Australian community responses to the use of genetic testing for personalised health promotion

Abstract

Personalised genetic health promotion may soon be available and affordable. To explore its likely public acceptance in Australia, a community sample (N = 800) provided quantitative and qualitative responses to a vignette scenario about a hypothetical expert who could test their genes and, based on this genetic profile, provide personalised health promotion advice. Three theoretical models were tested to explicate the process by which cognitive–affective factors of risk beliefs, benefit beliefs, and trust judgements influenced behavioural intentions. Results supported an expert trust model, where general beliefs about the risks and benefits of medical advances and general medical trust had indirect influences, while trust in a specific medical expert had a direct influence, on health promotion intentions. Subjective reasons for intentions included moral concerns, fear, trust, mistrust and a desire to maintain health at any cost. The advent of personalised genetic health promotion may heighten the need for specialised health psychologists.

Key words:

• behavioural intention
• genetic testing
• health psychology
• personalised health promotion
• risks and benefits
• trust

Abbreviated reports of this study were presented at the International Genomics and Society Conference held in Amsterdam, The Netherlands, 17–18 April 2008, and the Annual Meeting of the Society for Risk Analysis, held in Boston, MA, USA, 7–10 December 2008.

Recent advances in genetic technology suggest that complete DNA sequencing will soon be affordable, allowing widespread application in the area of personalised health promotion (Aldhous, 2009; Pushkarev, Neff, & Quake, 2009; Skipper, 2009). The role of genetics in health promotion and disease prevention is well recognised (Wang, Bowen, & Kardia, 2005), but community acceptance has been mixed (Calnan, Montaner, & Horne, 2005) and public attitudes have been somewhat contradictory (Jallinoja et al., 1998).

Australians have been quick to accept new technologies (Swinburne National Technology & Society Monitor [SNTSM], 2003–2010) and generally endorse the use of genetic testing for purposes of paternity, forensic investigation, and adult stem cell research (Critchley, 2007; Critchley & Turney, 2004). Recent research suggests that many Australians would be willing to adopt a genetically based personalised nutrition plan to maintain health and prevent illness (Pin, Critchley, & Hardie, 2008). Apart from that study, relatively little is known about Australians' willingness to use genetic testing for health promotion purposes.

Theory and research on the factors that influence people's intentions to engage in health‐promoting behaviours have often been based on rational cognition models such as the theory of planned behaviour (TPB; Ajzen, 1991), but many theorists acknowledge the equally important role of affect. For example, some argue that an affective state, with or without consciousness, can occur rapidly and automatically when forming intentions or making choices (Kahneman, 2003; Slovic, Finucane, Peters, & MacGregor, 2007). Health intentions can reflect cognitive–emotional judgements based on available cognitive schemas and affective heuristics (Kahneman, 2003; Slovic et al., 2007).

Three cognitive–affective factors, namely, trust, risks, and benefits, have been consistently used to examine public intentions about the use of genetic advances (hereafter referred to as ‘genetic intentions’). These factors, alone or in combination with other sets of variables, have been widely studied in relation to intentions to buy genetically modified products (Siegrist, 2000), accept new genetic applications such as nutrigenomics (Ronteltap, van Trijp, Renes, & Frewer, 2007), and engage in genetic screening for specific diseases (Bosompra et al., 2000).

Several studies demonstrate the importance of risk and benefit beliefs when forming genetic intentions. Both benefitand risk beliefs predict intentions to undergo genetic testing for hereditary diseases (Bosompra et al., 2000; Nordin, Bjork, & Berglund, 2004; O'Connor & Cappelli, 1999), while benefit beliefs are particularly strong predictors of intention when there is uncertainty about the test results (Frost, Myers, & Newman, 2001).

Some studies highlight social trust as another key factor in forming genetic intentions. For example, recent research showed that low levels of trust in medical experts, as well as greater perceived risks (negative beliefs about science) and doubts about the benefits of technology were associated with lower acceptance of new medical technologies such as genetic testing (Calnan et al., 2005). Siegrist (2000) found that trust influenced perceptions of risks and benefits which, in turn, influenced respondents' intentions to buy genetically altered foods.

Despite consistent empirical findings linking trust, benefits, and risks to health‐related intentions, most studies are not directly comparable. Research in the area of genetic advances has often been undertaken by multidisciplinary teams and/or framed by a range of theoretical assumptions. Slovic (1999) noted that even in the specialised area of risk research, there were conflicting views between risk assessment, risk management, and risk communication approaches. When risk researchers join with economists, sociologists, and psychologists, their studies can span many disciplines and theoretical perspectives (Ronteltap et al., 2007). Consequently, compromises are sometimes made to accommodate the complexity of multiple approaches. Disparities in the conceptualisation and measurement of key constructs have not only reduced comparability across studies, they have also undermined theoretical and practical advances in understanding of how people make choices or form intentions. Close examination of research on genetic intentions reveals several theoretical and measurement issues which may hinder understanding of how and why people form intentions to engage in health promotion activities.

