How Germany flattened the curve

Screenshot 2020-06-30 at 07.40.59
By Georg von Graevenitz
Tue 30th June 2020 | Blog

CORE links modern economic methods to pressing policy challenges: mounting inequalities, climate change, concerns about power in the workplace, and financial instability. COVID-19 has highlighted inequalities in new ways, demonstrated the risks of ignoring pollution linked to climate change and underscored the role of governments in stabilising economies, coordinating responses and preparing for adversity. The pandemic also highlighted the role of science and trust in protecting society against adversity. Differences in preparedness, often the result of many years of incremental policy developments, have been particularly significant in this fast moving crisis. This post describes how scientists, preparedness and luck combined to simplify crisis management in Germany. But neglect of the exploitation of workers in the meat-processing industry has created unexpected external effects as new lockdowns are now being declared.


There is no Glory in Prevention


By CORE author Georg von Graevenitz, Queen Mary University of London

Successful prevention undermines the willingness to support tough measures and attracts little praise, this is the paradox of prevention. Former US presidents are no strangers to this paradox as outlined by Gibbs and Duffy in The President’s Club. It has also been a regular feature of a remarkable series of podcasts produced by Korinna Hennig with Professor Christian Drosten, director of the Institute of Virology at the Charite hospital in Berlin. The podcast made Drosten a star and a hate figure in equal measure as Germany has gone through the lockdowns brought about by the Coronavirus SARS-CoV-2. Christian Drosten is a SARS (severe acute respiratory syndrome) virus specialist: he identified the first SARS virus and developed a test for it in 2003. So he was alert to the dangers of the new Coronavirus SARS-CoV-2 emerging in Wuhan in December 2019. He coordinated the medical response, warned policy makers in Germany and was present in the media early on. Eventually he settled on the podcast as an effective means of communicating the complexity of the evolving science to a broad public. His interventions in the science and policy community made early high volume testing possible (de – linked text in German).  This has made Drosten a target for those angered by the lockdowns. He is recognised, even when wearing a mask, his contact details are no longer public, significant volumes of hate mail include death threats.

This blog looks at how Germans collectively prevented the significant loss of life that the arrival of this new virus threatened to bring about. The post contains links to data sources, government announcements and publications. The focus is primarily on Germany, with some references to developments in England & Wales. This is primarily the result of the author’s experience with these countries. An infographic comparing the UK to several additional countries is provided by the BMJ.

If there has been glory in prevention, it was primarily abroad. Germans have not necessarily all realised their good fortune, although they have collectively fared better than people in some other similarly sized countries. The comparison of excess deaths in Germany with those in England & Wales illustrates this:



Note: Figures based on data released by Germany’s statistical office [data] and the Office for National Statistics [data] on registered deaths. Excess deaths are calculated for each calendar week of 2020, relative to the average of deaths in the same calendar week in 2015-2019. A discussion of different approaches to calculating excess deaths can be found here.  I report P-scores in the above figures.

The comparatively low death rate in Germany has not gone unnoticed abroad, with excellent and detailed reporting on how Germany has kept excess mortality low (FTBBCNew York Times). Three aspects are noted in these reports: initial conditions, management of the pandemic and good fortune. Initial conditions noted include research institutes, funding of health care, testing laboratories, capacity to produce equipment and environmental factors such as airborne pollutants. Initial conditions and management of the pandemic were not independent anywhere: worse initial conditions complicated management, e.g. efforts necessary to obtain ventilators for the NHS.

Before turning to details, it is worth dwelling a bit more on the figure above. Excess mortality in Bavaria and Baden-Wuerttemberg reached 20% in the early part of April, while Sachsen-Anhalt, like much of the former GDR, experienced almost no excess mortality. This strong regional disparity is also present in Italy [data], where the early lockdown protected many regions and cities, e.g. Rome and Naples, from significant outbreaks of the virus. Meanwhile the experience of England & Wales has been quite different. London was affected much more than Wales, but even Wales saw excess mortality twice as high as Bavaria at its peak. This suggests the virus had spread widely in England & Wales before lockdown. Work recently published by Flaxman, S. et al. (2020) in Nature corroborates this: working back from consolidated data on deaths reported by the European Centre of Disease Control they show that COVID-19 likely had infected 710,000 people in Germany and 3.4 million people in the UK by 4 May 2020. At this time testing in Germany had identified 164,807 positive cases, whereas testing in the UK identified 153,364 positive cases.

Good fortune

The first cases of Coronavirus to have been identified reached Germany in late January 2020. A Chinese employee of Webasto, had contracted the virus in China, travelled to Munich between 19 and 22 January and infected a number of colleagues. She was diagnosed on return to China and informed her employer. This led to contact tracing in Munich and quarantining of a number of employees of Webasto. The case was widely reported in German newspapers, contributing to the first spike of interest in Coronavirus (see figure below) around this time. The figure shows intensity of search for the keyword “Coronavirus” for each country as obtained from Google Trends. Google normalise each time series relative to the country maximum in the period reported. This means that we can compare the variation within each time series to that of the others, but it makes no sense to compare the absolute level of different series at specific dates across countries.


