'We always support each other – no matter what': a qualitative analysis of work-related psychosocial demands, stressors and resources in general practice teams

Background Work-related stress has been associated with poor psychological and physical outcomes. A better understanding of work-related psychosocial demands, risks and resources can therefore contribute to the prevention of work-related stress. Compared to the general population, medical staff have reported a higher prevalence of chronic stress. Using general practices as an example for small enterprises, this study aims to gain a deeper understanding of work-related psychosocial demands, stressors and resources in the primary care setting. Methods We applied an ethnographic design, comprising a combination of participating observations, individual interviews with physicians (N = 6) and focus group discussion with practice assistants and administrative staff (N = 19) in ve general practices in Germany. Access to the practice teams was established via a university general practice network. A grounded theory approach was applied to analyze all data. Results Our results identied specic psychosocial demands, stressors and resources exemplied mainly along two typical tasks in GP practices: the issuing of medical prescriptions and blood sampling. Main psychosocial risks included factors related to work content and tasks (e.g. incompleteness of task), organization of work (e.g. frequent interruptions), and the working environment (e.g. noise). Work-related resources comprised staff’s inuence on aspects related to work organization (e.g. scope for action) and social support (e.g. positively perceived teamwork). These factors did not occur in isolation but were closely interrelated with each other. Conclusion regulations, work processes Therefore,


Abstract
Background Work-related stress has been associated with poor psychological and physical outcomes. A better understanding of work-related psychosocial demands, risks and resources can therefore contribute to the prevention of work-related stress. Compared to the general population, medical staff have reported a higher prevalence of chronic stress. Using general practices as an example for small enterprises, this study aims to gain a deeper understanding of work-related psychosocial demands, stressors and resources in the primary care setting.
Methods We applied an ethnographic design, comprising a combination of participating observations, individual interviews with physicians (N = 6) and focus group discussion with practice assistants and administrative staff (N = 19) in ve general practices in Germany. Access to the practice teams was established via a university general practice network. A grounded theory approach was applied to analyze all data.
Results Our results identi ed speci c psychosocial demands, stressors and resources exempli ed mainly along two typical tasks in GP practices: the issuing of medical prescriptions and blood sampling. Main psychosocial risks included factors related to work content and tasks (e.g. incompleteness of task), organization of work (e.g. frequent interruptions), and the working environment (e.g. noise). Work-related resources comprised staff's in uence on aspects related to work organization (e.g. scope for action) and social support (e.g. positively perceived teamwork). These factors did not occur in isolation but were closely interrelated with each other.
Conclusion Although work processes in general practices are complex and required to comply with legal regulations, the speci c organization of work processes is the responsibility of the general practitioner. Therefore, there are opportunities for practice owners and practice teams to establish working procedures in ways that reduce psychosocial risks and strengthen work-related resources.

Background
The majority of employees in Europe work in small and medium-sized enterprises [1] where the availability of nancial and human resources may be limited to develop strategies for the prevention or mitigation of work-related stress [2]. General practices are established micro-enterprises confronted with healthcare speci c challenges but also with aspects similar across other smaller and larger businesses. This includes, for example, a highly regulated working environment [3,4], as well as nancial competition from other practices and new emerging health business models (e.g. medical care centers), or increasing di culties to nd successors willing to pursue a career within a complex and demanding working environment [5][6][7]. Using general practices as an example for small enterprises, this study aims to gain an in-depth understanding of work-related psychosocial demands, stressors and resources in the primary care setting, an increasingly challenging working environment in Germany [8,9] .
Work-related stress in the primary care environment population and other professions, prior research has shown that the risk of reporting work-related stress is relatively high in health care staff [26], including general practitioners (GPs) and practice staff [8,9,27] (subsequently, the terms "practice staff" or "practice team" include practice assistants and administrative staff to distinguish between the physician and non-physician professions). The reasons are manifold and have, for example, been related to a very high workload, increasing working hours per week or practice characteristics (e.g. high number of patients with statutory health insurance) [9]. Furthermore, a shortage of skilled workers, the economization of the health care system and the increase of administrative tasks can affect job satisfaction of employees and practice owners [6,28]. Although GPs in Germany have reported dissatisfaction with their income, physical and psychological workload [29], they have reported to be generally happy to cooperate with colleagues and practice staff [30]. A qualitative study including GPs from eight European countries identi ed common work-related resources such as a positive occupational pro le, a long-term doctor-patient relationship, or autonomy in the workplace [31]. A study from Germany showed that the opportunity to apply their skills and expertise had the strongest associations with job satisfaction for GPs [32]. This study also showed that practice staff rated their job satisfaction higher than GPs. A good working atmosphere, opportunities for career development, clearly de ned responsibilities within a diverse spectrum of activities, recognition of their performance, the support among the practice team as well as social interactions with patients have been identi ed as major work-related resources for practice staff in Germany [33,34]. They have, however, also reported a number of work-related stressors including high workload, dealing with unpredictable incidents, or a lack of support and/or appreciation from supervisors and colleagues [34]. Like the GPs, they rate their income as rather low [29].
Although previous research has investigated psychosocial working conditions of GPs [29,30,35,36] and practice assistants in Germany [25,33,34,37], there is a lack of in-depth analyses of potential relationships between particular psychosocial work-related stressors and resources affecting entire health care teams. General practices in Germany provide a variety of therapeutic and preventive health care services and procedures, including medical check-ups, vaccinations, laboratory testing, and disease management programs. Regarding the organization and execution of working routines, general practices have to comply with legal regulations (e.g. hygiene standards, quality management [38]) although the speci c organization of work processes is the responsibility of the general practitioner. The practice team has therefore limited in uence on the content of activities regulated by law and speci c to particular occupational groups (speci c training is a requirement for blood sampling; the GP has to check and sign prescriptions). GPs can, however, delegate speci c medical tasks to the practice assistants (PAs) once legal requirements are met [35].

