments and a trough in summer months • Understanding causality in • Multi-disciplinarity • Significant association with complex situations • What is the treatment? suicides among: – Causal pathways – 25-24 age group – Time lags – 45-54 year old women (older women tend to overdose so search for details, men tend to hang themselves) Beware incentives to distort data Benford’s Law on the distribution of the first significant digit in a collection of numerical data 0.35 P(d) ≈ log 10 ([d + 1]/d) 0.3 Data availability & accuracy 0.25 P (d) 0.2 0.15 0.1 0.05 0 1 2 3 4 5 6 7 8 9 d 3

Benford’s Law applied to EU 27 and Greek NBER definition financial statistics (average 133 observations/ country/ year) “a significant decline in [the] economic activity spread across 0.35 the country, lasting more than a few months, normally visible 0.3 in real GDP growth, real personal income, employment, 0.25 P(d) industrial production, and wholesale-retail sales” P(d) EU27 0.2 P (d) P(d) Greece 0.15 0.1 0.05 0 1 2 3 4 5 6 7 8 9 d Source: Rauch, Göttsche, Brähler, & Engel, 2011 Defining and interpreting exposure The need for a multidisciplinary approach Who is publishing? Recession: a definition Different disciplines • Systematic search of • A rule of thumb.... Web of Science for • “a decline in the literature on recessions seasonally and calendar and health adjusted real gross • 461 articles & 14,401 domestic product (GDP) cited documents in at least two • Network analysis of co- successive quarters” citation pattern by disciplines, journals and backgrounds of the Julius Shiskin, NY Times 1974 authors Stuckler D, Reeves A, Karanikolos M, McKee M. The health effects of the global financial crisis: can we reconcile the differing views? A network analysis of literature across disciplines. Health Econ Pol Law 2015; 10: 83-89. 4

Few reading each other’s work Koch’s Postulates • The microorganism must be isolated from a diseased organism and grown in pure culture. Psychiatry Economics • The cultured microorganism should Public health cause disease when introduced into a healthy organism. • The microorganism must be reisolated from the inoculated, diseased experimental host and identified as being identical to the original specific causative agent. Medicine Addiction Cause and effect Bradford Hill’s criteria of causality • Strength of association • Consistency • Specificity • Temporal relationship • Biological gradient • Experiment • Analogy • Biological plausibility • Coherence A fair experiment Population Timing When might we expect to see an effect? Control Intervention 30 5

Some effects occur almost instantly: Injuries, poisoning and violence mortality (excluding acute alcohol poisoning) Russia: Men aged 30-59 Per capita consumption litres pure alcohol/year 600 16 Age-standardised rate per 100,000 15 Alcohol consumption 14 300 13 Mechanisms Mortality rate 12 11 0 10 1965 1970 1975 1980 1985 1990 1995 31 Smoking and lung cancer in British men A lagged effect 4500 1600 • We used longitudinal data to evaluate the impact of housing payment problems on health status among home-owners and renters in 27 EU states Age-standardised mortality rate per million 4000 1400 • Multi-level and fixed-effects models applied to a retrospective cohort drawn 3500 Cigarettes per year per capita 1200 from the EU-SILC data Cigarettes per capita 3000 per year 1000 • Sample of employed persons without housing arrears in base year 2008, 2500 followed through to 2010 (n=45,457 persons, 136,371 person-years) 800 2000 • Multi-variate models tested impact of transitioning into housing arrears on 600 1500 self-reported health Lung cancer 400 • Arrears associated with a significant deterioration in health of renters but not 1000 mortality rate owners, after adjusting for individual fixed effects 200 500 • Effect independent of and greater than the impact of job loss for the full 0 0 sample 1911- 1921- 1931- 1941- 1951- 1961- 1971- 1981- • Magnitude varied across countries – largest adverse associations in Belgium, Austria, and Italy. Data for men aged 15+ 32 Source : HOAB and Wald Time • Some things we can predict … • Why have rates of homelessness claims in England risen since 2010? – A change in drinking pattern can have an impact almost at once • We used variations in rates across local authorities to test the impact of economic downturns and budget cuts. • Some things are more difficult • Using cross-area fixed effects models of data from 323 UK local authorities between 2004 and 2012, we evaluated associations of changes in statutory homelessness rates with – Smoking rates today predict lung cancer mortality years in the future economic activity, unemployment, and local and central government expenditure. • Each 10% fall in economic activity was associated with an increase of 0.45 homelessness claims per 1000 households – We can only expect to detect effects with short lags • Increasing rates of homelessness also strongly linked with government reductions in welfare spending – Many other effects will be lost in the noise • Disaggregating types of welfare expenditure, we found that strongest associations with reduced homelessness claims were spending on social care, housing services, discretionary housing payments and income support for older persons. 33 6