The prospect of predictive testing for personal risk: attitudes and decision making

Abstract

As predictive tests for medical problems such as genetic disorders become more widely available, it becomes increasingly important to understand the processes involved in the decision whether or not to seek testing. This study investigates the decision to pursue the possibility of testing. Individuals (one group who had already contemplated the possibility of predictive testing and one group who had not) were asked to consider predictive testing for several diseases. They rated the likelihood of opting for testing and specified the reasons which they believed had affected their decision. The ratio of the numbers of reasons stated for testing and the numbers of reasons stated against testing was a good predictor of the stated likelihood of testing, particularly when the reasons were weighted by utility (importance). Those who had previously contemplated testing specified more emotional reasons. It is proposed that the decision process is internally logical although it may seem illogical to others due to there being idiosyncratic premises (or reasons) upon which the decision is based. It is concluded that the Utility Theory is a useful basis for describing how people make decisions related to predictive testing; modifications of the theory are proposed.

Introduction

The obvious medical benefits of health screening lie in the possibilities for early detection and prevention of serious disease. Less obvious are the potential psychological costs (Wardle and Pope, 1992) and the ethical and psychological dilemmas which are likely to arise from the availability of such testing. Ethical implementation of health screening programmes also requires a good understanding of the psychological processes involved in the initial decision whether or not to undergo testing, and the factors which influence these processes.

These issues are becoming more pressing as the next generation of health screening procedures develops. Genetic screening in particular is a source of great interest and concern as possible clinical applications extend beyond predictive testing for single gene diseases to risk assessment for multifactorial diseases such as breast cancer. Increasing numbers of people are therefore likely to be faced with the decision whether or not to seek or undergo genetic testing. Genetic testing is unlikely to be made available as a routine population screening procedure in the near future. It is more likely that procedures will require individuals to actively seek such testing, for example, in the way which people can seek HIV testing. Individuals who are believed to be at risk might also be actively sought through advertising or through family practitioners as part of clinical or research programmes. It is clearly important to improve our understanding of the way in which decisions to seek testing are made and the factors which may influence these decisions.

As Shiloh (1996) points out, ``although ties between decision theory and genetic decision making have occasionally been considered e.g. Vlek (1987), there is far too little collaborative work and literature on genetic decision making has remained largely nontheoretical''. We propose here to investigate decision making in predictive testing from a theoretical perspective.

Much of the work on decision making has been based on a framework of laws of probability and normative models, originating from the Utility Theory, which was proposed by von Neumann and Morgenstem (1947) in a book discussing `some fundamental problems of economic theory'. It was proposed as an attempt to `find an exact description of the endeavour of the individual to obtain a maximum of utility, or, in the case of the entrepreneur, a maximum of profit.' One of the basic assumptions is that rational decision making involves the systematic weighing up of alternatives and exploration of each alternative to produce judgements and choices, which maximises positive relative to negative outcomes. It requires the decision maker to implicitly estimate the probability and the utility of the various possible outcomes and to combine these by multiplication. The rule states that the action chosen is the one that maximizes expected utilities (EU), using the following equation:

$$\text{EU}{}_{\mathrm{i}}\text{=}\text{∑}\text{j=1}\text{n}\text{P}{}_{\mathrm{i}}\text{U}{}_{\mathrm{ij}}\text{,}$$

where P_j is an estimate of the probability of a given future state of the world, j, occurring; U_ij is the utility to the individual of the outcome of alternative i, should the jth state occur.

It is assumed that individuals implicitly make such calculations in their heads.

The Utility Theory can be seen as an attempt to specify more precisely the common-sense principle that people will choose the course of action which, at the time the decision is made, seems to be the best or most useful; i.e. the one which maximises utilities. However, unsurprisingly, deviations from this normative model have been shown to occur and no one consistently makes objectively optimum decisions. There is little doubt that, in health as in other areas, people often make what appear or turn out to be objectively unhelpful or inappropriate decisions. For example, people still smoke although they know that it is unhealthy and not everyone has a healthy diet. Experimental data incompatible with the `Objective' Utility Theory (e.g. Tversky, 1969; Lichenstein and Slovic, 1971; Lichenstein, 1973) has led to the proposal of other models. For example, Bell (1982) and Loomes and Sugden (1982) simultaneously proposed Regret Theory and Kahneman and Tversky (1979) proposed the Prospect Theory in which it is suggested that we distort probabilities and that we think about utilities as changes from a reference point. They found consistent differences between the decision outcomes when the options are described in different ways in each condition, despite the fact that the corresponding options in each condition are mathematically equivalent. They argued that subjective assessments of probability and utility explain the deviation from the statistical judgements which are assumed by any rational approach.

The present paper aims to apply Utility Theory to the initial decision about whether to pursue the possibility of predictive testing for personal risk. Several researchers have addressed the question of what factors predict those who want to undergo such testing. For example, Lerman et al. (1996b) looked at variables predicting the uptake of BRCA1 testing (e.g. sociodemographic factors, knowledge about hereditary cancer and genetic testing, perceptions of testing benefits, limitations and risks). However, relating factors assumed to influence decision making to the decisions themselves can be misleading, because these factors are not necessarily identical to the reason that people give for their choices (see Shiloh, 1996). In this paper we aim to use Utility Theory to investigate the factors which people report as influencing their decision whether or not to seek predictive testing. The primary focus is on factors affecting the likelihood that people will seek to initiate the process of testing. For this reason, the information given to the individuals about predictive testing, particularly regarding the complications attached to such tests, is limited. In later studies we examine factors more closely related the pre-test counselling itself; here, we consider those factors which may be involved prior to the person contacting professional services.

