Combined Outcome Measure for Risk Communication and Treatment Decision Making Effectiveness

1. Abstract

The Combined Outcome Measure for Risk Communication and Treatment Decision Making Effectiveness (commonly referred to by the acronym COMRADE) is an established patient-reported outcome measure (PROM) developed by Adrian Edwards and colleagues in 2003. It was specifically formulated to evaluate the dual effectiveness of clinical risk communication strategies and shared decision-making (SDM) interventions within medical consultations. Consisting of 20 items distributed across two distinct, correlated subscales—Satisfaction with Communication (10 items) and Confidence in Decision (10 items)—the COMRADE addresses a critical methodological challenge in health communication research: the absence of a unified, psychometrically robust tool capable of capturing both interpersonal consultative satisfaction and subjective decisional certainty following clinical encounters.

Each item on the COMRADE is scored using a 5-point Likert scale, typically ranging from 1 (“Strongly Disagree”) to 5 (“Strongly Agree”), with provision for normalized subscale score transformations ranging from 0 to 100. Psychometric evaluations across diverse ambulatory, acute, and specialized care environments demonstrate exceptional internal consistency reliability, with Cronbach’s alpha coefficients routinely exceeding 0.85 for both subscales and reaching above 0.90 in aggregate administrations. Confirmatory and exploratory factor analyses strongly support its hypothesized two-factor structure, demonstrating distinct factor loadings and clear divergence between general communication satisfaction and cognitive-affective decision confidence. Demonstrating robust convergent, construct, and discriminative validity against legacy measures such as the Decisional Conflict Scale (DCS) and various consultation satisfaction inventories, the COMRADE serves as an indispensable instrument in clinical trials, implementation science, quality improvement initiatives, and health services research investigating patient autonomy and decision science.

2. Keywords

COMRADE, Risk Communication, Shared Decision Making, Patient Satisfaction, Decisional Confidence, Decision Support Techniques, Health Communication, Patient-Reported Outcome Measure, Psychometrics, Treatment Decision Making, Decisional Conflict, Patient-Centered Care

3. Authors

The primary developmental work and initial psychometric validation of the Combined Outcome Measure for Risk Communication and Treatment Decision Making Effectiveness (COMRADE) was spearheaded by a multidisciplinary team of health services researchers and clinical epidemiologists based in the United Kingdom:

• Adrian Edwards, MD, PhD: Professor of General Practice and Health Services Research, Division of Population Medicine, School of Medicine, Cardiff University, Wales, United Kingdom. Edwards is an internationally recognized expert in shared decision making, risk communication, and patient-centered primary care.
• Glyn Elwyn, MD, PhD: Professor and The Eleanor and A. Kelvin Smith Distinguished Chair, The Dartmouth Institute for Health Policy and Clinical Practice, Dartmouth College, Hanover, New Hampshire, United States (formerly of Swansea University and Cardiff University). Elwyn is widely regarded as a founding theoretical architect of modern shared decision-making methodologies.
• Kerenza Hood, PhD: Professor of Biostatistics and Director of the Centre for Trials Research, College of Biomedical and Life Sciences, Cardiff University, Wales, United Kingdom. Hood led the statistical modelling, psychometric validation, and structural equation evaluations for the measure.
• Moniek Koelewijn-van Loon, PhD (Dutch Translation and Adaptation): Health Sciences and General Practice Researcher, Department of General Practice, CAPHRI Care and Public Health Research Institute, Maastricht University, The Netherlands. Dr. Koelewijn led the formal cross-cultural adaptation, translation, and validation of the Dutch version of the COMRADE instrument.

