Non-rational choice behaviour versus informed consent

Governments and policymakers’ views on how people make choices are often dominated by assumptions underlying the so-called ‘rational choice theory’.1 This theory is based on the premise that as long as individuals are provided with sufficient information, they will subsequently make the choice that best suits their preferences. The assumptions here are that rational decision-makers use all factual information offered, strive for the best result, make correct estimates of chance and risk, are stable in their preferences and leave emotions out of consideration.1 The rational choice theory seems to equate a ‘rational’ choice with a choice that is made exclusively and without error by means of the deliberative (‘slow’) system, instead of with the ‘fast’ intuitive system.ii This equation is not normatively neutral: choices should be made consciously and analytically, because that would lead to ‘good’ choices.

Empirical studies into decision-making and choice behaviour, however, put these normative assumptions in a different perspective. First, research shows that the way people make choices in practice often does not fit within the conception of the rational choice theory: we cannot and do not always want to choose according to the rational choice theory.1 2 Behavioural sciences show how divergent biases affect our decision-making. Additionally, psychological research nuances the assumption that choices based on deliberation are superior to choices based on intuitioniii: deliberativeiv and intuitivev choice processes offer—depending on the context—both advantages and disadvantages and are complementary.12 13

Whether such findings have implications for informed consent depends on the extent to which the normative assumptions from the rational choice theory are present in theories of informed consent. According to the influential works of Faden and Beauchamp and Beauchamp and Childress, informed consent must be defined as an ‘autonomous authorisation’, in which a person 1) intentionally; 2) with (substantial) understanding; 3) (substantially) non-controlled; 4) authorises a professional to perform a certain act.5 6 ‘Authenticity’ or ‘value-consistency’ is not an explicit condition in this conceptualisation. However, alternative views stress that authenticity should be included in an adequate conceptualisation of autonomous decision-making and autonomy.14 To be ‘authentic’, choices must be such that a person can identify with them. Authentic choices are choices in accordance with one’s values and identity.15 We agree that authenticity is an important condition for autonomy and for informed consent, in particular in medical practices in which significant personal norms and values are at stake and consequences of decisions are far-reaching, like for example, in predictive tests such as prenatal screening, or genetic testing for Alzheimer’s disease. In our view, informed consent is an intentional authorisation, with (substantial) understanding, (substantially) non-controlled and authentic, with the primary objective to protect and promote autonomous decision-making. This conceptualisation of informed consent will be our starting point in the remainder of this article.

Non-rational choice processes can interfere with all conditions of informed consent, but in particular the condition ‘with understanding’. This condition means that someone has an adequate appreciation of the nature, foreseeable consequences and possible outcomes that might follow by (non-)participation in a specific intervention.5 In practice, this condition is translated into legislation and procedural regulations that aim to have the decision-maker make an ‘informed’ and ‘well-considered’ choice. However, taking psychological insights into decision-making into account, we can conclude that these conditions are not always feasible and perhaps not always desirable in informed consent to protect and promote autonomous decision-making. We will explain these two aspects below.

First, the requirement of ‘with understanding’ imposes high conditions that we cannot and/or do not always want to meet in practice. To fulfil ‘with understanding’, the healthcare professional has the obligation to provide relevant information regarding what the decision-maker must reasonably know about the medical intervention. The information considered relevant and necessary for an ‘informed’ choice is often comprehensive and of various, sometimes complex, nature. In case information exceeds the capacity of our deliberative system or if deliberation costs too much investment in terms of time and/or effort, we use—consciously or unconsciously—heuristics.2 Heuristics are divergent strategies that ignore part of the information, with the aim of making faster, more efficient and/or more accurate choices than with the more complex, analytical methods.13 Although heuristics can lead to ‘good’ choices in certain contexts,13 16 they are also sensitive to—often predictable—bias. An example is the ‘availability heuristic’, in which people assess the probability of an event by the ease with which occurrence can be recalled.17 This leads to bias, if this recall is inconsistent with an epidemiological chance/risk and is used as the only source of information to assess the occurrence of an event. This misprediction, for example, regarding the risk of a disease, may lead to (non-)participation in a therapeutic, diagnostic or screening intervention that the decision-maker would not have made if objective risk information had been accurately processed by the deliberative system. With the use of heuristics incomplete and/or incorrect information may play a disproportionate role in decision-making, and one can wonder whether such a choice fulfils the condition of ‘substantial understanding’ necessary for an ‘autonomous authorisation’. Because we are also—certainly in unfamiliar situations—sensitive to ‘framing’ effects,2 making an ‘informed’ choice for (non-)participation in a medical intervention is even more challenging.

