Baseline characteristics are presented as mean ±SD or median (min—max), as appropriate. Categorical variables are provided as absolute numbers and percentages. In the main analysis, environmental variables (country of survey; size of town in which practice was located: <5,000/5,000 to <20,000/20,000 to <100,000/≥100,000) that may have influenced the responses, and the role of the GP within the practice (employed vs. owner), were analyzed using General Linear Models. The main effects and all two-way interactions were analyzed. Bonferroni correction was used to take account of multiple testing. Estimated means and 95% confidence intervals were used to present the results. For a better understanding of the results, responses to the items on each scale were also presented. In this presentation, the response categories “yes” and “yes, probably” and the response categories “no, probably not” and “no” were combined. No statistical correction was carried out to adjust for non-representative samples.

The study protocol was approved by the Ethics Committee of Bond University, Australia (AS200424), Goethe University Frankfurt, Germany (ID 20-619). The study required no ethical approval under Austrian, Italian, Slovenian, Swiss, and Hungarian law.

The survey was filled out by 1,642 GPs from Australia (n = 120), Austria (n = 434), Germany (n = 583), Hungary (n = 190), Italy (n = 112), Slovenia (n = 86), and Switzerland (n = 117). Mean age of the GPs was 51.9 years (SD: 10.8). The majority of GPs were male (52.1%) and practiced in a city with fewer than 20,000 inhabitants (56.6%). All demographic characteristics are provided in Table 1.

GPs gave high ratings to their self-confidence (7.3, 95% CI 7.1–7.5) and their efforts to control the spread of the disease (7.2, 95% CI 7.0–7.3). The decrease in the number of patient contacts (5.7, 95% CI 5.4–5.9), GPs' perception of risk (5.3 95% CI 4.9–5.6), the provision of information to GPs (4.9, 95% CI 4.6–5.2), testing of suspected cases (3.7, 95% CI 3.4–3.9) and their preparedness for a pandemic (mean: 3.5; 95% CI 3.2–3.7) were rated as moderate. GPs gave low ratings to their efforts to protect staff (2.2 95% CI 1.9–2.4).

Our 2020 survey of over 1,600 general practitioners from seven countries revealed that GPs generally had considerable self-confidence and went to great efforts to control the spread of the COVID-19 pandemic. In the first stages, GPs in Europe and Australia were not well prepared to face a pandemic, had to confront a fall in the number of patient contacts and had to make do with unsatisfactory testing procedures for suspected cases. GPs regarded the risks of the pandemic as moderate, but only received limited information from health care authorities on how to deal with suspected cases of COVID-19. Between the various countries, the responses of the GPs differed in all dimensions except for the protection of staff, which was consistently low. Differences are understandable as GPs have to deal with the specific challenges of acting within the primary health care system of their own country (7, 17–25).

In accordance with other previously published papers, the highest scores in preparedness to face a pandemic were observed in Australia (22), and the lowest scores in Italy (20). While in Australia, the role of GPs during a pandemic has been discussed in scientific papers for a long time (26, 27), the topic in central Europe was only a minor issue before the outbreak of COVID-19. As early as 2006, Shaw (28) said it was important that Australian GPs were in a position to continue working effectively during a pandemic, adding that to be able to do this, it was essential that they receive appropriate education, training and equipment. Furthermore, Australian experts discussed how important it is that GPs participate in surveillance systems to identify clusters, and are involved in adapting the system in preparation for a pandemic in good time (29, 30). In Europe, the role GPs might play in a pandemic was discussed in several countries [e.g., the Netherlands (31), Great Britain (32), Germany (33) and Hungry (34)]. However, this research activity in Europe was limited to isolated publications. As Xiao et al. pointed out, not only the appropriate education and training for pandemic situations, but also the general education for GPs is in crucial aspect of preparedness (35).

