Living organ donation: the effect of message frame on an altruistic behaviour

Introduction

The shortage of deceased organ donations is a major public health concern. As supply continues to fall short of demand, patients on transplant waiting lists are dying before a suitable organ is found (Abouna, 2008). In an attempt to improve this situation, medical procedures have been developed which allow a healthy, living person to donate an organ. Living donor transplant procedures originally required a genetic relationship between the donor and the recipient in order to minimize possible organ rejection, but advances in immunosuppressant medication now allow non-kin donations. Therefore, this study examines factors that influence peoples’ decision to be a living donor, and is to the authors’ knowledge the first to explore perceptions of risk and message framing within the context of living donation. We explore this for living donor kidney transplantation (LDKT) and living donor liver transplantation (LDLT).

LDKT was first performed in 1954 and involves a healthy individual donating one of their two healthy kidneys to a patient in need (Murray, 1982). LDLT in contrast is a relatively new procedure and involves a healthy individual donating a section of their liver. In adult LDLT, 60% of the liver is donated and both the part that remains in the donor and the part received by the recipient regenerate to approach the size of the individual’s original liver (Marcos et al., 2000).

While such procedures offer hope to patients, living donation involves a level of risk for the healthy donor. LDKT is considered to be of relatively low risk to the donor, with a mortality rate of approximately 0.03%, and a complication rate below 10% (Matas, Bartlett, Leichtman & Delmonico, 2003). A much higher risk to the donor’s health is associated with LDLT: a mortality rate of 0.5% and a corresponding morbidity rate of between 1 and 100% (median of 16.1%) is reported (Middleton et al., 2006). However, while both procedures are costly to the donor they are highly beneficial to the recipient due to: (1) reduced time on the transplant waiting list, (2) the chance to check and monitor the condition of the organ prior to transplant and (3) the possibility of conducting the procedure at a time suitable for both the donor and recipient (Griva et al., 2002; Belghiti and Durand, 2000). Consequently, many transplant centres around the world currently offer LDKT and have begun to offer LDLT as a treatment option to patients.

Living kidney/liver donation may be described as an altruistic act as the donor puts themselves at risk to benefit another. Therefore, it is essential that the donor is made fully aware of the risks, over and above the benefits to the recipient, before they consent to the procedure. However, the extent to which the risks are understood by potential donors and recipients has been questioned, and has recently been noted as a major deterrent to the uptake of LDLT specifically (McGregor et al., 2008; McGregor et al., 2010). Consequently, to know if interventions that focus on perceptions of risk can increase possible willingness to participate in living organ donation would be advantageous.

Message framing effects refer to the different reactions or behaviours elicited when information is presented either as gains (e.g. 40% chance of survival) or losses (e.g. 60% chance of death) (Kahneman and Tversky, 1979). People have been shown to be risk seeking (prefer to take a gamble over a certainty) when information is presented as losses and risk averse (prefer a certainty over a gamble) when presented with gains (Kahneman and Tversky, 1979). Developing this concept to the domain of health promotion, Rothman and Salovey (1997) suggested that frame valence is moderated by the type of behaviour. They suggest there is a loss frame advantage for detection behaviours (e.g. health screening) which are described as high risk (e.g. people may discover health problems they were previously unaware of). Conversely, they argue for a gain frame advantage for prevention behaviours (e.g. sun screen use) which are seen as low risk behaviours (e.g. a simple act carried out to prevent future health problems, with little chance of side effects or risk). There is evidence to support this basic prediction (Ferguson and Gallagher, 2007). However, in a series of meta-analytic papers, O’Keefe and colleagues have shown that the gain frame advantage for prevention behaviours (O’Keefe and Jensen, 2007) and the loss frame advantage for detection behaviours (O’Keefe and Jensen, 2009) are small, corresponding to rs of .03 and -.04 respectively. As a consequence, it has been argued that it is not the purpose of the behaviour itself that is important but rather perceptions of risk associated with that behaviour (Latimer et al., 2007).

With respect to the topic of organ donation, the behaviour cannot be simply categorized as a detection or prevention behaviour as the behaviour itself is not based on self-directed perceptions of benefits but rather is directed towards the benefit of another person. Reinhart et al., (2007) investigated message frames with respect to intention to donate organs and tissue after death, while Purewal and van den Akker (2009) considered women’s intention to donate their eggs to childless couples. Both found support for a gain-framed advantage despite the more unusual behavioural scenario. Such results perhaps support the argument that it is the level of personal risk that is the crucial factor, as both are low risk behaviours (Bartels et al., 2010).

