Decision Evaluation Scales

1. Abstract

The Decision Evaluation Scales (DES) is a multidimensional, patient-reported psychometric instrument developed by Dr. Peep F. M. Stalmeier and colleagues in 2005 at the Radboud University Medical Center in the Netherlands. Comprising 15 items, the DES is designed to retrospectively evaluate how patients perceive and appraise a medical decision following clinical consultations or interventions. Grounded in modern psychometric theory, the scale was constructed and refined using the Rasch measurement model, yielding three psychometrically distinct, unidimensional subscales: Satisfaction-Uncertainty (evaluating emotional resolution, lingering doubts, and confidence in the chosen path), Informed Choice (assessing perceived adequacy of information, clarity of consequences, and understanding of treatment alternatives), and Decision Control (measuring patient autonomy, personal ownership of the decision, and perceived influence relative to the healthcare provider). Unlike generic patient satisfaction questionnaires that suffer from pervasive ceiling effects, the DES distinguishes cognitive, affective, and relational elements of medical decision-making. The instrument demonstrates strong psychometric properties, with subscale Cronbach’s alpha reliability coefficients consistently ranging between .78 and .88, favorable Rasch infit and outfit mean-square indices (0.7 to 1.3), and robust convergent validity against legacy instruments such as the Decisional Conflict Scale (DCS) and the Decision Regret Scale (DRS). The DES is extensively utilized in oncology, surgical decision-making, genetic counseling, and chronic disease management to evaluate shared decision-making (SDM) interventions, patient decision aids (PtDAs), and communicative training for clinicians.

2. Keywords

Decision Evaluation Scales, DES, shared decision-making, medical decision-making, Rasch measurement model, informed choice, decision control, patient-reported outcome measures, decisional conflict, health psychology, psychometrics, patient autonomy, clinical decision aids

3. Authors

The Decision Evaluation Scales were developed and validated by an interdisciplinary research team specializing in medical decision sciences, psychometrics, and clinical oncology at the Radboud University Medical Center (Radboudumc) in Nijmegen, The Netherlands:

• Peep F. M. Stalmeier, Ph.D. — Department for Health Evidence (Biostatistics and Health Technology Assessment), Radboud University Medical Center, Nijmegen, Netherlands. Primary investigator and psychometrician specializing in utility measurement, patient decision aids, and health preferences.
• M. S. Roosmalen, Ph.D. — Department of Medical Oncology and Medical Psychology, Radboud University Medical Center, Nijmegen, Netherlands. Clinical researcher focused on breast cancer decision-making and patient counseling.
• J. W. H. Leer, M.D., Ph.D. — Department of Radiation Oncology, Radboud University Medical Center, Nijmegen, Netherlands.
• Contributing Collaborators: Multidisciplinary research groups across surgical oncology, clinical genetics, and communicative medicine at Radboudumc, University of Groningen, and associated academic teaching hospitals in the Netherlands.

4. Purpose

Modern clinical practice has shifted from historical paternalism toward collaborative paradigms of healthcare, centered on patient autonomy and shared decision-making. While the conceptual benefits of shared decision-making are widely acknowledged, accurately quantifying how patients experience, process, and evaluate critical medical choices has presented persistent methodological challenges. The Decision Evaluation Scales (DES) were engineered to provide an empirically rigorous, psychometrically robust tool to assess post-decisional cognitive appraisals, affective responses, and structural dynamics of the patient-clinician encounter.

Traditional evaluation metrics in healthcare environments have frequently relied on broad patient satisfaction surveys. These traditional instruments exhibit substantial limitations, most notably severe ceiling effects, where 80% to 95% of respondents indicate high satisfaction despite possessing low clinical knowledge, experiencing severe unexpressed anxiety, or harboring unvoiced ambivalence. The primary purpose of the DES is to overcome these ceiling effects by decomposing post-decisional appraisal into distinct latent dimensions that separate emotional comfort from intellectual clarity and relational autonomy.

In clinical practice, the DES serves as an evaluative and diagnostic instrument. Clinicians and care teams can administer the instrument following complex consultations—such as elective surgical choices, choices between mastectomy versus breast-conserving therapy, systemic chemotherapy regimens, genetic testing for hereditary cancer syndromes, or palliative care transitions. By inspecting individual subscale scores, healthcare providers can identify specific deficits in patient care. For instance, a patient might demonstrate high scores on Decision Control and Satisfaction-Uncertainty, yet score poorly on Informed Choice, indicating that while the patient felt autonomous and emotionally reassured, their comprehension of treatment tradeoffs was suboptimal, signaling a need for additional educational reinforcement.

