COVID-19 and scienti c research interests and ndings in epidemiology and social sciences: a systematic review

The emergence of COVID-19 has prompted an unprecedented scienti c publication with the aim of better understanding this new disease. This study assessed the scienti c impact and disciplinary priorities of the published papers on the pandemic by comparing epidemiological (EP) and social sciences (SS) research interests. Papers were identi ed via keywords searching using Google Scholar and Scopus databases. From an initial 1720 papers, we identi ed 597 5 relevant articles, of which 347 were for EP researches and 250 for SS studies. We extracted information, such as authors' countries, and research thematic related to EP and SS. The results revealed that most papers were authored by Asian (37.5%), European (30.5%) and American (19.6%) scientists. Only 10.1% and 2.3% of authors were a liated with African and Oceanian institutions, respectively, indicating that the regions most a ected by the pandemic 10 mainly contributed to the scienti c publications. In total, 26 research themes were recorded from both EP and SS studies. There was a high signi cant di erence among themes in both research elds (χ = 1204.3, df = 1, p-value < 0.001). EP papers mostly dealt with clinical trials (54.5%) and diagnosis (53.3%). These papers assessed the incidence and epidemiological characteristics of the disease (incubation period, symptomatic period, recovering or death), 15 testing tests developed, drugs and vaccines used. SS papers were mainly concerned with the sociocultural analyses (78%) and economic impact (55.6%) of the pandemic. They mainly focused on behavioral changes induced by the pandemic and strategies developed to mitigate its impacts. This study highlights the di erence between regions and gaps between scienti c disciplines concerning the proposed responses to control the pandemic. It is important to 20 promote collaborative and interdisciplinary studies for health emergencies.


Introduction
The novel coronavirus (SARS-Cov-2) causes COVID-19, a severe and acute respiratory syndrome discovered in Wuhan, China (Velavan and Meyer, 2020;Wu and McGoogan, 2020). By 1st March 2021, about 113,467,303 total conrmed cases and 2,520,550 deaths recorded worldwide (WHO, 2021) can be considered as proof of the ravages caused by the disease. On 11 5 March 2020, the World Health Organization (WHO) declared COVID-19 as a pandemic (i.e. a public health emergency of international concern), thereby underlining its global signicance (WHO, 2020). This measure gave rise to numerous recommendations and guidelines on prevention and management from WHO itself and other predominantly vulnerable countries (Phua et al., 2020). 10 To control the pandemic, Non-Pharmaceutical Interventions (NPI), such as travel or movement restrictions, social-distancing, wearing masks and regular hand washing with water and soap have been promoted (Gilmore et al., 2020;Ngonghala et al., 2020;Kaslow et al., 2020).
Meanwhile, scientists have been actively engaged in several research works to understand the dynamic of the pandemic (Cao et al., 2020;Gnanvi et al., 2020;Tovissodé et al., 2020). A 15 number of research works focused on the use of mathematical models (Taboe et al., 2020;Anastassopoulou et al., 2020;Ngonghala et al., 2020). In this category, several models have been developed to predict the course of the disease (Golinski and Spencer, 2020;Agosto and Giudici, 2020;Kosmidis and Macheras, 2020). These research works have a fundamental role in guiding public health authorities to better manage the pandemic. There were also studies 20 related to the analysis of conditions in which vaccines and other drugs can quickly lead to a community herd immunity to curtail the pandemic Gumel et al., 2021).
However, research works in social sciences (SS) have less been highlighted in the search for solutions against the pandemic (Leslie et al., 2020;Van Bavel et al., 2020). Epidemiological (EP) models and clinical treatments do not always take into account the socio-cultural behaviors of 25 the populations facing a pandemic (Abramowitz et al., 2018). In addition, social scientists are not always able to translate their knowledge on social behaviors into epidemiologically relevant insights (Abramowitz et al., 2018). Beyond bio-medical and EP issues, the COVID-19 crisis has also social and cultural impacts (UNESCO, 2020). Decision support tools to understand and guide behavior during this crisis should also go through the social sciences. An analysis of the 30 adaptability of society to the pandemic, a critical assessment of the solutions proposed by the political authorities, and relations between governments and populations during the pandemic are contributions that SS can provide (Leslie et al., 2020).
The research questions addressed in this study were: (1) What has been the relative productivity in scientic publications related to  what are the range of 5 thematic topics addressed by the published EP and SS papers regarding  do the themes prioritized in both research elds converge or diverge? The objectives of the study were to (1) assess the involvement of the scientic communities in research on  in the eld of EP and SS, (2) analyze the research themes addressed by the authors in EP and SS regarding the pandemic, and (3) summarize the main ndings from the selected papers in 10 both research elds.