The influence of message frame has not yet been investigated with respect to living organ donation, an altruistic act with significant personal risk for the donor. This study investigates the concept of message frame in its most basic form and assesses the impact of frame using a single attribute, the associated risk to the donor (Levin et al., 1998). Based on the work of Rothman and colleagues, the hypotheses tested within this study is that people will be more willing to proceed with living kidney donation (low risk) when presented with a gain frame and living liver donation (high risk) when presented with a loss frame.

A key variable known to enhance framing effects is issue involvement (e.g. Millar and Millar, 2000). Specifically, involvement is believed to increase the amount of central processing that individuals will pay to the information provided (Petty and Cacioppo, 1986). Within the context of blood donation, Ferguson et al., (2008) equated issue involvement to the degree of prior commitment to donate. In the context of organ donation in the UK, where an ‘opt-in’ system currently operates, this would be equivalent to registering on the NHS organ donor register. A survey conducted in 2005 indicated that 87% of the UK population would be prepared to donate their organs following their death, and yet only 28% had taken action to this effect joined the organ donor register. The majority of those not on the register had either not yet considered the possibility of joining the organ donor register (45%) or simply had not got around to joining (30%) (BBC press release, 2005). Those individuals who have actively joined the register may, therefore, be described as showing more commitment to organ donation and, by comparison to the majority of those not on the register, more interest and involvement with the topic. In the wider framing literature this type of expressed commitment is sometimes referred to as involvement. For example, current experience and future expectation about undertaking the target behaviour have been used as indices of involvement (see Donovan, 2000; Millar and Millar, 2000). Signing on the organ donor register indicates a future expectancy to perform the behaviour. Framing effects are often stronger for those who are committed or involved in the issue (e.g., Maheswaran and Meyers-Levy, 1990; Millar and Millar, 2000; Rothman et al., 1993). Consequently, we established whether or not participants were on the NHS organ donor register and predicted that the framing effects would be stronger for those who are registered.

The experiments reported here were designed to explore how message framing, within the context of living organ donation, influences willingness to donate to or receive from a relative. While altruistic donation of an organ can be to a stranger, in the UK donation to a close relative is more common and as such this study examines kin-based donations (NHS Blood and Transplant, 2009).

Method

Results

The perceived risk to the donor was rated as significantly higher for the liver (M = 5.2, SD = 0.96) compared to the kidney group (M = 4.31, SD = 1.17); t (64) = −3.35, p = 0.001. There was no significant difference with regards the perceived risk to the recipient and no significant differences between liver and kidney groups regarding perceived personal cost and benefits (for both recipient and donor). This confirms that LDLT is perceived to be a riskier procedure for the donor than LDKT. The costs in terms of health consequence was calculated by subtracting the participants’ response to ‘How likely is it that the donor will, as a result of the living donor kidney / liver transplant procedure, have deterioration in their health?’ from responses to ‘How likely is it that the donor will, as a result of the living donor kidney / liver transplant procedure have improved health?’. Positive scores indicate a relative improvement in health and a negative score, a relative deterioration in health. The results confirm that LDLT (M = −2.17, SD = 1.97) has a significantly higher cost in terms of donor health than LDKT (M = −1.19, SD = 1.86); t (64) = 2.06, p = 0.04. There were no significant differences with regards the health consequences for the recipient.

Design

A 2 (frame: gain vs. loss) by 2 (organ cost: liver vs. kidney) by 2 (involvement: involved vs. not involved) between-subjects design was employed. Participants were asked to complete one of four questionnaires. Questionnaires differed according to organ and frame and were collated in a random order for blind distribution to participants: gain/liver (n = 21); loss/liver (n = 24); gain/kidney (n = 22), and loss/kidney (n = 20). Involvement was classified according to whether participants indicated they were currently on the NHS organ donor register or not (see Table 1).

OPEN IN VIEWER

Table 1.

Additional group characteristics

	Liver		Kidney
	Yes	No	Yes	No
Are you currently on the NHS organ donor register?	47.7%	52.3%	48.8%	51.2%
Before taking part in this study had you ever heard of living liver/kidney donation?*	66.7%	33.3%	87.8%	12.2%
If yes, on a scale of 1 to 7, how much do you know about living liver/kidney donation? (mean (SD))	3.33 (1.30)		3.30 (1.27)
Is someone in your close friends or family currently on an organ transplant waiting list?	0%	100%	4.8%	95.2%
Has anyone in your close friends or family ever received an organ transplant?	13.3%	86.7%	2.4%	97.6%

*

χ², p<0.05

Framing manipulation

Each questionnaire began by instructing the participant to read a short vignette corresponding the relevant frame and organ combination. The vignette was constructed to give basic information about either LDKT or LDLT and was designed to initially persuade participants to consider the information from the point of view of becoming a potential donor to a relative. The vignettes presented to the liver and kidney groups are presented in Figures 1 and 2 respectively.

