Introduction

Due to the rapid spread and unforeseen course of the pandemic, the high contagiousness of the SARS-CoV-2 virus, and the high mortality rate, the COVID-19 pandemic presented itself as a worldwide crisis [1,2,3,4,5]. Health systems were heavily exposed to this pandemic and faced new challenges to medical-professional interactions [5, 6]. In Germany, this crisis revealed who or what was coined "system-relevant" [7, 8], i.e., necessary for the maintenance and continued existence of the healthcare system [6]. The first COVID-19 wave (March 2020 – May 2020) hit the German health system mostly unprepared [9]. The health policy, as well as the system, was not sufficiently prepared at that time. The general public first noticed the effects of the pandemic in the public health sector, however the first place where the effects became visible was the clinical sector, especially the intensive care units (ICU). Initially, research, media and particularly political focus were primarily on this sector in its fight against the virus [10,11,12,13,14]. Nevertheless, a comprehensive approach to managing the COVID-19 pandemic should also focus on primary care, as the outpatient sector was also caught unprepared by this pandemic: no or inadequate personal protective equipment (PPE), no instructions or measures for pandemic preparedness (PP), and initially hardly or insufficient support from health policies [15]. Albeit at this time, six out of seven COVID-19 patients in Germany were mainly treated in primary care practices first [2, 14, 16].

Accordingly, the prolonged COVID-19 crisis has triggered government and policy reactions that intended to protect the ICUs from being overburdened by maintaining the primary care sector functioning [17, 18]. Especially with the beginning of the second pandemic wave (September 2020 – February 2021) in Germany [9], primary care played an essential role in the care of infected patients and was put at the forefront of providing essential support in containing and diagnosing COVID-19 [19,20,21]. New regulations posed organisational, structural, and content-related challenges in German general practices [21,22,23]: e.g., the emergency service being used for COVID-19 testing, as well as the implementation of special infection consultation hours [24]. Primary care was set to be the first point of contact where decisions had to be taken on whether patients should be referred to the hospital for further treatment. Patients with less severe and uncomplicated symptoms were entirely cared for in primary care [25, 26]. Furthermore, preventive care, e.g., vaccinations, was widely accomplished by General practitioners (GPs) in their practices [19, 20, 27, 28]. In this pandemic, the primary care sector had to maintain basic health care structure, especially if other parts of the healthcare system were in danger of overloading [4, 21, 29]. Therefore, German GPs functioned as gatekeepers who provided authorisation for access to hospital care by mitigating the risk of overburdened clinics and delayed specialist care [13, 30, 31]. Former studies have shown that primary care and its actors were under intense pressure and faced new challenges and difficulties during the COVID-19 pandemic [23, 26, 32] and has thus presented new and fast-changing tasks for GPs to maintain outpatient care [19, 21, 26, 33]. Furthermore, studies regarding previous pandemics demonstrated a negative impact on the GPs’ perception of their role. Accordingly, previous research also showed that the COVID-19 pandemic, with its policy measures as well as its regulations, negatively affected the GPs’ workload and, subsequently, their wellbeing [34, 35].

The international collaboration study PRICOV-19 – Primary Health Care in times of COVID-19Footnote 1 aims to assess the impact of the COVID-19 pandemic on the organisation of primary health care. The German part of this study focuses on the impact of the pandemic measures and health policy regulations on GPs’ perceptions and wellbeing during the second wave of the COVID-19 pandemic in Germany (February 2021). It is assumed from the research aforementioned that pandemic-related changes due to associated political actions would manifest themselves in the role and tasks of the GPs, affecting their workload, as well as their mental wellbeing. Accordingly, this paper addresses the following research questions: (1) How were changes in tasks on primary care and patient treatment perceived by GPs?, (2) What was the role of GPs during the second wave of the pandemic, and how was their wellbeing?, (3) How did GPs perceive the health policy measures?, and, (4) What influenced the perceptions and attitudes of GPs on health policy measures?

