Research priorities of the Canadian chiropractic profession: a consensus study using a modified Delphi technique

Establishment of the Delphi panel

Delphi panel member recruitment

Eighty-five individuals representing these chiropractic organisations and groups were invited to participate in the Delphi panel. Invitations were sent by email in August 2015, with up to two reminder emails spaced 2 weeks apart. Administrative staff at the CCA office sent most invitation emails to potential participants on behalf of the research team. The CCA has a contact database that comprises email addresses of each of the people we approached for participation. If this was not the case, e.g. representatives of the CMCC and UQTR, the lead author sent these invitation emails directly. The invitation emails contained information about participating in the study, outlined the consent procedure for participation (consent implied by completion of the survey), and a link to the online survey. We also raised awareness of the study by making a presentation at a meeting where key chiropractic profession stakeholder groups were present.

Delphi study procedure

The Delphi study consisted of three online surveys (created via the online platform FluidSurveys™, http://fluidsurveys.com/), and a face-to-face workshop. For each Delphi survey round, email invitations were followed up by two email reminders to non-responders.

In Round 1, participants were asked to suggest research areas that they considered important, and to provide a rationale for their suggestions, within each of the following broad research themes: 1) Basic science: studies investigating theories or mechanisms; 2) Clinical research: studies investigating patient outcomes; 3) Health systems and services research: Studies investigating access to health care, or the quality and cost of health care; 4) Social, cultural, environmental and population health research: studies investigating the health of the general population; and, 5) Other research areas.

Participants were asked to suggest research areas that they judged to be important, and important to the members of the organisation that they represented. If relevant, participants were asked to consult with the constituent members who they represented in order to determine which research priorities were important to them, e.g. practicing chiropractors who were members of their provincial association.

Results from Round 1 of the survey were presented to some participants in a face-to-face facilitated workshop at the CCA National Convention in September 2015. All survey participants were invited to attend the workshop, and 28 individuals attended. Participants were presented the results of Round 1 of the Delphi survey. Small groups, containing representatives from each of the stakeholder groups listed above, discussed the Round 1 results facilitated by a member of the research team. The small group task included a discussion of whether the initial research areas identified were comprehensive, and whether they met the research priority criteria. Members of the small groups could also identify new research priority areas not already suggested. Facilitators ensured that all small group members were given an opportunity to contribute to the discussion. Following the discussion, the facilitator collated any new research ideas, and these were added to the list of ideas identified in Round 1 of the Delphi survey.

All research priority areas suggested by participants were then collated by theme and further categorised into sub-themes. Sub-themes from the Round 1 responses, and from the workshop, were generated independently by two authors (PJHB and PB), and then discussed with the lead author to achieve consensus.

In Round 2 of the Delphi survey, we presented all of the sub-themes and research ideas to the participants, along with stated rationales from Round 1. Participants were asked the question “How important is the following research area for the chiropractic profession in Canada?” and were asked to rate the importance of the sub-themes on a Likert scale (essential; very important; moderately important; slightly important; not at all important). Participants could also suggest new research ideas and sub-themes during Round 2, and these additional areas would be rated for importance in Round 3. Using the results of Round 2, we defined consensus as 70% of the participants agreeing that a research priority was “essential” or “very important”. Although there is no universally agreed proportion for consensus for Delphi studies, 70% consensus is a commonly used benchmark [18].

In Round 3, the results of Round 2 were provided to the participants showing each research sub-theme and the percentage results for sub-themes that reached, and did not reach, consensus. We asked participants to re-rate the importance of each research sub-theme, taking into consideration the panel members’ importance rating as a whole.

Prioritisation process

In the final survey round, all research priority areas that reached consensus in Round 3 were ranked by the participants via 1000Minds (1000Minds software, https://www.1000minds.com/) using conjoint analysis methodology [19]. Panel members were asked to rank pairwise statements by determining which of two randomly selected research sub-themes they considered more important for the Canadian chiropractic profession. The software eliminates other possible pairwise comparisons that have been implicitly ranked as corollaries of questions already answered, thus minimising the number of pairwise-ranking questions for each panel member. This was achieved by the ‘transitivity’ principle; for example, if a participant ranked sub-theme ‘A’ ahead of research area ‘B’ and also ‘B’ ahead of sub-theme ‘C’, then ‘A’ would be ranked ahead of ‘C’ and the software did not ask a question pertaining to this third pairwise ranking. Based on participants’ responses, 1000Minds uses mathematical methods (explained in detail in [19]) to arrive at overall rankings of the research sub-themes, both for each individual participant and also averaged across all participants. We present median and mean rank across all participants for each research sub-theme are presented.

Strengths and limitations

We used a robust consensus method to develop our research priorities, and invited representatives of the main stakeholders of the chiropractic profession in Canada to ensure wide representation. However, our overall response rate was low at 67%, leading to uncertainty about the representativeness of our sample. We did, however, achieve a high retention rate with 93% of original panel members completing the prioritisation activity.

The top two ranked sub-themes had a median ranking of 3, indicating that there did not appear to be a strong consensus for which sub-theme should be ranked highest. This is also highlighted with the percentage agreement of rated importance by the group not necessarily being consistent with the group ranking of individual research sub-themes. The order of the final ranked list of priority research sub-themes should be interpreted with caution.

We chose to not sample a random cohort of practising chiropractors in the Delphi study; we determined that the challenges of achieving a representative group of chiropractors in Canada was too difficult with our available resources. In order to gain as broad an input from the chiropractic profession as possible, participants were asked to canvass the views of their members, including practicing chiropractors; however, this was only a suggestion to participants and we did not measure if participants actually did this. Due to our recruitment process being conducted mainly by administrative staff at the CCA we do not have demographic information about non-responders, hence cannot determine if non-responders were different to responders. We did not implement any measures to include patients of chiropractors in the Delphi panel, so this is another limitation of our results; the research priorities of patients are unknown. Another limitation was the absence of patient groups as stakeholders. In an era where patient-centred care is a priority, including the voice of those who receive chiropractic care may have influenced the results of this study. Finally, we did not include people from outside of the chiropractic profession, e.g. research funders, health systems experts, and researchers from other disciplines.