CONCEPTUALISING AND MEASURING TRUST, BENEFITS, AND RISK

There is little consensus on the definition or measurement of trust. Studies on community acceptance of genetic advances tend to treat social trust as a distal factor involving general trust (or mistrust) of people and institutions which represent the industry (Ronteltap et al., 2007; Siegrist, 2000; Slovic, 1999). McKnight, Choudhury, and Kacmar (2002) distinguished such general trust in impersonal‐structural targets from specific trust in personal‐interpersonal targets such as known others or particular experts. Many theorists agree that trust judgements, like attitudes, involve cognitive–affective evaluative responses to a target (Eagly & Chaiken, 1993). Trust judgements may sometimes involve slow, rational cognitive deliberation, but they can also be quick, automatic evaluations based on visceral emotional response, personal experience, available cognitive schemas, intuition, and/or social contextual cues (Kahneman, 2003; Messick & Kramer, 2001; Slovic et al., 2007). Trust is generally seen as multidimensional, and many elements of trust have been measured, but McKnight et al. convincingly argued that benevolence, competence, and integrity are three key dimensions of trust judgements.

There is a good deal of variability in the ways that risks and benefits have been conceptualised in research on community acceptance of genetic and other technological advances. Slovic (1999) highlighted the many competing definitions among risk researchers, and the multidisciplinary literature is similarly diverse, treating benefits and risks as general beliefs (Frost et al., 2001), attitudes (Nordin et al., 2004), perceptions (Bosompra et al., 2000; Siegrist, 2000), affective evaluative judgements (Finucane, Alhakami, Slovic, & Johnson, 2000; Slovic et al., 2007), or comparative advantages/benefits in relation to disadvantages/barriers (O'Connor & Cappelli, 1999). Slovic noted that there is a subjective lay belief that risks and benefits are inversely correlated. People perceive greater benefit to be associated with lower risk (or greater risk with lower benefit). This subjective association is strengthened under conditions of pressure and can be further manipulated by providing information that highlights potential risks, thereby reducing benefit judgements; or information that highlights benefits, thereby reducing risk judgements (Finucane et al., 2000). While there is no consensus on how best to define and measure these two constructs, Slovic advised that perceived benefits and risks to society should be distinguished from perceived personal benefits and risks to the individual. It also seems prudent to distinguish general beliefs about the risks and benefits of an industry (e.g., health, agriculture) from the specific perceived risks and benefits of a particular application (e.g., genetic disease screening, genetically modified foods).

THEORETICAL MODELS OF FORMING GENETIC INTENTIONS

Although research suggests that trust, risks, and benefits each play some role in genetic intentions, the theoretical process is not clear. Several theoretical models have been developed, but few have been empirically verified. For example, Ronteltap et al. (2007) developed a complex rational cognition model loosely based on the TPB (Ajzen, 1991), with integrated elements from various consumer acceptance theories. The model posits a set of distal factors (background characteristics of the consumer, social system, and genetic innovation) which indirectly influence behavioural intentions through their direct influence on proximal factors (more immediate perceptions of risks, benefits, subjective norm, and control). This plausible theoretical process was not tested; however, Ronteltap, vanTrijp, and Renes (2009) later developed and tested another model based on a combination of Ajzen's TPB (1991) and framing theory (Levin, Schneider, & Gaeth, 1998). They used filmed vignettes to manipulate consumer information about nutrigenomics (genetically based personalised nutrition). The provision of different types of information (e.g., whether nutrigenomics would benefit science, industry, or the consumer; whether genetic testing would be obligatory or free choice) was found to meditate the influence of psychological factors (e.g., cost–benefit assessment) on respondents' preferences for nutrigenomics‐based personalised nutrition.

Ronteltap et al.'s (2009) study assessed a key variable, cost–benefit, with a single item (relative positivity towards nutrigenomics after weighing up the pros and cons) that was rated after respondents had viewed a film that manipulated information about the commercial, consumer, and scientific benefits of nutrigenomics. Moreover, this consumer choice model did not include any distal characteristics as posited in their earlier theoretical process model (Ronteltap et al., 2007). The inclusion of background characteristics such as the respondent's general trust in scientific advances, or specific trust in the genetic industry, would have allowed exploration of distal characteristics and affective judgements in forming genetic intentions.

Siegrist (1999, 2000) developed and tested a parsimonious theoretical model of the process by which trust, benefits, and risks influence genetic intentions. He found empirical support for his posited process whereby trust directly influenced perceptions of risks and benefits, but had only an indirect influence on intentions. However, close inspection of the indicator variables used to test this causal model suggests that the diversity of genetic applications rated and the differing levels of specificity (e.g., general trust, specific intentions) may have obscured the role of trust in this process.

In Siegrist's study, the latent variables of perceived risks and benefits were represented by items rating the risks and benefits of several specific genetic applications (altering plant foods, animal foods, producing new medical treatments, developing techniques for early disease diagnosis). Trust was represented by items rating general trust in industries involved in the production or handling of genetically engineered products (universities, agricultural, pharmaceutical, and food companies). Acceptance was represented by ratings of specific future intentions to buy a selection of five genetically modified foods (cheese, tomatoes, meat, non‐allergenic food, and chocolate). The final model showed that general trust in the industry had only an indirect influence on intentions to buy a specific set of products, but that general trust in the genetic industry directly influenced perceptions about the risks and benefits of various genetic applications.

