Observing Patient Involvement in Decision Making

Abstract

The Observing Patient Involvement in Decision Making scale (commonly designated as the OPTION scale or OPTION-12) is an observer-rated psychometric instrument engineered to evaluate the extent to which healthcare professionals involve patients in shared decision-making (SDM) during clinical consultations. Developed by Professor Glyn Elwyn and an international consortium of health communication researchers in 2003, the scale was created to address a critical methodological limitation in medical interaction analysis: the pervasive ceiling effects and subjective variance characteristic of patient-reported experience measures (PROMs/PREMs). Comprising 12 standardized behavioral items, the instrument requires trained external raters to assess recorded (audio or video) or directly observed clinical encounters using an anchor-defined 5-point Likert-type rating scale ranging from 0 (“The behavior is not observed”) to 4 (“The behavior is exhibited to a very high standard”).

Psychometrically, the OPTION scale yields a raw composite score between 0 and 48, which is linearly converted to a standardized percentage score (0 to 100). Extensive psychometric validation across primary care, oncology, cardiology, and chronic illness contexts indicates adequate to strong internal consistency (Cronbach’s alpha typically ranging from 0.70 to 0.88) and acceptable inter-rater reliability (Intraclass Correlation Coefficients, ICC > 0.65 following standardized rater training). While structurally designed as a unidimensional index of clinician-facilitated shared decision-making, exploratory and confirmatory factor analyses have occasionally identified bi-factor solutions parsing cognitive-informational exchange from affective-deliberative partnership. The instrument serves as a primary benchmark in implementation science, medical education, and clinical trials examining decision aids and health communication interventions worldwide.

Keywords

Shared decision making, OPTION scale, clinician-patient communication, observer-rated measurement, patient involvement, medical decision making, clinical communication analysis, psychometrics, inter-rater reliability, patient autonomy, decision aids, health services research.

Authors

The OPTION scale was conceptualized and empirically validated by a multidisciplinary team of medical researchers, primary care physicians, and psychometricians led by Glyn Elwyn, BA, MB BCh, MSc, PhD. The primary development cohort included:

• Glyn Elwyn – Department of General Practice, Centre for Health Sciences Research, Cardiff University, Wales, United Kingdom (subsequently Professor at The Dartmouth Institute for Health Policy and Clinical Practice, Dartmouth College, Hanover, New Hampshire, USA).
• Adrian Edwards – Department of General Practice, Centre for Health Sciences Research, Cardiff University, Wales, United Kingdom.
• Michel Wensing – Centre for Quality of Care Research (WOK), Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands.
• Kerenza Hood – Department of General Practice and South East Wales Trials Unit (SEWTU), Cardiff University, Wales, United Kingdom.
• Jacqueline Atwell – Department of General Practice, Cardiff University, Wales, United Kingdom.
• Richard Grol – Centre for Quality of Care Research (WOK), Radboud University Nijmegen Medical Centre, Nijmegen, The Netherlands.

Institutional correspondence regarding the early development of the OPTION instrument was managed through the Primary Care Sciences Research Group at Cardiff University and later through The Dartmouth Institute for Health Policy and Clinical Practice.

Purpose

The primary clinical and scientific purpose of the OPTION scale is to provide an objective, reliable, and standardized behavioral metric for measuring the execution of shared decision-making behaviors by clinicians during medical encounters. Prior to the establishment of the OPTION instrument, investigations into patient-centered care and collaborative medical choices depended almost entirely on post-consultation patient surveys. Methodologists recognized that patient-reported metrics regularly manifest severe ceiling effects; patients routinely report high levels of satisfaction and perceive high involvement even in consultations governed by classic paternalistic paradigms. Patients often lack the technical benchmark knowledge required to evaluate whether alternative therapeutic options, clinical equipoise, or the option of active surveillance (“doing nothing”) were omitted or presented incompletely.