Note: Data obtained from Google Trends. Data on government measures taken from the Oxford COVID-19 Government Response Tracker (OxCGRT). It should be noted that lockdown measures varied significantly across and within countries and in the case of Germany the severity of lockdown varied by state (Land). Flaxman, S. et al. (2020) also compare lockdown measures based on their own research.

By 10 January 2020 TIB Molbiol, a manufacturer of biomedical test kits in Berlin, developed a kit for the new virus. This was due to a conversation (de) between Chistian Drosten and the TIB founder the Charite hospital regarding events in China. Germany has a significant number of laboratories capable of processing test kits (see Initial Conditions below), which sped up local identification of cases. Once the Webasto employees had been quarantined public interest in the virus fell, visible in the figure above. A few days later the WHO declared a Public Health Emergency of International Concern leading to short-term spikes of interest in Italy, Sweden and the UK, also visible in the figure. The German government decreed that infections with SARS-CoV-2 were notifiable on 1 February 2020.

On 26 February 2020 new cases arose in Heinsberg (Northrheinwestfalia [NRW]) and by 4 March 2020 the Robert Koch Institut reported 113 cases there. Then schools and kindergartens in the region were closed, interest in Germany spiked again. A week later there were 484 confirmed infections and 3 deaths around Heinsberg. Bavaria and Baden-Wuerttemberg recorded significant infection counts, many linked to ski-breaks in Italy and Austria.

On 4 March 2020 three doctors from Milan sent a letter to their colleagues at the European Society of Intensive Care Medicine warning of the dangers of the new virus. This letter kickstarted preparations in Bavarian hospitals (de). Isolation wards were set up, initially not done in Milan. Nurses and doctors were prepared for treatment of the new disease and schooled in triage. Two days later the first patients arrived. That week Germany saw 10 deaths from Coronavirus and England saw 5, most not publicised until much later.

By the first two weeks of April between 200 and 250 people would die from COVID-19 related complications per day in Germany. Overall to date 8800 people have died (data), of whom 86% over the age of 70.45% of deaths have been attributed to people living in retirement-homes, prisons, care-homes and other places in which large numbers of adults are housed together. In this group to date the proportion of those that died, if they were diagnosed with the virus, has been 20%. This compares with 4% for all those with no contact to such homes, to hospitals or to schools.

The number of deaths attributed to Coronavirus SARS-CoV-2 in Germany by April 2020 are likely to be lower than those attributed to influenza in March of 2018. However, the lockdowns reduced deaths from influenza and spreading of the new coronavirus. Flaxman, S. et al. provide simulations that indicate deaths in Germany would have exceeded half a million by 4 May in the absence of any public intervention. An alternative rough calculation is to compare the experience of Sweden (10 million inhabitants, 5000 deaths to date), where lockdown was considerably lighter, to that of Germany. Under Swedish conditions Germany would now have around 40,000 deaths, an order of magnitude less than predicted by Flaxman et al., but still around 4.5 times more than the current count. Germans were fortunate, because initial arrivals of the virus in Germany were detected and spreading of the virus was limited by quarantines and because Germany lagged Italy. Large scale gatherings continued to take place until 10 March 2020 in Bavaria and Baden-Wuerttemberg, the same day Italy went into national lockdown. Many deaths in southern Germany can be traced to these gatherings and might have been prevented, had the Heinsberg event been taken more seriously.

Initial Conditions

On 16 March 2020 the WHO sent out the message to test, test, test. That same day Ute Teichert, the leader of Germany’s association of doctors in public health gave an interview  (de) explaining how overstretched the 375 Local Health Authorities (Gesundheitsaemter) had been even before the pandemic – she appealed for more resources. Before the pandemic Germany employed around 2,500 doctors and 14,500 others in these local authorities. They have a wide range of tasks among which: test, track and impose quarantines in a public health emergency. Ute Teichert’s appeal worked. On 25 March 2020 the federal government and the federal states agreed that there should be contact tracing teams of 5 people for every 20,000 inhabitants by 22 April 2020. 105 additional mobile teams were set up centrally. Many of these contact tracing teams were staffed by civil servants, teachers and medical students. A team of 40 people was created at the Robert Koch Institute [RKI] (details below) to support local teams. Crisis management built on existing institutions with the aim of emulating South Korea’s approach to tracing and isolating cases.

Contrast this with initial conditions for public health in the UK. Public Health England employs 5,500 people in total. Of these only around 300 were available for contact tracing. The government made SARS-CoV-2 notifiable on 5 March 2020. On 12 March 2020 the UK government moved its response to the delay phase. Tracing in the community was stopped, due to a lack of testing and tracing capabilities  (Pollock, A. et al. 2020 in BMJ; Abbasi, K. 2020 in BMJ).  Many commentators have linked this public health failure to public health reforms in 2012. An editorial in the BMJ in 2016 supports this view. Once tracking was reinstated, the solution adopted in the UK was to centralise the effort.

