The promise of "prepared" clinical decision models
Better clinical decisions, more effective guidelines, and vertical integration of shared decision making throughout the health care system.
In the December 16, 2022 Musing, I described how creating tangible, explicit clinical decision models could improve the quality of medical care and promote informed, shared decision making. In this week’s post, I followup on two other ideas I mentioned briefly: creating “prepared” maps at the clinical level and using clinical decision models to promote vertical integration of shared decision making throughout the different layers of the health care system.
“Prepared” clinical decision models
As shown in the figure below, the proposed hierarchical decision model has an explicit statement of the goal of the decision at the top, a list of options at the bottom, and a layer containing the considerations being used as criteria to define good choices, i.e., those likely to meet the decision goal, in the middle.
Although the goal and options will vary depending on the situation, most clinical decisions share a group of common decision considerations: effectiveness, risk of adverse effects, cost, and the logistics involved (time required, means of administration, etc.).
One way to meet the quality goal of improving the decision making skills of patients is to introduce patients to the concept of decision mapping in a brief preliminary session. In the session they would work through a small number of hypothetical examples designed to teach them about the decision mapping process and give them a deeper understanding of the generic decision considerations. A patient-specific starter model derived from this session could then be used as the starting point to address an important future decision.
During the training phase, it would also be possible to show the patients what their physician’s generic set of decision considerations are for the same sample scenarios. This step would set the stage for enabling both parties to use decision models as a new communication medium for discussing clinical decisions and sharing in the decision making process.
Vertical integration
Taking the idea of prepared decision maps a step further, I see no reason why a clinical guideline panel could not create a generic decision model that would be published along with their recommendations. The guideline panel model would include factors included in the panel deliberations and a summary of the information gathered about each recommended option. Importantly, when included in a decision model, this information would be transmitted to clinicians and patients in a format directly applicable to using it to make a decision, rather than a text-heavy summary of findings.
In addition to providing structure and format to the decision making scenario, the guideline decision model could, and should, include information about the relative priorities the panel assigned to their decision criteria. A full guideline model in this format would explicitly summarize the reasoning and information that are the basis for the panel recommendations and thereby clarify the information and advice provided. The resulting models could then be examined and adjusted as needed to meet the needs of individual patients in the clinical setting.
It is important to note that the avenue of communication provided by using a communal decision model runs both ways. In addition to the information going from the guideline panel level to the clinical level, the alterations made by the clinical decision makers, reflecting the needs and priorities of individual patients and clinicians, could be fed back to the guideline panel and used to further refine the guideline recommendations. In other words, using guideline decision models in this way would create a feedback learning loop, that could operate in a regular fashion without the need for special studies.
Musings
Many, if not most, difficult clinical decisions involve trade-offs between competing objectives. For this reason, the proposed default clinical decision model is designed to identify the objectives the decision makers want to accomplish and facilitate the process of making any necessary trade-offs.
The idea of creating a decision model is not radical. It is one of the major contributions of decision analysis to the world. In practical terms it is no different from using a map to travel from one place to another or assembling a piece of Ikea furniture. It is true that it has not been done before, but that is no reason to conclude it cannot be done and is not worth a try.