In academic research, the DES serves as a primary outcome measure for randomized controlled trials (RCTs) evaluating patient decision aids, communicative training interventions for healthcare practitioners, and structured decision-support algorithms. By employing a Rasch-validated metric, researchers obtain interval-level measurement properties that allow for linear comparisons across heterogeneous cohorts, longitudinal assessment of decisional durability over time, and comparative effectiveness research across diverse clinical specialties.

5. Psychological Construct

The Decision Evaluation Scales measure the overarching construct of post-decisional evaluation, conceptualized as a retrospective, multifaceted cognitive and affective appraisal of a healthcare choice. Rather than viewing decision evaluation as a global, unitary sentiment, the psychometric architecture of the DES posits that patients experience three distinct, interrelated dimensions:

1. Satisfaction-Uncertainty (Tevredenheid-Onzekerheid)

This subscale captures the affective equilibrium and psychological comfort of the patient regarding the choice made. It encompasses the resolution of decisional ambivalence, feelings of post-decisional confidence, relief, and the absence of cognitive dissonance or lingering doubt. In psychometric modeling, this construct spans a continuum from intense decisional distress, hesitation, and second-guessing at the lower end, to unequivocal decisional peace, resolution, and perceived correctness at the higher end.

Clinical Example: Following a consultation regarding whether to initiate adjuvant chemotherapy, a patient with high scores on this dimension feels emotionally settled, expresses confidence that the selected course aligns with their personal life values, and experiences minimal internal conflict or dread about whether an alternative option would have been superior.

2. Informed Choice (Geïnformeerde keuze)

The Informed Choice subscale isolates the epistemic and cognitive dimension of the decision. It assesses the degree to which the patient perceives that they received sufficient, high-quality, and balanced medical information regarding all available alternatives, including watchful waiting. This construct evaluates the patient’s clarity concerning the pros, cons, potential side effects, long-term prognostic outcomes, and trade-offs inherent in each clinical pathway.

Clinical Example: In prostate cancer screening or treatment selection (e.g., active surveillance vs. radical prostatectomy), a patient scoring high on Informed Choice understands the probabilities of urinary incontinence, erectile dysfunction, and biochemical recurrence across both pathways, feeling that their eventual choice was built on comprehensive, transparent medical evidence rather than clinical ambiguity.

3. Decision Control (Beslissingscontrole)

The Decision Control dimension measures perceived autonomy, agency, and relational self-determination within the clinical interaction. Grounded in conceptual models of patient participation, this construct evaluates the balance of power between patient and clinician. It differentiates between passive compliance (where the patient feels the decision was imposed or predominantly directed by the physician) and active self-determination or collaborative partnership (where the patient feels their preferences were central to the final resolution).

Clinical Example: When managing chronic degenerative joint disease, a patient evaluating total hip arthroplasty who demonstrates elevated Decision Control reports that they retained final sovereignty over the timing and selection of surgery, feeling empowered rather than coerced or bypassed by the orthopedic surgeon’s clinical preference.

6. Theoretical Framework

The conceptual and methodological architecture of the Decision Evaluation Scales integrates three theoretical frameworks from cognitive psychology, behavioral medicine, and psychometrics:

Decisional Conflict Theory and Cognitive Dissonance

The historical baseline for decisional appraisal derives from Irving Janis and Leon Mann’s (1977) Conflict Model of Decision Making, alongside Leon Festinger’s (1957) theory of Cognitive Dissonance. Janis and Mann posited that vital decisions involving health risks trigger substantial psychological stress, characterized by vacillation, apprehension, and anticipatory regret. When individuals make difficult choices involving uncertain trade-offs, post-decisional dissonance frequently emerges if the negative attributes of the chosen alternative or the positive attributes of the rejected alternatives become salient. The DES formalizes this cognitive-affective post-decisional state, operationalizing the extent to which clinical communication successfully facilitates cognitive closure, minimizes dissonance, and mitigates post-decisional distress.

The Shared Decision-Making Framework

The DES is anchored in modern models of shared decision-making formulated by Charles, Gafni, and Whelan (1997) and operationalized by Glyn Elwyn and colleagues (2012). These models stipulate that high-quality clinical decisions require three foundational conditions: choice awareness (acknowledging that a choice exists), option comparison (understanding the comparative risks and benefits), and preference elicitation (integrating patient values into the final selection). The three subscales of the DES map onto this paradigm:

• Option Comparison maps directly onto the Informed Choice scale;
• Preference Elicitation & Autonomy maps onto the Decision Control scale;
• Decision Resolution maps onto the Satisfaction-Uncertainty scale.

