Background
Methods
GMB compared with traditional research methods
Group Model Building | Focus Group | Delphi method | Grounded Theory interviewing | Interpretative Phenomenological Analysis | |
---|---|---|---|---|---|
Description Researchers selected and assembled… | A group of individuals discuss and comment on a problems etiology [23]. | A group of individuals answer questionnaires on a problems etiology anonymously after which a summary of the groups’ results is presented and individuals have the possibility to revise their answers in reaction to these results in multiple rounds [24]. | A single individual or a group of individuals is questioned individually on their view on a problems etiology [25]. | A single individual or a group of individuals is asked individually to reconstruct their lived experience of a problem [26]. | |
Aim(s) To better understand the problem by… | Breaking down the complexity of the problem, focusing on interaction between contributing factors, visualizing participants’ shared view on these interactions in a model and developing shared commitment. | Analyzing the exchange of views between participants. | Analyzing the often more consensus-based results or explanations of the final rounds. | Developing concepts during questioning and adjusting questioning along the way based on these developing concepts. | Analyzing individually reconstructions of lived experience. |
Group discussion | Yes | Yes | No | No | No |
Types of models created by GMB
Case study
GMB protocol
Sessions, scripts, and adaptations
Name | Description | Goals | Motivation for selection | Adaptation |
---|---|---|---|---|
‘Expert preparation’ (A)1 Session one | Background information on older persons’ ED visits in Amsterdam, system dynamics (GMB and CLD), and the online format are sent to the experts by email before the first session. Experts are asked to write down their ten most important factors contributing to the problem based on their gut feeling. This assignment is derived from Vermaak (18). The agenda and an explanation of the session focus are also included. | To enhance in-session performance by stimulating experts to prepare themselves for the first session. To have a list of factors on paper before the start of the GMB. | See goals. | - |
‘Expert preparation’ (A) Session two | The processed and clarified initial list of key factors contributing to the studied problem are sent to the experts by email before the second session. Experts are asked to reflect on the following questions: Do we need to adjust or add factors? How do factors contribute to the studied problem? The agenda and explanation of session focus are also included. | To enhance in-session performance by stimulating experts to contribute factors to the studied problem. To develop new insights to improve the CLD. | See goals. | - |
‘Expert preparation’ (A) Session three | The processed and clarified initial overview of causal relationships between factors and the studied problem are sent to the experts by email before the third session. Experts are asked to reflect on the following questions: Do we need to adjust or add relations? The agenda and explanation of the session focus are also included. | To enhance in-session performance by stimulating experts to contribute factors to the studied problem. To develop new insights that improve the CLD. | See goals. | - |
‘Expert preparation’ (A) Session four | The processed and clarified CLD are sent to the experts by email before the fourth session. Experts are asked to reflect on the following questions: Is the CLD clear and complete? The agenda and explanation of the session focus are also included. | To enhance in-session performance by stimulating experts to contribute factors to the studied problem. To develop new insights that improve the CLD. | See goals. | - |
‘Entrance’ (A) All sessions | Before starting every meeting, the facilitators ask experts to check for stable internet connection, turn cameras on, mute when not speaking, raise their hand if they have a question, and wait their turn to speak. | To check and set rules for optimal communication. | See goals. | - |
‘Welcome and introduction’ (A) Session one | The facilitators welcome and thank experts for their participation. They share the enthusiastic reactions received from experts during the recruitment. They present an overview of session themes and agenda for the first session. | To explicate researchers’ enthusiasm for the project as well as to share experts’ enthusiasm. Thereby aiming to connect experts by shared enthusiasm. To create a dosed overview of the GMB process for experts. | To have a good basis for the GMB goals in the first session. | - |
‘Hopes and fears’ (S)2 Session one | The facilitators ask experts to write down expectations for the GMB session or project on post it’s. They are then compiled and read out loud before they are added to a board. | To elicit and establish expectations as a group. | By addressing expert and facilitator expectations for the process before starting with the model building, the researchers aim to connect with the experts, make them connect as a group, and boost motivation. | The original script is adapted by asking researchers first and experts second in a round robin fashion to introduce themselves, describe their expertise, and name one hope and one fear. The researchers aim to lower the threshold for taking the floor by starting with themselves. Hopes and fears are directly placed on the Miro board, but shown to the experts only after all experts shared, to keep focus on the person instead of the Miro board. All hopes and fears are explicated by the facilitators. The original time frame was reduced to direct plenary discussion of one hope and one fear. |
‘Presenting the reference mode’ (S) Session one | The facilitators present the problem selected. The experts establish the problem to be studied. | To reach consensus on the dynamic problem. | The model construction starts by finding consensus on the problem and its behavior. | The script is adapted by adding a short PowerPoint presentation recapping the problem. The focus is on finding consensus on a clear definition of the problem. Less focus is on its dynamic behavior over time, because this attributes to the complexity for experts. No graphs over time are used. PowerPoint slides are shared in Zoom and experts vote on the studied problem by raising their hand if they agree. If a hand is not raised, an explanation is asked for. If new suggestions are made for the studied problem, the experts vote until consensus is reached. Consensus on the problem elicitation is expected quickly, so only five minutes are reserved for this script. |