Search strategy and papers selection
This study was conducted following the Systematic Reviews and Metanalyses (PRISMA) guidelines (Casals et al., 2014) (Figure 1). The search consisted of identifying all original published 15 articles on COVID-19 using Google Scholar and Scopus databases. To nd the papers, we used the following keywords: COVID-19, coronavirus, SARS-CoV-2, 2019-nCoV, n-CoV, and Pandemic. The following elds were included separately in the search in combination with the keywords: Epidemiology and social sciences. The search was restricted to published articles between 1 January 2020 and 31 December 2020. 20  174 studies excluded (studies covering other aspects than epidemiology and social sciences)  2.2 Data extraction and analysis 10 Data extraction was performed separately by two people following the approach used by Abramowitz et al. (2018). The two separate analyses were then merged and disagreements 4 between the themes and sub-themes were reconciled. From each included papers, the following data were extracted: geographical location (country and continent) of authors, funded or unfunded status of the study and source of funding. To analyze the range of research topics addressed by the EP and SS published articles, themes and subthemes were extracted from a review of full-text articles. Each theme was coded as a binary variable (0 = No, 1 = Yes) when 5 a related subtheme was mentioned or not in the full-text.
Count and relative frequencies were computed and barplots were used to describe provenance of the authors, funded or unfunded status of the study and source of funding. Moreover, proportions of themes were calculated for EP and SS and a comparison test of two proportions was performed to assess the dierence between the two research elds. The analyses were   The majority of the selected papers (66.8%) were not nancially supported (28.8% gave no information about funding and 38.0% stated that they did not receive any funding) ( Figure 3).
In total, 13.4% of the selected papers were funded by European, 12.4% by Asian, 6.4% by American and 0.5% by African and Oceanian institutions, respectively. Moreover, it is noted that, European funding mainly supported European authors (82.7%) and Asian funding mainly 5 supported Asian authors (78.5%).

Relative importance of themes considered
A total of 26 themes were identied with 350 subthemes. Table 2 presents the number (n) and percentage (%) of papers per themes for EP and SS, respectively.
Other themes related to prevention and response to the pandemic were considered. These themes were: ethics (53.1%), funerary practices and burials (53.1%) and political issues (51.8%).
All these proportions were signicantly higher (Prob < 0.05) than those for EP papers (Table 2). 10 Finally, the two research elds showed similar priority (Prob > 0.05) for the themes related to clinical characteristics, health systems, comparison of COVID-19 with past disease outbreaks and modelling (Table 2).

Summary of main ndings on COVID-19
Key results obtained from the reviewed papers are presented below.

Diagnosis
To detect COVID-19, several tests have been developed according to the reviewed papers.
Nucleic acid amplication tests such as real-time RT-PCR are the most widely used and rec-5 ommended test to conrm infection with SARS-CoV-2 . Rapid diagnostic tests detecting viral proteins with the potential to speed up and simplify the detection of active SARS-CoV-2 infection are also used . Studies have shown that in most cases, SARS-CoV-2 becomes detectable in the upper respiratory tract around 1 to 3 days before the symptoms onset and for several days or weeks after the symptomatic period (To 10 et al., 2020). The average time between exposure to SARS-CoV-2 and symptoms onset (the incubation period) is 5 to 6 days, but can vary from 1 to 14 days and it is estimated that in 17% to 25% of cases, the virus may be detectable without developing symptoms (Kronbichler et al., 2020).

Clinical trials 15
Several papers have presented results from clinical trials assessing the ecacy of drugs. The eect of drugs on 3 important outcomes in COVID-19 patients (mortality, need for assisted ventilation and duration of hospital stay) were assessed in most studies (Barnabas et al., 2020).
In the cases of remdesivir, hydroxychloroquine, lopinavir/ritonavir and interferon, little or no reduction in the mortality, need for assisted ventilation and duration of hospital stay were 20 observed (Barnabas et al., 2020). However, Spinner et al. (2020) stated that patients randomized to a 5-day course of remdesivir had a statistically signicant dierence in clinical status compared with the standard care, but the dierence was of uncertain clinical importance. Moreover, there are many ocial vaccine projects subjected to clinical trials (Polack et al., 2020).
The rst authorized and recommended vaccines to prevent COVID-19 are Pzer-BioNTech

Age groups/vulnerable populations
Several studies found a signicant impact of age on the clinical characteristics and outcomes of COVID-19 patients Zhou et al., 2020;Verity et al., 2020;Zhang et al., 2020). The symptoms of the aged patients were more atypical than those of the young patients 5 and were characterized by more comorbidities . Moreover, older patients had more severe inammation on admission and during hospitalization, they received oxygen therapy and experienced more complications with a signicantly higher mortality rate (Verity et al., 2020;Zhang et al., 2020;Zhao et al., 2020). However, more recent studies showed higher rates of severe forms of COVID-19 in younger populations due to the mutations of the virus 10 (Garvin et al., 2020).

