key: cord-0281039-m9uz3v2n
authors: Boender, T. S.; Cai, W.; Schranz, M.; Kocher, T.; Wagner, B.; Ullrich, A.; Buda, S.; Zoellner, R.; Greiner, F.; Diercke, M.; Grabenhenrich, L.
title: Using routine emergency department data for syndromic surveillance of acute respiratory illness before and during the COVID-19 pandemic in Germany, week 10-2017 and 10-2021
date: 2021-08-21
journal: nan
DOI: 10.1101/2021.08.19.21262303
sha: 9e6ae72d1999209bf541ea2454b28ecd7b7e7d1a
doc_id: 281039
cord_uid: m9uz3v2n

Introduction: To better assess the epidemiological situation of acute respiratory illness in Germany over time, we used emergency department data for syndromic surveillance before and during the COVID-19 pandemic. Methods: We included routine attendance data from emergency departments who continuously transferred data between week 10-2017 and 10-2021, with ICD-10 codes available for >75% of the attendances. Case definitions for acute respiratory illness (ARI), severe ARI (SARI), influenza-like illness (ILI), respiratory syncytial virus disease (RSV) and Coronavirus disease 2019 (COVID-19) were based on a combination of ICD-10 codes, and/or chief complaints, sometimes combined with information on hospitalisation and age. Results: We included 1,372,958 attendances from eight emergency departments. The number of attendances dropped in March 2020, increased during summer, and declined again during the resurge of COVID-19 cases in autumn and winter of 2020/2021. A pattern of seasonality of acute respiratory infections could be observed. By using different case definitions (i.e. for ARI, SARI, ILI, RSV) both the annual influenza seasons in the years 2017-2020 and the dynamics of the COVID-19 pandemic in 2020-2021 were apparent. The absence of a flu season during the fall and winter of 2020/2021 was visible, in parallel to the resurge of COVID-19 cases. The proportion of SARI among ARI cases peaked in April-May 2020 and November 2020-January 2021. Conclusion: Syndromic surveillance using routine emergency department data has the potential to monitor the trends, timing, duration, magnitude and severity of illness caused by respiratory viruses, including both influenza and SARS-CoV-2.

The response to the Coronavirus disease 2019 (COVID-19) pandemic expanded the need for timely information on acute respiratory illness on the population level, to trigger public health action, and support healthcare planning. Therefore, robust surveillance systems of respiratory illnesses are essential to monitor the burden of disease, and therewith inform public health to take appropriate and timely measures.

In Germany, acute respiratory illnesses (ARI) are monitored with syndromic surveillance systems at several levels of disease severity: internet-based syndromic monitoring of ARI in the general population (GrippeWeb, "FluWeb")(1); sentinel surveillance of ARI consultations in primary care, including virological surveillance (Influenza Working Group (2) ) and sentinel surveillance of severe acute respiratory illnesses (SARI) among hospitalised patients using ICD-10 diagnostic codes (ICOSARI) (3) . In addition, defined respiratory diseases as COVID-19 and influenza are monitored through mandatory notifications by clinicians and laboratories, within the framework of the Protection against Infection Act.

The effects of the COVID-19 pandemic, as well as the associated non-pharmaceutical interventions (NPIs) to contain the spread of the virus, have affected the transmission of not only seasonal influenza, but also other respiratory viruses in Germany (5, 6) . These changes in the epidemiological situation affect routine surveillance systems for monitoring the epidemiology and virology of respiratory illnesses (7). Therefore, further investigation of the impact on population health, as well as the healthcare system is required.

The availability of electronically collected routine medical data, such as emergency department data, allows for an extended approach in applied epidemiology and public health surveillance (8) .

Syndromic surveillance using routine emergency department data can be a fast approach to report and detect changes in healthcare utilization, without creating an additional administrative burden. In the United Kingdom (9), France (10) and other countries, syndromic surveillance systems based on emergency department data have been established (11, 12) , and have previously provided timely public health insight at extraordinary times such as extreme weather events (13) or mass gatherings (e.g. the Olympic Games in 2012 (14) ), as well as during the COVID-19 pandemic (15) .