The differing levels of specificity between general trust and specific intentions may have masked the influence of trust on intentions in Siegrist's study. To fully assess the role of trust, any future test of this model should include both a general measure of trust in the industy and a specific measure of trust in someone who is more closely aligned to the specificity of the intention criterion.

Using Ronteltap et al.'s (2007) distinction of proximal and distal factors, it could be said that Siegrist's model treats general trust in the industry as a distal factor which exerts an indirect influence on intention through the proximal factors of specific risks and benefits (see Fig. 1, Model 1). If distal and proximal distinctions are based on background generality or more immediate specificity, their positions in the theoretical model could be shifted. For example, if risks and benefits were conceptualised as general beliefs, they might be treated as distal factors which indirectly influence intention through the more proximal factor of medical trust (see Fig. 1, Model 2) or trust in a specific expert (Model 3). Testing of such alternative models may help to clarify the theoretical process of forming genetic intentions.

Figure 1 Three theoretical models of the process by which trust, risks, and benefits influence behavioural intentions.

THE CURRENT STUDY

The current study examined Australians' likely intentions to have their genes tested by a (hypothetical) medical expert in order to explore the roles of general trust in the medical industry, specific trust in the medical expert, and beliefs about the risks and benefits of medical advances on the formation of intentions. It was thought that Australians would readily accept the potential application of genetic technology to personalised health promotion, as they generally endorse the benefits of scientific advances (SNTSM, 2003–2010) and have high levels of trust in medical experts (Hardie & Critchley, 2008).

To advance theoretical understanding of the intention process, structural equation modelling was used to examine the respective roles of trust, benefits, and risks in forming an intention to have a genetic test. Since the researcher‐selected variables of trust, risks, and benefits may not fully account for respondents' own reasons for their intentions, an open‐ended question was included to explore the motives behind Australians' genetic health promotion intentions.

METHOD

Participants

A national sample of 800 Australian adults participated in a computer‐assisted telephone interview survey (see Hardie & Critchley, 2008 for details of sampling and procedure). Older adults and women were somewhat overrepresented. On average, respondents were 54 years of age (standard deviation (SD) = 16, range 18–91, with 70% aged between 35 and 70). The gender breakdown was 64% females and 36% males. Self‐rated health was generally good (M = 3.89, SD = 1.04, range 1–5), with most reporting moderate (57%) or very good (34%) health. The highest level of education completed was 32% university, 18% trade or technical school, 38% high school, and 12% primary school. In terms of employment status, 36% were retired, 29% were employed full‐time, 20% were employed part‐time, 9% did home duties, and 6% were unemployed.

Measures

The telephone survey interview script included demographic questions (age, gender, education, employment), subjective health status (‘Generally speaking, how would you describe your own health?’ 5 = very healthy, I rarely get sick, and generally feel fit and well; 4 = moderately healthy, I generally feel well apart from the occasional minor health problem; 3 = fairly healthy, I have some health problems, but I manage to generally feel well; 2 = Not very healthy, I generally feel a bit unwell and often have health problems; 1 = Unwell, I have some serious health problems and generally feel unwell), a vignette scenario describing a hypothetical genetic expert, quantitative ratings of intentions to have genes tested, an open‐ended question about subjective reasons for intentions, and sets of items measuring medical trust, expert trust, and beliefs about the risks and benefits of medical advances.

A gender‐neutral vignette scenario was used as a standardised stimulus to control for individual differences in respondents' previous experiences with medical specialists. To avoid reference to any actual, available genetic application, the vignette broadly described the use of genetic testing for health promotion. The following scenario was read to all participants:

Medical experts say that people will soon be able to promote good health and prolong life by having their genetic material analysed. A person's genetic profile could tell them in advance what diseases they are vulnerable to, what early treatments they can use to prevent illness, what types of foods best suit their genetic profile, and what they can do to promote good health and prolong life.

I want you to imagine a future situation where you could have your own genes tested. Imagine that this analysis could tell you how to maintain good health, avoid illness and live longer. In this hypothetical situation, Dr Benson is the expert medical specialist who would analyse your genes.

Dr Benson previously worked in a hospital in Melbourne. This is where Dr Benson spent 10 years developing the best techniques to analyse a person's genes. Dr Benson now works in a hospital in your nearest capital city. This is where you would go to have Dr Benson test your genes. Based on your genetic profile, Dr Benson would tell you how to improve your health and prolong your life.

After hearing the scenario, respondents were asked to rate their degree of intention to have their genes tested (‘How likely is it that you would have Dr Benson profile your genes? Rate your intention on a scale of 0–5, where 0 means it is not likely, and 5 means it is very likely that you would have your genes analysed), followed by an open‐ended question (What is your reason for this intention?’) to explore participants' own subjective motives for the decision.

