4. Research design
4.2 Data collection
4.2.1 Participant observation
As part of the ethnographic study I participated as an observer. A two months long observation study was conducted to provide general knowledge about the organisation, its practitioners, procedures, norms etc. At this scale video-observations are supplemented by observations of the physical surroundings and affordances of the departments, observations of everyday life activities in the department, both formal meetings and informal gatherings, which will be elaborated upon below. These observations and daily chats map the cultural dynamics of the ward, as the practitioners employed there phenomenologically experience these. Eventually, I came to understand how, in particular situations, cultural and organisational dynamics that operated beyond the individuals were enacted.
I was intrigued by the overall efficiency at the ward, the sublime coordination in the medical team as well as the frustration I experienced when something went wrong. To gain a deeper understanding of the underlying mechanisms in these processes, I initiated systematic, detailed observations that covered various data: standard procedures (how to dress, speak and follow instructions in diagnostic processes), work procedures in diagnostic events (writing the electronic medical record, physical examination of the patient, using the electronic patient board etc.) and the daily workflow and patient flow7. Further, I participated in numerous medical conference meetings and de-briefings concerning treatment situations. I followed a primary doctor during his workday to
7 Written instructions and guidelines for various professions were collected. Moreover, other data such as statistics on workload at the ward, patient-flows etc. were incorporated in my field notes.
understand the division of labour, the level of resources available at the ward, the multiple formal team constellations a healthcare practitioner engages in, and the tasks that need to be completed during a workday. Moreover, I received verbal and written information about overall procedures spanning from medical guidelines and instructions, hygiene procedures, an introduction on how the electronic patient board is managed, what the procedures are for calls from various sources, e.g. 112 (911/999). The observations resulted in written field-notes about general work procedures, rules, educational programmes and team constellations.
At regular intervals the ward receives newly educated doctors. I participated in a general introduction course for new doctors. Likewise, I was given a medical uniform, I learned to give first aid, and I participated as an observer in a number of different treatment situations, where I was introduced to the techniques the medical teams used in diagnostic situations. The intention was to provide me with just enough knowledge of what was expected from the professionals in specific treatment situations that I was recording. My lack of medical and practice specific knowledge prompted me to ask naïve and clarifying questions that challenged fundamental assumptions guiding their work practice. To get a first impression of the workflow, I participated in day and night shifts, had lunch with the healthcare practitioners and often discussed the project with secretaries, nurses, doctors, paramedics and patients. I was soon treated as an employee and the staff often forgot that I was ‘just’ a researcher and were keen on discussing particular medical issues with me.
Paraphrasing Becker (1963), I became a specific kind of practitioner who blended in at the ward.
This part of the study investigated how a cognitive system exploits organisational and cultural constraints (e.g. decision structures, role hierarchies, cultural norms and habits) and, by extension, how they influence how people speak and gesture as they choose between actions by orienting to material resources and organisational routines. The results have explanatory power in the investigations of how non-local dynamics that relate to, for instance, role hierarchies, norms and informal rules related to work practice etc. mesh with situated dynamics in local interaction. Methodologically, this part of the project exploits Connolly’s (2006) underdeveloped observation that the cultural dynamics of what Bourdieu calls a habitus (Bourdieu, 1977) can be studied as part of distributed cognition.
Beside invaluable knowledge about sociocultural dynamics, the aims were to gain trust and understanding amongst the employees and clear the way for regular video observations.
This primary observation study provided me with an authority to investigate naturalistic situations first hand. The ethnographic fieldwork opened up for a possibility to conduct the research project in a trustful relationship with the organisation, but it also provided me with invaluable knowledge about the slower timescales that shape local situational behaviour.
The ethnographic fieldwork provided the warrant for proper interpretation and a basis for optimal video-observation. I aimed for a fluid transition between participant observation and video-observation. The intention was to accomplish as trustful, natural and relaxed video-recordings as possible.
4.2.2 Video-observation
The cognitive ethnographic fieldwork further included video-observations and qualitative semi-structured interviews. Within the tradition of ethnomethodology, micro-ethnography and cognitive anthropology, video-observation is a commonly used method for investigating embodied interaction in general (Streeck et al., 2011, Heath, 2002; Mondada, 2008; Goodwin, 2007). Because video-observation serves as a framework for holistic analysis of interaction involving more than pure verbal utterances (Streeck et al., 2011, Goodwin, 2000a; 2002; 2007), it has obvious advantages. It produces data that serve as a permanent source for documenting and it allows for rich detailed analysis as well as for unlimited reviews of what happens. The approach allows for results based on evidence rather than (faulty) memory. Indirect data in the form of recalled, past incidents has limited value when it comes to naturalistic descriptions. Recorded data overcomes potential biases in the retrospective construction of past events (cf. Mackenzie and Xiao 2012:525).
Specifically, 17 diagnostic treatment situations were video-recorded with up to three cameras over a month. One of the cameras was handheld to cover blind angles or zoom in on specific aspects. I video-recorded alternately at two wards: a sub-emergency and an acute emergency ward. The set-up is illustrated below.