Methods

Study design and setting

The PRICOV-19 study uses an online self-reported survey to conduct a multi-country cross-sectional design in 38 European countries [36]. In Germany, this study was conducted by the Institute of General Practice of the Friedrich-Alexander University Erlangen-Nürnberg (FAU).

Development, translation and validation of the questionnaire

The overall online questionnaire for all participating countries of this study was developed at Ghent University in multiple phases, including a pilot study among 159 GP practices in Flanders (Belgium). This questionnaire was developed based on a literature review and theoretical framework on quality of care [37] and patient safety culture [38]. More details are described in the study protocol (Additional File 1), including the questionnaire as an additional file [36]. The validated English questionnaire was provided to each research partner to be translated into the country’s primary language. In Germany, we used a forward-backwards method to guarantee contextual conformity in translating the questionnaire into German. Two German GPs from the Institute of General Practice Erlangen independently reviewed the translated questionnaire and adapted it where necessary. Furthermore, each country could add up to ten items to the questionnaire to pursue its own research question and focus on the context of its own country, as well as primary care and health system. The amended German questions were tested for comprehensibility through cognitive interviews [39] with four GPs and two non-GPs (research associates).

Structure of the questionnaire and measures

The main questionnaire (Additional Files 2 and 3) consisted of 53 items comprising seven topics, as seen in Table 1. Participants could also add further comments, suggestions, and feedback in an open text field. The open text field was the last item of the questionnaire, its limit was 500 words. All participants gave written informed consent on the first page of the online survey.

Table 1 Main content areas and items of the questionnaire in the respective ordera
Full size table

Furthermore, another three topics with ten items were added to the German version of the questionnaire (Additional Files 4 and 5), which mainly dealt with the context of primary care in Germany (Table 2). The focus of these particular questions lies primarily on the impact of the German health policy measures on primary care from the perspective of the GPs.

Table 2 Items added for Germany in particulara
Full size table

Recruitment and sample

The recruitment in Germany occurred nationwide from 01.02.2021 to 28.02.2021. The target (n) for Germany was 200 practices, following Ghent University requirements on specific ratios per country (n = 200 for countries with more than 10,000 GPs). We assumed a response rate of 10%. The online questionnaire was directed to GPs, internists working in primary care, and GP trainees who work in single or group practices in rural, suburban or urban settings. For better readability we will refer in the following to all participants as GPs. We recruited in a snowball procedure to gain quick and broad access to the research field via anonymised e-mail dispatch lists from the general practice institutes’ research practices (FAU Erlangen), other general practice institutes (throughout Germany), regional German general practices networks, as well as the Bavarian general practitioners’ association. The participants could access the questionnaire through a link, which guided them to the online platform REDCap [40]. No incentive was given to the participants, and no reminders were sent. We sent out the survey link to 1,710 GPs. To track the data collection and the targeted n, weekly updates were sent from the REDCap server regarding the number of the already completed questionnaires. All data were collected anonymously and stored on the secure Ghent University data servers [41].

The reporting of this study is based on the STROBE (Strengthening the Reporting of Observational Studies in Epidemiology) recommendations [42] and the CHERRIES (Checklist for Reporting Results of Internet E-Surveys) checklist [43].

Data analysis

All statistical analyses were carried out with IBM SPSS Statistics 28. Boundary values of p < 0.05 were set for statistical significance. Data and syntax can be found in the supporting information files (Additional Files 6 and 7).

Firstly, descriptive analyses were conducted for the following topics: sample and cohort characteristics, perceiving the changes in tasks on primary care and patient treatment, perceiving the role of GPs and task shifting, perceiving the wellbeing of GPs (expanded 9-item WellbeingIndex, eWBI), as well as the assessment of the health policy measures (PolicyIndex, PI).