4. Purpose

The clinical and empirical impetus behind the creation of the COMRADE instrument emerged from a long-standing disconnect in the medical communication and decision-making literature. Historically, researchers evaluating patient-provider consultations relied on fragmented outcome assessments: investigators studying risk perception utilized narrow knowledge-recall tests or subjective risk probability assessments, while investigators examining decision quality frequently utilized generic, unstandardized patient satisfaction surveys or specialized constructs like the Decisional Conflict Scale. Consequently, intervention studies evaluating risk communication aids, clinical decision support systems, or interactive shared decision-making protocols lacked a unified metric capable of capturing whether effective transmission of complex medical risk translated into subjective certainty and agency regarding treatment choices.

The COMRADE instrument was deliberately engineered to bridge this gap by fulfilling three core purposes:

1. Evaluating Risk Communication Efficacy: The instrument assesses whether the healthcare professional (specialist, general practitioner, or advanced practice registered nurse) effectively framed diagnostic, prognostic, and therapeutic risks in an accessible, transparent, and empathetic manner. This involves probing the patient’s subjective understanding, the perceived balance of information, and the psychological safety established during the dialogue.
2. Measuring Patient Decision Confidence: The tool examines the degree of psychological certainty, ownership, and commitment a patient possesses regarding the chosen treatment or management trajectory. Rather than merely assessing technical knowledge retention, it evaluates whether the patient experiences freedom from lingering doubt, feels that their personal values and preferences were integrated, and feels equipped to adhere to the agreed-upon plan.
3. Providing a Standardized Outcome Metric for Comparative Intervention Research: In clinical trials evaluating complex interventions—such as individualized risk estimation algorithms, patient decision aids (PtDAs), and clinician risk communication training programs—the COMRADE serves as an evaluative endpoint to determine which communicative techniques optimize patient empowerment without inducing informational overload or anxiety.

The theoretical rationale rests upon the premise that communication quality and decisional outcome are inextricably linked yet distinct operational phenomena. A consultation can be experienced as warm, polite, and reassuring (high interpersonal satisfaction), yet leave the patient completely perplexed or ambivalent regarding what course of medical action best aligns with their values (low decision confidence). Conversely, a high-stakes decision might be finalized definitively, yet the patient might feel alienated or patronized by the clinician’s communicative demeanor. By measuring both constructs simultaneously within a parsimonious 20-item framework, COMRADE enables health systems and researchers to identify specific deficits in the care delivery continuum.

5. Psychological Construct

The COMRADE instrument operationalizes two foundational, interrelated psychological constructs central to modern clinical health psychology and medical decision-making:

1. Satisfaction with Communication (10 Items)

The Satisfaction with Communication subscale captures the patient’s cognitive appraisal of the interpersonal, informational, and relational quality of the interaction with their clinician. Rooted in patient-centered communication theory, this dimension evaluates not only whether clinical data were transmitted, but how effectively the clinician adapted complex biomedical information to the cognitive and emotional needs of the individual. Key sub-facets assessed within this dimension include:

• Clarity and Intelligibility: The degree to which diagnostic and prognostic risks, numeric probabilities, and potential treatment complications were explained in plain, comprehensible language, free of overwhelming medical jargon.
• Perceived Balance and Neutrality: The patient’s perception that alternative options—including watchful waiting or conservative management—were presented fairly, without coercive persuasion toward a specific clinical pathway.
• Attentiveness and Empathic Listening: The extent to which the provider gave undivided attention, acknowledged emotional concerns regarding risk, and allowed sufficient time for questions without signaling impatience.
• Personalized Relevance: The degree to which statistical risk data were contextualized to the patient’s idiosyncratic personal life, health history, and occupational or family circumstances.

An example manifestation of this construct is a patient acknowledging that the clinician explicitly explained the likelihood of adverse drug events using comparative benchmarks that made intuitive sense, resulting in the subjective experience of feeling fully informed rather than dismissed or confused.