Previous analyses of emergency department data in Germany successfully detected syndromes to monitor unspecific gastrointestinal infections, including their seasonal fluctuation (16) , as well . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted August 21, 2021. ; https://doi.org/10.1101/2021.08.19.21262303 doi: medRxiv preprint made a first description of ARI cases (17) . Since end of June 2020, the weekly Emergency Department Situation Report presents data from the routine documentation of up to 21 selected emergency departments in Germany (18) . These reports showed that in parallel to the COVID-19 pandemic and the associated NPIs starting mid-March 2020, attendances dropped, and partially increased again over summer (19) . During the resurge of COVID-19 cases and implementation of NPIs during the fall and winter of 2020-2021, the number of attendances decreased again to about one quarter below the 2019 average, and remained below 2019 average until the time of writing (April 2021) (18) .

We aimed to describe emergency department attendances for acute respiratory illness in Germany over time, for the purpose of the development and implementation of syndromic surveillance. Furthermore, we explore the impact of the COVID-19 pandemic, and its associated NPIs, on these presentations of acute respiratory illness.

We used routinely available emergency department data as a basis, to describe presentations of ARI in emergency departments in Germany for syndromic surveillance. Following a process of data curation, quality management and exploration, and defining syndromic case definitions, we described emergency department attendances over time to provide insight into presentation for acute respiratory illness. In addition, the impact of the COVID-19 pandemic on these trends was further explored.

Data used in this analysis was provided through two networks of German emergency departments which continuously transfer anonymized routine health data to the Robert Koch Institute (RKI, Germany's national public health institute), to explore their use for surveillance: the AKTIN -Emergency Department Data Registry (20) and the ESEG-project ("Erkennung und Steurung Epidemischer Gefahrenlagen"; translation: detection and control of epidemic situations http://www.rki.de/eseg). Emergency departments participate and share data on a voluntary basis.

At the RKI, using the NoKeDa (NotaufnahmeKernDatensatz) data model (21), these routine health data from the different sources are merged and processed, when the date of visit, age, and sex were available. Observations were recorded by visit (attendance) and not by person; multiple visits by the same person cannot be linked.

. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted August 21, 2021. ; https://doi.org/10.1101/2021.08.19.21262303 doi: medRxiv preprint Study population Emergency attendance data was exported from the RKI server on 13 April 2021, using SUMO-DB Release 3.0.3 (database population). In addition to the database release, data quality summary reports were generated, including information on completeness of attendances and variables. The data availability and quality could be assessed, based on the descriptive statistics and data visualisation of data quality summary. For the purpose of this work, a subset of emergency departments was included (study population): departments who continuously 

The following variables were included for analysis: emergency department/hospital, age-group (0-2 years, 3-4 years, followed by 5-year groups, and 80+ years), sex (male/female/other), triage (codes 1 to 5, higher code meaning less urgent), diagnosis (ICD-10 codes, when labelled confirmed/suspected), chief complaints (according to the Canadian Emergency Department

[MTS] (23)), referral, mode of transport, disposition (e.g. hospital admission) and isolation.

. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 21, 2021. ; https://doi.org/10.1101/2021.08.19.21262303 doi: medRxiv preprint Potential missing data refers to the absence of documentation; all data are based on routine emergency care at the respective sites, based on routine documentation practices.

We defined a set of case definitions in agreement with the established syndromic surveillance systems of acute respiratory infections at RKI and based on (inter)national standards (24, 25) .

Case definitions were based on a combination of ICD-10 diagnostic codes, or chief complaints, sometimes combined with disposition and age. Case criteria and case classifications (i.e. probable, confirmed) for acute respiratory illness (ARI), severe acute respiratory illness (SARI), influenza-like illness (ILI), respiratory syncytial virus (RSV) and Coronavirus disease 2019 Table 1 . Combined case definitions were used for the primary analysis; i.e. cases of ILI, RSV and COVID-19 were either probable and/or confirmed.

Attendances and cases were analysed as aggregated counts by week. Weekly numbers of all cases that met the case definitions (see above and Supplemental Table 1 ) were plotted as absolute counts, as well as proportions of cases per 1,000 attendances to adjust for fluctuations in attendances (i.e. a change in the denominator) over time. Cases of ARI and SARI were explored as individual indicators, as well as through calculating the proportion of severe cases (%SARI/ARI). Probable and confirmed cases of ILI, RSV-Disease and COVID-19 were explored as separate and combined overall case numbers (probable + confirmed cases).