Secondly, two additive indices were created: The eWBI describes the frequency of distress and wellbeing among healthcare professionals. In this paper, the validated expanded 9-item Wellbeing Index version was used according to Dyrbye [44]. Cronbach's α was used to check whether all variables could be combined into an additive index (Table 6, Items: 6.0—6.8). The internal consistency was satisfying, with Cronbach's α for a positive effect of 0.76 for the eWBI. Within the eWBI, seven items had to be responded to with a yes (scored as 1) or no (scored as 0). Two items that had to be answered on a 7-point Likert scale were converted into a new variable: response options of a scale of 1 or 2 (a low level of meaning in work) were assigned the value + 1, response options of 3 to 5 (a neutral level of meaning) was given 0, and response option of 6 or 7 (a high level of meaning in work) was assigned the value -1. The 5-point Likert scale was recoded to the same scheme: those responding ‘strongly disagree/disagree’ having + 1, those who responded ‘agree/strongly agree’ to having -1, and 0 for those with middle neutral responses. As in previous studies being at risk of distress is defined as a score of ≥ 2 [44]. The PI was built on the same basis as the eWBI. The PI was used to aggregate the GPs' perceptions and attitudes towards health policy measures on primary care during the pandemic. It was composed of the variables of a 5-point Likert scale (Table 8, Items: 8.0—8.3). The internal consistency was proved satisfying, with Cronbach's α for a positive effect of 0.69.

Thirdly, inferential analyses were conducted using correlation and multiple regression models. Bravais-Pearson [45] correlation was first calculated to see if the two indices correlate to ensure they can be used in the regression model. Regarding the multiple regression model analysis, we used the ones with the best model quality (R2). Three models were calculated, with the metric PI as the dependent variable. The independent variables were introduced in three steps: (1) variables that deal with the change of role and tasks of GPs (2), the eWBI, and (3) the practice-relevant influencing factors (cohort characteristics).

Results

Sample and cohort characteristics

Three hundred forty-nine GPs filled in the questionnaire (Item 3.0) with an estimated response rate of 20.4%. At least, we had 250 fully completed questionnaires. Participants who filled out the questionnaire were mainly GPs (59.6%, Item 3.1) with mean work experience of 15 to 20 years (Item 3.2). The majority worked in single practices (62.5%, Item 3.3). Most practices were located in rural areas or in small towns (53.3%, Item 3.4). For further details please see Table 3.

Table 3 German cohort characteristics
Full size table

Perceived changes in tasks on primary care and patient treatment

 GPs agreed that priority was given to suspected COVID-19 cases over other patients in terms of appointments (63.3%, Item 4.0). Also, a high proportion of respondents indicated that patient care for uncomplicated cases/diseases was ensured (85%, Item 4.1). GPs experienced that the impact of the pandemic increased recommending non-COVID-19 vaccinations, e.g. influenza or pneumococcus (70.7%, Item 4.8), and requests for COVID-19 testing by asymptomatic patients (62.2%, Item 4.9). For further results, please see Table 4.

Table 4 Perceived changes in tasks on primary care and patient treatmenta
Full size table

Perceived role of GPs and task shifting

Almost 40% of the GPs agreed that their role had gained more attention. However, an equal proportion also said that they did not perceive this to be the case (Item 5.0). 76% of the respondents perceived their work during the pandemic had a personal added value for them and became more meaningful to them (Item 5.1). Next, 80.3% of them reported that their role and tasks in primary care had changed since the beginning of the pandemic (Item 5.2), with increased responsibilities (70.6%, Item 5.3). Not even half of them felt prepared for the task shifting in their professional role (49,2%, Item 5.5). 42.2% stated that they were unhappy with the task shifting in their professional role (Item 5.4). Further results can be seen in Table 5.

Table 5 Perceived role of GPs and task shiftinga
Full size table

Perceived wellbeing of GPs (expanded 9-item WellbeingIndex, eWBI)

Half (50.5%) of the GPs stated that they felt burned out from their work (Item 6.0) and had less time for leisure time activities (38.0%, Item 6.8). 88.5% said their work during the pandemic was meaningful to them (Item 6.7). More details please see Table 6.