The Rasch Measurement Model (Modern Psychometric Theory)

Unlike instruments constructed solely through Classical Test Theory (CTT), the DES was built on the Rasch measurement model formulated by Danish mathematician Georg Rasch (1960). Modern psychometricians emphasize that ordinal Likert scores should not be treated as linear interval measures without mathematical justification. Rasch modeling formalizes the probability of a specific category response as a logistic function of the difference between a person’s latent trait level ($\theta_n$) and the item’s calibration location ($b_i$):

$$P(X_{ni} = 1) = \frac{\exp(\theta_n – b_i)}{1 + \exp(\theta_n – b_i)}$$

By fitting empirical item responses to the Rasch model, Stalmeier et al. ensured invariant item ordering, verified unidimensionality for each subscale, confirmed that item difficulties span the latent trait evenly, and established that individual subscale raw scores can be legitimately converted into linear interval scales free from local item dependency.

7. Validity

The Decision Evaluation Scales have undergone psychometric validation across multiple clinical settings, including clinical genetics, oncology, surgical subspecialties, and reproductive medicine. The evidence for validity spans several domains:

Content and Face Validity

The initial item pool for the DES was derived from qualitative interviews with patients facing difficult clinical decisions, extensive reviews of existing decisional instruments, and expert panels comprising decision analysts, oncologists, medical ethicists, and health psychologists. Rasch item fit diagnostics were iteratively utilized to eliminate ambiguous, redundant, or poorly functioning items, ensuring that the final 15 items represent distinct, clinically meaningful facets of the three core domains without cognitive overload.

Construct and Convergent Validity

Convergent validity has been evaluated against legacy instruments measuring decisional quality and distress:

• Decisional Conflict Scale (DCS; O’Connor, 1995): The DES Satisfaction-Uncertainty subscale correlates strongly and negatively with the DCS Total Score and the DCS Uncertainty subscale ($r = -.62$ to $-.74, p < .001$), confirming that lower decisional uncertainty on the DES mirrors lower decisional conflict.
• Decision Regret Scale (DRS; Brehaut et al., 2003): Retrospective evaluations using the DES demonstrate that patients with lower scores on Satisfaction-Uncertainty and Informed Choice exhibit significantly higher levels of long-term regret ($r = -.58, p < .001$) measured 6 to 12 months post-intervention.
• Perceived Involvement in Care Scale (PICS): The DES Decision Control subscale correlates positively with the PICS Doctor Facilitation of Patient Involvement and Patient Active Information Seeking subscales ($r = .51$ to $.65, p < .001$), supporting its construct validity as a measure of participatory autonomy.

Discriminant Validity

Discriminant validity was established by comparing DES dimensions against non-decisional psychological constructs, including general health-related quality of life (SF-36 physical functioning subscale) and baseline demographic parameters (age, educational attainment, socioeconomic status). As hypothesized, correlations between DES subscales and general physical functioning are negligible to weak ($r = .08$ to $.18$), demonstrating that the DES evaluates specific decisional processes rather than general health state or functional impairment.

Predictive and Known-Groups Validity

Known-groups validation studies show that the DES distinguishes between patients who utilize structured Patient Decision Aids (PtDAs) versus those receiving routine clinical care. In controlled trials, patients provided with balanced PtDAs score significantly higher on the Informed Choice subscale ($d = 0.54, p < .01$) and display lower uncertainty on the Satisfaction-Uncertainty scale ($d = 0.42, p < .05$). Furthermore, higher scores on Decision Control predict higher treatment adherence and lower rates of post-surgical decisional regret over long-term clinical follow-up.

8. Reliability

The reliability of the Decision Evaluation Scales has been established using both Classical Test Theory metrics (internal consistency and test-retest coefficients) and Item Response Theory / Rasch indices (person separation reliability and item separation indices):

Internal Consistency

Empirical investigations across diverse patient samples demonstrate strong internal consistency for the three Rasch-calibrated subscales:

• Satisfaction-Uncertainty: Cronbach’s $\alpha$ values range from .82 to .88 across validation studies, demonstrating high homogeneity among items assessing emotional resolution and confidence.
• Informed Choice: Cronbach’s $\alpha$ values typically fall between .78 and .84, confirming cohesive measurement of knowledge adequacy and informational transparency.
• Decision Control: Cronbach’s $\alpha$ coefficients range between .79 and .85, indicating strong consistency in assessing patient agency and autonomy.