Aims and objectives
To address current health system challenges at the local level -e.g. managing new infectious diseases and dealing with the increasing morbidity due to chronic diseases effectively and sustainably -it is important to gain a better understanding of issues related to occupational health and safety for personnel working in the primary care setting [19] in order to also ensure the quality of health care [25]. This study is the rst work package (WP 1) of the transdisciplinary research network IMPROVEjob (supplementary material A), funded by the German Federal Ministry of Education and Research (FKZ-01GL1751A, 01GL1851D). In the research network, researchers from medical, social and economic disciplines address questions concerning work-related stress and job satisfaction in general practice teams, using the primary care setting as an example for small enterprises [40,41]. The results of this study (WP 1) will contribute to the development of subsequent work packages of the IMPROVEjob study comprising the development and testing of a participatory intervention for the improvement of job satisfaction and the prevention of work-related stress within primary care teams (WP 2 and 3). The intervention is expected to improve job satisfaction and reduce and prevent work-related psychological distress [42,43]. Finally, the IMPROVEjob-Consortium will evaluate options for the transfer of the results into small enterprises of other economic branches (WP 4) which may eventually promote the prevention and reduction of work-related stress in small enterprises.
In this rst work package of the IMPROVEjob study, we aimed to gain an in-depth understanding of how speci c psychosocial demands, associated stressors as well as organizational and social resources in general practice teams are interrelated. During our analysis, we discovered that the interconnectedness of tasks and responsibilities was a relevant factor. We chose the issuing of medical prescriptions and blood sampling as two core tasks in primary care to elaborate upon in this article. The different levels of interconnectedness already become apparent particularly in the responsibility of the practice staff for the care and treatment of their patients. Furthermore, both examples are subject to regulations [44] and speci c guidelines [45] and are therefore not only relevant nationally but also in an international context. Similar to other European countries, in Germany there are over-the-counter and prescription-only drugs. The latter have to be acquired in pharmacies and require medical prescriptions issued by the treating physician only, including regular monitoring and medical supervision of the patient [46]. According to legal regulations, blood collection may be delegated to [44,47]; however, the decision as to whether blood collection is necessary and which tests are required, remains the responsibility of the physician. However, legal regulations alone do not give us information on how the two procedures are actually organized on site by practice teams and how this affects their perception of work-related stress.
Consequently, the following research questions will guide this paper: How do general practice teams organize working procedures related to the issuing of medical prescriptions and blood sampling?
What can we learn from these two examples (issuing of medical prescriptions and blood sampling) about the interrelation of speci c work-related psychosocial demands and associated stressors and resources in general practice teams?

Methods
In this manuscript, we focus on the most important aspects of the research methods applied. Further detailed information on background, research design and methodology have been published in the research protocol of this study [19].

