The impact of lockdowns during the COVID-19 pandemic on work-related accidents in Austria in 2020

Austria provides an interesting case in studying the impact of lockdowns in 2020 on public life, in particular work-related accidents due to a few unique features.

At first glance, restrictions to social contacts, both private and at work, were similar to other developed countries in terms of timing and scope. Between 29th of February and 22nd of March most European [1] and Asian countries [2] had implemented decisive restrictions of movement, social interaction and gatherings. In Austria, such policies came into place on Monday, March 16th, 5 days after the WHO declared COVID-19 (Corona-Virus-Disease 19) a global pandemic and 3 days after Europe was declared the epicenter of this pandemic [3, 4]. The first cases in Austria, however, were detected between 24th and 26th of January 2020 [5]. The last of the initial restrictions was lifted in Austria on May 29th [6]. The following lockdowns in the 4th quarter (Q4) of 2020 had two different characters. Between 3rd and 16th of November 2020 and between 6th and 25th of December 2020 social distancing policies similar to the first lockdown period were implemented, with the exception of retail and services staying open [7]. These periods were colloquially called “lock down light” as opposed to the stricter measures of the earlier “hard lockdown”. In reaction to and anticipation of surging cases, another two hard lockdowns were imposed between and after the lock down light phases [8]. This is where Austria is different from other European countries, where lockdown measures remained fairly homogeneous and constant over a longer period of time. This provides the opportunity to study a situation similar to a repeated experiment regarding different lockdown measures in Austria (Fig. 1).

Due to these lockdown restrictions, approximately one third of the working population in Austria and other European countries switched in Q1 2020 to working from home [10]. For this form of teleworking the term “home office” was used in German-speaking countries. Home office was, of course, not possible in all industries and some sectors had to reduce operations significantly. Consequently, production declined in all European countries on average 3.1% and in Austria about 2.8% in Q1 2020 [11]. In Q1 and Q3 2020 people worked about 8% less hours compared to the same quarters from 2019 while in the Q2 the decrease in hours worked was 19% and in Q4 11% [12]. This difference in decline seems to coincide with lockdown phases but data are not available detailed enough to draw conclusions. It is clear, however, that Austrians worked less in total and more often from home during the pandemic, similar to their European neighbors. Newly introduced was a Home Office Bill alongside measures to curb the economic outfall of the lockdown measures in March 2020. The new law extended the definition of work-related accident to all private rooms, including the living room and the kitchen. Moreover, working time was defined more flexibly and the range of covered activities broadened to all activities necessary to satisfy essential needs, including cooking [13]. The combined effect of the increase in telework alongside a broader definition of work accidents at home on the one hand and reduced hours worked and reduced production on the other has not been studied before.

Work accidents in Austria are defined as all sudden events that influence health from outside the body and that are related to the activity that is insured under the § 175 Abs (1) ASVG (Arbeits und Sozialversicherungsgesetz, “Work and Social-Security Law”) [14]. This quite unique formulation is a product of a historically grown distinction between traumatologic or toxicologic workplace accidents and chronic, internal diseases that may arise from an occupation. As a result, the data set used in this study, which was derived from Allgemeine Unfallversicherungsanstalt (AUVA, General Accident Insurance Institution), consists of only patients with acute traumatological needs. This is important, since most European countries, including Austria, accepted a COVID-19 infection as work-related disease in their insurance catalogue. These cases are, however, not included in the AUVA database. The following analysis of the impact of covid-related measures on work-related accidents only covers classical accidents without the COVID cases.

Several sequential, intensity-modulated lockdown periods, novel legislation on and a surge in working from home workplaces, and a unique way of recording work-related disease make Austria an interesting case in analyzing the effects of lockdowns on numbers of work-related accidents.

The overview over the data in the run-sequence plots on daily work-related accidents in 2020 and 2019 exhibits a similar range, a seasonality with a frequency of about 51 per year and a “dent” in the holiday month August (Fig. 2) during both years. Clearly visible is a decline in numbers of registered accidents in 2020, roughly where one would assume the effects of the lockdowns taking place at the end of Q1 and beginning of Q4.

In total, the number of registered and accepted work-related accidents in the analyzed time period in 2019 amounted to 97,149 and 70,985 cases in 2020, a minus of 26,164 cases or 26.93%. Clear differences in the data of 2019 and 2020 become visible in the initial statistical analysis (Table 2). In 2019, the daily average of registered work-related accidents amounted to 272.3 cases. In 2020 this number was significantly lower with 199.4 cases. The range of registered cases in 2020 was also significantly lower with 407 cases compared to 511 in 2019. An F‑test on the difference of variances suggests a significantly different distribution of the number of registered cases in both years. These inferential findings are graphically supported by a violin plot (Fig. 3).