The OPTION scale circumvents these biases by utilizing third-party observer assessment. The scale measures the actual communicative conduct of healthcare providers rather than the subjective affective response of the patient. The theoretical rationale rests on the principle that shared decision-making is an overt clinical competency composed of observable, teachable, and measurable micro-behaviors. These behaviors include framing a decision problem, asserting that multiple management pathways exist, clarifying the pros and cons of viable alternatives, exploring patient values and preferences, and negotiating the timing of a definitive choice.

In research contexts, the instrument is utilized as an endpoint measure in randomized controlled trials (RCTs) evaluating patient decision aids, clinician communication workshops, and health system structural reforms. In medical education and clinical quality assurance, the OPTION scale provides a diagnostic rubric for formative feedback during residency training, objective structured clinical examinations (OSCEs), and continuing professional development, identifying specific communicative gaps where clinicians fail to invite patient agency or clarify clinical equipoise.

Psychological Construct

The psychological construct operationalized by the OPTION scale is Clinician-Facilitated Shared Decision Making. In clinical psychology and behavioral medicine, shared decision-making is understood as a dynamic interpersonal process wherein healthcare providers and patients collaborate to reach healthcare choices supported by clinical evidence and patient preferences. Within the OPTION framework, this construct is conceptualized primarily as a clinician-driven communicative competency, reflecting the ethical reality that healthcare providers wield structural power and medical knowledge in the clinical dyad and must actively structure the communicative space to facilitate patient engagement.

The construct encompasses several interrelated behavioral and cognitive dimensions:

1. Problem Framing and Equipoise Awareness

The preliminary phase of the construct involves explicitly signaling to the patient that a diagnostic, management, or therapeutic juncture has been reached that necessitates a decision (Item 1). Central to this dimension is the communicative presentation of clinical equipoise (Item 2)—the explicit acknowledgement by the practitioner that medical science or individual trade-offs do not dictate a single mandatory treatment path, and that legitimate alternative strategies exist.

2. Informational Tailoring and Choice Architecture

The second dimension addresses the cognitive processing of medical risk and uncertainty. It operationalizes how clinicians calibrate information delivery to the patient’s individual cognitive preferences (Item 3), enumerate valid options including conservative management or watchful waiting (Item 4), and systematically dissect the comparative risks, benefits, and uncertainties of each pathway (Item 5). This dimension prevents clinicians from presenting a pre-selected management plan disguised as open consultation.

3. Elicitation of Patient Values, Expectations, and Fears

Shared decision-making requires that patient internal states be explicitly integrated into the clinical deliberation. The OPTION construct requires the clinician to explore patient ideas, expectations, and subjective illness models (Item 6), while proactively eliciting affective concerns, phobias, and emotional barriers regarding possible treatments (Item 7). Rather than assuming patient priorities, the practitioner must actively elicit what matters most to the individual patient.

4. Comprehension Verification and Dialogic Symmetry

This dimension reflects communicative reciprocity and metacognitive monitoring. The clinician must check patient comprehension of complex probabilities and medical concepts (Item 8) and provide clear, unhurried opportunities for questions and clarification (Item 9). This prevents informational asymmetry and reduces the power gradient between expert and layperson.

5. Autonomy Negotiation and Temporal Structuring

The final dimension focuses on negotiating decision agency and timing. The clinician must explicitly ask what level of decision-making role the patient desires (e.g., autonomous, shared, or delegated) (Item 10). The clinician must also define a clear decision-making stage, acknowledging that choices may be deferred (Item 11), and formalize a mechanism for longitudinal follow-up and review (Item 12). This establishes that healthcare decisions are iterative processes rather than irreversible clinical events.

Theoretical Framework

The OPTION scale is grounded in normative and descriptive models of interpersonal communication, bioethics, and cognitive decision theory. Its primary structural foundation emerges from the normative model of shared decision-making formulated by Charles, Gafni, and Whelan (1997, 1999). Charles and colleagues differentiated shared decision-making from both the paternalistic model (where the clinician decides autonomously on the patient’s behalf) and the informed/consumerist model (where the clinician functions as an information vendor, leaving the patient to choose in isolation). In the Charles framework, shared decision-making requires:

• The active involvement of both clinician and patient;
• Reciprocal information exchange involving clinical evidence from the clinician and personal values from the patient;
• Simultaneous participation in the deliberation process; and
• A jointly agreed-upon treatment decision.