‘Variable elicitation’ (S) Session one | The facilitators ask experts to write down as many key factors contributing to the problem and vice versa. They are then added to a board. | To facilitate divergent group discussion about factors contributing to the studied problem. | The most logical step after reaching consensus on the problem is using a script that asks experts to name key contributing factors to the studied problem, thereby providing a basis for a CLD. | This script is adapted using the preparation assignment for session one as initial input. This gives experts have extra time and resources to come up with a clearly described list of possible factors that are ready to be shared directly after the script starts. The facilitators ask experts to name the factors one by one in round robin fashion. Experts are asked not to name factors that have already been named. Factors are shown directly in Miro. Facilitators and experts are asked to clarify the factors. |
‘Dots’ (S) Session one | The facilitators ask experts to divide three or five votes over factors and place them on a board. | To highlight the most important factors for the experts. | This script is selected to give the experts a first impression on a selection of most important factors for the group as a whole before starting to describe relationships. | Experts are asked in round robin fashion to assign five dots in random distribution to the most important factors. These dots are added to the Miro board directly. |
‘Introduction and recap’ Sessions two, three, and four (A) | The facilitators ask experts for their reflections on the preparation information and assignments. | To check for unclarities, new ideas, and agreement on the CLD. | This allows the facilitators to check and address any unclarities about the CLD. This gives experts a jump start for the next script. | - |
‘Initiating and elaborating a CLD’ (S) Sessions two and three | The facilitators add identified contributing factors to the studied problem one by one to a board. Then the causal relationship with the problem is described. | To form an initial CLD. | This script was chosen because the researchers believed it provides an overview of the problem, which was one of their main goals. | An individual exercise is added to the original script because it is hypothesized that a lot of factors will be identified and experts will find it difficult to describe relationships in a plenary ad hoc fashion. In the exercise, each expert is assigned a factor and has to describe the relationship to the problem and all other mentioned factors. After this, the results are discussed plenary one by one and put on the online whiteboard. The time frame is widened because of the comprehensiveness of the number of relationships expected. |
Closing (A) Sessions one, two and three | The facilitators close the session by summarizing the progress made in the CLD construction and explaining the focus of the following session. | To motivate experts on progress and outline the following step in the CLD construction. | To close sessions effectively. | - |
Model review (S) Session four | The facilitators give a plenary demonstration of the CLD formed. | Experts reflect on adequacy of the CLD formed and the need for alterations. | The script is selected because it clearly demonstrates, stresses, and reflects on the CLD. | Because of its good fit, no adjustments are made except for the online format. The adjustments include screen sharing of the CLD in Miro, capturing feedback from experts in Miro, and reserving more time for feedback. |
Action ideas (S) Session four | The facilitators ask experts to come up with as many ways to improve the system modeled and share their most important one. Experts’ suggestions are placed in a quadrant illustrating difficulty to achieve and size of effect. | To prioritize actions after a model has been developed. | The script was selected because it lets experts think about solutions from a system dynamics effect viewpoint. | Because of its good fit, no adjustments are made except for the online format. The adjustments include screen sharing of the CLD and quadrant in Miro, capturing feedback from experts in Miro, and reserving more time for feedback. |
Next steps and closing (S) Session four | The facilitators present the next steps that will be taken after the GMB session. | To inform the experts on follow up. | This script is chosen because it provides clear closure. | Because a detailed agenda of the study is shared with the experts repeatedly, less time is reserved for this script. |
Roles and facilitation manuals
Expert selection
Data collection, data analyses, and model validation
Results
Session implementation and adjustments
Name of script | Adjustments |
---|---|
‘Expert preparation’ (A) Session two | The information experts receive is paraphrased and categorized in an overview of key contributing factors. The assignment includes the question: Do we need to adjust or add factors? The question How do factors contribute to the studied problem? is excluded. |
‘Expert preparation’ (A) Session three | The information experts receive is paraphrased and categorized in an overview of causal relationships between factors and the studied problem. The assignment includes the question: Do we need to adjust or add factors and relations? The question about the factors is additional. |
‘Expert preparation’ (A) Session four | The information experts receive is paraphrased and categorized in the CLD clarified in session three. The assignment includes the questions: Do we need to adjust or add factors and relations? Is the CLD clear? The question about the factors is additional. |
‘Introduction and recap’ Sessions two, three, and four (A) | The primary focus of these scripts was shifted to plenary reflection on factor definitions. |
‘Initiating and elaborating a CLD’ (S) Sessions two and three | Extra time was scheduled for this script. Extra time was saved by processing some experts’ suggestions after the session ended. |
Model review (S) Session four | Extra time was scheduled for this script and it was merged with the ‘introduction and recap’ script. |