A procedure called “seasonal decomposition“ [16] provides a more detailed look at the seasonality of work-related accidents in the years 2019 and 2020. The top panel (year 2019) and bottom panel (year 2020) in Fig. 4 show the final estimate of how work accidents in the respective years would look like if seasonality is removed. After subtracting the weekly component from the original data several outliers in these error values that occur in both years were identified as national holidays in Austria where also a sharp drop in work-related accidents is observed.

A test whether the estimated model showed better fit compared to intercept-only model (“Omnibus Test”) suggested a highly significant improvement to fit of the estimated model (p < 0.001). The estimates of the difference-in-differences model are listed in Table 3. Neither the year nor the months corresponding to the pandemic as a whole nor their interaction variable were estimated to have a significant effect on work-related accidents. Both lockdowns, however, had a statistically highly significant effect on work-related accidents. The hard lockdown is estimated to have reduced the average number of daily registered work-related accidents by 39.9%. The light lockdown phases reduced this number by an average of 51.1%. Weekends and holidays were estimated to have the greatest impact on work-related accidents with a reduction of 68.8% and 72.9%, respectively.

Government measures in reaction to the pandemic around the world have raised several questions including economical [18] and ethical [19] discussions alongside the medical debate. Accidents at work touch on all three of these subjects and this study aims to provide a puzzle piece in the discussion with its analysis of the development of work-related accidents during the lockdown periods in Austria. The data analyzed in this study show a significant decline in registered work-related accidents in Austria in 2020 of 26,164 cases (26.93%) compared to the previous year. This drop is very evident when comparing the panels in Fig. 4, where a sharp decline in accident-numbers can be observed shortly after the first lockdown was introduced.

When compared to the available data from other countries, the drop in work-related accidents in Austria in 2020 is immense. Kwon et al. [20] provided an excellent overview on workplace accidents during the pandemic in Korea and elsewhere. They summarized that the number of patients injured at the workplace declined in 2020 by 1664 or 1.8% in Korea and by 2429 cases in Singapore. Japan, on the other side, saw an increase in such cases of 4.4% and in the UK the case count remained fairly stable. Kwon et al. correctly stated that with fewer people going to work, the remaining employees might face tougher working conditions and are more prone to injuries. In most European countries, these numbers are not available until a few years after the year in question. In past years, Austria was with 2.87 injuries per 100,000 persons employed, well above the EU average of 1.77 [21].

However, due to the unique definition of work-related accidents in Austria, an international comparison is difficult. Another possible reason for the discrepancy in the dynamics of work-related accidents during the pandemic could be owed to the difference in lockdown strategies between the aforementioned countries. Korea started already on February 4th with mass testing and implemented a test-and-trace system with the individual isolation of infected patients. No lockdown was implemented in 2020 and the strictest measure was a mandated 2.5 m distance in Q4 2020 [22]. During this time, Austrian authorities were still sceptical about a mass testing strategy and the scientific background inconclusive [23, 24]. Japan, on the other hand, implemented a rather complex COVID-19 response. In accordance with recent legislation, there was no nationwide lockdown in 2020 but only regionally restricted mobility. Moreover, authorities could only request businesses to comply with reduced opening hours and working force but never imposed a mandatory lockdown [25].

It could be argued that government-imposed lockdowns are only valid on paper if not strictly enforced and people still went out, got injured and reported that injury as a “home office” accident, potentially distorting the numbers in this study. Moreover, the lockdowns could have only a delayed effect on people’s behavior, rendering the dummy variables for lockdown phases less significant. A look on Google Mobility Data in 2020 [26] suggests, however, that workplace, transit, and retail mobility measures dropped significantly during the pandemic and that these drops coincided with the implementation dates of the lockdowns (Supplementary Fig. 1). These measures come with several caveats, such as baseline value calculation, selection bias and the fact that many background calculations and definitions by Google remain hidden to the authors of this study. It can therefore only serve as a rough indicator of lockdown effectiveness and work activity.

A very clear takeaway is the immense overall reduction in work-related accidents in Austria in 2020 of 26,164 registered accident patients. According to the Austrian Labor Inspectorate and the AUVA, each workplace injury is attributed an average of 12,500 euros of economic costs [27]. In terms of injuries at the workplace, an estimated 327 million euros in economic costs have been saved. This might not seem much compared to other costs the pandemic has caused and what the rehabilitation of COVID patients might still cost [28]. It is a sum, however, that can be used as an incentive to further investigate the lessons from the pandemic work environment.