Power and complexity in obstetric teamwork: an observational study.

Background In recent years, increasing attention has been paid to team processes in peripartum care settings with the aim to improve fetomaternal outcomes. However, we have yet to understand how the perception of teamwork in peripartum care is shaped in a complex, multi-disciplinary environment. Methods The aim of this study was to approach the question using a qualitative social-scientific methodology. The theoretical foundation of the study was that obstetric teamwork is the result of a balancing act in which multiple goal conflicts are continuously negotiated and managed right at the boundary of acceptable performance in a complex adaptive system. We explored this theory by gathering lived experiences, from multiple perspectives, of successful management of peripartum emergencies in the setting of a German university hospital. Based on our analysis we generated an understanding of teamwork as a phenomenon emerging from interpersonal relationships, complex relations of power, and the enactment of current quality management practices. Results Caregivers often define teamwork through the quality of their collaboration, defined by respect and appreciation, open communication, role distribution, and shared experiences. However, teamwork also becomes the framework for negotiation of many conflicts that originated elsewhere. Power, in various forms, was the core theme that emerged in the analysis of our participants’ narratives, which is in stark contrast to the otherwise promoted egalitarian rhetoric of team training. While our participants generally reverted to explanations based on their professional identities, traditions or cultures, interesting dynamics become visible when work is viewed through the power lens. Conclusions Our study paints the convoluted picture of a work environment with all its intricacies, constraints, interpersonal relations and hierarchical struggles that are much more representative of a complex system rather than the easily tractable environment

that so many stakeholders would like healthcare practitioners to believe in. The issue of power emerged as a decisive factor in the social dynamics at the workplace, revealing hidden agendas in the teamwork discourse.

Background
The dynamic nature of Obstetrics, the risks involved, and the need for constant adaptation on part of the care providers inevitably lead to increased interest in team processes involved in the provision of peripartum care 1 . As the physiological event of giving birth is similar all over the world, it is probably due to this 'common physiologic ground' that different traditions, myths, taboos and practices are socially constructed around childbirth 2 . While often a "medical/social-model of childbirth" 3 is dichotomized as the contrast between obstetrical and midwifery practice, reality usually resembles a much more nuanced negotiation in between: A complex web of interpersonal, inter-professional and hierarchical relationships presents in obstetric care 4 , and success and resilience as a result of this diversity can be considered a 'petri dish' for further examination of team processes.
Current approaches to teamwork in obstetrics resemble the predominant behavioristic, normative approach generally found in medicine 5 , both regarding the scientific basis used for the argument and the discursive language. A trend of simplification can be traced, albeit often disguised by professional pragmatism. As an example, one can consider the use of the concept of Situational Awareness 6 in an obstetric context: "Situational awareness is a concept that was first defined in aviation. In obstetrics, it has been difficult to define and reliably measure […]; however, put simply, it refers to knowing what is going on" 7 . This is in stark contrast to ongoing debates about the general viability of the concept in human factors research [8][9][10] . Moreover, simplicity is explicitly desired: "In order to achieve better outcomes […], it is essential that team-training interventions are simple and relevant to the maternity care setting" 7 . Nonetheless, one can also find a critical discourse around team training in obstetrics, where crew resource management (CRM)-based interventions have failed to demonstrate the desired improvements 11 . Furthermore, in a review of teamwork performance measurement tools in obstetrics, Fransen et al. 12 evaluated six different frameworks available for simulated settings and noted both limited evidence for their psychometric properties as well as the general lack of a "gold standard" for teamwork performance. This mimics other critical reviews on performance measurements in healthcare 13,14 . We question whether this current approach to teach and 'improve' teamwork addresses the needs of practitioners in dynamic, resource-constrained and goal-conflicted work environments 5 .
The knowledge about teamwork in obstetrics remains ambiguous and conflicted. 'First, do no harm' remains the guiding moral imperative for frontline staff to constantly reflect upon and improve their work practices in an effort to increase fetomaternal safety in the perinatal setting . The relevant questions then become both what constitutes, and what shapes the caregiver's perception of, 'good teamwork'? The aim of this study was to contribute to the debate using a qualitative social-scientific approach to teamwork, and explore the theory that obstetric teamwork is the result of a balancing act in which multiple goal conflicts are continuously negotiated and managed right at the boundary of acceptable performance 15 in a complex adaptive system. We do so by gathering lived experiences, from multiple perspectives, of successful management of peripartum emergencies in the setting of a German university hospital. Based on our analysis we generate an understanding of teamwork as a phenomenon emerging from interpersonal relationships, complex relations of power, and the enactment of current quality management practices.