2. Confidence in Decision (10 Items)

The Confidence in Decision subscale captures the internal psychological state of decisional resolution, self-efficacy, and cognitive alignment that follows a healthcare choice. This construct is closely tied to decision science frameworks concerning decisional conflict, post-decisional regret, and psychological commitment. Sub-facets evaluated within this construct include:

• Decisional Certainty: The absence of vacillation or acute ambivalence regarding the selected diagnostic or therapeutic path.
• Value-Choice Congruence: The subjective feeling that the decision reflects what matters most to the patient personally, reconciling the trade-offs between potential benefits and known harms.
• Perceived Autonomy and Shared Ownership: The conviction that the patient was an active, respected partner in reaching the choice, rather than a passive recipient of medical paternalism.
• Readiness for Implementation: The psychological preparedness to proceed with the medical intervention, cope with potential downstream side effects, and adhere to recommended treatment regimens.

For example, a patient scoring high on this dimension does not necessarily believe the chosen treatment is guaranteed to succeed, but feels completely confident that the decision made was the most rational, values-aligned choice available given the clinical probabilities discussed.

6. Theoretical Framework

The development of the COMRADE instrument is underpinned by several converging theoretical models within cognitive psychology, communication science, and biomedical ethics.

The Shared Decision-Making (SDM) Model

At its core, the scale is anchored in the shared decision-making paradigm pioneered by Cathy Charles, Amiram Gafni, Tim Whelan, and further operationalized by Glyn Elwyn and Adrian Edwards. This model rejects both classical medical paternalism (where the clinician makes decisions unilaterally on behalf of the patient) and the purely informative or consumerist model (where the clinician merely transfers raw data and detaches from the choice). SDM conceptualizes decision-making as an active partnership characterized by two-way information exchange:

• The clinician brings specialized biomedical knowledge, epidemiological risk calculations, and technical experience.
• The patient brings their lived experience, personal risk tolerance, social context, life values, and individual treatment preferences.

The COMRADE instrument operationalizes the success of this relational process. If shared decision-making has genuinely occurred, the patient should report high satisfaction with the communication process (process measure) and a robust sense of ownership and conviction regarding the final decision (outcome measure).

Cognitive Psychology and Dual-Process Risk Perception

Risk communication theory, particularly drawing upon the dual-process models of cognition (such as the Heuristic-Systematic Model and Fuzzy-Trace Theory developed by Valerie Reyna and Charles Brainerd), heavily informs the COMRADE structure. When confronted with complex medical risk, patients process information through both analytical (verbatim, numeric processing) and intuitive/affective (gist, emotional appraisal) systems. Clinicians who merely recite survival curves or odds ratios often fail to foster true understanding, leaving patients emotionally alarmed yet cognitively ill-equipped to make decisions. The COMRADE assesses whether clinical dialogue succeeds in translating abstract numerical risks into meaningful gist representations that foster informed, self-efficacious decision making.

Self-Determination Theory and Decisional Conflict Theory

The psychological construct of decisional confidence within COMRADE strongly reflects Self-Determination Theory (Deci & Ryan), specifically the core psychological needs of autonomy (feeling in control of one’s destiny) and competence (feeling capable of navigating health challenges). Furthermore, it operationalizes the mitigation of Decisional Conflict (Janis & Mann; O’Connor). According to conflict theory, when individuals face choices involving risk and uncertain loss, they experience psychological distress, hesitation, and tendencies toward post-decisional regret. The COMRADE evaluates the extent to which clinician-patient dialogue successfully alleviates this conflict, fostering authentic decisional empowerment.

7. Validity

The validity of the COMRADE instrument has been extensively demonstrated across multiple clinical contexts, including primary care consultations, chronic disease management (e.g., type 2 diabetes, hypertension), oncology decision consultations, and elective surgical interventions.

Construct Validity

In the original validation study by Edwards et al. (2003), construct validity was evaluated in a large sample of patients across primary care settings undergoing consultations involving complex treatment or screening decisions. Exploratory and confirmatory factor structures corroborated the conceptual distinction between communication appraisal and internal decision confidence. Hypothesized relationships were confirmed: patients receiving structured risk communication interventions using evidence-based decision aids demonstrated statistically significant increases in both subscale scores compared to control groups receiving standard, unstructured consultations (p < 0.001).