Risk and benefit beliefs

To measure general beliefs about medical risks and benefits, six items were adapted from previous research (Critchley & Turney, 2004; SNTSM, 2003–2010). The original items were reported to be psychometrically sound measures of general beliefs about science and technology. For the current study, items were reworded to reflect beliefs about medical science and technology. Three items represented risk beliefs (‘Advances in medical science will create new, unexpected problems’, ‘Medical science and technology has improved quality of life for patients and their families’, ‘Medical science and technology has improved quality of life for doctors and health professionals’), and three items represented benefit beliefs (‘Advances in medical science will solve most human problems’, ‘I think medical science gives too much control over life and death’, ‘I believe that medical science and technology are out of control’). Items were rated on a 6‐point scale (0 = strongly disagree to 5 = strongly agree), mean scores were calculated for the five‐item risk scale (α = 0.59) and the three‐item benefit scale (α = 0.51), with higher scores representing greater benefits or risks.

Trust in the medical industry

General trust in the medical industry was measured with a set of items representing three medical targets (see Hardie & Critchley, 2008). To assess the general perceived trustworthiness of family doctors, medical specialists, and hospitals, items were worded as follows: ‘On a scale of zero to five, where 0 = no trust at all and 5 = a great deal of trust, how much do you trust your GP/medical specialists/hospitals?’. Mean scores were calculated for the three‐item medical trust measure (α = 0.63), with higher scores representing greater trust in the medical industry.

Trust in the expert

To measure specific trust in Dr Benson, the hypothetical genetic expert, a set of nine items was devised in accordance with McKnight, Choudhury, and Kacmar's (2002) conceptualisation of trust as a multidimensional construct composed of judgements about the perceived integrity, competence, and benevolence of a target. The integrity items (‘I would describe Dr Benson as honest’, ‘I feel that Dr Benson is reliable’, ‘I think that Dr Benson is dependable’), competence items (‘I think that Dr Benson does a good job’, ‘I feel that Dr Benson is knowledgeable’, ‘I would describe Dr Benson as capable and competent’), and benevolence items (‘I feel that Dr Benson is concerned about my welfare’, ‘I think that Dr Benson is motivated to help me and my family’, ‘I feel that Dr Benson acts in society's best interest’) were each rated on a 6‐point scale (0 = not at all to 5 = a great deal), with higher ratings representing greater agreement. The nine items were subjected to principal components analysis. A three‐factor solution accounted for 86.3% of the variance, with three items (loadings > 0.50) representing each dimension. The three dimensions were highly correlated (r = 0.70–0.81), and a second‐order factor analysis showed that the three subscales loaded above 0.79 on a single trust factor which explained 75% of variance. Therefore, the integrity, competence, and benevolence subscale scores were combined to yield a single expert trust score. Internal consistency of this measure was very good (α = 0.90).

Statistical analysis

Analyses of quantitative ratings, including sample descriptives, psychometric analyses, t‐tests, correlations, and analyses of variance, were conducted using SPSS version 17 (SPSS Inc., Chicago, Illinois, USA). AMOS 17 (SPSS Inc., Chicago, Illinois, USA) was used for structural equation modelling (SEM). Three models were tested and compared for comparative fit: Model 1, a benefit–risk model based on Siegrist (2000); Model 2, an alternative trust model that reversed the proximal and distal positions of trust, risks, and benefits; and Model 3, a revised expert trust model that gave prominence to trust in a specific expert as the key proximal determinant of intention (see Fig. 1). Each model contained four latent variables (benefit beliefs, risk beliefs, medical trust, and expert trust) and one observed variable (intention). The three benefit items and three risk items were treated as indicators of benefit beliefs and risk beliefs, respectively. Three items measuring trust in the medical industry were treated as indicators of medical trust. For expert trust, the three subscale scores reflecting judgements about the perceived integrity, competence, and benevolence of Dr Benson were used as indicators of this latent variable.

Analysis of open‐ended responses (reasons for genetic intentions) involved identification of themes, using a combination of coding down and coding up strategies (Bowling, 2002). A preliminary coding scheme was devised by the primary researcher for initial thematic classification of open‐ended responses by the team of trained telephone interviewers who conducted the survey. Uncoded verbatim responses were later examined by two senior researchers to identify and add emerging themes to the classification system and cross‐check the original thematic coding of each response.

RESULTS

For this sample, the average likely intention rating was 2.64 (SD = 1.83, range 0 to 5), with 42% rating their likely intention to have ‘Dr Benson’ profile their genes at 2 or below, and 58% rating their intentions at 3 or above. There were no significant effects of education, employment, or gender on intention to have genetic tests (p > .05 for all comparisons using t‐tests and one‐way analyses of variance) and no association between subjective health ratings and intention (r = 0.01, p > .95). There was a modest, but statistically significant negative association with age (r−0.12, p < .001), showing that intention declined with age.

Examination of the sample means for all key variables showed that beliefs about the benefits of medical advances were moderately high (M = 3.72, SD = 0.76), and beliefs about the risks were moderately low (M = 2.49, SD = 0.99). General trust in the medical industry was moderately high (M = 3.79, SD = 0.71), as was trust in ‘Dr Benson’, the hypothetical medical expert (M = 3.52, SD = 0.90).