Methods
We conducted a qualitative, exploratory case-study 16 at a large University Hospital in Germany. Participants engaged in a semi-structured face-to-face interview following a preapproved interview guide (see Table 1 and Appendix I). What, in your mind, made it successful? 3.
What makes work successful in general?

4.
Consider a colleague that you perceive as good and successful in working together with others, which qualities makes you put them in high regard? 5.
In your opinion, is there a correlation between good work and good outcome for mother and child?
Interview language was German. The interviews were audio-recorded, and recordings were later transcribed using f4transcript for Mac® (dr. dresing & pehl GmbH, Marburg, Germany) and translated into English. All personal or identifying information was removed during the transcription process. Data was analyzed using Applied Thematic Analysis (ATA). Drawing from a multitude of theoretical and methodological perspectives, its " […] primary concern is with presenting the stories and experiences voiced by study participants as accurately and comprehensively as possible" 17 . Guest,MacQueen 17 describe data analysis in ATA as "locating meaning in the data". Distinct to the oftenencountered idea of sensemaking in qualitative research, this reinforces a measured approach that is cautious of highly imaginative over-interpretation of problematic data.
Our analytic strategy was therefore designed according to recommendations by Guest, MacQueen 17 with the main goal of providing an "audit trail" of the process rather than an analytic "black box" that leaves many questions up to the reader's imagination (for a complete description of the methodology, see Appendix II). Quantitative data analysis is restricted to descriptive analysis of the participants (occupation, experience) and the interview process (number of interviews, average length) and was performed using

Participants and demographics
A total of 13 healthcare professionals were interviewed between June and November 2018. Table 2 provides an overview of participants' roles and experience. Mean interview duration was 29.6 minutes (SD 6.4 min.). All interviews showed the structure specified in the interview guide. As the interview questions #1-3 built on one another, on two occasions participants already intuitively gave answers to subsequent questions during the conversation, which led the interviewer to not explicitly ask these questions again. All instances were reviewed by the research team, which concluded that the interview content followed the logic of the pre-structured script and that structural reliability was not compromised. Quotes are referenced by profession (OB -obstetrician, MW -midwife, ANA -anesthesiologist) and participant number. In the transcriptions, […] denotes that a part of the quote was left out or changed to preserve context or anonymity, while … signifies a pause on the recording.

Results
What is good teamwork?
When analyzed by existing frameworks that describe relevant aspects of teamwork in healthcare 18 , the most frequently addressed theme concerned the "quality of collaboration". For participants, this aggregate construct is defined by the degree of mutual respect and support, trust, and reliability. This was especially relevant for midwives when describing their daily collaboration on the ward. Mutual respect was evoked as prerequisite for the clear definition of boundaries. The concepts of "teamwork" and "team play" appeared heavily interwoven for the participants: From the midwives' point of view, the lack of accountability is not only acknowledged, but even sometimes welcome.

And like I said I am glad that I can say here's the point where I am allowed to hand things off […] Where I can say "this is not my job anymore, that's over now and I don't want to be responsible for this". (MW1)
The aspect of accountability, while partly acknowledged by the midwives, turns into questions of trust within the team.

