Three physicians, two nursing study coordinators, a human subjects protection coordinator, and a geriatric emergency medicine study coordinator performed interviews and focus groups.
All facilitators were employees of the University of Michigan. PAS, AMS, SMF received salary support from the cited NIH grant.
All physician facilitators were male. All other facilitators were female.
5. Experience and training
Half of the facilitators had prior focus group experience. A marketing professional with experience in qualitative research provided in-person training in focus group conduction and analysis to WJM, AMS, SMF, and PAS in February of 2007.
6. Relationship established
A prior relationship between the facilitator and participants did not exist in most cases. PAS conducted recruitment of site Principal Investigators; AMS and SMF performed site monitoring and in some cases had a prior relationship with participants.
7. Participant knowledge of the interviewer
Prior to any data collection, we provided all potential participants with an overview of the INSTINCT study prior to signing of the informed consent. The facilitator provided a brief introduction prior to beginning an interview or focus group.
8. Interviewer characteristics
PAS, WJM, AMS, and SMF are in favor of tPA use when local conditions ensure that it can be delivered in accordance with published guidelines. PAS is a co-author of the 2007 American Stroke Association/American Heart Association guidelines for the Early Management of Acute Ischemic Stroke. The other facilitators recognized in the acknowledgments section have no specific opinion for or against the treatment.
Domain 2: Study Design
9. Methodologic orientation and theory
We used the taxonomy described by Cabana to categorize barriers to behavioral change from the perspective of the physician.8 To further describe our findings, we used grounded theory to inductively derive additional themes that characterized the transcripts.
All emergency physicians and nurses at each site were invited to participate in the focus groups. Representatives from administration, radiology and neurology were approached based on the recommendations of the local site PIs, thus this was purposive sampling.
11. Method of approach
Participants were identified by the local site investigators and their coordinators.
12. Sample size
Our goal was to achieve participation from several stakeholders in acute stroke care at each site. Our sampling method allowed for prioritization and customization of targeted educational interventions at each site; this was the main objective of the qualitative analysis.
One physician decided not to participate in a focus group after the informed consent was explained as he was concerned that his participation in another stroke study represented a possible conflict.
14. Setting of data collection
We used a large conference room at a hotel during the initial site investigator meeting. At each of the participating hospitals, we utilized ED conference rooms, offices, and classrooms.
15. Presence of non-participants
Our protocol did not allow non participants. Any non-participants were immediately asked to leave if they entered the room where an interview or focus group took place.
16. Description of sample
Our protocol only allowed for collection of occupation. Demographic characteristics of the participants were not collected.
17. Interview guide
This was pilot tested during phase 1 and improved for phase 2. The focus group discussion guide is available as additional file 2 (appendix_focus_group_script.doc).
18. Repeat interviews
Our protocol did not specifically allow for this. In some cases, participants from the initial site investigator meeting also participated at the site barrier assessment focus groups or interviews. As the transcripts were de-identified to protect the subjects, the exact number of times this occurred is unavailable, but is approximately in the 5-10 range.
19. Audio/visual recording
Digital audio recordings were made and transcribed verbatim.
20. Field notes
Field notes were not taken in real time. Digital recordings were rapidly reviewed upon the return of the study team to the clinical coordinating center. In the event of recording failure, the facilitator created field notes based on memory and the discussion guide. Recording failure occurred in only two interviews and in no focus groups.
The initial site investigator meeting focus groups were about 90 minutes. The on-site focus groups were approximately 45 minutes. The individual interviews lasted 20 - 30 minutes.
22. Data saturation
While achieving data saturation is an important aspect of qualitative research, our design did not allow for repeat site visits.
23. Transcripts returned
We did not return transcripts to participants for comment and/or correction. Our protocol did not allow for this and the transcripts had personal identifiers removed to protect the participants in the event of a security breach.
Domain 3: Analysis and Findings
24. Number of data coder:
WJM and JJM performed all coding.
25. Derivation of coding tree:
The coding tree used for initial assignment into the 9 major themes is adapted from Cabana and described in the methods.
26. Derivation of themes
Major themes were derived in advance; minor themes were derived from the data inductively.
NVivo 7 was used for data analysis and management.
28. Did participants provide feedback on the findings
At educational interventions later, although protocol did not allow for collection of this data.
29. Quotations presented
See results section
30. Data and findings consistent (was there consistency between the data presented and the findings)
Questions 30-32 of COREQ address the evaluation of the findings of a qualitative study and are intended for readers of qualitative research. They are included here for completeness. We have attempted to present our findings in this work clearly in a manner that was consistent with the data collected.
31. Clarity of major themes
32. Clarity of minor themes