Development and Validation of a Patient-Centered Scale to Evaluate Decision Aid Usability and Information Overload

Abstract

Introduction

Decision aids aim to improve the quality of and satisfaction with decision-making. Few scales exist that directly evaluate decision aids from the patient’s perspective, particularly with respect to information overload.

Methods

We developed a Decision Aid Evaluation Scale with domains assessing acceptability, satisfaction, cognitive load, and helpfulness for decision-making. Cognitive interviews and iterative testing were performed. Convergent and discriminant construct validity were assessed by comparing the novel scale with domains from the validated Decisional Conflict Scale (DCS; values clarity, uncertainty, feeling informed subscales). Internal consistency was measured using Cronbach’s alpha. The final scale consisted of Likert-type scale items assessing information amount, values clarity, ease of use, cognitive effort required, nervousness, and overall helpfulness. Participants (any sex or gender, aged 40-60) completed the scale after viewing a cancer screening decision aid within a hypothetical screening decision scenario.

Results

We surveyed 1249 participants; 778 completed the Decision Aid Evaluation Scale. As hypothesized, a larger proportion of participants who reported low cognitive load had higher DCS values clarity subscale scores than those with high cognitive load (P=0.003). Participants who found the decision aid helpful had significantly higher DCS values clarity (P<0.001), informed (P=0.004), and certainty subscale scores (P<0.001). Scores on the acceptability questions were higher among those who found the decision aid helpful (P<0.001), except for questions related to “length,” “relatability of photo,” and “photo helped make decision.” Participants reporting nervousness had lower DCS certainty subscale scores (P<0.001). Cronbach’s alpha for the acceptability and satisfaction domains were 0.74 and 0.75, respectively, indicating internal consistency.

Conclusions

We developed a patient-centered scale that captures users’ perceptions of decision aid usability and information overload. The scale demonstrated acceptable construct validity and internal consistency, supporting its use in evaluating decision aids from the patient perspective.

Introduction

Shared decision-making (SDM) is an essential component of modern clinical practice, in which patients and clinicians collaborate to make informed health decisions that align with patients’ values and the best available evidence.^1,2 Patient decision aids play a key role in SDM, particularly for preference-sensitive medical decisions such as prostate cancer or breast cancer screening.^3-6 Substantial research has shown that decision aids improve the quality of patient decision-making and enhance patient satisfaction.^7-10

However, some decision aids, in a well-intended attempt to ensure patients are well-informed, may inadvertently overwhelm patients with excessive and complex information. This overload often arises when decision aids include complicated quantitative data. Medical decisions, particularly those involving cancer screening or management, are inherently complex and require patients to express preferences related to unfamiliar health states.¹¹

Cognitive overload occurs when the volume of information exceeds a patient’s ability to process it effectively, potentially impairing decision-making ability.^12-14 Rather than viewing complexity as inherently negative, it is important to acknowledge that striking a balance between simplifying information and providing enough detail to support patient autonomy is critical. Studies have consistently shown that patients with low health literacy or numeracy may struggle to comprehend risk estimates or multiple quantitative data points, leading to confusion and decision fatigue.^15-17 We recently developed “gist-based” decision aids and demonstrated that simplifying information and focusing on key “gist” messages reduces cognitive burden.^18,19

Information overload remains an underexplored area in decision aid design that many existing frameworks do not explicitly address.^9,20-22 Existing instruments used during early decision aid development such as the Ottawa Acceptability Questionnaire²², Preparation for Decision-Making Scale²³, IPDAS-based usability checklists⁹, and reporting frameworks such as SUNDAE²⁴, typically assess feasibility, satisfaction, and usability but do not evaluate cognitive burden or information overload. As a result, there remains no brief, patient-centered measure that directly assesses the cognitive effort required when interacting with a decision aid.