And that's what [the doctors] are afraid of, I think. To trust the midwife if they know that [if it goes wrong] the next morning their boss will give them [a hard time]. (MW3)
In summary, pursuing questions of accountability proved revelatory of more profound ontological differences. As a midwife explains:

Discussion
When viewed from a more meta perspective, many aspects of our participants' narratives that seemed confusing and contradictory at first revealed two logical paradoxes.
Although glad to be able to hand off responsibility, midwives often voice notions of not feeling valued or appreciated, especially if those responsible disagree with a proposed assessment or treatment option. They feel conflicted about the involvement of additional medical professionals in the delivery process; while generally perceived as outsiders, they might also represent the required resource to avoid adverse outcomes and increase the systems' overall adaptive capacity.
Clinical work is organized in a way that physicians, while formally responsible, are occupied with a multitude of tasks, and can therefore only be superficially involved in the delivery process. At the same time, their performance is not judged with considerations of their local rationality 19, which would take into account what made sense to them at the time, but measured according to outcome-based quality indicators. To deal with these idiosyncrasies, obstetricians not seldomly revert to a "play-it-safe" approach to decision making that defeats any advantages from an integrative, interprofessional, teambased approach. In an exemplary fashion, this shows how the quality agenda is effectively able to hijack safety efforts. This is obfuscated by an accompanying rhetoric constantly appealing to the individual actors' belief that "good teamwork" will benefit the patient and solve any potential problems.
Consequently, teamwork becomes the framework within which many conflicts are negotiated that, in reality, originated elsewhere. The implications for teamwork under these circumstances become increasingly visible when the notion of power is considered.

Power in obstetric teamwork
It has been previously noted that a representation of organizational life without the consideration of power may result in serious shortcomings 20  Power, in its most obvious form, is easily identified in our participant's narratives: Obstetricians have the authority, and thus power, to direct the course of a delivery by indicating medical interventions, e.g. a Caesarean section. Coercive power is executed by superiors when residents have to justify their decisions in hindsight during morning rounds. In more general terms, this dimension of power is perceived as a positive, or driving, force behind successful teamwork. However, it is not tied to hierarchical status, as midwives have the power to involve physicians from a variety of domains, and transfer authority, thus shedding themselves of formal responsibility.
Subtler, the second and third dimensions of power are represented in covert and latent conflict between obstetricians and midwives. This is not only reinforced, but created in the first place by a system that differentiates between formal hierarchy and experience, and places the former in the hands of junior obstetricians on the journey of acquiring the latter. It becomes apparent whenever decisions have to be made by residents against the will of midwives, which hold much informal power through their knowledge and obstetric experience. One resident explains the situation where a caesarean section was delayed, against the resident's better medical judgement, due to this covert power struggle, resulting in an unwanted fetal outcome: The resident also offers a glimpse into the complexity of social interactions taking place at the workplace. Interestingly, these processes rarely fit into frameworks and formal descriptions of how work should best be performed, or organized.
Another source of much informal power lies in information about the course of the delivery, and the power to decide on when to involve the obstetricians on part of the midwives. Although this interaction has been targeted through a multitude of approaches, be it "best practice", guidelines, or team training, such approaches are at the heart of what Dekker, Bergström 4 characterize as essentially meaningless normative rhetorical commitments ignoring the complexity of obstetrics.
If the problems associated with obstetric intervention were merely complicated, the solution would lie in optimizing, through best practice guidelines, the intervention criteria, and sensitivity to evidence of those closest to the obstetric process. But a complex system cannot be reduced to the behavior or compliance of individual components. It is about understanding the intricate web of relationships they weave, their interconnections and cross-dependencies, and the constantly changing nature of these as people come and go and technologies get adapted in use. 4 To connect the aspect of complexity to power and teamwork, it is important to remember the high emphasis our participants placed on relationships and personal knowledge when describing successful teamwork, a factor where proponents of standardization and team training would strongly disagree as to its importance. Again, teamwork rhetoric provides the framework for much organizational negotiation, while more dormant issues like resident training and workload, or discussions about suitable quality indicators and their connection to departmental funding remain below the surface.