Model testing: influence of trust, risks, and benefits on intentions

The sample size of 800 allowed generous parameter to case ratios of at least 1:24 for SEM model testing. The asymptotically distribution‐free estimation method was used to calculate parameter estimates and fit statistics. Model fit was assessed via conventional criteria (goodness of fit (GFI) and adjusted goodness of fit (AGFI) > 0.90), but in light of the sensitivity of the chi‐square (χ²) statistic to sample size, fit was deemed acceptable if the value of χ²/df fell between 2 and 3 (see Bentler, 1990; Byrne, 2001 on fit indices). For all three structural models, the measurement model was found to be sound, with all measured indicators showing significant loadings (p < .001) on their respective latent factors.

For Model 1, the benefit–risk model which posited an indirect pathway from trust to intention via direct pathways of risks and benefits, the overall fit of the structural model seemed good (χ²/df = 2.85, GFI = 0.96, AGFI = 0.94); however, only two pathways were statistically significant. There was a direct pathway from medical trust to benefits (standardised path coefficient 0.53, p < .001) and a direct path from benefits to intention (coefficient 0.35, p < .001). The pathways from medical trust to risks, and from risks to intention, were not significant (p > .05). Although these results partly supported Model 1 by showing that trust in the medical industry indirectly influenced intention through benefit beliefs, it appeared that risk beliefs were not influenced by trust and did not play a role in the process of forming an intention.

Model 2, an alternative trust model, reversed the positions of latent factors, positing indirect influences of risks and benefits via trust, which was expected to have a direct influence on intentions. When this alternative process was tested, the overall fit was similar to the first model (χ²/df = 3.11, GFI = 0.94, AGFI = 0.91), but there were four significant pathways. There were direct paths from benefits to medical trust (coefficient 0.64, p < .001) and from benefits to trust in the expert (coefficient 0.50, p < .001). There was a pathway from risks to medical trust (coefficient = −0.36, p < .001) but not to expert trust (p > .05). There was no path from medical trust to intention (p > .05), but there was a direct path from expert trust to intention (coefficient 0.48, p < .001).

Modification indices for Model 2 suggested that inclusion of a bidirectional pathway between risks and benefits could improve the model. This was consistent with Slovic's (1999) assertion of an inverse relationship between risks and benefits, so this additional path was added to Model 3. Since general medical trust had shown no direct influence on intention in Model 2, this path was dropped in Model 3. This revised Model 3 was expected to provide a more parsimonious assessment of a process whereby general beliefs about the risks and benefits of medical advances, and general medical trust, are distal factors with only indirect influences on intentions. Only trust in a specific expert was expected to exert a proximal direct influence on intentions. Model 3, including a covariance pathway between risks and benefits, is shown in Fig. 2. The fit statistics were good (χ²/df = 2.93, GFI = 0.94, AGFI = 0.92), and there was a significant improvement on the previous model (Δχ² (1) = 9.87, p < .01). Every posited pathway in Model 3, the expert trust model, was found to be significant (p < .001).

Figure 2 Final expert trust model showing indirect influences of general benefits, risks and medical trust, and direct influence of expert trust, on intentions (N = 800, χ²/df = 2.9, GFI = 0.94, AGFI = 0.92, all pathways p < .001).

Thematic content analysis: subjective reasons for genetic test intentions

Most participants (n = 761) provided reasons for their intentions. One reason per participant was recorded (sole response or first response). Content analysis revealed 11 distinct themes: wellness, not for me, mistrust/trust, morality, cost, Benson's reputation, fear, family history, curiosity, public health system, and private health system.

Three themes referred to the health of the respondent or their family, namely wellness, not for me, and family history. Wellness responses included ‘I'll do all I can to maintain health and prolong my life’ and less enthusiastic responses such as ‘I don't want to, but I would have the test if my family insisted’. Not for me reasons included ‘I'm too old’, ‘I'm not interested’, and ‘Maybe not for me, but I would have it done for my children’. Family history involved such reasons as ‘I would only have the test if I knew that certain illnesses run in my family’.

Five cognitive–emotional themes were identified. Mistrust/trust reasons included ‘I don't trust new technology’, ‘I would worry about who could access my genetic material’, ‘I trust the medical system’. Morality reasons included religious beliefs (e.g., ‘Doctors should not play God’, ‘My health is in God's hands’), moral and ethical dilemmas (‘How long should we prolong human life?’, ‘There could be unintended eugenic consequences’). Fear reasons included negative cognitions (‘I don't want to know’), emotions (‘Knowing could make me panic’), and attitudes (‘I prefer not to know bad things so I can maintain a positive attitude’). Curiosity reasons were simply ‘I'd have the test to satisfy my sense of curiosity’ or ‘I'm curious about my genes’. Finally, some respondents commented on the hypothetical Dr Benson's reputation as the reason for their intention, with some citing his (imagined) track record, knowledge, competence, and experience, while others were more cautious, saying they would seek testimonials and evidence from other patients before making a decision.