Daily updates on the progression of the pandemic in Germany have been provided by the Robert Koch Institute (RKI) starting on 4 March 2020, based on data notified by Local Health Authorities. These have been supported by press conferences, led by the institute’s director, Professor Wieler. The Robert Koch Institute has a similar mission as Public Health England: to protect the population from disease and to improve public health. RKI has 1100 employees. They leveraged a number of existing networks linking RKI to laboratories in university hospitals, research centres and to accredited independent laboratories. Within 3 weeks the institute mobilised over 200 laboratories (de) to support testing of the population. The distributed nature of testing in Germany has drawn attention as it contrasts with the centralised approach taken in the UK. This has been important for a reason that is not widely acknowledged: many tests created to detect Coronavirus SARS-CoV-2 are novel. The usual process of establishing precision and reliability of diagnostic tests could not be followed. Doctors and lab specialists in Germany often relied on direct communication to determine how individual tests should be interpreted. This communication between the doctors taking the swabs and the lab technicians becomes very difficult when laboratories are not local.

As the epidemic unfolded in Italy it became apparent that intensive care units and ventilators could quickly become a significant bottleneck. The result was a slew of reports comparing per capita ICU beds in various countries. Germany did very well in these league tables. This is the result of a significantly higher per capita expenditure on health in Germany (5986 US$) relative to the UK (4070 US$) (OECD). This higher outlay does not result in significant improvements in life expectancy in Germany compared to the UK, nor are health risks very different across the two countries. The UK scores significantly worse on access to primary and preventive care though. One implication of the much more efficient UK health system was the need to free beds in ICU units for large numbers of expected patients. This seems to have led to transfer of infected patients into the community and social care homes (Godlee, F. 2020, BMJ).

Since the coronavirus affects the lung it has been proposed that intensity of an outbreak in a location will be affected by the levels of airborne pollution. A recent study from Italy underscores the correlation between pollution and the severity of outbreaks there.  Another paper using data from the Netherlands also confirms this correlation. This study suggests the importance of particulate matter (PM) classified as fine particles (pm 2.5). Here is a map showing the average annual concentration of pm 2.5 in Europe in 2016. It shows comparatively high concentrations in Lombardy and neighbouring areas.

Management Efforts

Germany is a federal state, responsibility for health care is largely devolved as is responsibility for imposition and policing of lockdowns. Predictably there was competition among political leaders of the federal states to be the first to move into and out of lockdown. An interesting aspect of this competition has been that some of the principal figures, i.e. the federal health minister (Jens Spahn), the leader of NRW (Armin Laschet), the leader of Bavaria (Markus Soeder) are also key figures in the race to succeed Angela Merkel. One odd result was that while borders to Austria, Switzerland and France were closed (Bavaria), the border to the Netherlands remained open (NRW). Communication about the virus and political measures has rested on many shoulders, sometimes undermining cohesion of the message, but generally ensuring that public trust remained high. No particular scientist or group of scientists was ever co-opted into regular communications on behalf of government policy or positions.

In spite of the multiplicity of voices the federal government managed to coordinate effectively. The new virus was decreed notifiable on 1 February 2020. Between 23 February 2020 and 27 March 2020 the cabinet, the Bundestag and the states (Bundesrat) passed two laws providing financial support to the health care system and transferring powers to the federal level to manage provision of health care resources and adjust rules and regulations in the health care system. Between 29 April 2020 and 15 May 2020 a further law was passed to fund the measures for testing, tracking and isolation following the South Korean approach. This law also provides funds for treatment of patients from other EU countries in Germany. On 16 June 2020 the government released an app to help the people establish whether and when they had contact with someone carrying the virus. The app was downloaded 12.2 million times over the following 6 days.

It is a truism that the virus is putting a spotlight on existing social inequality. This is true in as much as it has affected minorities and migrants in the UK more, including health professionals. People with lower incomes typically have greater exposure to airborne pollution and are more likely to be affected by health conditions that make the infection with the virus more serious. In Germany, a particular spotlight is currently falling on the working conditions of employees in the meat processing industry. This industry has increasingly relied on migrant workers, who are housed in run-down communal buildings and who work in conditions that encourage infections to spread. The population is getting a crash course in the effect of labour market reforms that have enabled these workers to be exploited through opaque chains of subcontracting. Those who are unlucky enough to live close to meat processing plants with outbreaks are now finding themselves back in lockdown.

Since 26 February 2020 Korinna Hennig and Professor Drosten have produced 49 podcasts, each up to an hour in length covering established and emerging science regarding the new virus. Their podcasts have topped Germany’s podcast charts and by 8 May 2020 it had been downloaded 41 million times. On Google Trends searches for Merkel, the RKI and Drosten all outstrip searches for job-centres from mid February. While there have been demonstrations against the lockdowns and criticism of the government’s communication style, the general mood in the country has remained broadly supportive [ongoing survey , data] of the measures taken thus far. Germany has been fortunate and the response well managed in equal measure, hopefully this will continue to be recognised. The contribution of scientists to the management of the public response should not be underestimated. This is also evident in other countries, such as Greece.


I’d like to thank Stephan Ganter, Horst von Bernuth, Reiner Leidl, Christopher Schwarting and Dietmar Harhoff, Wendy Carlin, Stella Ladi, Brigitte Granville and Pietro Panzarasa for their comments regarding this post.