As part of the ethnographic work, I attempted to cover the general workflow and patient flow at the ward. A recording plan was made in order to secure a broad section of treatment situations. Multiple variables were taken into consideration and the plan ensured that the recordings embraced an accurate representation of diagnostic situations in relation to (a) night and day shifts, (b) different workdays, (c) novice and experienced practitioners, and (d) emergent and sub-emergent patients.
I was present during all video-recordings as a silent observer (Phellas et al., 2012) and I took field-notes during all sessions.8 Ideally, a whole treatment situation was recorded from the moment the medical team waits for the patient to arrive (911 calls) or from the moment the patient enters the ward and up until the doctor informs the patient about future scenarios (hospitalisation at another specialist unit, patient handover to another hospital or returning home). Unfortunately, some recordings are incomplete in length (for instance, in cases where the patient arrives before it has been reported to the key personnel at the ward). However, overall, the recordings contain coherent, diagnostic situations.
The medical team includes the specific configuration of medical practitioners – in the widest sense (e.g. doctors, nurses, porters, administrative workers, medical students, lab and x-ray workers, paramedics, police, public health and safety workers) – who deal with the patient. Beside the medical team, a recording includes the patient, and in some cases the patient’s relative(s).
To supplement the analysis of what happened in diagnostic and treatment situations, I interviewed key practitioners afterwards when possible. The qualitative interviews were highly flexible and loosely structured and used structured free recall (Phellas et al., 2012), which encouraged the participant to talk freely about their own immediate view and reflection on a given diagnostic situation. Qualitative interviews are particularly useful, as a method for accessing a subject’s own life world, which includes a set of beliefs, values and attitudes (Seale, 2012:209). As values form and affect the reflective processes and phenomenological experience, it becomes important not only to define what practitioners do and think in situ, but also how they make sense of a situation retrospectively. Often, their reflections diverge from analysis of the actual performance (see also Pedersen 2010, 2012), which will be elaborated in the analyses. Understanding this discrepancy is important to understand how culture is maintained and develops. The perceptions - or the practitioners’ own perspectives - were compared with results based on systematic, detailed analysis derived from an observational perspective. It is, thus, in the light of this frequent inconsistency, that the key to learning is hiding.
In total, the project ended up with a large amount of rich data. The project uses, as primary data; video-recordings and as secondary data; fieldwork notebooks, interview data, and organisational material from the ward. The secondary data serves as a prerequisite for understanding how the sociocultural non-situational dynamics mesh with situated action.
To gain knowledge about the subject and the work practice, and to understand how
8 Momentarily I would place the handheld camera on a table or a chair, so I could write down my observations in a diary.
sociocultural constraints are enacted in situated interaction, I expanded video-ethnography to a broader study of the overall work setting.
4.2.3 Research ethics and anonymity
Before the research study was initiated, the chief physician for education, the executive chief physician and the executive head nurse at the department, and I formed a consultant committee. The aim was to discuss and manage practical, ethical and legal issues in corporation. A 70 pages long research protocol that covered relevant practical, ethical and legal aspects of the project as well as relevant documents, e.g. documents concerning written consent, information papers to all involved participants were worked out by me and sent to relevant key practitioners at the hospital. The project was registered with the Danish Data Protection Agency and the National Ethical Committee was informed about the project. After the project was registered, all departments at the hospital were informed verbally at department meetings as well as in writing, for instance on information boards and on the Intranet. Furthermore, health and service workers, and the ambulance service were informed about the project, as the cameras would be recording their arrival with patients. In such cases it was only possible to get retroactive consent from the involved parties afterwards.
I was responsible for providing informed written and verbal consent from all recorded participants. In relation to patients, I asked for permission in advance when possible. If they were too ill when they arrived, I asked afterwards. If they did not want to participate, I deleted the data immediately. If a patient died during the recordings, the data were also destroyed immediately. I did not record patients under the age of 18. As a silent observer I was present during all recordings and took field-notes, collected written consent from all participants involved and managed all technical issues.
The data made public are anonymised and all patient and practitioner identifiers are removed. Data are only kept physically secure under a triple set of locks. To preserve confidentiality the management could under no circumstances gain access to the recordings. Such access was a big concern amongst the practitioners before they gave consent. The project uses non-invasive methods only and the patients are not contacted after they have been recorded. They were informed that they could contact me with questions or regret their participation at any time as long as their data had not yet been made public, though this matter only came up once.
During the three months I spent at the hospital, I built good rapport with many of the practitioners, and I am aware of how trust, friendship and personal engagement in the daily work could bias my conclusions. However, it is my conviction that there have not been any problems, since all data are anonymised and no parties are being personally confronted with either good or bad behaviour. The aim is to achieve a general knowledge of working dynamics to better understand the systemic dynamics that constrain effective treatment, rather than create a list of specific practitioners that may turn out to perform in an (in)effective manner within particular situations.