While existing tools provide essential guidance for ensuring clarity and values alignment, they do not fully address the complexity of managing information overload during decision-making. The present study responds to this gap by developing and empirically validating a novel Decision Aid Evaluation Scale (DAES), a patient-centered measure with an emphasis on cognitive load.

Methods

Study Design and Recruitment

Recruitment for this study was conducted via ResearchMatch, a national health volunteer registry supported by the U.S. National Institutes of Health as part of the Clinical Translational Science Award (CTSA) program. ResearchMatch provides access to a large population of volunteers who have consented to be contacted for health studies for which they may be eligible. Review and approval for this study, along with all procedures, was obtained from the Institutional Review Board at Memorial Sloan Kettering Cancer Center. Participants were eligible if they were of any sex or gender and between 40 and 60 years old. All participants first reviewed a cancer screening decision aid and subsequently completed our DAES.

Development of Decision Aid Evaluation Scale

The development of the DAES involved a multistep process, leveraging both expert input and patient feedback. A multidisciplinary research team, including clinical, behavioral, and decision science experts, initiated the drafting of the scale. Items were adapted from established instruments, including the Ottawa Acceptability Questionnaire⁸ (with permission from the developer) to assess critical domains such as values clarity, acceptability, and certainty. These items were tailored to align with the goals of the current study, focusing on assessing patients’ experiences with cancer screening decision aids.

The development process also incorporated qualitative data from 3 series of cognitive interviews with patients recruited from ResearchMatch and Kings County Hospital (Brooklyn, NY): Phase 1 concept elicitation interviews (n=32), Phase 2 Round 1 cognitive interviews (n=32), and Phase 2 Round 2 cognitive interviews with new participants (n=16). The iterative nature of these interviews allowed the research team to refine both the decision aids and the evaluation scale simultaneously. This process of development and the results of the qualitative analyses have been described in detail elsewhere, as we simultaneously developed gist-based cancer screening decision aids.^18,25 Notably, the qualitative interviews informed the scale’s ability to capture patients’ perceptions of decision-making values clarity, certainty, and the acceptability of decision aids.

The initial items generated for the evaluation scale underwent rigorous review and refinement to ensure patient comprehension and relevance, based on feedback from both content experts and interview participants. Specific attention was given to capturing information overload and cognitive effort during item generation, including questions addressing the amount of information presented and the effort required to understand it. Cognitive interviews probed whether these items reflected participants’ experiences of being overwhelmed or cognitively burdened, and wording was refined accordingly. The DAES that emerged from this process consists of questions across several domains (e.g., acceptability, values clarity, certainty), each designed to capture key dimensions of decision aid quality. The final version of the scale, question text, available response options, associated domain, and the short-hand reference term used throughout the remainder of this article are detailed in Supplementary Table 1.

Validation of the Decision Aid Evaluation Scale

The design and reporting of this validation phase are in accordance with relevant COSMIN principles for evaluating the measurement properties of patient-reported instruments²⁶. The DAES domains align conceptually with IPDAS criteria; specifically, items addressing clarity and acceptability correspond with IPDAS standards for comprehensibility and presentation of probabilities, while the helpfulness and cognitive load items complement IPDAS by capturing patients’ perceived effort and information burden, dimensions not explicitly measured within current IPDAS checklists.

This study presents the validation of the DAES, building on our previous work that utilized the novel scale in a randomized controlled trial focusing on the impact of increasing the amount of presented information in decision aids.¹⁹ In this earlier trial, the scale was employed to assess the impact of decision aids varying in length, informational depth, and literacy demands on patient outcomes related to SDM and information overload. The present study focuses on the psychometric validation phase, whereas the prior randomized trial informed feasibility and wording but was not designed to evaluate measurement properties.