Pandemic preparedness is a global issue. One year before the outbreak of COVID-19, Gupta et al. recognized that many countries were not adequately prepared to face a pandemic (36). More than 50% of countries scored inadequately in terms of almost 90% of the indicators used to gauge preparedness for outbreaks of infectious diseases. Oppenheimer et al. also found that many countries were not prepared for a pandemic (37). Furthermore, both studies found that an association existed between preparedness and the economic strength of a country. Therefore, countries in Europe and North America were most prepared and independent of region a higher GDP per capita, more public health expenditures as a percentage of GDP and higher density of skilled health professionals per 10.000 persons were associated with a better preparedness (36, 37).

Self-confidence, or the belief in one's ability to successfully accomplish specific goals, and perception of risk, both play an important role in GPs' professional and private lives. We could observe differences between GPs in different countries in terms of both risk perception and self-confidence. Similar results for COVID-19 risk perceptions have already been reported for overall populations, and among health care workers and GPs (38, 39). Furthermore, it is not only the overall level of risk perception that differs between countries, but also the predictors of risk perceptions. While political ideology was an important predictor of COVID-19 risk perceptions in South Korea and the United States, social amplification was significantly more important in Australia, Germany, Spain, Japan, Sweden and the United Kingdom (39, 40). It is also interesting that COVID-19 risk perceptions are not associated with local epidemic severity (38). Risk perceptions are, however, an important predictor of self-confidence (40). In GPs and health care workers, it was found that a higher level of confidence is associated with lower levels of emotional exhaustion, anxiety in general, COVID-19 anxiety, and concerns about one's family, as well as with higher levels of self-perceived preparedness (41–43). On the other hand, such aspects as social support and quality of sleep, and pre-crisis education and training programmes, can also increase self-confidence (44, 45).

The provision of primary healthcare services decreased in all countries, and especially in Germany and Austria, where GPs said that many patients avoided coming to the practice. In Switzerland and Australia, the number of patient contacts decreased the least. Drawing on data from over 80 million visits in 2019 and 2020, the INTernational ConsoRtium of Primary Care BIg Data Researchers (INTRePID) confirmed that a fall in patient contacts was a global phenomenon. Nevertheless, primary care physician consultations remained stable in Australia (19). One longitudinal observational study from Germany comparing the pre- COVID period with April–July 2020 described a significant decrease in GP consultations per week for conditions relating to the total and upper gastrointestinal tract, vertigo, spinal disorders, general fatigue and weakness, as well as in 12 further services (house calls, stool tests, referrals to a specialist, check-up 35, urine analysis, pain therapy, skin cancer screening, electrocardiograms, blood tests, pulmonary function tests, sonography, and wound management) (25). A decline in practice visits was accompanied by a clear shift to telemedicine appointments, as highlighted in a narrative review published by Kichloo et al. (46). The number of telemedicine appointments was particularly high at the beginning of the pandemic, but decreased later (16).

GPs in Switzerland and Australia rated the amount of information they had been provided with more highly than other countries, which is probably because they received guidelines on how to deal with suspected COVID-19 cases earlier (Supplementary Table S1). As in Switzerland and Australia, Slovenian GPs also appear to have enjoyed a more productive exchange of information with health authorities than other countries. Nevertheless, the majority of GPs in our study said that information was available on public media before it was officially provided by official institutions such as health insurers. The problem of inadequate communication with the health authorities was described in an Australian publication as soon as 2015 (27), and further confirmed by Rawaf et al., who concluded from a worldwide survey that primary care professionals were poorly informed by policymakers (47). As insufficient information was available from public stakeholders, many GPs creatively established and used regional networks to share information during the first phase of the COVID-19 pandemic (13, 18, 22, 24, 47).