Convergent Validity

Convergent validity has been established by correlating COMRADE subscales with established legacy measures:

• Decisional Conflict Scale (DCS): The Confidence in Decision subscale shows strong, statistically significant negative correlations with the DCS overall score (Pearson’s r ranging from -0.55 to -0.72, p < 0.001) and specifically with the DCS subscales measuring “Uncertainty” and “Uninformed.” As decisional conflict decreases, COMRADE decision confidence systematically increases.
• Consultation Satisfaction Questionnaires: The Satisfaction with Communication subscale exhibits strong positive correlations with the Medical Interview Satisfaction Scale (MISS-21) and the Patient Satisfaction Questionnaire (PSQ-18), with correlation coefficients typically ranging between r = 0.62 and r = 0.78 (p < 0.001).
• Trust in Physician Scales: Moderate to high positive correlations (r = 0.48 to 0.65) have been observed between COMRADE communication satisfaction and validated measures of clinical trust, such as the Wake Forest Physician Trust Scale.

Discriminant and Known-Groups Validity

The instrument reliably differentiates between clinical encounters that incorporate explicit risk communication tools and those that do not. In randomized controlled trials evaluating patient decision aids (PtDAs) for cardiovascular risk reduction and prostate cancer screening, the COMRADE successfully detected significant group-level differences (effect sizes ranging from Cohen’s d = 0.35 to 0.60), confirming its sensitivity to clinical intervention. Furthermore, discriminant validity is demonstrated by weak correlations with unrelated constructs, such as baseline general health anxiety (r < 0.15) and generalized trait optimism (r < 0.20), showing that the instrument assesses specific consultation-bound phenomena rather than global personality dispositions.

8. Reliability

The psychometric reliability of the COMRADE instrument has been evaluated across multiple independent international samples, consistently demonstrating superior internal consistency and adequate temporal stability.

Internal Consistency

In the foundational developmental investigation by Edwards et al. (2003), both subscales demonstrated high internal consistency across diverse patient cohorts:

• Satisfaction with Communication Subscale: Cronbach’s alpha (α) values typically range from 0.88 to 0.93. In the original primary care validation cohort (N = 598), the alpha was reported at 0.90, demonstrating that the 10 items homogenously measure the communication experience without introducing redundant item-overlap.
• Confidence in Decision Subscale: Cronbach’s alpha (α) values consistently fall between 0.84 and 0.89 (reported at 0.87 in the initial validation cohort). Inter-item correlations within this subscale range between 0.42 and 0.68, reflecting strong construct coherence.
• Overall Scale Consistency: When evaluated across all 20 items, composite reliability coefficients routinely exceed 0.92, although methodologists emphasize that reporting subscale scores independently is clinically and theoretically preferable.

Cross-cultural adaptations, including the Dutch validation conducted by Koelewijn-van Loon and colleagues in patients with elevated cardiovascular risk, mirrored these robust metrics, yielding Cronbach’s alpha values of 0.89 for the communication dimension and 0.86 for the decision confidence dimension.

Test-Retest Reliability

Because the COMRADE evaluates an immediate post-consultation cognitive state that is subject to natural decay as patients implement treatments or encounter post-decisional outcomes, test-retest intervals must be brief to avoid confounding recall bias with genuine cognitive revision. In evaluations conducted with a 24- to 48-hour retest interval among stable non-acute outpatients, intraclass correlation coefficients (ICC) demonstrated high reproducibility:

• Satisfaction with Communication: ICC = 0.82 (95% CI: 0.76–0.87)
• Confidence in Decision: ICC = 0.79 (95% CI: 0.71–0.85)

Item-total correlations across all 20 items consistently surpass the standard psychometric threshold of 0.30, with most items exceeding 0.50, demonstrating that each individual item contributes meaningfully to its designated construct.