Second, it is debatable whether current informed consent procedures really protect and promote autonomous decision-making. The normative assumptions from the rational choice theory seem to be contained in theories and current practice of informed consent. It seems to be assumed that the decision-maker comprehends, correctly processes and uses all the information offered and makes a choice that suits him/her based on a logical integration of this information. As in the rational choice theory, it seems that within the practice of informed consent, it is expected, and considered desirable, that a choice should be made exclusively by means of the deliberative system. This is what the notion of ‘well-considered’ seems to hint at. Information which is processed and/or obtained by the intuitive system seems to be considered inferior or is ignored.

However, this emphasis on choice through deliberation can also limit autonomous decision-making. By resorting to deliberative choice processes, there is a risk that the decision-maker will focus on choice attributesvi that are clear, plausible, easily accessible and formulable,18 but are less important to him/her personally.12 This may lead to a decision that is not actually fully in line with his/her values and that, even though the decision-maker can substantiate it, he/she would not have made if less use had been made of deliberation, and more of intuition.12

Intuitive choice processes offer specific advantages which may be necessary for autonomous decision-making and which cannot be provided by deliberation. The intuitive system can simultaneously process and integrate a larger amount of information—including affective information—than the deliberative system.3 19 Affective information is, among other things, processed and integrated through the ‘affect heuristic’. In a certain choice situation, we consult our ‘affect labels’ that are associated with the mental representations we have of a similar situation.20 The affect heuristic functions as a mental short-cut, in which we do not weigh all the advantages and disadvantages, but rather base ourselves on emotions and feelings that evoke certain choices for us. The affect heuristic can provide information and insight into one’s personal values and life goals. By means of intuitive processing, the decision-maker thus has more aspects regarding the choice (unconsciously) in mind, which may be necessary for ‘substantial understanding’ for informed consent.

Possible solution: the use of PDAs

The ‘rational’ way of interpreting the condition ‘with understanding’ and its translation into the conditions ‘informed’ and ‘well-considered’ choice may not always be feasible for human beings nor conducive to autonomous decision-making. Intuitive and emotional reactions, feelings, (irrational) beliefs and previous experiences are also a source of information that can—and sometimes must—be used to make an autonomous choice. If informed consent is intended to protect and promote autonomous decision-making, then—depending on the nature of the choice and the decision-maker as a person—there should be adequate recourse to deliberative as well as intuitive choice processes. We can bridge the gap between the informed consent doctrine and real-world choice behaviour, by sustaining and promoting authentic, autonomous decision-making of the ‘limitedly rational’ decision-maker. The increasingly used PDAs11 seem suitable interventions to support this, because they have the same purpose: promoting autonomous decision-making. Moreover, PDAs can be designed in such a way that they are also applicable in decision-making where there is no healthcare professional who can support the decision-maker in his/her decision-making process, such as (non-)participation in screening interventions. In other settings, PDAs can help overcome communicative limitations and errors of healthcare professionals.

PDAs are described as interventions that seek to provide support in making informed, deliberated, value-consistent choices in complex medical decisions regarding treatment, diagnostics or screening, where there is no unequivocal ‘best’ choice.11 21 PDAs cover a broad category of interventions, with great diversity in design: leaflets, videos, computer-controlled decision aids; for individual use or in the context of a counselling interview.11 To realise their goal, the various PDAs are often focused on three aspects, namely11:

1. Explaining the choice problem;

2. Providing evidence-based information on the choice options and the associated (dis)advantages (including uncertainties);

3. Clarifying personal values associated with the choice by means of a ‘values clarification method’ (VCM). VCMs can be categorised into implicit and explicit VCMs.22  Implicit VCMs are, for example, detailed information provision about the options, or personal experience stories. These VCMs enable one to imagine what the choice made would mean for them personally or to see how others made the choice.22 23 The more frequently used explicit VCMs11 concern interactive methods, whereby decision-makers must rank or trade-off the different offered choice attributes.22 23 Active processing can give insight into personal values, which must be weighed up and the personal significance of each choice attribute, for example, by hypothetical scenarios or visualisation in balance scales of personal advantages and disadvantages regarding the choice options.22 23

The extent to which attention is paid to the above aspects varies. Almost all PDAs provide relevant information regarding the choice that has to be made. Further support in the decision-making process takes place in 66% of PDAs,11 although it is not exactly clear and specified what this support entails.

Whether PDAs contribute to autonomous decision-making is a question that can only partly be answered by scientific research. A recent Cochrane systematic review shows that PDAs lead to an increase in knowledge about the choice, more accurate risk perceptions and a better understanding of the decision-makers’ personal values that play a role in their decision-making.11 A few studies in this systematic review support that the increase in knowledge and better understanding of personal values also lead to more ‘informed’ and ‘value-consistent’ choices.11 PDAs do not lead to less or more satisfaction, anxiety and depression, nor to worse health outcomes.11

However, the existing empirical research into the effectiveness of PDAs in achieving value-consistent choices is problematic, in part because it cannot fully be determined when PDAs contribute to ‘value-consistent’ choices.24 Insights into (non-rational) choice processes can provide direction for the further development and use of PDAs in order to protect and promote informed consent manifesting autonomous decision-making. In our opinion, these insights seem to be insufficiently considered in current PDAs, and thereby may not support autonomous decision-making. We will clarify this in the next section by explaining how current PDAs are designed, how this—considering psychological insights—can be problematic for the realisation of autonomous decision-making and how psychological insights could be implemented in PDAs.

