Methods
We used a case study approach, based on Yin (2008), to study communication via team based care
delivery on a palliative care unit. The tenets of a case study approach are the use and triangulation of
different data sources that represent varying perspectives, such as between non-participant
observations and interview data (Denzin and Lincoln, 1998).
Our study focuses on communication around the implementation of palliative sedation therapy (PST).
PST is suitable to study communication because it is a collaborative intervention that requires the input
of all team members who possess knowledge that will influence the decision to initiate PST (Cornett and
Kuziemsky, 2015).
Data sources
Three data sources were used in our study. The first data source was 76 h of non-participant observation
done of all three nursing shifts (day, evening and night), and all major team activities (including priority
setting each morning and weekly team rounds). Written notes
of observation were created using a template that included information on the individuals involved in
the communication act and the purpose, resources used, and method of communication (e.g. face-to-
face, telephone, written, etc.).
The second data source was 17 semi-structured interviews, ranging in length from 14 min to 1.5 h with
an average of 30 min in length. The interview participants were representative of the healthcare team
and included members from multiple different disciplines (physicians, nurses and allied health).
The third data source was charting documents used on the ward – interdisciplinary progress notes,
medication administration record (MAR), nursing assignment sheet, physician priority board. Documents
also included the Champlain PST Guidelines and Protocols, which can be considered as the ideal “best
practice” document for understanding PST as no unit-specific guidelines had been implemented at the
time.
In total, 404 pages of observation data and 169 pages of interview data were transcribed. The
transcription yielded 314 pages of observation data and 95 pages of interview data that were reviewed
and coded prior to saturation being reached.
Ethical approval for this study was obtained from the hospital’s Research Ethics Board (REB), as the
primary location of the research being conducted. Ethics approval was also received from the University
of Ottawa’s REB.
Analysis
Our overall objective was to identify different aspects of the clinical communication space in order to
model it. Qualitative content analysis (Hsieh and Shannon, 2005) was used for data analysis with two
sensitizing concepts. First is the clinical communication space. Coiera (2000) highlights the need to
understand the relationship between information and communication tasks that lead to medical errors
and other adverse events and suggests that improvement must begin by understanding the activities
within the communication space. A particular need is an understanding of how communication occurs
across different providers.
The second sensitizing concept was an approach for modeling clinical workflows (Malhotra et al., 2007).
The approach emphasizes the need to understand how individual provider workflows integrate into a
main (group) workflow model and how contexts such as time or shiftwork influences goal achievement.
Coding and analysis were performed using QSR NVivo 9 qualitative data analysis software and continued
until theoretical saturation was reached, determined to be the point when a comprehensive
communication space model was developed and no new themes were identified. Analysis was led by the
first author and checked and validated by the second author. The validation process led to revisions or
reframing of different parts of the communication space model, one example being how the different
types of common ground were defined. Different means of external validation of study findings was also
done including discussion with peers, member checking of interview data, and presentation of research
findings to selected study participants.
Results
Figure 1 shows our overall model of the healthcare communication space, incorporating both of the
sensitizing concepts. The workflow modeling approach structured the model into five distinct stages of
purpose, healthcare communication practices and workflows, structure, implementation and common
ground, while the communication space concepts provided details for each stage. We used stages to
emphasize that a communication act is developed sequentially where each stage builds upon the
previous one. Staged models have been used to represent other healthcare management concepts such
as transformation (Golden, 2006) and quality improvement (Kellogg et al., 2017).