Research Design
To study everyday working life and work-related stress in the primary care environments we applied a team-based ethnographic approach comprising participatory observation, individual interviews with practice owners, and focus group discussions with PAs and other staff (e.g. administrative staff, trainees). We treated each practice as a case that we tried to capture holistically through different data formats and perspectives. Participatory observation allowed to capture aspects of the working day in real time, and point out patterns the participants themselves are not aware of or would be unlikely to disclose in an interview (e.g. interdependence between working procedures and the structural design of the working environment) [19]. The individual interviews with the GPs included particular challenges in the areas of leadership, team and patient care. The focus group discussions captured the collective view on work contexts, working conditions and work processes within the practice team. As established previously [48], the triangulation of methods and data formats [49] allowed us to provide a focused and detailed evaluation of psychosocial demands, stressors and resources in the primary care environment and point out causal relationships which would not become apparent through surveys or interviews alone. In line with the COREQ-criteria [50], the study design conforms to the guidelines for qualitative papers (supplementary material B).

Setting, recruitment and ethical considerations
Prior research has shown that work-related stressors and resources differ within the primary care environment; for example, a recent study [8] reports that female physicians showed a higher risk for emotional exhaustion. In contrast,, the burnout prevalence was higher in general practitioners working in group practices compared to single practices, but group practice employees were more likely to report burnout symptoms than group practice owners. We expected each practice to have unique characteristics, nevertheless sharing strong similarities. To represent a variety of primary care settings, we used a purposive non-random sampling frame [51]. We looked for differences in the characteristics of the practices such as the location, number of doctors and PAs, or the sex of the practice owners. All practice teams agreed to participate in the entire team-based ethnographic approach.
We included three general practices in urban and two practices in more rural areas of North Rhine-Westphalia, a densely populated federal state in Germany. The practices comprised single and group practices (up to 6 physicians) owned and managed by male and female GPs. Practice staff comprised between 5 and 29 mostly female employees including PAs, administrative staff and trainees. We estimate the age of the practice owners between 40 and 60+ years whereas the practice staff mirrored the whole age range of occupational life. While most practice owners were white, a higher proportion of persons of color could be found among the PAs. The gender balance between the practice owners was almost equal, whereas most of the PAs were female.
All practices recruited were part of a general practice network associated with the Institute for General Medicine, University Hospital Essen (Germany) and had signed a letter of intent to participate in this project during the initial funding application for this project. The practices were recruited via postal invitation and telephone. For reasons of con dentiality, we did neither collect any patient-related data nor any personal information about the general practice teams. Practice staff and patients were informed about the participatory observation, individual declarations of consent were signed by the practice staff, and each participant had the possibility to revoke their consent at any time over the course of the study.
Furthermore, each of the three researchers collecting the data (ET, SH, ER) signed a declaration of con dentiality. Ethical approval for this study was obtained from the responsible Ethics Committee of the Medical Faculty, University Hospital of Tuebingen (MR, SH, ET, ER; reference number: 640/2017BO2).

Preparation of the eld work, data collection and management
In preparation of the eldwork, the transdisciplinary research network IMPROVEjob, including all authors, designed an observational framework based on established recommendations [22] with a speci c focus on psychosocial demands potentially relevant in the primary care environment. The framework helped the observers to identify relevant context and to better understand "what was going on". [34]; For example, we used the framework to identify and to structure factors that were likely to be observed (e.g. actual interruptions) or that would rather be discussed during the interviews and the focus group discussions (e.g. different perceptions of interruptions).
Prior to the participatory observations, the observers attended a workshop -conceptualized and conducted by BW -to gain further insight into the particular norms and culture of the setting. Additionally, the researchers (ET, SH, ER) conducted a two-day trial observation in different general practices to gain rst impressions of the setting, its facilities and organizational structures which were developed and organized by BW. The female observers with different professional backgrounds (health sciences, health care and sociology), could thereby explore their own role in the eld and identify suitable areas for the participatory observation where they would attract as little attention and disruption as possible.
Data collection (ET, SH, ER) commenced in February 2018. Each practice was visited daily (Monday to Friday) and in turn by two researchers for 2 to 4 hours [52] to cover a variety of situations and procedures over the course of one working week (e.g. at the reception desk, in the waiting area, the laboratory or consulting room). If appropriate, the researchers took eld notes on site, and observation protocols were written subsequently. The researchers also conducted semi-structured interviews (questioning routes: supplementary material C) with each practice owner (n = 6) and ve focus group discussions with members of the practice teams (n = 19). The interviews and focus group discussions lasted about 45 min -60min. They were recorded and transcribed word-by-word according to a simpli ed system [53] by a professional company. Quality checks, de-personalisation and pseudonymisation of all data were carried out by the team conducting the eldwork (ET, SH, ER). MAXQDA 2018 [54] was used for data management. Subsequent to the analysis, all quotations included in this study were translated from German into English (ET, ER). As we did not apply a conversation analysis approach [55] but focussed on the overall content and meaning of the data collected, we do not expect any signi cant loss of meaning due to the translation.

