A new graphical format to communicate treatment effects to patients—A web‐based randomized controlled trial

Abstract Objective Patients making treatment decisions require understandable evidence‐based information. However, evidence on graphical presentation of benefits and side‐effects of medical treatments is not conclusive. The study evaluated a new space‐saving format, CLARIFIG (clarifying risk figures), aiming to facilitate accuracy of comprehension. Methods CLARIFIG displays groups of patients with and without treatment benefits as coloured sectors of a proportional bar graph representing in total 100 patients. Supplementary icons indicate the corresponding group's actual condition. The study used an application showing effects of immunotherapy intended to slow disease progression in multiple sclerosis (MS). In a four‐arm web‐based randomized controlled trial, CLARIFIG was compared to the reference standard, multifigure pictographs (MFP), regarding comprehension (primary outcome) and processing time. Both formats were presented as static and animated versions. People with MS were recruited through the website of the German MS society. Results Six hundred and eighty‐two patients were randomized and analysed for the primary end point. There were no differences in comprehension rates (MFP static=46%, CLARIFIG static=44%; P=.59; MFP animated=23%, CLARIFIG animated=30%; P=.134). Processing time for CLARIFIG was shorter only in the animated version (MFP static=162 seconds, CLARIFIG static=155 seconds; P=.653; MFP animated=286 seconds, CLARIFIG animated=189 seconds; P≤.001). However, both animated versions caused more wrong answers and longer processing time than static presentation (MFP static vs animated: P≤.001/.001, CLARIFIG static vs animated: P=.027/.017). Conclusion Comprehension of the new format is comparable to MFP. CLARIFIG has the potential to simplify presentation in more complex contexts such as comparison of several treatment options in patient decision aids, but further studies are needed.


| BACKGROUND
Patient involvement is particularly indicated in medical decisions comprising more than one option usually including the option of watchful waiting. 1 Medical reasoning might be capable of comparing treatment efficacy with regard to a defined outcome parameter. The patient's opinion is needed to weigh up the values of different outcomes with potential side-effects. This applies even more for complexly structured decisions and/or for decisions associated with pronounced scientific uncertainty such as in the case of multiple sclerosis treatments.
Multiple sclerosis (MS) is a chronic inflammatory and degenerative disease starting predominantly in young adults. Apart from symptomatic therapies, the range of treatments comprises an increasing variety of immunotherapeutic options. Making decisions amongst them is challenging with regard to putative risks and uncertain benefit. 2,3 Comparison of drugs is a complex endeavour as few comparative studies exist and even less evaluating treatment escalation series or longterm effects of immunotherapies.
To be able to make informed choices about immunotherapies, MS patients need information prepared in line with the criteria of evidence-based patient information. 4,5 These criteria require communication of benefits and harm for each option presented as changes of absolute risk together with an estimation of the information's trustworthiness. Furthermore, the criteria include presenting event rates by the additional use of graphical frequency formats. Previous studies have shown that different graphical formats visualizing probabilistic information using bar graphs, survival curves and pie charts 4,6 improve patients' understanding 7 and even the quality of physician patient communication 8,9 when compared to text-only risk information. Frequently, multiple-figure pictographs (MFP) (also called icon arrays) are used in evidence-based patient information as, for example, in decision aids (DA). 4,10 MFPs show proportions of patients with effects and no effects of a medical intervention using a given reference number of stick figures or smileys (N=100 or N=1000) ( Figure 1). MFPs have been proven effective in establishing sustainable comprehension of the difference between relative and absolute risk reduction in MS patients. 11 Compared to bar graphs, MFPs lead to equal comprehension of the proportions shown. Qualitative evidence suggests that MFPs are better suited to conveying the message of uncertainty about whether or not an individual will belong to the benefit group. 12 There are, however, practical drawbacks associated with using MFPs, particularly in multiple-option decisions like those addressed in our previous studies. [13][14][15] As the number of three consecutive MFPs needed to present the benefit of a single option ( Figure 1) multiplies with the number of outcomes reported for benefit and harm and the number of available options, information materials easily become long and difficult to comprehend. 16 Also, elements of MFPs, that is stick figures or smileys, do not indicate the nature of clinical outcomes (eg in the MS example "disease progression" or "relapses") and therefore need additional explanations in the graphic's legend. Based on the elaborate qualitative design methodology, 17  This article reports on an investigation aiming to evaluate the new presentation format's efficacy with regard to communicating study effects comprehensibly. Comprehension was defined in terms of accuracy of understanding the given quantities and time needed to process and complete the task. The first research question was: Does CLARIFIG lead to better comprehension and faster processing compared to MFP as the gold standard? Considering the increasing importance of making patient information tools feasible for web-based presentation, we also aimed at elucidating possible advantages of a stepwise animation.
Our second research question was: Does animated presentation lead to better comprehension and faster processing than static presentation? benefit of two years' interferon treatment in terms of progression:

| Design
The study used a web-based four arm randomized controlled trial ( Figure 2) using a basic information example considering the effect of interferon-beta treatment in slowing disease progression in MS. 18 The previously tested basic example of CLARIFIG ( Figure 3) was compared with a corresponding application of the MFP reference standard ( Figure 1) and with animated versions of the two graphs, respectively.
The study was part of a research project within the German Multiple Sclerosis Competence network on decision coaching on immunotherapies in MS, which was approved by the Ethics Committee of the Hamburg Chamber of Physicians (PV4576). The information displayed in Figure 3 can be summarized by saying that nine of 100 patients benefit (blue bar segment/thumb up) and

| Intervention
another 91 do not benefit (yellow segment) but present in two conditions, stable (hands behind the back) and progressed (icon with stick).
The study tested the identical application of the CLARIFIG graph previously used during its development.

| Sample
To allow for a representative sample of people with MS, we used only two self-reported inclusion criteria: age ≥18 and a confirmed diagnosis of MS. The sample size was calculated based on the results of the pre-test. Accordingly, N=143 participants were needed in each group to detect a difference between 10% and 25% of the participants meeting the primary end point. The calculation was based on two-sided testing with a 5% alpha error and a 90% power. Compensating a 20% dropout rate, this calculation resulted in a proposed sample size of N=686 participants.

| Procedure
Web presentation of the study was programmed using Unipark software 19 and accessed from the starting page of the German MS Selfhelp Society website (DMSG). It included the following components: invitation teaser, study instructions, the actual intervention consist-  the questionnaire, the system registered a participant as a finisher.
However, before the procedure was officially finished, participants were additionally asked to fill in a numeracy questionnaire.

| Measurements
The primary end point was previously developed and tested as a measure of accurate comprehension of the given quantitative information. 17 The score was dichotomized, defining four correct an-

| Analyses
Descriptive statistics were used to characterize the sample and the four study groups (Table 1)

| Primary end point
The two formats did not differ with regard to frequencies of comprehension, neither in the static nor in the animated presentation (MFP static =46%, CLARIFIG static =44%; P=.59; animated MFP animated =23%, CLARIFIG animated =30%; P=.134) ( Table 2). Single correct answers within the four-item comprehension questionnaire were more frequent; 85% of the participants identified the correct number of patients benefiting from treatment ( Table 2).
For the static presentation, the animated formats led to significantly less comprehension and longer processing time (MFP: P≤.001). However, compared to the static presentation, the animated formats led to significantly less comprehension and longer processing time (MFP: p <= .001 / .001, CALRIFIG: p = .027/.017) (   By only looking at two end points (comprehension and processing time), the present study failed to investigate the new graph's possible impact on a number of reasonable end points, such as perception of uncertainty, motivation to take an active role in the decision-making process, memorability of the information and transfer competence.

| DISCUSSION
Most importantly, however, its impact on the decision-making process in terms of facilitating shared decision making, informed choices and realistic expectation should be focused in further studies.
Effects of frequency formats on risk perception are not yet fully understood, 12,33 and the optimal format has not yet been found. 6 Moreover, as the context of the information, the target group and even the numerator size itself moderate the formats' suitability, current evidence is far from being able to inform systematic recommendations for developers and users of frequency formats. 6 In this respect, our study responds to a persistent lack of comparative studies and systematic developments in the field of communication and understanding of frequency formats. 6 In summary, the new format is promising because it has undergone a sound development process involving patients and a rigorous evaluation within a randomized controlled trial. As is immediately evident, CLARIFIG complies with the criteria of evidence-based patient information, 4 but also shows practical advantages with regard to multipleformat arrangements in limited space.

| CONCLUSION
Comprehension and processing speed of the new format, CLARIFIG, is comparable to commonly used multifigure pictographs (MFPs). The new format is advantageous with regard to space requirements and will facilitate the comparison of different treatment options in comprehensive patient information. This trial is considered exploratory as it compared the methods in a limited application using information from just one isolated study. Having found low comprehension rates irrespective of the experimental condition, the study demonstrates the gap between recognizing and fully understanding the information on the rate of benefit. This result implies that further research is needed on strategies to establish realistic expectations regarding the disease's natural course. Moreover, further studies are needed to prove the format's advantages in more complex contexts such as patient decision aids presenting information on various treatment options in parallel and in other medical domains.