Scoring of the Decision Aid Evaluation Scale

Scoring procedures are described below to facilitate reproducibility. In summary, domain scores were calculated by taking a weighted average of questions in the domain, where missing any of the questions yielded a missing domain score, and domains were scored so that higher scores corresponded with positive agreement. Of note, the first 2 questions of the acceptability domain regarding the “information amount” and “length” of the decision aid were non-monotonic, with the response levels “Much less than I would like,” “Less than I would like,” “About right,” “A little more than I would like,” and “Much more than I would like.” As the ideal response was “About right,” scoring for these 2 questions was based on the difference in magnitude from the ideal response, rather than direction compared to ideal response (i.e., “much less” and “much more” were grouped together vs “a little more” and “less”). Items related to photos, quotes, and diagrams are optional and may be omitted when not applicable, without altering calculation of core domain scores.

Construct Validity and Hypothesis Testing

To assess the construct validity of the DAES, we explored associations between related constructs using questions from the Decisional Conflict Scale (DCS),²⁷ which was also answered by patients. The DCS captures established decision-making outcomes that the DAES is expected to influence, thereby providing a test of convergent validity. These questions, along with available response options and corresponding domains, are detailed in Supplementary Table 2. Based on theoretical underpinnings, we hypothesized the following associations: 1) participants who reported on the DAES that less effort was needed to read the decision aid would report higher scores on DCS domains related to values clarity, feeling more informed, and feeling greater certainty regarding their choice; 2) participants who reported on the DAES that the decision aid helped them decide whether to undergo testing would report higher DCS scores corresponding to greater values clarity, feeling more informed, and feeling greater certainty regarding their choice, and would also report higher scores on questions related to “acceptability”; and 3) participants who reported on the DAES increased nervousness from the decision aid would report on the DCS less certainty about their choice. We refer to the “uncertainty” domain from the DCS as the “certainty” domain in the text henceforth, given that higher scores reflect increased certainty about decision-making. Formal hypothesis testing utilized the Wilcoxon rank sum test.

Reliability Analysis

The internal consistency of the DAES was assessed using Cronbach’s alpha for each domain. A Cronbach’s alpha value greater than 0.7 was prespecified as indicative of internal consistency.

All analyses were performed using R Statistical Software (v4.3.2) with the ggrain (v0.0.4), tidyverse (v2.0.0), and gtsummary (v1.7.2) packages.^28-30

Results

Scale Development: Cognitive Interviews

The development process included Phase 1 concept elicitation interviews (n=32), followed by Phase 2 cognitive interviews conducted in two rounds (round 1: n=32; round 2 with new participants: n=16). The cognitive interview phase therefore included 48 participants in total. In the first round, participants highlighted the need for simpler, more concise language and clearer instructions for assessing decision aids. Ambiguity in questions related to perceived usefulness and the rating scale for information overload led to revisions, including the adoption of a Likert-type scale to improve consistency and reduce subjectivity. Additional examples and instructions were added to help users accurately rate content comprehensibility and visual presentation.

In the second round of interviews, participants responded positively to these changes, particularly the revised rating scale, though minor adjustments were made to align terminology with patient-centered care frameworks. Feedback also prompted the rewording of response options to reduce cognitive burden and clarify how to evaluate the balance of risks and benefits in decision aids. These modifications streamlined the scale, making it more intuitive and effective in capturing user evaluations.

Participant Characteristics and Response Rates

Within our study, 1249 participants were given the DAES, of whom 778 completed the scale (62% response rate). Table 1 presents patient characteristics based on completion status (where only the characteristic of age was statistically significantly different; P = 0.011) and Table 2 shows the distribution of responses to the individual questions on the DAES.

Table 1.

Participant characteristics based on completion status of the Decision Aid Evaluation Scale.