Table 6 Perceived wellbeing of GPs (eWBI components)a
Full size table

The risk of distress is defined with a score of ≥ 1 (median) according to Dyrbye [44]. Considering the eWBI, the mean is at + 1.2 (SD = 2.6), with 53.4% of the GPs situated above + 1.0. This indicates that more than half of our GPs surveyed were at risk of distress due to COVID-19 during the second wave of the pandemic in Germany.

Perceived health policy measures (PolicyIndex, PI)

Around 44.5% of the respondents agreed that the government's policies and regulations on primary care practices threatened the good practice organisation (Item 7.0). Besides that, 43.3% reported that the policies threatened the personal wellbeing (Item 7.1). Furthermore, the majority agreed that no adequate government support was given (75.2%, Item 7.2), and 66.4% considered that the policies overwhelmed and negatively affected the daily life of their practice (Item 7.2). Please see Table 7.

Table 7 Perceived health policy measures (PI components)a
Full size table

The PI index has a range from -4.00 to + 4.00 points. The lower this score, the more GPs agreed with the health policy measures; the higher this score, the less GPs agreed with the government’s policies and were more dissatisfied. The mean was + 0.7, and the median was + 1.0. We defined being less in agreement with the policy measures as a score of ≥ 1 (median) according to Dyrbye [44]. Almost 60% of GPs were above the cut-off point of + 1.0, meaning that there was a rather high degree of dissatisfaction among the German GPs with the health policies during the second wave of the COVID-19 pandemic.

Correlation of eWBI and PI

Table 8 shows a highly significant two-sided relationship between the eWBi and PI indices. R = 0.44 corresponds to a medium negative effect, e.g., the higher the eWBI score (and the lower the wellbeing), the more negative the GP’s health policy attitude. In the context of our previous results, this means that in the second COVID-19 wave in Germany, we found that when GPs were in greater distress, they rated the health policies more negatively.

Table 8 Bravais-Pearson-Correlation between eWBI and PI
Full size table

Multiple linear regression models

All three models (Table 9) were highly significant (p < 0.001). In the first model, a significant influence on GPs' health policy attitudes was exerted above all by the perception of whether one felt happy with the task shifting (β = -0.249, p < 0.001) and whether one felt unprepared (β = 0.223, p < 0.001). The most substantial negative influence was dissatisfaction with the change of tasks in practice, connecting a lower satisfaction to a higher PI score. In the second model, adding the factor eWBI, which had a highly significant influence on the PI (β = 0.347, p < 0.001), reduced the impact of the previously significant variables. This means that the more GPs were exposed to the risk of distress during the pandemic, the worse they assessed the government's measures. In the third model, the significance of the eWBI (β = 0.352, p < 0.001) did not change by adding the cohort characteristics, which did not influence the PI. We therefore assume that the cohort characteristics have no influence on the PI, neither directly nor indirectly mediating via the eWBI. It can be assumed that the GPs assessed the health policy measures irrespective of their function, practice location, type of practice, or work experience. The respondents' wellbeing predominated as the most influential variable.

Table 9 Results of linear mixed model analysis of potential predictors for the GP’s perception of health policy measures (PI score, PolicyIndex) during the pandemic
Full size table