The availability of diagnostic COVID-19 testing varied substantially across countries. Our analyses revealed that GPs from Italy, Hungary and Austria said that in the early stages, access to laboratory testing for COVID-19 in GPs practices was inadequate and rarely carried out. These GPs increasingly demanded that they should be permitted to decide who should be tested and who should not (see Supplementary Table S1). Since there are differences between countries in the number of tests per million inhabitants, it is to be expected that the GPs assessments would vary (48). The number of tests not only differed between countries, but also developed very differently over time (48). It should be kept in mind that our survey was conducted before routine diagnostic COVID-19 testing was available, i.e., only PCR testing was possible and rapid antigen screening tests did not exist. Circumstances have therefore changed substantially, and there appears to no longer be any shortage of test kits in the countries under review.

GPs in our survey complained that they could not protect their staff, but such complaints were not only heard from GPs, with the British Medical Association, for example, also warning that doctors were at “considerable” risk due to a lack of personal protective equipment (49, 50). Furthermore, the Global Forum on Universal Health Coverage and Primary Health Care, which represents 29 countries, also said they were insufficiently equipped to provide care to protect staff (47). In contrast, a survey of 361 Chinese GPs (of whom 54 worked in hospitals) (43) found that GPs were overall well-equipped and supported during the outbreak of the COVID-19 pandemic, particularly in large Chinese cities. In order to deal with the virus, nearly all clinics provided training and seminars, and gave talks, not only to healthcare workers but also to the general public. Almost all GPs provided information to their patients during consultations, on social media and the telephone, and by means of posters and leaflets in clinics. However, despite all these efforts, only 15% of hospital GPs thought their clinic provided sufficient support to protect staff.

Our study has some limitations. Firstly, the questionnaire was developed in a very short time in order that it could be answered when the situation was most acute. Even though we tried to include all relevant topics, some issues may therefore have been missed. Secondly, we could not calculate the response rate because a systematic area-wide survey was not possible in the time frame we permitted ourselves. Nevertheless, the number of responses far exceeded our expectations, especially considering the difficulties that are usually encountered in recruiting GPs for research. Thirdly, as the recruitment process was conducted through regional networks and professional associations, the participants may not have been representative of GPs as a whole. Fourthly, we know that online surveys are not as suitable for the collection of in-depth information as interviews. Fifth, this cross-sectional survey was carried out in the first half of 2020, since when many changes have occurred in pandemic management. For this reason, we also carried out a longitudinal survey in Germany and Austria, which revealed that physicians in primary care have adapted quickly to new situations and have gained experience in telemedicine, enabling them to overcome changes in the delivery of routine health care (16). Despite involving seven countries, our study was not truly representative. However, the PRICOV-19 study that is being carried out in 37 European countries and Israel is currently ongoing and should provide more information on how GP practices have functioned during the COVID-19 pandemic on a European level (51). One further limitation is that our survey was only carried out among GPs and did not involve other team members from a primary care setting. Nevertheless, the study succeeded in providing an insight into the challenges of providing care in the early stages of the pandemic in a wide range of countries, and it identifies substantial differences between them. A further limitation is that sample sizes vary between countries. These unbalanced sample sizes may have an impact on the results.

Last but not least, the questionnaire was used in different languages. It cannot therefore be ruled out that differences between countries reflect discrepancies in the translation. We used differential item functioning (DIF) to investigate the extent of any discrepancies. This analysis investigates whether GPs from different groups (e.g., countries; pairwise comparisons) that are being assessed using the same scales (e.g., self-confidence) answer individual items in a similar way. The analysis revealed that on two scales (self-confidence, preparedness) and 16 further items, no DIF could be found, while only one DIF was found in 6 out of 39 investigated items. Since the number of DIFs is comparable in countries where the same language is spoken (e.g., Austria – Germany: n = 4; Switzerland- Germany: n = 4) and countries using different languages (e.g., Australia–Germany: n = 3), the DIFs are assumed to be due to aspects other than language (e.g., differences in primary health care systems). The highest number of DIFs could be observed in a comparison between Germany and Slovenia (n = 8) (Supplementary Table S2).