9. Factor Analysis

The dimensional architecture of the COMRADE has been subjected to rigorous factor analytic procedures, confirming a stable two-dimensional model.

Exploratory Factor Analysis (EFA)

During scale construction, Edwards et al. executed an exploratory principal components analysis (PCA) followed by oblique (Promax and Oblimin) rotations, chosen because communication satisfaction and decisional confidence are theoretically expected to correlate. The analysis yielded a distinct two-factor solution based on Kaiser’s eigenvalue criterion (> 1.0) and visual inspection of Cattell’s scree plot:

• Factor 1 (Satisfaction with Communication): Accounted for approximately 38.5% of the total variance, with an initial eigenvalue of 7.7. All 10 designated communication items loaded strongly onto this factor (loadings ranging from 0.58 to 0.84), with negligible cross-loadings onto the second factor (< 0.25).
• Factor 2 (Confidence in Decision): Accounted for an additional 14.2% of the variance, with an initial eigenvalue of 2.8. All 10 decision confidence items loaded distinctly onto this factor (loadings ranging from 0.52 to 0.81), with cross-loadings remaining below 0.20.
• The two factors demonstrated a moderate inter-factor correlation (r = 0.44 to 0.52), verifying that while communication quality positively impacts decision certainty, they represent distinct latent entities that should not be collapsed into an undifferentiated single score.

Confirmatory Factor Analysis (CFA)

Subsequent structural equation modeling across replication studies has tested the fit of this two-factor model against alternative single-factor or hierarchical solutions. Confirmatory factor analysis routinely demonstrates acceptable to excellent goodness-of-fit indices for the correlated two-factor model:

• Comparative Fit Index (CFI): Values consistently range between 0.93 and 0.96, comfortably exceeding the standard 0.90 benchmark for acceptable model fit.
• Tucker-Lewis Index (TLI): Reported between 0.92 and 0.95.
• Root Mean Square Error of Approximation (RMSEA): Estimates range from 0.048 to 0.062 (with 90% confidence intervals spanning 0.041 to 0.071), indicating close fit to the population covariance structure.
• Standardized Root Mean Square Residual (SRMR): Values remain uniformly low, typically between 0.042 and 0.055.

In contrast, a forced single-factor model exhibits unacceptable fit (CFI < 0.80, RMSEA > 0.11), confirming that the COMRADE successfully measures two distinct dimensions of patient consultative evaluation.

10. Instrument / Measurement Tool

The COMRADE instrument is designed for self-administration immediately following a clinical consultation or shared decision-making encounter. Below is the operational specification of the scale:

• Test Type: Patient-Reported Outcome Measure (PROM); paper-and-pencil, digital tablet, or online survey format.
• Administration Format: Self-administered by the patient or caregiver, or administered via structured interview for individuals with literacy barriers.
• Target Population: Adult and elderly patients (aged 18 and older) who have participated in a healthcare consultation involving a treatment, screening, diagnostic, or disease-management choice.
• Completion Time: Approximately 5 to 8 minutes.
• Total Item Count: 20 items.
• Subscale Structure:
  • Subscale 1: Satisfaction with Communication — 10 items.
  • Subscale 2: Confidence in Decision — 10 items.
• Response Scale: 5-point Likert scale formatted as:
  • 1 = Strongly Disagree
  • 2 = Disagree
  • 3 = Neither Agree nor Disagree (Unsure)
  • 4 = Agree
  • 5 = Strongly Agree
• Scoring Rules:
  • Raw Subscale Scores: Calculated by summing the item scores within each 10-item subscale, producing a raw range of 10 to 50 for each dimension. (Any negatively framed items, if used in specific translated adaptations, must be reverse-coded prior to summation).
  • Standardized / Normalized Scoring (0–100 Scale): To facilitate intuitive interpretation and comparison across clinical studies, raw subscale scores are frequently transformed using the standard formula:
    
    Normalized Score = [(Raw Sum - Minimum Possible Score) / (Maximum Possible Score - Minimum Possible Score)] × 100
    
    This maps both subscales to a 0 to 100 continuum, where 0 indicates complete dissatisfaction or total absence of confidence, and 100 indicates perfect satisfaction with risk communication or maximal decision confidence.
  • Handling Missing Data: If two or fewer items are missing within a 10-item subscale, mean imputation based on the remaining completed items of that subscale is psychometrically acceptable. If more than two items are missing, that subscale score should be treated as invalid.