The OPTION scale also draws upon Relational Autonomy Theory from contemporary bioethics (Mackenzie & Stoljar, 2000). Relational autonomy rejects the individualistic assumption that patients can independently evaluate complex interventions without relational support. Instead, it posits that autonomy is fostered within supportive interpersonal relationships. Under this framework, clinicians carry an affirmative professional duty to create the conversational conditions under which a patient’s capacity for self-determination can develop.

From cognitive psychology and behavioral decision theory (Kahneman & Tversky; Edwards), the OPTION scale integrates concepts of heuristic bias, framing effects, and decisional conflict. Because human decision-makers under stress revert to cognitive shortcuts, clinicians must structure choice architecture carefully to prevent framing biases. By systematically defining options, exploring trade-offs, and explicitly separating the informational deliberation phase from the commitment phase, the OPTION framework operationalizes a clinical process designed to reduce decisional conflict and post-decisional regret (Janis & Mann, 1977).

Validity

Extensive international validation studies have evaluated the construct, criterion, convergent, and discriminant validity of the OPTION scale across diverse clinical contexts.

Construct Validity

Construct validity has been established by testing hypotheses regarding clinical scenarios where shared decision-making is functionally required versus scenarios where emergency or protocolized care dominates. In early validation studies by Elwyn et al. (2003, 2005), the OPTION scale showed clear sensitivity to the presence of multiple preference-sensitive treatment options. Clinicians dealing with chronic disease management, elective surgery, or prostate-specific antigen (PSA) testing systematically exhibited higher OPTION scores than those managing acute conditions with standard clinical indications. Furthermore, randomized controlled trials of educational interventions targeting clinician communication demonstrated statistically significant increases in total OPTION scores post-intervention, confirming that the scale is sensitive to behavioral changes in clinician practice.

Convergent Validity

Evaluating convergent validity presents unique challenges in shared decision-making research due to the known divergence between subjective patient ratings and objective observer ratings. When compared against other observer-rated instruments—such as the Roter Interaction Analysis System (RIAS), the MAPPIN’SDM system, and the subsequently developed OPTION-5 scale—the OPTION-12 scale demonstrates strong convergent correlations (Pearson’s r typically ranging from 0.62 to 0.79). Conversely, correlations between OPTION scores and patient-reported measures (such as the Decisional Conflict Scale or the CollaboRATE scale) are frequently low to modest (r = 0.12 to 0.30). Far from demonstrating psychometric failure, this divergence confirms that observer ratings capture specific communicative behaviors that patient satisfaction surveys consistently fail to measure due to positive reporting bias.

Discriminant Validity

Discriminant validity has been demonstrated by showing that OPTION scores do not correlate with unrelated clinical variables, such as general medical knowledge, physician technical diagnostic accuracy, or purely structural consultation variables. While consultation length does show a weak-to-moderate positive correlation with OPTION scores (longer consultations provide more opportunity for shared decision-making behaviors; r = 0.25 to 0.45), multivariable analyses confirm that high OPTION scores can be achieved in time-constrained encounters when clinicians use efficient, structured communication techniques.

Reliability

The psychometric evaluation of the OPTION scale focuses heavily on inter-rater reliability (IRR) and internal consistency, given its design as an observer-rated tool requiring human scoring of recorded interactions.

Internal Consistency

The scale consistently exhibits acceptable to high internal consistency across independent studies and linguistic adaptations. In the initial validation study by Elwyn et al. (2003), the overall instrument yielded a Cronbach’s alpha (α) of 0.79. Subsequent international validation investigations in German, Dutch, French, Spanish, and Mandarin Chinese clinical settings have yielded alpha coefficients ranging between 0.72 and 0.88, demonstrating that the 12 items form a coherent behavioral measurement scale.