Data analysis and data validation
Applying a grounded theory approach [56], data collection and data analysis were carried out alternatingly including open, axial and selective coding. The grounded theory approach was chosen to connect the different data formats to each other and to unveil content and topics as well as procedures and practices. It is a process of continuous abstraction of the data: open coding helped getting a rst and creative access to the data. During axial coding, connections between codes where analysed and a linking concept was worked out during selective coding. The analysis was conducted by the researchers carrying out the eldwork (ET, SH, ER) with an interdisciplinary team of researchers of the IMPROVEjob collaboration with expertise in general (BW), psychosomatic (TSD, FJ, FS), and occupational medicine (MR) as well as sociology (CP). This promoted intersubjectivity in determining key topics and core concepts emerging from the data which were included in the subsequent analysis. We completed the alternating process of data collection and analysis when no new conceptual insights occurred during the discussion of the material from the fths practice [57]. To validate the ndings presented in this study, ET, SH, and ER conducted a workshop with three GPs and two PAs who were not previously involved in the research process [49].

Results
In our study, we analyzed how speci c psychosocial demands, associated stressors as well as organizational and social resources in general practice teams were interrelated. Subsequently, we will present how general practice teams organize working procedures related mainly to the issuing of medical prescriptions and blood sampling. Both are subjected to speci c regulations and guidelines and come along with different forms of interconnectedness of procedures and responsibilities. We will describe which work-related stressors and resources emerge during these processes according to the relevant psychosocial demands. Although we aimed to present speci c examples for each of the psychosocial demands included, it becomes evident that many of the quotations could also be categorized into other subsections highlighting the complexity of the working procedures described.

Work content and task
Responsibilities for the issuing of prescriptions as well as for blood sampling are largely determined by legal regulations and required training. In comparison to GPs who are responsible for the entirety of the procedures, the practice staff takes over primarily preparatory and executive tasks. For example, all PAs need to be able to prepare prescriptions and take blood from patients, whereas trainees need to be supervised and other employees (e.g. administrative staff) may support the medical staff, but should not execute tasks requiring medical training.
Two aspects became evident during the preparation of medical prescriptions: (1) the speci c responsibilities of the GPs, and (2) the teamwork required to complete the task. Prescriptions can be issued or reissued directly as a result of the consultation with the GP, or they may be reissued without an appointment. Frequently, the latter occurs during consultation hours when patients ask for a renewal of prescriptions by telephone, email or directly at the registration desk. Before the prescription can be handed to the patient, the GP is required by law to check and sign the prescription. Hence, the GPs can either complete the entire tasks by themselves or delegate parts of the process -the preparation and handing over of the prescription to the patient -to trained personnel. Although each practice observed had slightly different ways of handling prescriptions, staff was always involved in the administrative part highlighting the division of this task through delegation and the dependency of the practice team on the GP to complete the whole procedure: Q1, observation protocol, single practice: "A MA returns to her work station where a waiting patient requests a prescription. The MA immediately begins to prepare and print off the prescription. The patient is asked to wait in the waiting area until the prescription has been signed [by the doctor]." The practice staff (and the patients) had to wait frequently for the doctors to sign prescriptions. Due to the high work intensity observed in all practices, the practice team usually turned to another task (e.g. patient registration, answering the phone) which was interrupted when the prescription was signed and had to be handed over to the patient. This example also highlights that staff are constantly meeting the needs of patients, colleagues or superiors.
Compared to the process of issuing prescriptions, MAs carried out both administrative and medical procedures during the process of blood sampling. The blood sampling procedure usually comprised a prior consultation of the patient with the GP who orders a speci c blood test. Subsequently, the doctors can either complete the whole procedure by themselves or delegate particular tasks to a MA, including the preparation of the necessary administrative and medical equipment, the collection of the blood, and the preparation of the blood sample for the transport to an external laboratory usually undertaken by a laboratory transport service. Some tests (e.g. blood sugar) were carried out directly in the practices whereas the comprehensive analysis of blood samples was completed by an external laboratory. Usually the MAs could execute their part of the procedure without any further consultation of the GP until the results from the laboratory arrived. Then, the GP communicates the results to the patient and is responsible for the nal documentation. Compared to handling prescriptions, the MAs had a larger scope of action during the process of blood sampling including preparatory as well as executing tasks, more in uence on the sequence of the work (e.g. the preparation of the equipment was usually undertaken the day before), and the probability of completing one task at a time was greater.