Quality vs. Safety
In an exemplary fashion our study into teamwork revealed how the quality-management agenda can effectively influence clinical teamwork in unforeseen ways. In 2015, Germany founded the Institute for Quality Assurance and Transparency in Healthcare (IQTiG) with the aim of evaluating, overseeing and assuring healthcare quality. Although an independent, academic institution, it has a federal mandate to develop instruments that measure healthcare quality and enable the transparent comparison of different hospitals 22 . A standard of care is defined through a set of quality indicators, and hospitals are required to provide associated data that are compared using statistical instruments.
Deviations from the norm result in official inquiries with possible implications for future hospital certification and funding 23 .
For obstetrics, the quality indicators include the APGAR score after 5 minutes, as well as Base Excess (BE) and pH-values of the umbilical cord as surrogates for fetal depression and distress. Repeatedly, our participants described how clinical decision making was influenced by the requirement to monitor, report and justify these parameters. Midwives not seldomly scoff at doctors' decisions to perform caesarean sections, writing them off as pre-emptive, based on an exaggerated need for safety and indication of a lack of traditional obstetric experience. Furthermore even the quality of teamwork is affected, as midwives often see their own competency undervalued and questioned when their suggestions to continue a natural delivery are overruled in favor of a caesarean section.
On the other hand, the reasoning of junior doctors having to "face the heat" in the morning is equally understandable, especially with the consideration of organizational and federal peculiarities.
While these repercussions were certainly not intended when federal requirements for quality reporting were introduced, it shows the (perhaps unintended) consequences of quality management operationalized as an effort to standardize and quantify, thereby ignoring qualitative, interpersonal and social intricacies of a complex workplace. The delivery of care ultimately comes down to the one-on-one interaction of humans that will always require a degree of compassion and dedication that cannot be automated or standardized. It would almost amount to cynicism to state that the system that relies on these properties being exerted by its practitioners, is unable to extend to them the same amount of individual consideration in return. It also reinforces the intractable quality of healthcare as a complex adaptive system 24 .

Limitations
Our study has several limitations. The overall scope of the project establishes clear boundaries in terms of generalizability of our findings. We were always aware that it is an exploratory case-study that might prove to be hypothesis-generating at best, and descriptive of local, insulated and circumstantial phenomena at least. However, through the use of additional data sources and triangulation, we try to provide some frame of reference for the reader to put our findings into perspective. This is also the reason why we try to harness another source of potential bias, our own domain knowledge in healthcare and medical simulation, and for one of us (CN), our continuing involvement in the active management of peripartum emergencies as part of our professional duties.
Interviewing study participants, no matter how well scripted, will elicit biased responses, based on a multitude of factors. All study participants were at least vaguely acquainted with the investigator, and some had previously worked with him on several occasions.
When reading through the interviews, however, the honest, critical and open responses to our questions can be seen as testament to the participants' professionalism, and to the atmosphere and setting the interviews were conducted in.
During qualitative data analyses, it was helpful for the process of reflexivity that the other researcher (DEL) was impartial, and the joint analysis and coding could help reveal and deal with biased interpretation. This is also one reason why we chose to do all the coding and analysis work together, thereby accepting the fact that we would not be able to calculate a Cohen's Kappa as formal inter-rater reliability score, but had to revert to a mere subjective assessment. Also, our analysis work was not subject to external review.

Conclusion
Instead of providing easy answers to our research question, our participants' narratives paint the convoluted picture of a work environment with all its intricacies, constraints, interpersonal relations and hierarchical struggles that are much more representative of a complex system rather than the easily tractable environment that so many stakeholders would like healthcare practitioners to believe in. The issue of power emerged as a decisive factor in the social dynamics at the workplace, revealing hidden agendas in the teamwork discourse. We could show how ultimately, local team processes can become influenced by the pursuit of system-wide quality agendas in a tightly-coupled, complex system. This should lead us to question the kind of support and training obstetric teams require to perform their everyday jobs of providing safe peripartum care, despite sometimes adversarial structural, organizational, logistical or political conditions.