	Complete, N = 778	Incomplete, N = 471
Age
40–44	171 (24%)	25 (25%)
45–49	174 (25%)	12 (12%)
50–54	166 (23%)	22 (22%)
55–60	199 (28%)	41 (41%)
Unknown	68	371
Gender
Male	369 (48%)	60 (56%)
Female	395 (51%)	47 (44%)
Non-binary	2 (0.3%)	0 (0%)
Prefer to self-describe	3 (0.4%)	0 (0%)
Prefer not to say	3 (0.4%)	0 (0%)
Unknown	6	364
Education
Some or no high school	3 (0.4%)	0 (0%)
Graduated high school/GED	37 (4.8%)	6 (5.5%)
Completed some college	117 (15%)	16 (15%)
Associate’s degree	73 (9.4%)	18 (17%)
Bachelor's degree	251 (32%)	35 (32%)
Graduate school or professional degree	293 (38%)	34 (31%)
Unknown	4	362

Table 2.

Responses to individual questions on the Decision Aid Evaluation Scale.

	N	N = 778
What did you think about the amount of information in the decision aid?	775
Much less than I would like		46 (5.9%)
Less than I would like		157 (20%)
About right		397 (51%)
A little more than I would like		134 (17%)
Much more than I would like		41 (5.3%)
What did you think about the length of the decision aid?	776
Much less than I would like		20 (2.6%)
Less than I would like		90 (12%)
About right		446 (57%)
A little more than I would like		170 (22%)
Much more than I would like		50 (6.4%)
Was the decision aid…Visually appealing?	773
Not at all		60 (7.8%)
Somewhat		438 (57%)
Very much so		275 (36%)
Was the decision aid…Attention-grabbing?	770
Not at all		143 (19%)
Somewhat		424 (55%)
Very much so		203 (26%)
Was the decision aid…Easy to read?	768
Not at all		36 (4.7%)
Somewhat		238 (31%)
Very much so		494 (64%)
Was the decision aid…Educational	758
Not at all		29 (3.8%)
Somewhat		264 (35%)
Very much so		465 (61%)
Did you lose interest while reading the decision aid?	770
Not at all		406 (53%)
Somewhat		284 (37%)
Very much so		80 (10%)
Cognitive Load - How much effort did it take for you to read the decision aid?	774
1 - Not much effort		339 (44%)
2		215 (28%)
3		137 (18%)
4		63 (8.1%)
5 - A lot of effort		20 (2.6%)
Did the decision aid make you feel nervous?	773
Not at all		571 (74%)
Somewhat		177 (23%)
Very much so		25 (3.2%)
Photos of the people on the decision aid - Could you relate to them?	775
Not at all		114 (15%)
Somewhat		387 (50%)
Very much so		149 (19%)
There were none		125 (16%)
Photos of the people on the decision aid - Did they help you make your decision?	774
Not at all		466 (60%)
Somewhat		146 (19%)
Very much so		65 (8.4%)
There were none		97 (13%)
Diagram(s) in the decision aid - Did you understand them?	770
Not at all		15 (1.9%)
Somewhat		140 (18%)
Very much so		477 (62%)
There were none		138 (18%)
Diagram(s) in the decision aid - Did they influence your decision?	773
Not at all		206 (27%)
Somewhat		297 (38%)
Very much so		158 (20%)
There were none		112 (14%)
Helpful - Did this decision aid help you decide (to have the test or not)?	777
Not at all		107 (14%)
Somewhat		430 (55%)
Very much so		240 (31%)
Overall, did the decision aid make you feel more confident about your decision?	775
Not at all		110 (14%)
Somewhat		390 (50%)
Very much so		275 (35%)
How likely are you to recommend this decision aid to a friend or relative?	770
Not at all likely		130 (17%)
Somewhat likely		367 (48%)
Very much likely		273 (35%)