Organization of work
As noted previously, how work is organized is partly determined by legal regulations and recommendations as well as by the management preferences of the GPs and arrangements negotiated within the team. The subsequent example also highlights the importance of acknowledging the broader organizational context including an assessment of whether work content and task t the training or abilities of the staff: Q2, interview, single practice: "We used to work with ladies who have retired. For them it was a nightmare to work with the computer. Or we used to have somebody […] who was almost deaf. She said: 'I can do anything, but I can't answer the phone.' Ok, here I have to show consideration. But I have to have worked at another work station to appreciate the work of the others and to understand -we do that in team meetings when we have timewhich part do I play in the whole system? And if I do not play my part which processes are interrupted or blocked?" An interruption is a temporary suspension of the current activity which is to be continued at a later time [58]. Across all practices we observed high levels of work intensity. Interruptions through colleagues and patients were part of many work processes although not always consciously perceived as being particularly stressful: I3: Yeah, that is stressing you out, I think so, that is stressful, isn't it? So, hopefully I haven't forgotten anything. Register something, prepare a bill, this and that." Subsequently, a MA describes an approach to deal with interruptions: Q4, focus group discussion, group practice: "Especially at the reception desk, you cannot nish a thing, you really have to put your notepad next to you and write stuff down, bullet points, because otherwise you'll forget stuff. Of course, this shouldn't happen, so everything should be written down immediately and if possible one thing should be completed before the next thing." On the one hand, interruptions due to incoming calls or inquiries from patients were likely to be put on hold till the previous task was completed: The procedure of blood sampling is an example of mostly "do[ing] the same thing" (Q 8). Usually, blood sampling took place during early consultation hours in a speci cally equipped laboratory undertaken by one MA assigned to that task, sometimes for the entire week. As the process of blood sampling was largely prepared and executed by the MAs, we observed interruptions more frequently during the issuing of prescriptions where the practice team is more dependent on the GPs to complete the procedure. On the other hand, we also observed that prescriptions were signed by GPs during ongoing treatments of patients interrupting the consultation process, potentially resulting in spending more time, concentration and energy to return to the original task. For the MAs, however, this is an opportunity to complete their task quickly and hand the signed prescription to a waiting patient: In some of the practices observed, there were particular areas for prescriptions to be signed (e.g. trays, shelf space), some of which were in the immediate vicinity of the consulting rooms. The doctors were able to nish an appointment or a series of consultations before signing documents such as prescriptions stored in designated trays. Doctors were thus able to control their ow of work, and interruptions in the consulting room were reduced.