Discussion

Our main findings show that the German GPs experienced a considerable change in their work and practice organisation, as well as rapid changes in tasks. They felt a high responsibility within the COVID-19 pandemic and found that their work has become more meaningful since the pandemic. On contrary the GPs also saw a lack of political support, and many felt that they were not well prepared by the government. Furthermore, they found that the measures taken by the government have overburdened the daily practice. However, the wellbeing index showed that many respondents did not feel very strongly affected by their psychological wellbeing, but rated the implementation of the policies and especially the impact on the organisation of the practice rather negatively. The GPs' wellbeing is significantly related to how the respondents rated the German health policies: The more dissatisfied the GPs were with the task shifting, the more negative they felt about the policies decided from the government. It follows, that from the perspective of the German GPs, the health policy measures had a negative influence on everyday practice rather than being supportive. If the practice team and the GPs felt unprepared, experienced little support by the government and were dissatisfied with the changes in tasks, they evaluated the political actions more negatively. It could be deduced that if GPs were dissatisfied and overwhelmed with the measures, they may be less able to implement them. Accordingly, active inclusion of the GPs in health policy decisions and recommendations for action and a participatory approach should be considered.

A study by Wangler et al. [46] showed that GPs, based on their own insights and experiences, do not just want advice, action and suggestions from other experts, such as policymakers. Therefore, it seems advisable to involve GPs and their experience in the planning, implementation and evaluation of measures to preparedness planning and in contingency plans [46]. Other studies [15, 21] have also highlighted the need to involve GPs in preparedness planning and showed that GPs gave low ratings to their preparedness for a pandemic. According to Siebenhofer et al. (2021), primary care is important for dealing with pandemics like COVID-19. GPs in that study said that they were confident and willing to take an active role, but they need to be provided with the appropriate ambient conditions, and information and especially be included in communicative exchange [21]. Other research [10, 47] also demonstrated that the COVID-19 pandemic had negatively affected healthcare workers in the primary ambulatory sector. Those seemed to be the most vulnerable group due to increasing responsibilities and transforming tasks. Similar findings [34, 35, 48] showed that GPs experienced a new role marked by a high-stress level combined with a deleterious work environment caused by the COVID-19 pandemic and its health policy measures. Collinset al. [34], concluded within the international comparison of the PRICOV-19 study, that GPs with less experience, in smaller practices, and with more vulnerable patient populations had higher scores in the eWBI, thus being at a higher risk of distress. Furthermore, Groenewegen et al. (2022) found, that all over Europe, GPs and their staff members were more involved in giving information and recommendations to patients contacting the practice by phone, they were more engaged in triage of whom to refer to the hospital, and that they experienced more responsibilities [26].

Strengths and limitations

The strengths of our study are that it represents a gap in research on German primary care during the second wave of the COVID-19 pandemic in February 2021. It gives an important snapshot of the primary care practices within this time frame with a validated questionnaire and a high response rate. Furthermore, regarding our data and as it is an international cross-sectional study design, comparison with other countries becomes possible and could be an incentive for further evaluation and analysis steps.

We also see several limitations to this work. It can be assumed that the GPs who took part in the survey were rather more dissatisfied with the situation than the basic population. We therefore assume the existence of a selection bias. However, the demographic data collected does not indicate a clear major shift in the composition of the sample and the basic population. Additionally, we expect further biases, e.g. social desirability and dependence on the familiarity of the respondent with the practice organisation. Moreover, within our study, no comparison can be made between the situation before and after the pandemic and its waves.

Conclusion

The COVID-19 pandemic led to new challenges, barriers and responsibilities for German primary care. During the second wave of the pandemic, the German GPs of this study perceived their work as relevant and felt confident that they could fulfil their tasks. Nevertheless, they noticed that health policy initially hardly reacted or was delayed in reacting and supporting the outpatient sector. This led to an increased risk of distress and dissatisfaction on the part of the GPs, as an overload and high workload became apparent in the primary care setting as well, and not only at ICU level. High-quality outpatient care during the pandemic could initially only be maintained through improvisation and extreme flexibility on the part of the GPs. What was needed was a transparent pandemic plan, including the primary care sector. Consequently, it is recommended that health policy and its actors ought to take an active interest in understanding the complexity of GPs' work. They could increase their competence in relation to the primary care service, ensure that the needs of primary care are considered in preparedness plans (pandemics, climate change etc.), and include the GPs expertise concerning health policy questions.