Modelling
Modelling techniques have been profusely used in the scientic papers as a tool to assess the COVID-19 disease transmission dynamics and to predict its future course. The most used models were compartmental and statistical models (Tovissodé et al., 2020;Anastassopoulou 15 et al., 2020;Agosto and Giudici, 2020;Taboe et al., 2020). The compartmental model divided the population into dierent sub-populations, such as Susceptible, Exposed, Infected, Quarantined, Recovered, and Dead . Classical and improved versions of the compartmental models have been considered in many studies (Cao et al., 2020;Ngonghala et al., 2020;Taboe et al., 2020). Statistical models included growth models, spatial models, 20 time series models, Poisson models and their alternatives (Agosto and Giudici, 2020;Tovissodé et al., 2020). Another class of models widely used are machine learning models (Gupta et al., 2020;Farooq and Bazaz, 2020). These models have contributed among others to (i) assess the impact of control interventions; (ii) generate short and long-term forecasts; (ii) determine epidemic peak time and size, epidemic size and duration (Agosto and Giudici, 2020;Tovissodé 25 et al., 2020).

Social and cultural analyses
Findings of several studies revealed that communities responded to the pandemic in various ways (Sutin et al., 2020;Croll et al., 2020;Dong et al., 2020;Lesser and Nienhuis, 2020).
Sociocultural factors, such as behaviors, beliefs and practices, aect the responses of the communities (Duan et al., 2020;Jerey, 2020). For instance, while in some countries control and 5 preventive measures, such as lockdown, closure of non-essential establishments and businesses were successfully respected, in other countries these measures were not or partially accepted (Doogan et al., 2020). Other papers have shown several negative psychological eects of social isolation of COVID-19 patients, such as high levels of anxiety, stress, or even the presence of depressive symptoms that can persist after the pandemic (Antunes et al., 2020;Duan et al., 10 2020). Antunes et al. (2020) found that women presented higher levels of state anxiety and trait anxiety when compared to men. An age-related variation was also found, among the youngest (18-34 years) groups showing higher levels of trait anxiety (Antunes et al., 2020).
Economics and political issues  has not only caused a health crisis but also a general slowdown in economic activities, 15 especially for small and medium-sized enterprises (Kim et al., 2020;Song et al., 2020;Xie et al., 2020). It has serious impacts on trade, as well as on public and international policies (Bruns et al., 2020;Miller, 2020;Motta Zanin et al., 2020). The management of the pandemic has given rise to many doubts about the ability of leaders and administrative systems to manage the crisis (Cohen et al., 2020;Shatri et al., 2020). Some countries have undertaken measures 20 to support workers who lost their jobs during the pandemic and to oer help to vulnerable people (Bruns et al., 2020;Meisner et al., 2020;Blustein and Guarino, 2020;Carroll et al., 2020). In addition, policy makers have not only been faced with the arduous task of nding viable solutions to respond eectively to the health crisis but also to the economic emergency to support vulnerable businesses and maintain nancial stability (Cohen et al., 2020;Jerey, 2020;25 Li et al., 2020;Miller, 2020). However, despite its negative impacts on economy, the pandemic is also seen as an opportunity in some studies. Indeed, the current pandemic situation can increase the development of newer technologies (Okyere et al., 2020). These innovations may contribute to ecient ways and means of productions and low-cost productions (Karunathilake, 2020).
Following the onset of the pandemic, several countries faced serious ethical challenges (Jerey, 2020). These include, but are not limited to: resources allocation, rights and duties of workers (Jerey, 2020;Miller, 2020;Cohen et al., 2020;Carroll et al., 2020;Ogden, 2020;Shatri et al., 2020;Sorokowski et al., 2020). These challenges were complicated by a health system and a socio-economic and cultural context of each country (Dong et al., 2020;Lesser and Nienhuis, 5 2020).

COVID-19 Research Perspectives
Although eorts are being made by the scientic communities worldwide to understand and ght against COVID-19, many unknowns remain regarding this pandemic. Some research subjects raised in the reviewed papers that must be addressed in future studies are presented 10 below.
Use of Articial Intelligence (AI) techniques to understand COVID-19 dynamic. AI methods have been moderately used in the battle against COVID-19. AI can help to address many issues posed by COVID-19. For instance, mathematical foundations of AI can be used for real-time spread tracking, early warning and alerts for particular geographical 15 locations and to provide accurate forecasting.
Modelling optimal vaccine allocation strategies within and between countries to maximize health under constraints on dose supply.
Modelling the public health impact of the COVID-19 vaccines.
Assessing positive eects of COVID-19 on the environment and natural ecosystems.
Incorporation of social behavior during the COVID-19 pandemic into mathematical models.
Modelling the impact of environmental factors on the spread of COVID-19 according to 25 the main climatic zones of the world.
Assessing the impacts of COVID-19 on small and medium-sized businesses in Africa.
Assessing the determinants of COVID-19 vaccines acceptance in African countries.