Working environment
The previous example highlights the interrelation between work organization and working environment, the latter including, for example, the spatial design of the waiting area, the treatment and consultation rooms and the associated work stations including any equipment.
During consultation hours, the registration desk was the most exposed work station, the center of various activities including, for example, short consultations between physicians, the PAs and administrative staff, the registration of patients, handling of prescriptions, arrangements of appointments, and the dealing with a multitude of phone calls: Q11, observation protocol, group practice "Today, the registration desk and the waiting area appear to be busier [than yesterday]. In the registration area, trainee 1 and 2 are making phone calls at the same time, one of the PAs is talking to patients or to a doctor, the waiting room is full, and the patients seated in front of the laboratory talk to each other. . Now, the waiting room is completely packed, and three patients have to stand. Again, I notice that only a few patients say 'hello or good day' at the registration desk, but usually mention their concern immediately. PA 1 is always friendly and mostly replies 'What can I do for you?' PA 2 appears at the registration desk […], she is wearing surgical gloves, looks around and I ask whether she had time to talk about her tasks in the laboratory. She replies that she has to take care of an electrocardiogram rst and leaves the registration area." In some of the practices, the registration area was designed in such a way that a multitude of tasks could be carried out at the same time. In other practices, however, the design of the working environment aimed at the separation of tasks. For example, there were workplaces at the registration desk not equipped with a telephone. There were also practices where separate workplaces were set up in a back-o ce area to handle, for example, administrative tasks or to take telephone calls. Similar to working in the laboratory, patient contact in these areas was limited, interruptions occurred less frequently, the parallel processing of several tasks was less likely, and the noise level was lower. At the same time, the staff working at the registration desk did not have to deal with particular administrative tasks (e.g. scanning of laboratory results) or were relieved of taking phone calls reducing the noise level during consultation hours considerably. Overall, noise levels at the registration desk were, however, not mentioned by practice staff without direct inquiry from the observers.
Another factor discussed was the concern with the feeling of the staff of being under constant observation. Working at the registration desk, all personnel was continuously exposed to inquiries from patients, colleagues or superiors, and micro breaks were important to relax the body and the mind: Q12, observation protocol, group practice "[The employee mentions to the observer] that she is "on display" at her work station. The patients can watch her continuously and she has to be on the spot all of the time. Therefore, she occasionally has to leave [her work station] for a short time to see something else and take a breath." Working in the laboratory, the privacy of the PAs was more protected because the perceived social control through third parties (e.g. patients or superiors) did occur less frequently or not at all as the PA was responsible only for one patient at a time.
The working environment also includes the availability and functioning of technical equipment at different work stations, such as computers, printers and work speci c software (e.g. lab management software). For example, the advantage of investing in several printers at different work stations facilitates smooth working processes, particularly evident in the laboratory environment. All but one of the laboratories visited were equipped with a computer, whereas a printer was available in only one lab. This printer, however, could not be used to print the required forms used in the lab: Furthermore, appropriate software can facilitate laboratory processes including information on particular blood tests or medical prescriptions when staff can access patient data directly in the lab. For example, we observed that PAs checked current medication plans directly at the laboratory computer before choosing appropriate equipment to take blood from patients taking drugs affecting clotting time. This was possible if electronic patient records were available which we observed only in some of the practices.
The laboratories were usually equipped with all utensils necessary to take blood samples (e.g. surgical gloves, antiseptics, used needle containers or disinfectants); nevertheless, we observed that hygiene regulations were implemented to varying degrees of consistency in several practices: Q14, observation protocol, group practice "[The PA] goes to the sink, takes some disinfectant into her right hand to rub a little on her hands. Shortly afterwards, she takes paper towels and wipes her hands again. [The observer] wonders whether this was the entire procedure of hand disinfection. In any case, this did not comply with the hand hygiene protocol poster displayed at the wall […]. Furthermore, neither did she use surgical gloves during the blood sampling nor has she disinfected or washed her hands between different patients." As our project did not intend to evaluate workplace hygiene, the implementation of hygiene regulations was not discussed further in any of the interviews or focus group discussions. During one observation only, one staff member mentioned that hand hygiene gives the patient a sense of security whereas selfprotection was not mentioned at all. As we wanted the staff to discuss factors related to occupational stress as openly as possible, we aimed to avoid any criticism related to hygiene regulations. Spatial design, missing or faulty equipment resulted in additional work for staff, particularly evident in the laboratories and at the registration desk. The availability of additional printers and appropriate software in the lab could facilitate work processes, avoid additional noise or work at the registration desk; from an economical point of view, however, providing additional equipment also creates additional costs.

Social relations
Whereas some of the GPs emphasized their special and long-term relationships with patients, the PAs participating in the focus group discussions highlighted particularly the value of mutual support and team work that we observed across all practice teams: Q15, focus group discussion, group practice Across all practices observed, there were several work stations at the registration desk where PAs could work simultaneously, facilitating mutual support, but also increasing the likelihood of interrupting each other. In the laboratory, work stations were more isolated which sometimes complicated direct communication and support. We observed, however, that PAs took particular effort and care to support each other during busy laboratory hours which also included communication across different work stations and rooms, and additional support from staff designated to different work stations than the lab: Q16, Observation protocol, group practice "[After trying twice] the PA stops the blood sampling and informs the patient that she will get a colleague. The patient says that this is not necessary. The PA insists and says that she only tries twice [to puncture a vein].
[…]. She leaves the laboratory and after a short time a colleague appears, who […] grabs a pair of surgical gloves from one of the two boxes, and sits down at the table [to continue with the procedure]." Over the course of the eld work, all practice teams highlighted the dependency on each team member to deal with the work intensity and ensure the quality of patient care. It was also discussed that insu cient team work can result in a serious burden for both individual team members as well as for the entire staff.