Construct Validity: Cognitive Load and Decisional Conflict Scale Subscales

Figure 1 presents the distribution of the DCS subscales based on the DAES “cognitive load” question. Based on the distribution of the “cognitive load” score, with only 2.6% reporting a value of “5” that has an anchor text of “A lot of effort” (Table 2), this was dichotomized so that values of both 4 and 5 were considered as having a high cognitive load. DCS subscales for values clarity, feeling informed, and certainty are each scaled from 0 to 1, and reverse-coded so that higher scores reflected greater values clarity, feeling more informed, and higher certainty about one’s decision. A larger proportion of participants with low DAES cognitive load had higher DCS values clarity subscale scores, higher DCS informed subscale scores, and higher DCS certainty subscale scores compared to participants with high DAES cognitive load, though the latter two were not statistically significant (P = 0.003, P = 0.4, and P = 0.3, respectively). For example, a larger proportion of patients with a low DAES cognitive load versus high DAES cognitive load had a value of 1 on the DCS values clarity subscale (69% vs 52%), informed subscale (64% vs 59%), and certainty subscale (68% vs 63%), indicating greater values clarity, feeling informed, and certainty, respectively.

Figure 1.

Distribution of subscales from the Decisional Conflict Scale, based on whether participants had high “cognitive load” (orange corresponds to “yes” and blue corresponds to “no” ; i.e., low cognitive load) on the Decision Aid Evaluation Scale. Each of the Decision Conflict Scale subscales range from 0 to 1, with higher scores corresponding to greater (A) values clarity; B) feeling more informed; and C) higher certainty.

Construct Validity: Perceived Helpfulness and Decisional Conflict Scale Subscales

Figure 2 presents the distribution between the DAES “helpful” question (the response of “Very much so” was considered as being helpful, whereas the responses of “Not at all” and “Somewhat” were considered as not being helpful) and the DCS subscales. A larger proportion of participants who rated the decision aid helpful on the DAES had higher DCS values clarity subscale scores (P < 0.001), higher DCS informed subscale scores (P = 0.004), and higher DCS certainty subscale scores (P < 0.001) than participants who rated the decision aid as not being helpful on the DAES. Similarly, a larger proportion of participants who rated the decision aid helpful versus not helpful on the DAES had the highest possible DCS values clarity score (78% vs 62%), informed score (71% vs 60%), and certainty score (86% vs 60%).

Figure 2.

Distribution of subscales from the Decisional Conflict Scale, based on whether participants found the decision aid “helpful” (orange corresponds to “yes” and blue corresponds to “no”) on the Decision Aid Evaluation Scale. Each of the Decision Conflict Scale subscales range from 0 to 1, with higher scores corresponding to greater (A) values clarity; B) feeling more informed; and C) higher certainty.

Construct Validity: Certainty and Nervousness

Figure 3 presents the distribution of the DCS certainty subscale scores based on the response of “Somewhat” or “Very much so” to the DAES question asking about nervousness (“Did the decision aid make you feel nervous?” ; dichotomization due to the low rate of response of “Very much so” [3.2%]). A larger proportion of participants who reported nervousness had lower DCS certainty scores than participants who reported no nervousness (P < 0.001).

Figure 3.

Distribution of the “certainty” subscale on the Decisional Conflict Scale, based on whether participants felt “nervousness” related to the decision aid on the Decision Aid Evaluation Scale (orange corresponds to “yes” and blue corresponds to “no”).

Construct Validity: Acceptability Domain and Helpfulness Ratings

Supplementary Figures 1a-1j present the distribution of the individual questions in the acceptability domain of the DAES based on whether participants found the decision aid helpful. We consistently found that scores on the individual acceptability questions were higher among participants who found the decision aid helpful. All associations were statistically significant at P < 0.001 other than for the questions related to “Length” (P = 0.12), “Relatability of photo” (P = 0.3), and “Photo helped make decision” (P = 0.5). Figure 4 presents the distribution of the overall acceptability domain score, where much like the individual questions from the acceptability domains, there are higher scores among participants who found the decision aid helpful compared with those who did not find it helpful (P < 0.001).

Figure 4.

Distribution of the “acceptability” domain score on the Decision Aid Evaluation Scale, based on whether patients found the decision aid could “help decide” on the Decision Aid Evaluation Scale (orange corresponds to “yes” and blue corresponds to “no”).

