Different formats for communicating surgical risks to patients and the effect on choice of treatment

Abstract

Effective communication of treatment risks is important to enable patients to make informed decisions. This study aimed to determine the effects of different risk formats on participants’ evaluation and interpretation of risk information and on their treatment choice. Participants (N=44) were recruited among patients who had undergone surgery for an abdominal aneurysm and were asked to evaluate treatment risks (surgery or an observation policy) of two hypothetical cases presented in one of three risk formats (numbers, vertical bars or icons). Risk information presented in vertical bars was evaluated as the most difficult to comprehend, and the perceived threat of this information was evaluated as higher than that of the other risk formats. Risk information presented as icons was evaluated as more helpful for making a decision, but resulted in a lower percentage of participants choosing for surgery than when risks were presented in the other formats. In conclusion, this study showed that different risk formats have different effects on participants’ evaluation of the information and on their choice. Doctors should therefore be careful in choosing the format in which they present treatment risks.

Introduction

An active role of patients in the decision-making process with regard to their treatment is becoming increasingly important, and discussing the risks and benefits of treatment options is therefore an essential part of modern health care. Understanding the risks is considered to be crucial for appropriate decision-making [1]. Providing patients with objective (i.e. based on empirical evidence) and useful information in the most accessible manner is not easy. In particular, the communication of risks is a complex matter. It has been shown that patients find it difficult to adequately comprehend risk information [2], [3].

While many studies focus on interventions concerning risk communication to optimize decision-making, fewer studies are concerned with comparing various formats for effective risk communication. Studies, in which relative and absolute risks were compared, showed that risks were considered to be greater when information was expressed as relative risks than when it was presented as incidence rates [4], [5] and that more patients agreed to treatment than when the risks were presented as absolute differences [6]. Risks presented as frequencies were also perceived as higher than when the same risks were presented as percentages [7]. Although having the attraction that common words are used that seem to be generally understood, the use of verbal labels (e.g. “a great chance”) for communicating risks appears to be less effective in clearly explaining risks, because people differ greatly in their interpretation of verbal labels [8], [9].

The visual presentation of risk information has also been explored. We know little about how (and even whether) graphical presentations are superior to other formats, such as numerical or verbal formats, in terms of effective risk communication. The findings of some empirical studies suggest that many patients prefer the simple bar charts to other formats such as icons (for example, showing how many people per 100 are affected), survival curves or pie charts [10]. In non-medical domains it has been shown that a combination of graphs and tables produced more accurate performance [11], [12] and that subjects using simple bar graphs needed less time for making a judgment than when using pie charts [13], [14]. Elting et al. [15] showed that displaying the results of clinical trials in icons resulted in more accurate decision-making by physicians than displaying the results in tables, pie charts or bar graphs. Stone et al. [16] found that participants were less risk averse when the risks were presented with numbers than with any of the graphical formats, such as bar graphs or icons. They argue that a graphical presentation of risks increases the subjective estimate of the risk (either the perceived chances or the seriousness of adverse outcomes), because the visual salience facilitates the comparison of risks. Most of the experimental research that has been carried out suggests that combining graphics with numerical information does affect several outcomes, such as the perceived helpfulness of the information and the perceived threat of risks. There is still little experimental research testing whether people’s perception and understanding of risk vary by graphical format and whether the addition of graphical displays significantly improves comprehension compared to the numerical presentation of risks only.

As there is agreement that effective risk communication is necessary for the involvement of patients in the decision-making process, there is no agreement on what constitutes effective risk communication. One criterion of effectiveness is that risk communication should increase patients’ understanding of the risks in order to enable them to use this information in their decision-making. Cognitive outcomes are often used, such as an accurate recall of objective risk estimates, assuming this is related to an adequate comprehension of the information. A better understanding of risk information, may thus lead to better decisions. Good decisions are often defined as decisions based on relevant empirical evidence and patient’s values, thereby maximizing patient’s expected utility [17], [18]. In addition to these aspects, attitudinal criteria (such as satisfaction with the decision), and more affective outcomes (such as threat induced by the information) are applied [19]. These aspects are assumed to reflect the quality of the information and the decision-making process [20], i.e. effective risk communication should contribute to the satisfaction with and confidence in the decision that has been made, and the risk information should be evaluated as helpful in decision-making. Further, information presented in a risk format which induces less threat than when presented in other risk formats might be preferred. All in all, the effectiveness of risk communication can be evaluated on many dimensions. The evaluation which risk format is most effective therefore depends on the evaluation of the importance of the different aspects of effectiveness [19].

The findings of some studies suggest that the effectiveness of a format not only depends on the characteristics of the information (e.g. magnitude of the risks to be communicated), but also on user characteristics (e.g. cognitive style [10]). While patients in general have a great desire for information, this varies substantially among patients due to, for example, the level of education and age [21], [22], [23]. Sutherland et al. [21] showed that the desire for information is related to the desire for decision involvement. These individual differences in desire for decision involvement might be a moderating factor in the effectiveness of different formats for risk communication.

The objective of the present study was to investigate various effects of risk communication concerning a surgical treatment using three different formats. Numerical risk information only was compared with two graphical formats (vertical bars and icons) in addition to numerical information. Bars and icons are the most frequently used graphical formats for presenting risks in health care, and have been shown in a few studies to be superior to other graphical displays especially for comparing risks [14], [15]. The criteria we used to measure the effects of risk format are related to the risk information, i.e. comprehensibility of risk information, helpfulness for making the decision, and threat induced by the risk information (i.e. the described situation). The decisions of participants were further compared with the outcome of an empirically based decision analytic model [25], [26] and participants were asked to report their confidence in the decision.

In this study participants were asked to compare several mortality risks and to make a hypothetical treatment decision. Since the ultimate goal of our study was to select the appropriate format for information provision to aneurysm patients, we chose to use this group as participants in this experiment. The magnitude of the risk was varied in two hypothetical decision problems, because it was assumed that the effect of risk communication in a certain format interacts with the magnitude of the risks, in the sense that numbers would be better to quantitatively compare risks when differences between risks were small while graphs and icons would be better when differences between risks were larger. The following hypotheses were formulated. Based on the results of previous studies, it was assumed that numerical risk information would be evaluated as more complex, less helpful for decision-making, but as less threatening than graphical information, and in particular than risk information presented as icons. These effects may be moderated by factors such as desired decision involvement and level of education. Greater decision involvement might be related to smaller differences between formats in the evaluation of the risks. Participants with a higher level of education might have fewer problems in interpreting numerical risk information. With respect to the participants’ choice of treatment, it was expected that risks presented in a graphical format would lead to more risk-averse behavior (cf. Stone et al. [16]). In the present study it was expected that this would lead to more risk-averse behavior regarding the short-term risks, i.e. risks of surgery, as research results suggest that the attention of patients is more focused on the short term risk and they are more inclined to avoid this short term risk of surgery and thus to choose for the observation policy [24]. The visual salience of a graphical display would make especially the short-term risk of surgery more hazardous or threatening, and therefore fewer participants would choose this option.