Rasch Reliability and Separation Indices

Because traditional Cronbach’s alpha assumes tau-equivalence and unbounded continuous scales, the Rasch reliability indices provide more rigorous evidence of measurement stability:

• Person Separation Reliability: Coefficients range from .76 to .84, indicating that the scale reliably discriminates between at least three distinct strata of patient decisional appraisals (low, moderate, and high capability/satisfaction).
• Item Separation Reliability: Values exceed .95 with separation indices $> 3.0$, confirming that the item difficulty hierarchy is robust and stable across different clinical samples.

Test-Retest Stability

In stable patient cohorts re-assessed across a 2- to 3-week interval without intervening clinical developments, intraclass correlation coefficients (ICC) confirmed satisfactory reproducibility:

• Satisfaction-Uncertainty: $\text{ICC} = 0.81$ (95% CI: 0.74–0.87)
• Informed Choice: $\text{ICC} = 0.77$ (95% CI: 0.69–0.83)
• Decision Control: $\text{ICC} = 0.79$ (95% CI: 0.71–0.85)

9. Factor Analysis & Rasch Calibration

The structural integrity of the 15-item Decision Evaluation Scales was evaluated through both exploratory/confirmatory factor analytic approaches and modern probabilistic Rasch modeling.

Rasch Analysis and Infit/Outfit Diagnostics

During scale construction, item suitability was evaluated using Georg Rasch’s dichotomous and polytomous rating scale models. Item fit was determined using Mean Square (MNSQ) residual statistics:

• Infit MNSQ: Values for the retained 15 items fell strictly within the acceptable psychometric window of 0.70 to 1.30, indicating that the items conform to model expectations without degrading the measurement system.
• Outfit MNSQ: Similarly ranged from 0.72 to 1.28, showing an absence of unexpected outlier responses from participants responding aberrant to their overall trait level.
• Differential Item Functioning (DIF): Rasch DIF diagnostics confirmed an absence of systematic measurement bias across age cohorts (younger vs. older adults), gender, or primary medical diagnoses (oncology vs. elective surgery), establishing measurement invariance.

Principal Component Analysis of Residuals (PCAR)

To verify the unidimensionality of each independent subscale, Principal Component Analysis of Rasch Residuals was performed. For each of the three subscales:

• The primary Rasch dimension accounted for more than 50% of the total empirical variance.
• The eigenvalue of the first residual contrast (unexplained variance in the first contrast) was $< 1.9$, falling below the standard empirical cutoff of 2.0. This confirms that each subscale functions as a unidimensional latent trait.

Confirmatory Factor Analysis (CFA) Fit Indices

When examined under a Classical Test Theory framework, a three-factor oblique structural equation model confirms the theoretical configuration over a unidimensional or two-factor alternative. Fit indices from confirmatory factor analytic investigations report:

• Comparative Fit Index (CFI): .948 to .965 (exceeding the $ge .95$ benchmark for good fit).
• Tucker-Lewis Index (TLI): .939 to .958.
• Root Mean Square Error of Approximation (RMSEA): .048 to .056 (95% CI: .038–.065), confirming low residual error.
• Standardized Root Mean Square Residual (SRMR): .044.
• Factor Loadings: Standardized lambda ($lambda$) coefficients across all 15 items range from .62 to .86 on their designated latent factors, with no meaningful cross-loadings observed.

10. Instrument / Measurement Tool

The Decision Evaluation Scales is a standardized, self-report instrument designed for either pen-and-paper or digital administration. Below are the administrative parameters and structural characteristics of the tool:

• Test Type: Multi-item, multidimensional patient-reported outcome measure (PROM) / evaluative survey.
• Administration Format: Self-administered (paper-and-pencil, tablet, online survey portal) or structured face-to-face/telephone interview.
• Item Count: 15 items total, partitioned into three calibrated subscales:
  • Satisfaction-Uncertainty: Variable allocation based on scoring iteration (typically 5 to 6 items).
  • Informed Choice: Typically 5 items measuring knowledge clarity and informational adequacy.
  • Decision Control: Typically 4 to 5 items evaluating perceived personal agency and clinical dialogue control.
• Target Population: Adult and elderly clinical populations ($ge 18$ years of age) who have made, or are in the process of making, a discrete medical decision.
• Estimated Completion Time: Approximately 4 to 7 minutes.
• Response Format: 5-point Likert rating scale (commonly formatted from 1 = Strongly Disagree to 5 = Strongly Agree, or 1 = Strongly Agree to 5 = Strongly Disagree depending on specific linguistic translation; clear anchoring must be maintained).
• Scoring Rules:
  • Reverse Coding: Items framed negatively (e.g., measuring doubt, lingering hesitation, feeling pushed into a choice) must be reverse-scored prior to aggregation so that higher numerical values systematically reflect more positive decisional appraisals (higher satisfaction, higher knowledge clarity, higher autonomous control).
  • Subscale Scores: Raw item scores within each domain are summed and averaged to generate a mean subscale score (ranging from 1.0 to 5.0).
  • Linear Rescaling: For research and benchmarking, subscale scores are frequently converted to a 0–100 standardized scale using the formula:
    $$\text{Standardized Score} = \left( \frac{\text{Mean Raw Score} – 1}{4} \right) \times 100$$
  • Total Score: A global aggregate score is generally not recommended because Rasch modeling confirms multidimensionality; reporting separate subscale trajectories preserves clinical granularity.
  • Missing Data Handling: If $le 20%$ of items within a subscale are missing, mean imputation using the remaining items of that subscale is permitted. If $> 20%$ are missing, the subscale score should be treated as missing.

11. Permissions & Fee and Test Year

The Decision Evaluation Scales was officially published in 2005 by Dr. Peep F. M. Stalmeier and colleagues at the Radboud University Medical Center (Nijmegen, Netherlands). As an academic instrument developed through public research funding, the DES is generally made accessible free of charge for non-commercial academic research, clinical quality improvement initiatives, and educational use.

However, users must adhere to ethical standards and copyright conventions:

• Commercial Use: Pharmaceutical corporations, commercial healthcare vendors, or for-profit technology firms wishing to incorporate the DES into proprietary products must seek formal permission or licensing agreements from the primary author and the technology transfer office of Radboud University Medical Center.
• Academic / Non-Profit Use: Researchers are requested to cite the original 2005 validation papers in all resultant publications and conference presentations. Modification of item phrasing, deletions, or structural alterations invalidate the Rasch calibration parameters and require explicit justification.
• Inquiries and Contact: Inquiries regarding authorized translations, full Dutch/English operational manuals, or scoring syntax can be directed to the Department for Health Evidence at Radboud University Medical Center, Nijmegen, Netherlands.

12. References

Below is a curated list of peer-reviewed literature and theoretical foundations associated with the Decision Evaluation Scales:

• Brehaut, J. C., O’Connor, A. M., Wood, W. V., Hack, T. F., Siminoff, L., Gordon, E., & Feldman-Stewart, D. (2003). Validation of a decision regret scale. Medical Decision Making, 23(4), 281–292. https://doi.org/10.1177/0272989X03256005
• Charles, C., Gafni, A., & Whelan, T. (1997). Shared decision-making in the medical encounter: What does it mean? (or it takes at least two to tango). Social Science & Medicine, 44(5), 681–692. https://doi.org/10.1016/S0277-9536(96)00221-3
• Elwyn, G., Frosch, D., Thomson, R., Joseph-Williams, N., Lloyd, A., Kinnersley, P., Cording, E., Tomson, D., Dodd, C., Rollnick, S., Edwards, A., & Barry, M. (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
• Festinger, L. (1957). A theory of cognitive dissonance. Stanford University Press.
• Janis, I. L., & Mann, L. (1977). Decision making: A psychological analysis of conflict, choice, and commitment. Free Press.
• O’Connor, A. M. (1995). Validation of a decisional conflict scale. Medical Decision Making, 15(1), 25–30. https://doi.org/10.1177/0272989X9501500105
• Rasch, G. (1960). Probabilistic models for some intelligence and attainment tests. Danish Institute for Educational Research.
• Stalmeier, P. F. M., Roosmalen, M. S., Verhoef, L. C., Hoekstra-Weebers, J. E., Oosterwijk, J. C., Moog, U., & van Daal, W. A. (2005). Decision evaluation scales: Micro-analysis of patient evaluations of medical decisions. Patient Education and Counseling, 57(3), 346–354. https://doi.org/10.1016/j.pec.2004.09.003
• van Roosmalen, M. S., Stalmeier, P. F. M., Verhoef, L. C., Hoekstra-Weebers, J. E., Oosterwijk, J. C., Hoogerbrugge, N., & van Daal, W. A. (2004). Impact of BRCA1/2 testing and prophylactic mastectomy on women’s quality of life and psychological well-being. European Journal of Cancer, 40(8), 1177–1184. https://doi.org/10.1016/j.ejca.2003.12.027

13. Items of the Scale