Limitations of PDAs

The design of current PDAs may not always support autonomous decision-making. In our opinion, this can be explained by two assumptions underlying PDAs.

First, an assumption that many PDAs seems to rely on is that personal values and preferences regarding the decision are already formed and present, and only need to be clarified by means of the PDA.24 However, individual values and preferences are not as present and stable as was probably always assumed.23–25 This may imply that a PDA—and specifically the VCM—does not always serve as a ‘value clarification method’ of individual values but may lead to a constructed preference that does not fit with one’s actual personal values. This probably mainly concerns explicit VCMs, wherein the choice attributes are introduced explicitly by the designers of the VCM. This feature of explicit VCMs may hinder decision-makers more from freely associating and interfering in what information offered means to them personally. The risk of a constructed preference that does not match with someone’s values is possibly reinforced by graphics, for example, a scale, that visualise the ‘weighing’ of choice attributes for the decision-maker.25 If the attributes chosen are all given the same weight on the scales, as in the ‘tallying’-heuristic,vii the decision-maker is presented with recommendations that may not reflect their own weighing of pros and cons. If the decision-maker is not aware of this simplification with which choice attributes are weighed and has insufficient insight into the personal significance of each choice attribute, there is a risk that the decision-maker will not make the choice that suits him/her. The same applies to recommendations indicating which option is the most appropriate for the decision-maker’s answers.25 In these cases, PDAs do not do justice to the sometimes divergent, conflicting values and the personal significance of each choice attribute and may hinder autonomous decision-making. To avoid this, it is advisable not to add such normative, directing graphics and recommendations, and/or to supplement some information that raise the decision-maker’s awareness of the simplification of the choice problem if a heuristic-based VCM is used.

Second, current PDAs are often focused on a deliberative processing, in which personal values are made explicit and weighed up,11 assuming that this leads to ‘better’ decision-making than if this is done through intuitive choice processes.24 However, as previously mentioned, the emphasis on deliberation can also limit autonomous decision-making. We have argued that there must be adequate recourse to deliberative as well as intuitive choice processes in informed consent to protect and promote autonomous decision-making. Therefore, PDAs should rely on deliberation and intuition, in which the advantages and disadvantages of both are considered. de Vries et al argue how these different choice processes can be applied to the practice of PDAs design.12 They advise to articulate the decision-maker’s preferences and motives late in the decision-making process in order to prevent the decision-maker from concentrating his/her attention on choice attributes that are clear and easy to formulate, but which do not actually play an important role for him/her.12 In addition, they recommend to incorporate a moment in a PDA at which the decision-maker distances himself/herself from the choice problem,12 so that generated emotions—which can indeed be informative and support authenticity—can subside, and the intuitive system is given the opportunity to (unconsciously) process and integrate factual and affective information.18 These recommendations are a valuable starting point for integrating deliberative and intuitive choice processes into a PDA and may support informed consent.

Further research into the design and evaluation of the combination of deliberative and intuitive decision processes is ongoing and highly necessary. For instance, a remaining question is whether ‘fast and frugal heuristics’, like the aforementioned ‘tallying’-heuristic, or the ‘take the best’-heuristic,viii are suitable tools for intuitive decision processes and should be implemented in PDAs. The use of fast and frugal heuristics is a simple—but not yet widely used—25 method to operationalise a form of intuitive choice processes. Preliminary research into using these heuristics seem promising and could be used to improve informed consent processes.26 However, although there is evidence that heuristics can lead to ‘better’ outcomes, this is especially true for the use of heuristics in everyday choices and choices with which someone has experience.12 13 Since medical choices are usually not of this kind, it should be further investigated whether the evidence in which heuristics have been shown to be effective is indeed generalisable in the context of medical decision-making.27 Moreover, more empirical research is needed to ascertain how PDAs using heuristic-based VCMs could best be designed and integrated into the informed consent process. Should, for example, information regarding the choice options be followed or preceded by a heuristic-based VCM? And would it be necessary to supplement the PDA with a counselling interview, in order to support truly informed and value-consistent choices, or could a well-designed PDA stand on its own?

In summary, it seems that PDAs can at most mitigate, but not completely address the problems we identify. To overcome this, further research into PDAs is necessary, but we also call for a more fundamental rethinking of the nature and goals of informed consent.