Tuesday, April 20, 2010

Using Delphi Technique in a Consensual Curriculum for Periodontics

Fried and Leao conducted an experiment using the Delphi technique to obtain a group consensus on the curriculum for periodontics. The goal of this study was to use the Delphi technique to identify a consensus about what topics should be included in a periodontics curriculum for undergraduate dental students.

Nine periodontics faculties at dental schools in two Brazilian cities participated in this curriculum planning process. Forty undergraduate professors, all dentists, were invited to participate in the study. For the first phase of the experiment, lecturers were asked to list, in writing, items that should be included in a periodontics curriculum for dental students. Suggested items were split into two groups. The first group involved theory-related items associated with foundational concepts and basic principles of periodontics. The second group of items, related to dental practice, was associated with laboratory training and clinical experience with patients.

For the second phase of the experiment, previously obtained responses were scrutinized and collated into items. Then, a comprehensive Likert-scale questionnaire was compiled for submission to the panel. The questionnaire offered the following options for rating the importance of proposed curriculum items: "indispensable"; "important"; "relatively important"; "of little importance"; and "should not be included." The researchers sent the questionnaire to all forty respondents and evaluated the answers as follows:
  • item was "kept" if it reached a "50 percent plus one" consensus classification as "essential/important"
  • item was eliminated if it reached a "50 percent plus one" consensus classification as "of little importance/should not be included"
  • item was included in the next round of the questionnaire if it did not reach a "50 percent plus one" consensus agreement or a "50 percent plus one" consensus rejection
During stage three, the only participants were individuals who provided the most extreme positive ratings or the most extreme negative ratings during the previous stage for items where no clear consensus was reached.

For the fourth and last stage, for each non-consensual item singled out in the third stage, the positive and negative justifications required from third-stage participants were transcribed into a new questionnaire. All the original participants were then asked to for inclusion or against inclusion of each pending item. The format used for the fourth stage questionnaire was that of a sequence of items, each of them carrying its associated positive and negative justifications, followed by corresponding "yes" and "no" answering options. Frequencies for the answers thus obtained were calculated with the SPSS statistical package.

Out of the forty initial participants in the study, six (15 percent) eventually dropped out. Two moved abroad, and the other four were unable to stay involved until the end of the experiment. In accordance with the literature, this 85 percent response rate may be considered as "good."15 Out of the thirty-four lecturers who participated in each of the four Delphi technique stages, thirteen (38 percent) were females, and twenty-one (62 percent) were males. They had an average of ten years (±8.19) teaching experience.

Throughout the four stages of the consensus-building Delphi technique approach used in this study, participants identified various up-to-date scientific research trends within periodontics. However, in spite of that, items associated with some recent technological advances were not included in the final set of "indispensable" items. Items related to technological advances in odontology, such as human molecular genetics and DNA probing, were initially listed as important to a periodontics curriculum. However, half way through their study, they were excluded from the final list of selected items. The results of their study indicate that some issues such as a public health-oriented vision, were not addressed by the panel. This suggests that the majority of the individuals involved are focused on the treatment model rather than on health promotion. However, such a drawback may also be credited to limitations of the Delphi technique, which does not allow an interpersonal discussion of topics. Overall, use of the Delphi process allowed accomplishment of the primary goal of the study, which was to identify a consensus about what items should be included in the periodontics curriculum for undergraduate dental students based on the perspectives of a panel of faculty who teach periodontics at several different institutions.

Source: http://www.jdentaled.org/cgi/content/full/71/11/1441


  1. I really enjoyed this article. While periodontics may not be the most riveting subject, I appreciate their scientifically rigorous approach to the technique. While a previous article mentioned removing the upper and lower 25% of responses but didn't mention criteria, use of a Likert scale seems like an excellent way to accomplish this. I also thought that removing the participants who were most generally in agreement was a unique approach which makes quite a bit of sense.

  2. Very interesting study in a niche field, however one thing that jumped out at me was their response/attrition rate. How long did this study last in general and while 85% is still rather good, is there any way the researchers could have expedited the study in some fashion to provide for a better response rate?

  3. This study keep getting better and better This is an excellent resource! Thanks for the great write up!