The hypothesis under investigation is that people make decisions in the way described by the Utility Theory, and that the reason that research has failed to support it is because the model has been applied in too narrow a fashion. To use this approach in psychological research, it is crucial to specify the nature of the elements involved in combination to make the expected utility. We propose then that Utility Theory can serve as a useful basis for understanding the decision whether to seek predictive testing, but we propose a modification from a perspective based on Beck’s cognitive model (Beck, 1976). That is, individuals appraise a situation in different ways according to those beliefs about that situation which are active at that time. Particularly threatening situations can activate pre-existing danger schemata which modulate subsequent reactions (Salkovskis and Rimes, 1997). We propose then that the Utility Theory can be applied with the added crucial provision that the particular reasons (and the probabilities and weights assigned to them) are idiosyncratically determined according to the persons prevailing internal logic. Such reasons will be strongly influenced by the particular beliefs which are active during the decision process. For example, the notion of risk testing for cancer may activate memories of a relative who died of the disease. The impact of this would in turn be influenced by the persons belief about the value (or cost) of early detection and intervention. As suggested by Kahneman and Tversky (1979) in the Prospect Theory, we propose that the probabilities and utilities assigned to particular reasons are also subjective.

Utility Theory may therefore still apply if `objective' logic is differentiated from internal, idiosyncratic `subjective' logic. It seems reasonable to expect that decisions will be made based on a number of subjective premises which may or may not be shared by others, and which the person believes to be sensible at the time the decision is made. These starting premises may be generally regarded by others as illogical, but they can still form the basis of a logical conclusion by the person making a decision. For example, someone may have firm beliefs in the value of astrology; when their horoscope indicates that it would be unwise to take an irreversible decision at a particular time, and the person decides not to have a major surgical procedure, this decision can be regarded as having an internal logic although others may regard it as irrational. Similarly, the person who refuses to have a medical test because they believe that the test itself may cause the illness being tested for, is behaving logically from their own perspective. Sometimes, this type of apparently `irrational' belief may be based on compelling evidence. In an interview, a patient who was refusing to undergo a mammography told one of the authors that her experience convinced her that mammography caused breast cancer. She could list family and acquaintances who had and had not had mammography, and it was objectively true that a high proportion of those who had undergone mammography had been diagnosed as having breast cancer, compared with none of those who had not. This type of example highlights the importance of considering the internal logic of decision making and of assessing idiosyncratic beliefs in relation to the specific decisions under consideration.

This modification of the theory requires also that the reasons influencing the decision include only those on which the decision maker focuses at the time of he decision. It is likely also that emotion related reasons such as anxiety influence the factors on which the decision maker focuses at that time. It is known that mood, including anxiety and depression, influences the factors which people recall and the factors to which they attend (Williams et al., 1988). Many researchers have investigated how emotion influences decision making, some of whom have taken the line that emotion causes conflict and is an irrational agent in an otherwise rational decision process, e.g. Janis and Mann (1977); Sjoberg (1980). Others have questioned this common assumption, for example Toda (1980) proposed that emotion actually aids the decision process. It is likely that emotions at the time of the decision influence the decision outcome whether to seek predictive testing. For example someone may want testing because he/she feels that the uncertainty of whether or not they will develop the disease is very distressing, or someone may not want to know his/her genetic risk as they are so anxious about the disease. Another likely influence on the decision process is anticipated emotion, such as anticipated depression, anticipated anger following a bad result from the test, and anticipated relief and anticipated guilt following a good result from a test. We propose that a potent source of motivation lies both in the associated anxiety and in anticipated emotions. The study described in this paper considers general health anxiety and emotions specifically related to the disease as well external factors.

It is proposed then that people do make decisions implicitly by weighing up the pros and cons, where these pros and cons, (i) consist only of those that the decision maker focuses and considers relevant at the time of the decision, (ii) include emotional factors such as anticipated anxiety. It seems unlikely that, even with appropriate pre-test counselling, all relevant pros and cons will be included in the decision process and irrelevant reasons are likely to intrude. Instead, it is likely that people will tend to focus on subsets of pros and cons, and not necessarily on the most relevant ones.

Research addressing the question of how people decide whether to opt for predictive testing needs to address different stages in the decision making process (cf. Prochaska and DiClemente, 1982). Research should include separate consideration of (i) the decision making process involved in the option to seek the pre-testing procedure, and (ii) the decision process when people are actually required to accept or reject testing. The research presented in this paper relates to the first of these stages. Our later work involves experimental studies of the decision making process when people consider the option of testing.

The study described here involves both a student sample, who were presented with the issue of predictive testing and a second sample who define themselves as having already actively considered the option of predictive tests. The research reported in the present paper has several aims. The first aim was to determine the range of subjective reasons (including emotional ones) cited by people as influencing their decision whether to seek such health screening, and to develop a system for categorising and quantifying these reasons. The second aim was to compare the reasons given by people for whom testing was a completely hypothetical issue raised by the experimenter (the student sample) with those of people who reported that they had actively been contemplating such testing (subsequently known as the `contemplators'). The third aim was to investigate how the reasons that people stated as being relevant to their decision predicted the decision outcome. In the contemplators sample, a more detailed analysis in terms of Utility Theory was carried out by evaluating the pros and cons which were weighted according to their utilities (relevance).