Study Design

A five-round modified Delphi study was conducted from August to December 2006. The Delphi technique is a research tool designed to address complex problems with a high level of uncertainty that are not suited to statistical methods or open deliberation.^36,37 Its goal is to obtain the most reliable consensus among a group of experts on a particular topic.³⁶ The technique involves recruiting a group of experts to participate in an iterative process of answering questionnaire, receiving feedback regarding group responses, and revising their opinions in light of this feedback.³⁷ The distinguishing characteristics of the Delphi technique are anonymity, iteration (processes occur in rounds), controlled feedback, (showing the distribution of the group's response) and statistical group response (expressing judgment using summary measures of the full group response). This method was chosen over other consensus techniques due to its ability to allow all group members equal participation and influence, even when separated geographically.³⁸³

Study Participants

Participants were seven individuals who acted as members of the Technical Expert Panel (TEP) for a report on the state of the research of meditation practices in healthcare. Each member lived in the United States or Canada and had expertise and training in meditation practices.

Development of Questionnaires

An initial list of potential criteria for an operational definition of meditation was generated from a preliminary list of key articles.^9,98 Similarly, a list of potential meditation practices was developed based on an initial review of the literature. The items were refined through an iterative review process, until a set of nine criteria to define meditation was found. The first-round questionnaire consisted of two parts. In the first part of the survey, participants were asked to rate the importance of the nine criteria as “not important at all,” “important but not essential,” or "essential. They were also asked to suggest any other criteria that they felt were essential for a working definition of meditation. In the second part of the questionnaire, participants were given a list of 41 interventions and were asked to indicate which interventions qualified as meditation practices based on the essential criteria rated in the first part of the questionnaire. The participants were also asked to indicate any other intervention that they believed involved meditation but that was not represented in the list.

In round two, feedback was provided on the group responses from round one. The participants were asked to reflect on their responses from the first round in light of the peer responses and to either confirm or change their responses accordingly. Based on round one comments, the wording of some criteria was modified, and three practices were added to the list of interventions. The second-round process was repeated until consensus was reached in round five. Round three helped to establish consensus on items for which disagreement persisted. Round four aimed to determine if the criteria considered “essential” to meditation in previous rounds were, in fact, a necessary part of the practices. Participants were asked to indicate which of the “essential” and “important but not essential” criteria applied to each potential intervention. The list of potential meditation practices was refined until consensus was reached.

Study Procedures

The TEP members received a personalized letter describing the Delphi process, and the expectations regarding their participation. The questionnaires were sent electronically. Participants were given up to 1 week to respond to each questionnaire, and nonresponders were sent one reminder. Although participants were aware of the identity of other responders, they were blind to individual responses, ensuring anonymity throughout the process.

Data Analysis

Data from electronic questionnaires were exported into Microsoft Excel^™ (Microsoft Corporation, Redmond, WA) spreadsheets and analyzed with Statistical Package for the Social Sciences for Windows (SPSS^™ version 14.1, SPSS, Inc., Chicago, IL). Categorical data were collected from each survey round and expressed as frequencies. The frequency of endorsement was tabulated for each criteria and practice. A priori, it was established that a frequency of endorsement of five out of seven would be considered consensus.