Three practical themes involving cost and access to public or private health‐care systems were identified. Cost responses included ‘Whether or not I have the test depends on how expensive it is’, ‘I want to have the best services at any cost’, or ‘I probably couldn't afford it’. Reasons involving the public health system included ‘I would only have it if it was a public health service’ or ‘I would have it if it's offered by Medicare’, while private health system responses ranged from ‘I probably wouldn't have it because I don't have private health insurance’ to ‘I would only have the test if it's available in a private hospital’. Public and private reasons sometimes implied financial issues, and could arguably be located within the cost theme, but the stated reasons centred on public/private health‐care systems and services without explicitly mentioning cost.

Table-1 shows the most frequently cited themes (and example quotes) reported by those with low (0–2) intention ratings (n = 322 non‐intenders) and those with high (3–5) intention ratings (n = 439 intenders).

Table 1 Frequencies of Australians' subjective reasons for intending or not intending to have their genes tested for health promotion purposes

Theme	Frequency (%)	‘Example quote’
	n = 439 Intenders
Wellness	321(73.1%)	‘I'll do anything to maintain my good health.’
Cost	36(8.2%)	‘I would want the test at any price.’
Mistrust/trust	19(4.3%)	‘I trust the medical system.’
Benson's reputation	19(4.3%)	‘I think Benson has a good track record.’
Not for me	17(3.9%)	‘Maybe not for me, but I'd have my children tested.’
Curiosity	9(2.1%)	‘I'd like to know more about my own genes.’
Public heath system	6(1.4%)	‘I'd have the test if it was a public health service.’
Private health system	5(1.1%)	‘I would only have the test done in a private clinic.’
	n = 322 Non‐intenders
Not for me	116(36.0%)	‘It's too late for me—I'm already ill’
Wellness	68(21.1%)	‘If my family insists—I owe it to them to stay healthy’
Mistrust/trust	62(19.2%)	‘I don't trust new technology’
Morality	31(9.6%)	‘My health is in God's hands’
Fear	15(4.7%)	‘I would be afraid to know about my genes’
Cost	12(3.7%)	‘I probably couldn't afford it’
Family history	10(3.1%)	‘I wouldn't be tested unless there was an inherited risk’
Private health system	5(1.5%)	‘I don't have private health insurance’

The most frequent reasons given by intenders were health‐related reasons of wellness, a disregard for practical considerations of cost, cognitive–emotional themes of trust and Benson's reputation, and not for me intentions to have tests for their children but not themselves. The reasons most frequently given by non‐intenders were not for me (too old, too late), reluctant (family‐driven) wellness, and cognitive–emotional themes of mistrust, morality, fear, and cost. Only a small number based their intentions on family history, curiosity, public or private health systems.

DISCUSSION

More than half of this large community sample of Australian adults indicated a willingness to have their genes tested, suggesting that when genetic profiling becomes readily available and affordable, many Australians will adopt this personalised approach to health promotion. Results of both the quantitative and thematic analyses highlighted the important role of cognitive–affective trust judgements in forming genetic intentions. Trust ratings for the (hypothetical) medical expert were shown to have a direct influence on intentions in the SEM analysis, while the content analysis showed that trust/mistrust themes consistently emerged in the top three of respondents' subjective motives for their intentions. These results have a number of theoretical and practical implications.

Theoretical process of forming intentions

Three theoretical models were tested to explicate the process by which trust, risks, and benefits influence intentions. There was partial support for Model 1, but only two pathways reached statistical significance. Trust in the medical industry indirectly influenced intention through beliefs about the benefits of medical advances; however, medical trust did not influence risk beliefs. Indeed, risk appeared to play no role in the process of forming an intention in Model 1. Thus, there was minimal support for Siegrist's (2000) benefit–risk model.

There was better support for Model 2, an alternative trust model which gave prominence to expert and industry trust via four significant pathways. This model supported a process whereby general beliefs about the benefits and risks of medical advances directly influenced general medical trust, but general medical trust had no direct influence on intention. Beliefs about benefits, but not risks, influenced trust in the specific expert which, in turn, had a direct influence on likely intentions.

Although Model 2 was promising, Model 3 was a significant improvement on the two previous models. In Model 3, the expert trust model which posited a modified theoretical process, every specified pathway was significant. The inversely related factors of general beliefs about the risks and benefits of medical advances directly influenced general trust in the industry but exerted only indirect influences on intentions. General trust had a direct influence on the more proximal factor of trust in the specific expert, ‘Dr Benson’. In turn, expert trust was the only direct influence on genetic intentions. These findings support Ronteltap et al.'s (2007) distinction between background, distal factors, and more immediate, proximal factors. Background beliefs about the benefits and risks of medical advances and general trust in the medical industry had only an indirect impact on intention through their contributions to a more immediate factor, trust in the expert who would perform the genetic testing. These results also supported Slovic's (1999) assertion of a negative relationship between subjective lay beliefs about risks and benefits.