The number of weekly attendances were described and summarised for the period before and during the COVID-19 pandemic: before/after 2 March 2020 (week 10-2020), and the different pandemic phases (26) . Because the number of weekly attendances were not normally distributed, Wilcoxon rank sum test was used for simple comparisons of the distribution of weekly cases per 1,000 attendances before and during the pandemic. Next, time series were decomposed to visually investigate the seasonality and trend over time; a decomposition of the COVID-19 time series of one year was not possible because the series was too short (less than two periods).

Data were analysed the data using R statistical software, using the tidyverse, xts and zoo packages (27) . The aggregated data of the attendances and cases (ARI, SARI, ILI, RSV, COVID-19) by week used for the analysis, are provided with this manuscript.

. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 21, 2021. ; https://doi.org/10.1101/2021.08.19.21262303 doi: medRxiv preprint

Data from eight emergency departments that transferred attendance data between 6 March 2017 up to and including 7 March 2021 with ICD-10 diagnostic codes for >75% of the attendances, were included in this study. Consequently, the study population included data of 1,372,958 attendances from 6 March 2017 (week 10-2017) up to and including 13 March 2021 (week 10-2021) ( Table 1 ). The included departments provided comprehensive (5 departments) and extended (3 departments) emergency care in seven federal states of Germany (Baden-Württemberg, Bavaria, Berlin, Brandenburg, Hesse, Lower Saxony, Saxony).

The general characteristics of the emergency department attendances are summarized in Table   1 ( Table 1) . The overall sex (52% female) and age (6% under 2 years, 8% 3-9 years; 8% 10-19 years; 17% 20-39 years; 25% 40-59 years; 21% 60-79 years; 15% 80+ years) distributions remained similar over the full observation period.

Over time, the absolute number of weekly emergency department attendances dropped significantly during the first and second wave of the COVID-19 pandemic in Germany (Figure 1) . and RSV-disease cases per week, per 1,000 attendees (Figure 2) , as well as per absolute weekly case counts (Supplemental Figure 2) . Seasonality of ARI, SARI, ILI, and RSV-disease was confirmed by the decomposed time series analysis (Figure 3) .

. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 21, 2021. ; https://doi.org/10.1101/2021.08. 19.21262303 doi: medRxiv preprint In the winter of 2020/2021, both the absolute and relative ARI case counts were lower than in previous years, and showed a downward trend (Figure 2a and 3a) . Conversely, the trend of SARI cases remained fairly stable over time, with a small downward trend in the final months of winter (Figure 2b and 3b) . This pattern was also visible when assessing the proportion of severe and hospitalised ARI cases (i.e. SARI cases) among ARI cases (Figure 4) (Figure 2c-d and 3c-d) is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 21, 2021. ; https://doi.org/10.1101/2021.08.19.21262303 doi: medRxiv preprint based on sentinel surveillance in Germany (28) , as well as in Denmark, Norway and Sweden (29) ,. The striking absence of a flu season in 2020/2021 (5, 30) , was also visible, in parallel to the resurge of COVID-19 cases during the fall and winter in 2020/2021. In 2020 and the beginning of 2021, the proportion of SARI cases among ARI cases peaked twice for five subsequent weeks in April-May 2020 and November 2020-January 2021, which was the same time the first wave and resurge of COVID-19 cases peaked in Germany (31) . The reduction in ILI and RSV cases during the pandemic period could, at least in part, be explained by the influence of NPIs on overall disease transmission, as observed by other surveillance systems in Germany (5, 6) , and internationally (32) .

Overall emergency department attendances dropped substantially following the start of the COVID-19 pandemic in March 2020, increased during summer, and declined again during the resurge of the pandemic during the autumn and winter of 2020/2021. The shift in attendances required a baseline adjustment for case reporting (change in denominator), by standardising the case count per 1,000 attendances. This reduction in attendances could have been partially due to the advisories to reduce physical contact with healthcare services, and to use telephone services instead (at least before visiting), when possible. This observation fits the overall reduction in hospitalisation as reported by other German hospitals (33, 34) . In addition, a reluctance to visit health care facilities due to a fear to be exposed to SARS-CoV-2 (35), could also be a reason for a reduction in emergency department attendances. Similar trends were observed in emergency departments in other countries, including the UK (36) and USA (37) .