Scientic response to the COVID-19 pandemic
In response to the COVID-19 pandemic, scientic research is emerging at an unprecedented rate.
Between 1 January and 30 April 2020, more than 4,000 publications on COVID-19 were reported on PubMed, with an average of about 33 publications everyday (Sarkis et al., 2020). It would 5 have taken 24 months to reach the same number of articles during the 2009 H1N1 Inuenza pandemic (Sarkis et al., 2020). Between 1 January and 30 November 2020, more than 38,000 Web of Science and 78,000 PubMed articles on COVID-19 were identied (Shapira, 2020).
Behind this abundant scientic publication, our results revealed disparities between regions.
There is a correlation between the regions and countries most aected by the pandemic and the 10 scientic contribution. The regions with the highest number of publications and authors were Asia, Europe, and America while Africa and Oceania had lower published scientic papers. The same trend was observed for research funding and regional collaborations. These dierences between regions may have several explanations. First, the most aected countries are also the richest in the world. Consequently, authorities and research funding institutions quickly 15 understood the issues and provided research institutions with funds. Thus, the dierences may be due to the inuence of rapidly available added COVID-19 research funding (Shapira, 2020).
Second, COVID-19 has also changed the traditional way of working. A new way of working has been clearly observed thanks to the increased use of new technologies, which requires expertise (IAU, 2020). Third, in some regions, there are not enough resources and equipment 20 to conduct some kinds of studies as they were not enough prepared for this unexpected situation (Abramowitz et al., 2018). Moreover, due to the border closure for several months, researchers who should have traveled to conduct their research in equipped laboratories have been blocked.
This prevented several experiments which should be performed.
These results call for strengthening regional collaboration. COVID-19 is completely a new 25 situation, which invites reconsidering the existing forms of collaborations between regions. The lessons learned from previous epidemics, like Ebola or Inuenza can guide. In addition, the creation of new research networks between dierent regions can promote exchanges of knowledge and also pooling resources and providing nancial support. By funding infrastructure and research projects in the poorest regions like Africa, researchers can gain more autonomy. Skills will be gained and long-term collaborations between regions and sub-regions can be eective.
This will help smooth out the knowledge dierences between regions in order to be better prepared for future pandemics.

Epidemiological and social sciences research ndings
Our research showed that the two investigated disciplines (EP and SS) have approached the 5 pandemic in dierent ways and on dierent themes. While EP addressed themes related to clinical trials, diagnosis, incidence, mortality and outbreaks, SS prioritized sociocultural themes and economic impacts. Abramowitz et al. (2018) obtained similar results from their study on Ebola. They pointed out that the approaches used by epidemiological, social and behavioral sciences often seemed diametrically opposed. They found that epidemiology is often based on 10 population data (eg, age, sex) to make general inferences without incorporating local insights (eg, cultural practices, traditional structures) while behavioral sciences used small samples to make sweeping inferences (Abramowitz et al., 2018).
There is a need to combine the two disciplines for more eective responses to COVID-19 or other future pandemics (Moon et al., 2015). Eorts must be made to develop new approaches 15 for interdisciplinary research. Data collection systems integrating both EP and SS variables need to be developed. Where as much of the literature in the epidemic space situates social scientists as cultural brokers (Leslie et al., 2020), this study shows that no area of research should be overlooked when faced with a pandemic, like COVID-19. Knowledge and experiences of specialists from other elds, such as economy and nance, environment or politics need to 20 be integrated through interdisciplinary cooperation and setting up of collaborative projects.

Conclusion
This study highlights the disparities between regions of the world when dealing with a health emergency, such as COVID-19. It also underlines the gaps between scientic disciplines concerning the proposed responses to control the pandemic. Based on these results, creation of 25 international cooperation and collaboration networks between national research centres of infectious diseases is required for an ecient and global response to pandemics. The objectives of these networks will be to prepare sub-regional and national research centres to equip themselves in resources and skills to eectively respond to present and future pandemics. We also suggest the establishment of new interdisciplinary and integrated research mechanisms and strategies in the sub-regional and national research centres. All experts who can bring relevant local contextual, medical, epidemiological, environmental and political information on global health emergencies must be involved. 5