Inter-Rater Reliability (IRR)

Because the OPTION scale relies on observer assessment, inter-rater reliability is a central psychometric requirement. In the foundational validation studies involving pairs of trained raters evaluating audio-recorded primary care consultations, inter-rater reliability measured via Intraclass Correlation Coefficients (ICC) ranged from 0.62 to 0.77 for total scores. Studies utilizing generalizability theory (G-theory) indicate that rater variance accounts for a minority of total variance when raters complete standardized training manuals and calibration sessions (typically requiring 8 to 15 hours of baseline practice with standardized anchor tapes).

Item-level reliability varies across the 12 items. Behaviors with overt linguistic markers—such as Item 1 (identifying that a decision must be made), Item 4 (listing options), and Item 5 (explaining pros and cons)—consistently achieve high inter-rater agreement (weighted Cohen’s kappa κ > 0.70). Conversely, subtle interpersonal behaviors, notably Item 3 (assessing patient information preferences) and Item 10 (eliciting preferred involvement level), show lower item-level agreement (κ = 0.40 to 0.58), reflecting the difficulty of detecting nuanced conversational cues without explicit verbal prompts.

Test-Retest Stability

Intra-rater reliability (temporal stability of an individual rater scoring identical recorded consultations at two separate time points) is generally high, with ICCs regularly exceeding 0.85. This demonstrates that scoring drift can be minimized when raters adhere closely to the operationalized coding manual.

Factor Analysis

The latent structural dimensions of the 12-item OPTION instrument have been evaluated using both exploratory factor analysis (EFA) and confirmatory factor analysis (CFA).

Exploratory Factor Solutions

Although Elwyn et al. designed the OPTION scale as a single, unidimensional behavioral index reflecting clinician effort toward shared decision-making, early exploratory factor analyses across diverse datasets frequently yielded two distinct, correlated latent factors accounting for 45% to 58% of the total variance:

• Factor 1: Information Provision and Problem Structuring – Comprising Items 1, 2, 4, 5, 8, and 11. This factor represents technical communicative tasks: identifying decisions, stating equipoise, detailing comparative risks/benefits, checking comprehension, and outlining decision stages.
• Factor 2: Relational Deliberation and Value Elicitation – Comprising Items 3, 6, 7, 9, 10, and 12. This factor represents affective and partnership-building tasks: assessing information preferences, exploring fears/expectations, inviting questions, eliciting desired involvement levels, and arranging longitudinal reviews.

Confirmatory Factor Analysis and Structural Fit

Confirmatory factor analytic investigations evaluating the unidimensional model against a two-factor correlated model have found that the two-factor model often demonstrates superior fit indices. Representative fit statistics across structural equation modeling (SEM) evaluations include:

• Comparative Fit Index (CFI): 0.91 – 0.95 (two-factor model) versus 0.83 – 0.88 (unidimensional model);
• Tucker-Lewis Index (TLI): 0.90 – 0.94;
• Root Mean Square Error of Approximation (RMSEA): 0.055 – 0.071 (with 90% confidence intervals within acceptable ranges);
• Standardized Root Mean Square Residual (SRMR): 0.048 – 0.062.

Item loadings across factor analyses are consistently highest for Items 1, 4, and 5 (λ > 0.70), confirming that presenting treatment choices and outlining pros and cons are central to observer evaluations of shared decision-making. Factor loadings are lower for Items 3 and 10 (λ = 0.35 to 0.52), primarily because clinicians rarely exhibit these two behaviors in routine practice, leading to floor effects at the item level.

Instrument / Measurement Tool

The structural characteristics, administration requirements, and scoring algorithms for the OPTION scale are summarized below:

• Instrument Name: Observing Patient Involvement in Decision Making (OPTION / OPTION-12).
• Target Population: Healthcare professionals (physicians, nurse practitioners, clinical specialists) conducting clinical consultations with adults or older adults across primary, secondary, and tertiary care.
• Measurement Paradigm: Objective third-party behavioral rating scale based on direct observation or audio/video recordings of medical consultations.
• Number of Items: 12 behavioral statements.
• Response Format: 5-point anchor-defined rating scale:
  • 0 = The behavior is not observed
  • 1 = A minimal attempt is made to exhibit the behavior
  • 2 = The behavior is clearly observed
  • 3 = The behavior is exhibited to a high standard
  • 4 = The behavior is exhibited to a very high standard
• Administration Duration: Real-time consultation duration plus approximately 15–25 minutes of coding time per recorded consultation by a calibrated rater.
• Scoring Protocol:
  • Each item is scored on an ordinal integer scale from 0 to 4.
  • No reverse scoring applies; higher scores consistently reflect greater demonstration of shared decision-making behaviors.
  • Total Raw Score: Sum of all 12 items, yielding a range from 0 to 48.
  • Standardized Rescaled Score: The raw score is linearly transformed into a standard percentage scale ranging from 0 to 100 using the following formula:
    Standardized Score = (Total Raw Score / 48) × 100
• Rater Training Requirement: Reliable administration requires raters to complete a calibration protocol using the official OPTION manual, typically evaluating 5 to 10 benchmark consultations to achieve acceptable inter-rater agreement (ICC ≥ 0.70) prior to formal study coding.

Permissions & Fee and Test Year

The OPTION scale was first published in 2003 by Glyn Elwyn and colleagues in Quality and Safety in Health Care (now BMJ Quality & Safety). The instrument was placed in the academic and scientific domain to advance healthcare communication research, education, and implementation science. It is accessible for non-commercial research, academic, and clinical educational purposes without licensing fees, provided formal citation and attribution are given to the original development literature.

Subsequent iterations, such as the shortened 5-item version (OPTION-5, published in 2014), are similarly available for health services research through The Dartmouth Institute for Health Policy and Clinical Practice. Researchers conducting funded commercial investigations or institutional quality-improvement programs should review Dartmouth Institute guidelines and contact the corresponding authors to verify adherence to current coding manuals and established scoring rubrics.

References

• 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
• Charles, C., Gafni, A., & Whelan, T. (1999). Decision-making in the physician-patient encounter: Revisiting the shared treatment decision-making model. Social Science & Medicine, 49(5), 651–661. https://doi.org/10.1016/s0277-9536(99)00145-8
• Couët, N., Desroches, S., Robitaille, H., Vaillancourt, H., Leblanc, A., Turcotte, S., Elwyn, G., & Légaré, F. (2015). Assessments of the extent to which health-care providers involve patients in decision making: A systematic review of studies using the OPTION instrument. Health Expectations, 18(4), 542–561. https://doi.org/10.1111/hex.12054
• Elwyn, G., Edwards, A., Wensing, M., Hood, K., Atwell, J., & Grol, R. (2003). Shared decision making: Developing the OPTION scale for measuring patient involvement. Quality and Safety in Health Care, 12(2), 93–99. https://doi.org/10.1136/qhc.12.2.93
• Elwyn, G., Edwards, A., Wensing, M., & Grol, R. (2005). Shared decision making: Identifying the criteria for evaluating the quality of patient involvement (OPTION). Zeitschrift für Ärztliche Fortbildung und Qualitätssicherung, 99(8), 517–524.
• Elwyn, G., Hutchings, H., Edwards, A., Rapport, F., Grol, R., & Wensing, M. (2005). The OPTION scale: Measuring the extent that clinicians involve patients in decision making tasks. Health Expectations, 8(1), 34–42. https://doi.org/10.1111/j.1369-7625.2004.00311.x
• Elwyn, G., Tsulukidze, M., Edwards, A., Légaré, F., & Newcombe, R. (2013). Using a ‘talk’ model of shared decision making to propose an update to the OPTION item text: OPTION5. Patient Education and Counseling, 93(1), 102–107. https://doi.org/10.1016/j.pec.2013.04.019
• Makoul, G., & Clayman, M. L. (2006). An integrative model of shared decision making in medical encounters. Patient Education and Counseling, 60(3), 301–312. https://doi.org/10.1016/j.pec.2005.06.010
• Stiggelbout, A. M., Pieterse, A. H., & De Haes, J. C. (2015). Shared decision making: Concepts, evidence, and practice. Patient Education and Counseling, 98(10), 1172–1179. https://doi.org/10.1016/j.pec.2015.06.022

Items of the Scale