The findings of this study should be interpreted taking the following limitations into account. First, the selected data are a convenience sample and are not representative for emergency departments in Germany as a whole. Differences on the level of individual departments, as well as differences between emergency departments, played a limited role in the analysis and interpretation. This study aimed to explore the possibility and feasibility to use these routine data for public health surveillance. In addition, structural changes at the department, including active interreference with patient flows with respiratory complaints, could have affected reported attendances and case counts. To fully understand trends in the data, being it due to a change in infection dynamics, healthcare seeking behaviour, the COVID-19 pandemic and NPIs, or structural changes at the emergency department, more information is needed. This type of investigation is, however, beyond the scope of this work. The strength of the current approach of passive surveillance using routine documentation has its own advantages being fast (timely reporting) without creating an additional burden on the healthcare workers. Furthermore, case . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 21, 2021. ; https://doi.org/10.1101/2021.08.19.21262303 doi: medRxiv preprint definitions were based on routine documentation, designed for triage and clinical documentation in the emergency department, not designed for surveillance purposes. The syndromic case definitions heavily rely on clinical presentation, and laboratory diagnosis is often not available, or not foreseen. Shortness of breath, for example, may be a cardiac symptom as well, and is not always caused by a respiratory pathogen (38) . Furthermore, information on hospitalisation (disposition), a key variable for SARI cases, was not always available, leading to potential underreporting of SARI cases. In addition, ICD-10 codes for COVID-19 diagnosis were not available from the beginning of the pandemic, leading to underreporting of the true number of cases.

Moving forward, the current ongoing, systematic collection, analysis, interpretation, and dissemination of emergency department data for use in public health action, needs to be evaluated, to make targeted recommendations to improve quality, efficiency, and usefulness of Moving forward, increased COVID-19 vaccination roll-out and subsequent (de-escalation of) NPIs, will have effects on circulation of other respiratory viruses as well. In the near future, this type of almost real-time syndromic surveillance can help to timely inform public health decision making, moving forward in the next phases of the pandemic.

The current analysis of four years of emergency department data identified syndromic cases of (severe) acute respiratory illness, and could differentiate between ILI, RSV and COVID-19 cases, which should be further investigated. Clear patterns of seasonality and trends over time were identified, for all case definitions, in line with findings reported from established ARI, SARI, ILI and virological surveillance systems (41) . Syndromic surveillance of ARI and SARI in the emergency department could contribute to the overall burden estimates of acute respiratory infections in the upper and lower respiratory tract. Syndromic surveillance using routine emergency department . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 21, 2021. ; https://doi.org/10.1101/2021.08.19.21262303 doi: medRxiv preprint data has the potential to complement established indicator based, syndromic and virological surveillance systems used to monitor the timing, duration, magnitude and severity of epidemics caused by respiratory viruses, including RSV, influenza virus and SARS-CoV-2.

. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity.

The copyright holder for this preprint this version posted August 21, 2021. ; https://doi.org/10.1101/2021.08.19.21262303 doi: medRxiv preprint

LG, MS and TSB coordinated the ESEG-project and established the surveillance using emergency department data at RKI. TSB developed the study design, performed the data analysis and wrote the manuscript. FG and RZ ensured coordinated data sourcing from the AKTIN and ESEG part respectively. TK and BW ensured data provision and data quality, under supervision of AU. WC developed the case definitions, supported by TK, MS, BW, MD and SB who provided input on the methods, development and interpretation of the work. All authors critically revised the manuscript and approved its final version.

None declared. 

The aggregated data of the attendances and cases (ARI, SARI, ILI, RSV, COVID-19) by week used for the analysis, will be provided.

We are very grateful for all emergency departments in the AKTIN and ESEG networks, who contributed to the establishment of the emergency department surveillance in Germany, and shared their data.

. CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review) . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted August 21, 2021. ; https://doi.org/10.1101/2021.08.19.21262303 doi: medRxiv preprint Figure 2A : acute respiratory illness (ARI) Figure 2B : severe acute respiratory illness (SARI) Figure 2C : influenza-like-illness (ILI) Figure 2D : respiratory syncytial virus (RSV) . CC-BY 4.0 International license It is made available under a is the author/funder, who has granted medRxiv a license to display the preprint in perpetuity. (which was not certified by peer review)

The copyright holder for this preprint this version posted August 21, 2021. ; https://doi.org/10.1101/2021.08.19.21262303 doi: medRxiv preprint

Internet-based syndromic monitoring of acute respiratory illness in the general population of Germany, weeks 35/2011 to 34/2012

Bericht zur Epidemiologie der Influenza in Deutschland Saison

Establishing an ICD-10 code based SARI-surveillance in Germany -description of the system and first results from five recent influenza seasons

Gesetz zur Verhütung und Bekämpfung von Infektionskrankheiten beim Menschen (Infektionsschutzgesetz -IfSG)

Trends in respiratory virus circulation following COVID-19-targeted nonharmaceutical interventions in Germany

Impact of COVID-19 pandemic and associated non-pharmaceutical interventions on other notifiable infectious diseases in Germany: an analysis of national surveillance data during week 1-2016 -week 32-2020. The Lancet Regional Health -Europe. 2021;Accepted for publication. 7. WHO. Interpreting influenza surveillance data in the context of the COVID-19 pandemic. Geneva: World Health Organisation

Glossary for public health surveillance in the age of data science

Surveillance des urgences et des décès SurSaUD®: Santé publique France

R/bulletins-sursaud-R-sos-medecins-oscour-mortalite

Emergency department syndromic surveillance systems: a systematic review

Using an Emergency Department Syndromic Surveillance System to investigate the impact of extreme cold weather events

Establishing an emergency department syndromic surveillance system to support the London 2012 Olympic and Paralympic Games

Emergency department use during COVID-19 as described by syndromic surveillance

Syndromic surveillance using emergency department data for the monitoring of unspecific acute gastrointestinal infections

Vergleich von akuten respiratorischen Erkrankungen (ARE) in der Notaufnahme eines Schwerpunktversorgers mit Daten der ARE-Surveillance des Niedersächsischen Landesgesundheitsamtes zwischen

Inanspruchnahme deutscher Notaufnahmen während der COVID-19-Pandemie -der Notaufnahme-Situationsreport (SitRep)

Das AKTIN-Notaufnahmeregister -kontinuierlich aktuelle Daten aus der Akutmedizin

Revision of the Canadian Emergency Department Information System (CEDIS) Presenting Complaint List version 1.1

Modern Triage in the Emergency Department

):122-8. 25. WHO. Global epidemiological surveillance standards for influenza: World Health Organisation

Retrospektive Phaseneinteilung der COVID-19-Pandemie in Deutschland bis Februar 2021

Surveillance: An R package for the monitoring of infectious diseases

Erste Ergebnisse zum Verlauf der Grippewelle in der Saison 2019/20: Mit 11 Wochen vergleichsweise kürzere Dauer und eine moderate Anzahl an Influenza-bedingten Arztbesuchen

Abrupt termination of the 2019/20 influenza season following preventive measures against COVID-19 in Denmark, Norway and Sweden

Very little influenza in the WHO European Region during the 2020/21 season

COVID-19) Weekly Situation Report from the Robert Koch Institute: Calendar week 17/2021

The effect of seasonal respiratory virus transmission on syndromic surveillance for COVID-19 in Ontario

Medical Emergencies During the COVID-19 Pandemic

Reduced Rate of Inpatient Hospital Admissions in 18 German University Hospitals During the COVID-19

Non-COVID-19 visits to emergency departments during the pandemic: the impact of fear

Impact of the SARS-CoV-2 pandemic and associated lockdown measures on attendances at emergency departments in English hospitals: a retrospective database study. The Lancet Regional Health -Europe

Impact of the COVID-19 Pandemic on Emergency Department Visits -United States

Chief complaints in medical emergencies: do they relate to underlying disease and outcome? The Charité Emergency Medicine Study (CHARITEM)

Updated guidelines for evaluating public health surveillance systems: recommendations from the Guidelines Working Group

The COVID-19 pandemic: a new challenge for syndromic surveillance

5%) 292,950 (21.3%) 80+ 156,049 (14.4%) 43,988 (15.2%) 200,037 (14.6%) Sex Female 524,790 (48.5%) 140,161 (48.3%) 664,951 (48.4%) Male 557,971 (51.5%) 149,937 (51.7%)