The current results were consistent with previous research suggesting that greater community acceptance of a new medical technology such as genetic testing is strongly associated with trust in medical experts and less strongly with beliefs about the potential risks of medical advances (Calnan, Montaner, & Horne, 2005). These findings also mirror recent research showing that trust plays a key role in Australians' acceptance of scientific advances (Critchley, 2008).

The current findings give prominence to trust in the process of forming intentions. In contrast to Model 1, a parallel of Siegrist's (2000) model which proposed that risks and benefits had direct influences on intentions, Model 3, the expert trust model, showed that risks and benefits had only an indirect influence through trust factors. There are, of course, several points of difference in the conceptualisation of key variables which should be noted. Siegrist assessed general trust in the industry, which would be construed as a distal factor. The current study assessed both distal and proximal forms of trust: general trust in the industry and specific trust in an expert. Siegrist measured perceived risks and benefits about specific genetic applications and respondents' likely intentions to buy specific products related to those same applications. The current study assessed general beliefs about the risks and benefits of medical advances, but measured the more specific intention of having a genetic test. The current findings give prominence to trust, while Siegrist's research gives prominence to benefit and risk perceptions in the process of forming intentions. It seems likely that distal‐proximal distinctions and the generality/specificity of measurement instruments could have influenced these incongruent findings. These issues must be addressed in future studies in order to advance theoretical and empirical understanding of the processes by which people make health and medical decisions and form behavioural intentions.

Subjective motives for intentions

While the quantitative findings highlighted the importance of trust judgements and, to a lesser extent, beliefs about risks and benefits in people's intentions to have their genes tested, they did not adequately explain why some people were more willing than others to have such tests. A majority (58%) said they were likely to have their genes tested by ‘Dr Benson’, but a substantial minority (42%) did not intend to do so. Responses to the open‐ended question provided some insight into the subjective motives behind people's intentions.

Intenders were overwhelmingly concerned about their own wellness. For nearly three quarters of this group, intention was motivated by a wish to maintain or improve health, or extend the lives of themselves or their family. Some were willing to do so at any cost. Others were willing to try a new technological approach because they trust the medical system that (presumably) endorses such an approach. For a small number, curiosity about their own genetic profile provided sufficient reason to have the tests. Interestingly, information gleaned from the vignette led some intenders to cite confidence in the reputation and competence of ‘Dr Benson’ as the reason for their intention, even though they had been clearly informed that this was a hypothetical scenario.

In contrast, over one third of the non‐intenders felt it was too late for health promotion due to their age or ill‐health, but said they might consider genetic testing if their family insisted on it. It should be noted that this was only partly consistent with the quantitative results, which showed a significant negative correlation between age and intention, but no association between subjective health status and intention. Some non‐intenders cited general mistrust of new technologies or mistrust about the confidentiality of the system, with concerns about who could access their genetic information or use their genetic material. This supports literature noting that genetic privacy is an important public concern (Anderlik & Rothstein, 2001).

Nearly 10% of non‐intenders cited moral concerns or ethical dilemmas based on religious beliefs and personal values. Such moral dilemmas have been previously noted in an analysis of contradictory public attitudes to genetic testing (Jallinoja et al., 1998). A small proportion of the non‐intenders seemed to favour avoidance, stating that fear or anxiety about what the test might reveal was the main reason for not being tested. Others cited practical reasons for not being tested. These included the (imagined) high cost of the test or the perception of little benefit when there was no known risk or family history of disease.

These findings highlight some of the factors which might underpin public acceptance of new genetic technologies. They may also help to explain why some people are reluctant to engage in health promotion activities, while others are willing to do ‘whatever it takes’. Public health messages endorsing health promotion behaviours abound, yet many people continue to make risky lifestyle choices. If health promotion strategies are to be effective, then the subjective motives that underpin people's decisions to promote or risk their own health must be better understood.

Acceptance of personalised genetic health promotion in Australia

One of the most striking findings in the current study was the influence of the hypothetical ‘Dr Benson’ on people's decisions about having their genes tested. Both the quantitative ratings and subjective themes verified the power of perceived medical expertise. Many Australians said they were willing to undergo genetic testing on the basis of trust in the imagined competence of this ‘expert’. This was consistent with previous research suggesting that physicians play an influential role in patients' intentions to undergo genetic testing (Bosompra et al., 2000) and recent research showing that Australians have high levels of trust in medical doctors (Hardie & Critchley, 2008).

Many Australians who intended to have their genes tested also reported a strong desire for wellness that seemed to make them particularly receptive to ‘expert’ advice. These findings suggest that a reasonable proportion of the community may be very receptive to personalised health promotion advice, although it is not yet clear if such optimism about personalised genetic health promotion is warranted. Previous research has shown that Australians' genetic intentions were guided by positive affect (Pin et al., 2008). In contrast, that study showed that Dutch intentions were tempered by worry. The common Australian attitude of ‘no worries’ may contribute to unrealistic expectations about how genetic advances will help to promote health and prolong life.