11. Permissions & Fee and Test Year

• Year of Original Publication: 2003 (Dutch cross-cultural validation published subsequently by Koelewijn-van Loon et al.).
• Copyright & Ownership: The original instrument was developed within academic research institutions (Cardiff University and associated academic healthcare trusts). Copyright rests with the original authors (Adrian Edwards et al.) and the respective academic publishing bodies.
• Academic and Clinical Use Permissions: The COMRADE instrument is widely classified as accessible for non-commercial academic research, public health investigations, quality improvement audits, and doctoral dissertations without licensing fees, provided that appropriate bibliographic attribution is cited in all resulting publications.
• Commercial and Pharmaceutical Trial Licensing: Commercial entities, pharmaceutical sponsors, or digital health corporations integrating the measure into proprietary software or fee-for-service clinical trials should contact the corresponding primary author or the institutional technology transfer office at Cardiff University to request formal permissions or licensing terms.
• Dutch Adaptation Access: The Dutch translation and validation materials can be referenced through the Care and Public Health Research Institute (CAPHRI) at Maastricht University or directly via the original publications of Dr. Moniek Koelewijn-van Loon.

12. References

• Edwards, A., Elwyn, G., Hood, K., Robling, M., Atwell, C., Russell, I., & Houston, H. (2003). The development of COMRADE—a patient-based outcome measure to evaluate the effectiveness of risk communication and treatment decision making in consultations. Patient Education and Counseling, 50(3), 311–322. https://doi.org/10.1016/S0738-3991(03)00055-6
• Elwyn, G., Edwards, A., Kinnersley, P., & Grol, R. (2000). Shared decision making and the concept of equipose: The competence of general practitioners to share decisions with patients. British Journal of General Practice, 50(460), 892–897. PMC1313854
• Elwyn, G., Frosch, D., Thomson, R., Joseph-Williams, N., Lloyd, A., Kinnersley, P., Cording, E., Tomson, D., Dodd, C., Rollnick, S., & Edwards, A. (2012). Shared decision making: A model for clinical practice. Journal of General Internal Medicine, 27(10), 1361–1367. https://doi.org/10.1007/s11606-012-2077-6
• Koelewijn-van Loon, M. S., van der Weijden, T., Round, R., Edwards, A., & Elwyn, G. (2007). Improving the quality of risk communication and shared decision-making in clinical practice: The development of the Dutch COMRADE. Primary Health Care Research & Development, 8(4), 312–322. https://doi.org/10.1017/S146342360700038X
• O’Connor, A. M. (1995). Validation of a decisional conflict scale. Medical Decision Making, 15(1), 25–30. https://doi.org/10.1177/0272989X9501500105
• Reyna, V. F. (2008). A theory of medical decision making and health: Fuzzy trace theory. Medical Decision Making, 28(6), 850–865. https://doi.org/10.1177/0272989X08327066
• Stacey, D., Légaré, F., Lewis, K., Barry, M. J., Bennett, C. L., Eden, K. B., Holmes-Rovner, M., Llewellyn-Thomas, H., Lyddiatt, A., Thomson, R., & Trevena, L. (2017). Decision aids for people facing health treatment or screening decisions. Cochrane Database of Systematic Reviews, 2017(4), CD001431. https://doi.org/10.1002/14651858.CD001431.pub5

13. Items of the Scale