Discussion
Using the examples of issuing prescriptions and blood sampling, we analyzed how general practice teams organize working procedures, focusing on the interrelation of speci c work-related psychosocial demands and associated stressors.
Key organizational differences in working procedures and related psychosocial stressors and resources All professional groups were involved in the handling of prescriptions at various workstations throughout the medical practices. The practice staff usually prepared the prescriptions for the doctors to check and sign resulting in a division of labor and dependency of the practice team to complete the task. In comparison, individual steps of the blood sampling procedure were less intertwined between the professional groups, and the PAs could usually complete their responsibilities in the laboratories independent of the GPs. As summarized in gure 1, key psychosocial risks observed comprised stressors related to work content and task (incompleteness of tasks), organization of work (frequent interruptions, high levels of work intensity, simultaneous processing of several tasks, tightly coupled work processes), and the working environment (noise, missing, unsuitable or unused/incorrectly used equipment/software, the feeling of being under constant observation). Key resources related to work content and task comprised an appropriate scope for action (in uence on the sequence of activities) and su cient patientrelated information, particularly during the process of blood sampling. Furthermore, factors regarding work time (possibility of mini-breaks), and e cient communication and cooperation within the team (clearly de ned areas of responsibilities particularly in the laboratory) were important resources related to work organization. In terms of social relations, positively perceived teamwork was an important resource, and a supportive working environment included access to suitable workstations, equipment and software. Aspects related to "new forms of work" (e.g. atypical forms of employment, geographic mobility, no clear division between work and private life) [7] were not relevant in the material chosen for this analysis. were not always perceived as particularly stressful by the study participants, prior research has shown that a continuously high level of noise, frequent interruptions, ineffective communication or the parallel processing of multiple tasks have a negative effect on psychological and physical well-being [58]. At the reception we also observed the compensatory effect of working together as a team, a well-established work-related resource mitigating factors associated with an intense working environment [16,59]. We also point out the uniqueness of the doctor-patient relationship, comprising long-term rich relationships with patients as well as mutual trust and respect, previously described as an important resource for GPs in Europe [31].
Furthermore, our study highlights that the process of dealing with interruptions affects the staff involved differently. We observed, for example, that doctors were interrupted by practice staff during ongoing consultations or other tasks (e.g. at the registration desk) to check and sign prescriptions for patients waiting outside the consultation rooms. For the GPs, the interruption was a potential stressor, including loss of time and additional effort required to nish their primary task [21,60]. For the PAs, however, the interruption of the doctor resulted in a broader scope of action, a speedier nalization of dealing with a prescription and waiting patients, and a decreased likelihood of being interrupted at another task while waiting for the GPs signature. Interestingly, none of the practices observed decided to issue repeat prescriptions after consultation hours which could reduce the frequency of interruptions for both, the physicians and the practice staff probably mitigating the negative effects associated with interruptions [60].
Work design measures to reduce work-related psychological stress and strain We observed that in some practices the working environment was designed in a way that allowed for areas and periods during which the signing of prescriptions could be undertaken without being disturbed (e.g. speci c desk or shelf where doctors can sign prescriptions between consultations). This also provided the possibility of restorative breaks which have been shown to have a positive effect on physical and mental health in health care staff [61].
Depending on nancial, spatial and human resources, the design of the workplace can mitigate the negative effects of an intense working environment and facilitate the implementation of effective and correct working procedures. In comparison to single work stations, for example, a working environment with several work stations can be louder, bustling and buzzing. Spatial proximity can, however, facilitate mutual support and speed up work processes. Additionally, other factors need to be considered regarding the correct implementation of working procedures. In terms of blood sampling, for example, all the necessary equipment was usually provided in the laboratory areas, but sometimes not handled in the correct way (e.g. disinfectant) or not used at all (e.g. surgical gloves). Our observation is consistent with prior research showing that only 20% of the medical staff in hospitals applied correct hand hygiene procedures [62]. Interestingly, this inadequate implementation of hygiene and occupational safety measures was not actively addressed by the study participants in the interviews. The main reasons for poor hand hygiene reported by Erasmus et al. [62] were a lack of positive role models and of convincing evidence that hand hygiene is one of the most important factors to prevent cross-infection. Other reasons included, for example, insu cient time, skin problems, or not being aware of the right protocols.
Conducting a psychosocial risk assessment in general practices, which is mandatory in Germany [22], could help to identify the underlying reasons for poor hand hygiene and improve the acceptance and consequent implementation of hygiene procedures. Finally, quali cation and job rotation could be a good measure in some settings to equally distribute the different psychological demands, stressors and resources among the PAs.