For the substantial minority who said they were unwilling to have their genes profiled, concerns about issues such as genetic privacy, the morality of seeking to prolong human life, fear, and mistrust of technology contributed to the decision‐making process. If future gene‐based health promotion strategies are to be broadly accepted, then these cognitive–emotional concerns of the more reluctant will need to be addressed. Health professionals will be needed to help individuals tackle moral dilemmas, fear, mistrust, unrealistic optimism, and prescribed behavioural and lifestyle changes which might accompany personalised genetic health promotion advice.

A future role for genetic health promotion psychologists?

For both groups of Australians, the intenders and the non‐intenders, the availability of personalised genetic health promotion may create a need for health promotion psychologists. Health promotion in the not too distant future could involve multidisciplinary teams of health professionals that profile a person's genes, identify their strengths and vulnerabilities, and make highly personalised recommendations for proactive diet, exercise, and other lifestyle changes, plus defensive options for prevention or pre‐emptive treatment of diseases which pose a genetic risk. Given the current constraints on our health‐care system and the limited resources available for illness prevention, the health professionals who are best trained to help individuals make proactive lifestyle changes and/or accept unpleasant news about their susceptibility to disease are health psychologists. Although genetic counsellors can help with some issues surrounding genetic testing, and clinical psychologists are highly skilled in the diagnosis and treatment of mental health problems, it is health psychologists who are trained to focus on cognitive, emotional, behavioural, and social aspects of physical health and medical illness. Those who specialise in health promotion have the unique professional skills needed to help clients with various medical decisions and lifestyle changes prescribed by their personal health promotion profiles of genetic strengths and vulnerabilities.

Limitations of this study

This study had some limitations that must be acknowledged. Although the sample size was large, women and older adults were overrepresented. This is a common problem with Australian telephone surveys (see, e.g., SNTSM, 2003–2010). Women and older adults tend to be more likely to answer landline telephones and answer survey questions. Future studies might use alternative methods (e.g., online or mobile phone surveys) to access more young adults, particularly younger men. There was also a bias towards reasonably good self‐reported health in the current sample. Access to participants experiencing a wider range of health conditions and use of more objective health status measures (e.g., medical assessment) may provide a clearer picture of Australians' genetic intentions in regard to health promotion.

Another limitation was the cross‐sectional nature of this study. The factors which predicted intentions in this study may not remain stable over time. Various life changes in health and family circumstances may influence the process of forming intentions; therefore, longitudinal studies are needed to explore temporal aspects of the intention process.

It must also be noted that actions do not always follow intentions. Recent studies on the intention‐behaviour gap suggest that volitional factors such as planning, self‐efficacy, and control can mediate between intention and behaviour (Sniehotta, Scholz, & Schwarzer, 2005) and that action is more likely to follow when intentions are aligned with moral norms (Godin, Conner, & Sheeran, 2005). It remains to be seen if intentions will result in actual genetic testing when complete DNA sequencing becomes available in Australia; however, future studies would benefit from the inclusion of volitional and moral factors when examining genetic intentions and subsequent actions.

This survey used a vignette to describe a hypothetical genetic expert, but no manipulation check was included to ensure that the vignette did indeed elicit expert trust. Although this sample's moderately strong mean ratings for the integrity, competence, and benevolence of Dr Benson would suggest that the vignette did elicit perceived trust, future studies would benefit from closer examination of this type of methodology.

The qualitative component of this study was minimal, using a simple thematic content analysis of participants' initial responses to a single open‐ended question. More comprehensive qualitative studies involving longer interviews and in‐depth analyses are clearly needed to explicate Australians' motives for having (or not having) their genes profiled.

CONCLUSION

Previous research suggested that a deeper understanding of people's contradictory attitudes was needed to understand health decision‐making in relation to genetic testing (Jallinoja et al., 1998). The current study attempted to expand that understanding by exploring people's rated intentions and also their subjective motives for decisions about personalised genetic health promotion. Many past studies examined only researcher‐selected quantitative variables. The current study extended the quantitative analysis by also exploring participant‐driven subjective reasons for genetic intentions. Taken together, the quantitative and qualitative findings of the present research showed that cognitive–emotional factors such as trust in medical experts, trust in the industry, beliefs about the risks and benefits of medical advances, desire for wellness, mistrust of the system, curiosity, fear, and moral concerns can contribute to health decisions. In particular, these results highlight the importance of trust judgements in health‐related decision‐making. It remains to be seen if Australians will embrace personalised genetic health promotion when it becomes readily available and affordable. These findings suggest that many will do so. If that is the case, there may be an impending need for health psychologists with specialised skills in the area of genetic health promotion.

ACKNOWLEDGEMENTS

This study was part of a larger project supported by a Swinburne University Research Development Grant. I would like to thank Dr Christine Critchley for her many contributions to the project and to early versions of this manuscript. I would also like to thank Gordana Bruce and her dedicated team of computer‐assisted telephone interviewers for conducting the telephone surveys.

Notes

Abbreviated reports of this study were presented at the International Genomics and Society Conference held in Amsterdam, The Netherlands, 17–18 April 2008, and the Annual Meeting of the Society for Risk Analysis, held in Boston, MA, USA, 7–10 December 2008.