Reliability Analysis

Within the domains of the DAES, Cronbach’s alpha was not evaluated for the domains of “cognitive load” and “help decide” since there was only 1 question in both of those domains. The acceptability domain had a Cronbach’s alpha of 0.74 and the satisfaction domain had a Cronbach’s alpha of 0.75, both of which are above the prespecified threshold of 0.70 and therefore deemed internally consistent.

Discussion

The purpose of this study was to develop a novel Decision Aid Evaluation Scale (DAES) and then explore construct validity and internal consistency across several key domains relevant to SDM. We observed that, as hypothesized, people who reported lower cognitive load on this measure while using the decision aid also reported greater clarity in their values, supporting the notion that cognitive load plays a crucial role in the decision-making process. Additionally, patients who found the decision aid helpful as rated on this novel measure were more likely to have higher scores in feeling “informed,” and “certainty” subscales of the validated and widely used DCS. These findings suggest that both the perceived helpfulness of a decision aid and the cognitive effort required to engage with it are important to its effectiveness.

Our analysis also demonstrated the internal consistency of the DAES, with Cronbach’s alpha values exceeding the 0.7 threshold in the acceptability (0.74) and satisfaction (0.75) domains. This suggests that the items within these domains reliably measure the intended constructs, thereby supporting the scale’s use in evaluating decision aids.

Our findings are important because of a common yet underexplored limitation of decision aids: the concept of information overload. Previous research has demonstrated that decision aids, particularly values clarification methods, are often at odds with contemporary scientific understandings of psychology.¹² Indeed, the contemporary standard practice for developing a decision aid as outlined by both the International Patient Decision Aid Standards (IPDAS) and the Ottawa Decision Support Framework does not explicitly consider informational or cognitive overload during design.^9,11,20-22 However, for patients to make truly informed decisions, optimal delivery and processing of information are paramount. Our novel evaluation scale incorporates appropriate elements into the assessment of a decision aid. In a recent publication, our group presented two “gist-based” decision aids for prostate and breast cancer screening, both of which were found to be acceptable, feasible, and promoted effective decision-making while limiting information overload.^18,25 The present study builds on these findings by empirically validating our DAES and providing evidence that lower cognitive load and perceived helpfulness are associated with better decision-making outcomes.

There is substantial evidence of the adverse effects of information overload, which include increased confusion, delay in decision-making, and inferior satisfaction with decisions.^31-34 Related to the concept of information overload is the reported “cognitive load” of a decision aid. Cognitive load itself is multifactorial, and includes characteristics of the individual (e.g., cognitive abilities), the task (e.g., its complexity), and the environment (e.g., noise levels).^35,36 Decision aids are often complex, inundating patients with quantitative data, and frequently written at a literacy level that surpasses that of the average U.S. adult. For example, a meta-analysis of 11 patient decision aids found that none were written at a sixth-grade level or below (the average American reads at a seventh to eighth grade level).³⁷ These findings underscore the complex trade-offs that decision aid developers must navigate: balancing the desire to provide complete and accurate information with the need to present it in a way that is cognitively manageable for users with varying health literacy and numeracy levels. Our results highlight the practical difficulty of integrating these considerations into decision aid design, as well as the importance of tools like the DAES to identify areas where cognitive burden may hinder effective decision making. Our observed association between lower cognitive load and higher clarity of values supports the importance of using decision aids that are both easy to read and comprehend.

It is important to note that while excessive cognitive load can undermine decision quality, cognitive effort itself is neither inherently undesirable nor always avoidable. Many medical decisions – particularly cancer screening and treatment decisions – are complex and require deliberation. The concern, therefore, is not cognitive load per se, but the point at which informational demands exceed an individual’s processing capacity. The DAES is intended not to eliminate cognitive effort, but to help identify when informational burdens become counterproductive to effective decision-making.