Strengths and limitations
The study bene ted signi cantly from the cooperation within the transdisciplinary research collaboration IMPROVEjob [41]. For example, the development of the observational framework included the GPs and PAs of the research support group, the members of the scienti c advisory board and the interdisciplinary research group of the IMPROVEjob consortium. The combination of academic and practical expertise contributed to the development of a comprehensive observational framework relevant in the primary care setting [see 63 for comprehensive discussion of transdiciplinarity in health research], facilitated the recruitment process (as we had access to a local GP network), and mitigated challenges of the ethnographic approach (e.g. disruption of daily routines) through comprehensive familiarization with the setting. Furthermore, intersubjectivity was achieved over the course of the analysis (by triangulation of methods and an interdisciplinary analysis team), con rmed through communicative validation by an independent research support group [49]. The applied ethnographic design is a methodological strength of this study, as it was possible to analyze (a) which work characteristics were observed, (b) which were actively addressed in the interviews and (c) which were only addressed upon a narrative request during the interviews. Thus, applying only interviews or even only a standardized questionnaire to assess psychological demands, stressors and resources would have resulted in a less complex understanding of the interrelation between the factors described in this study.
All practices visited were part of a general practitioner network comprising practices that are involved in teaching and training of medical students. This may have resulted in a sample including participants being particularly open and re ective concerning our study approach. Due to the ethnographic approach, we assume that the issue of social desirability is rather low in the data. Over the course of the eldwork, it became evident that especially practice staff are used to being observed at work. Our participatory observation t into a daily routine, but was surely more intimate and more intensi ed as we were also allowed access to sections of the practice where patients were not allowed. Still, full adaption of behavior would not have been possible for a whole week. The participants were simply too busy to constantly pay attention to the observers. The incidents in our data in which participants did not follow protocol or regulations (e.g. hand hygiene) show that participants were either not able or not intending to hide critical aspects of their routines. Furthermore, the focus groups and interviews were conducted after the observations and the participants were already familiar with the researchers.
It should also be noted, that the presentation of our results has intended and unintended "blanks" [64]. For example, the different professional backgrounds of the observers (sociology, health sciences, health care) have contributed to a comprehensive view on reoccurring themes (e.g. interruptions or hygiene standards). Nonetheless, unintended "blanks" have occurred over the course of the observation; for example, we did record the technical equipment or the organization of particular working procedures to varying degrees because the researchers had different priorities depending on the actual situation observed or discussed during the eldwork. Intentional gaps arose where, for example, we observed issues related to the implementation of hygiene regulations but did not further explore motives for particular practices to avoid any sense of occupational hazard control which was not the aim of our research. For reasons of patient con dentiality, we also predominantly observed situations outside the treatment rooms; hence, the results focus rather on the perspectives of the practice staff and to a lesser degree on those of the GPs.
Finally, we chose to report work-related psychosocial demands and related stressors and resources exempli ed mainly along the preparation of medical prescriptions and the procedure of blood sampling.
This means we did not capture all of the psychosocial demands and risks summarized in the observational framework based on the recommendations for implementing a comprehensive psychosocial risk assessment [22]. We therefore also did not capture all of the psychosocial risks or resources relevant in the primary care setting. For example, we did not discuss appointment scheduling within the context of this paper which is, however, a well-known organizational task that has a signi cant impact on the management of consultation hours [65]. In Germany, new regulations for drop-in consultation hours particularly relevant for patients with statutory health insurance came into effect in January 2020 [66]. The impact of putting new law into practice remains to be seen, but issues concerned with an improved accessibility to health care, additional administrative workload and intervening with workplace autonomy are currently being discussed [67]. Yet, many of the resources and work design measures described here have general positive effects on reducing psychosocial stress in primary care practice teams.

Conclusions
The psychosocial risks reported in this study including high work intensity, frequent interruptions, simultaneous processing of several tasks or tightly coupled work processes have been reported relevant in small and medium-sized enterprises of other economic branches [11]. Prior research has also highlighted the importance of positive social relationships in small enterprises where many workers describe their working environment as a place where they were "treated as a person" or "where the boss 'cares'" and "where people 'look out for one another'" [68]. We emphasize that the mutual social support we observed across all practices mitigates work-related stress to a certain degree, but should not be taken for granted for coping with an intense working environment. Putting laws and recommendations into practice, general practice teams may also choose different approaches such as the organization of work or the design of the working environment. Implementing the mandatory psychosocial risk assessment in all smaller and larger enterprises should therefore not be seen as "another administrative burden", but as an opportunity to nd different ways to deal with work-related psychosocial risks. individual declarations of consent were signed by the practice staff, and each participant had the possibility to revoke their consent at any time over the course of the study. Furthermore, each of the three researchers collecting the data (ET, SH, ER) signed a declaration of con dentiality.

Consent for publication
Not applicable.
Availability of data and materials