Table 3 Summary of the key findings of each section

Definitions of Patient Preferences

Provides an overview on the various definitions of patient preferences that the publications contain.

Utility-based definitions offer numerical clarity but risk oversimplification. Trade-off focused definitions highlight balancing risks and benefits but need more nuanced models. Elicitation-based definitions systematically capture priorities but lack adaptability. Context-specific definitions address variability across conditions, while value alignment definitions emphasize aligning decisions with broader patient goals.

Use Cases

Summarizes the specific applications and medical conditions (e.g., oncology, cardiology) where patient preferences are incorporated.

Oncology and cardiology are the most common use cases, with applications focused on high-incidence, complex conditions where patient values are highly relevant. Chronic disease management tools prioritize patient preference to manage treatment adherence and quality-of-life aspects, especially in diseases like multiple sclerosis and diabetes.

Type of Decision Support

Identifies types of decision support tools such as models, decision aids, and decision support systems (DSS) employed.

Decision aids and DSS tools are widely used to facilitate shared decision-making; they help both patients and clinicians navigate complex decisions with high patient involvement. Models aiding in more accurate and personalized medical planning.

Methods Used

Outlines statistical and MCDA methods as well as specific models like Markov models, Bayesian networks, etc.

MCDA is frequently applied to balance clinical outcomes with individual preferences, especially in chronic diseases. Statistical models like Markov chains for tracking disease progression and Bayesian networks for probabilistic risk modeling are commonly used to integrate patient preferences, improving alignment with personalized treatment goals.

Frameworks & Treatment Contexts

Describes how decision-making tools are matched to different clinical contexts (e.g., chronic disease, preventive care) with framework examples.

In oncology, Markov models accommodate sequential treatment needs, adjusting for changing patient conditions and preferences over time. Preventive care uses frameworks like cost-effectiveness analysis (CEA) and QALY-based models to weigh patient quality of life and longevity in health recommendations, supporting screenings and lifestyle interventions.

Elicitation of Patient Preferences

Details various techniques (e.g., Standard Gamble, Conjoint Analysis, DCEs) used to capture individual patient preferences.

Standard Gamble quantifies patient risk tolerance, while Conjoint Analysis and Discrete Choice Experiments (DCE) capture relative importance of treatment attributes, allowing personalized risk-benefit profiles. Analytic Hierarchy Process (AHP) and Time Trade-Off (TTO) are used for prioritizing long-term health states and outcomes.

Patient Preferences Incorporated

Lists the specific patient preferences, including risk tolerance, treatment attributes, and outcome priorities integrated into models.

Patient preferences span risk tolerance, treatment invasiveness, side effect concerns, and quality of life. Preferences vary by condition, with cancer patients often valuing survival outcomes, while chronic disease patients may prioritize treatment simplicity and minimizing side effects to improve adherence.

End User Involvement

Identifies the primary end users targeted by these tools, such as patients, clinicians, or both, and their roles in decision-making.

Most tools aim to enhance shared decision-making between patients and clinicians, with features designed to address clinician concerns around time efficiency and usability. Some tools target patients exclusively, focusing on empowerment through easy-to-understand visuals and interactive options for preference adjustments.

End User Acceptance

Summarizes levels of acceptance of these tools by patients and clinicians, highlighting any pilot or beta testing outcomes.

Studies show high acceptance rates among patients in pilot testing, where tools often lead to greater satisfaction and engagement in decision-making. Clinician acceptance, however, requires additional support, as barriers like workflow integration and complexity persist, calling for tailored training programs.

Sources of Heterogeneity

Explains the main sources of variability in preference integration, such as differing model types, elicitation methods, and user involvement.

Heterogeneity arises from variations in decision-making models, ranging from basic statistical tools to complex MCDA and utility-based frameworks. Differences in preference elicitation methods, such as DCE versus Standard Gamble, and the varying roles of patients and clinicians further add complexity.