A central objective of any decision aid is to support patients in making informed medical decisions that align with their values and preferences. Our study underscores the importance of patients’ perceptions of a decision aid’s “helpfulness” in achieving this goal, demonstrating a clear association between perceived helpfulness and improved decisional outcomes (e.g., values clarity, feeling informed, and certainty). Our findings emphasize that an important component of the effectiveness of a decision aid is its perceived helpfulness by the patients who use it. Yet, a Cochrane Review on decision aids, which included 209 studies involving 107,698 participants, primarily assessed decision-making outcomes such as knowledge, decisional conflict, and engagement in SDM.³⁸ There are few if any standardized measures of how patients perceived the decision aid’s helpfulness or usability.

The primary strength of this work is that it addresses an unmet need for a validated, patient-centered scale that evaluates critical dimensions of decision aid usability and effectiveness, including a decision aid’s ability to mitigate information overload. While the Ottawa framework emphasizes key elements necessary for high-quality decision aids – including values clarity and informativeness – our scale expands upon these principles by incorporating measures that directly assess how patients experience cognitive burden and information processing challenges when using decision aids. Importantly, our scale is intended to inform early-stage and iterative decision aid refinement rather than replace later-phase IPDAS-endorsed evaluations of decision quality and decision process. This scale provides decision aid developers with critical metrics to design and evaluate their aids more effectively. Additionally, it facilitates better customization to meet individual patient needs. For example, domain scores can highlight when specific groups of patients – such as those with lower numeracy, lower health literacy, or higher reported cognitive load – experience greater difficulty, allowing developers to tailor the amount, format, and presentation of information accordingly. This research deepens our understanding of how patients comprehend information presented in decision aids and how this comprehension relates to information overload and their decision-making ability. By optimizing decision aid design based on these insights, there is significant potential to enhance contemporary SDM practices. Additionally, the large and diverse sample enhances the generalizability of our findings. Lastly, the use of a well-established instrument such as the Decisional Conflict Scale further strengthens the validity of our results.

There are several limitations to consider. The lack of a comprehensive evaluation for domains with only one question, such as “cognitive load” and “help decide,” means that we were unable to assess the internal consistency of these domains. The use of single items reflected a deliberate trade-off between precision and respondent burden. We prioritized brevity for early-stage feasibility, with future work planned to evaluate multi-item expansions of these domains. Future studies should focus on developing more comprehensive measures for these domains to allow for more thorough evaluation. Additionally, some DAES items assess features such as photos, quotes, and diagrams that may not be present in all decision aids. These items are optional and may be omitted without affecting calculation of the core domain scores, though their absence may limit evaluation of certain experiential aspects of decision aid presentation. Furthermore, our evaluation scale was developed and validated based on cancer screening decision aids. While the core principles of decision aids are generally applicable across different disease states,¹⁰ the scale’s effectiveness and applicability may vary in other contexts. Future research will focus on testing the effectiveness of our evaluation scale by developing a novel decision aid and evaluating its effect on SDM outcomes when compared to the standard of care in a randomized trial.

Conclusions

We present and validate a novel patient-centered Decision Aid Evaluation Scale that emphasizes the importance of managing cognitive load to improve decision-making outcomes. Decision aids with lower cognitive load and which are perceived as helpful are more likely to enhance patients’ values clarity, feeling of being informed, and certainty in their healthcare decisions. Informational load considerations should therefore be integrated into the design of decision aids. Future research should explore this scale’s effectiveness in improving decision aid design through a randomized trial.

Highlights:

• We developed and validated a novel Decision Aid Evaluation Scale that directly assesses decision aid usability and information overload from the patient’s perspective.

• Our Decision Aid Evaluation Scale provides a critical metric for improving decision aid design, ensuring that patients can make well-informed, values-based decisions.

Data availability:

Data will be shared upon reasonable request.