OPEN IN VIEWER

Figure 1.

Vignette for liver condition

OPEN IN VIEWER

Figure 2.

Vignette for kidney condition

After reading the vignette, participants were asked to imagine themselves as ‘a relative of a person with kidney/liver disease’ and answer a question on their willingness to donate. Participants were also asked to imagine themselves as ‘a patient with kidney / liver disease’ and answer a similar question regarding their willingness to accept a donation. The order of each role was counterbalanced.

Measures

Results

A series of three-way between-group ANOVAs (frame × transplant × involvement) was conducted to establish the impact of message frame, transplantation type, and issue involvement on consideration to donate and receive an organ.

Considering donation

With regards willingness to become a living donor, there was a significant main effect for NHS registration, (F(1, 77) = 8.66, p = 0.004 ( η p 2 = 0.101)). Those currently on the NHS organ donor register were more likely than those not currently on the NHS organ donor register to talk to the doctors about donating to a sick relative (M = 6.34, SE = 0.17 and M = 5.66, SE = 0.16 respectively).

In addition, the two-way interaction between transplantation type and frame was also significant, (F(1,77) = 4.68, p = 0.034 ( η p 2 = 0.057)). Decomposing this interaction showed that for those exposed to a gain frame, the kidney (low risk) group (M = 6.36, SD = 0.95) were more willing to donate compared to the liver (high risk) group (M = 5.67, SD = 1.24) (t (41) = 2.07, p = .04) (see Figure 3).

OPEN IN VIEWER

Figure 3.

How likely would you be to talk to the doctors about donating one of your kidneys / part of you liver to your sick relative? (Organ * Frame)

With respect to willingness to be a recipient of a living organ transplant there were no significant main or interactive effects for frame, involvement or organ type.

Discussion

This is the first study to show that willingness to become a living organ donor to a family member is influenced by the interaction between the type of donation (high versus low cost donation) and how the information about the donation is framed. Specifically, the results show that objectively low risk donations are influenced by gain frames, which supports our hypothesis based on Rothman and Salovey’s (1997) theory. However, partial rejection of this hypothesis is also warranted as high risk donations are not influenced by loss frames. This partial framing effect may suggest an important role for perceived personal benefit. Personal risk, such as that involved in health screening (i.e. detection behaviours), also has the potential for personal benefit. However, the personal benefit is less tangible with organ donation and focuses more on a sense of personal reward or warm glow from helping (Andreoni, 1990). When the risk is high, a more favourable reaction from a loss frame may only be possible if personal benefit is also high allowing the behaviour to be personally encouraged.

It should be noted that this study focused on kin selection i.e. altruism toward a related individual rather than altruism to a stranger (Griffin and West, 2003; Nowak, 2006), which is a strong and pervasive force for altruism among kin (Griffin and West, 2003). Hamilton’s rule suggests that helping among kin occurs when the degree of relatedness exceeds the cost-benefit ratio (where cost is to the donor and benefit to the recipient). Thus an added importance of this study is in showing that a simple intervention focusing on one aspect of Hamilton’s rule i.e. the costs to the donor, is sufficient to increase the donors’ willingness to donate. It is more than probable that this alters the cost-benefit ratio, but this needs direct testing. We focused on costs to the donor as focus on the actor rather than the recipient of an action has been the focus of framing studies and health psychology in general (Ferguson, 2011). Furthermore, it allows generalization to the wider organ donation literature. That is, if a person is considering donation to a stranger then, according to Hamilton’s rule, personal costs will be more important than benefits to the recipient because the person is not related to the donor. Therefore, future framing studies on organ donation may want to consider focusing on donation to strangers and examine the mediating effect of perceived benefits to the donor. In our study benefits to the donor are implied by the type of donation. The relative benefits to the liver recipient are likely to be greater than to the kidney recipient.

Even when personal benefit is low, additional information may help to highlight benefits to others and self (especially if it is a positive frame as this highlights risk avoidance and moves the person towards acting in the direction of a low risk behaviour) and this should encourage helping. The pattern of results reported here support this. When risk is high people’s altruistic choices may require different frames such as those that emphasize more egoistic outcomes (Ferguson et al., 2008). It is also of note that the moderating effects of frame were observed for objective rather than subjective risk.

There was no framing effect for the recipient, regardless of the type of organ transplant. This may reflect that for the recipient the decision making process is more complex than for the donor. The donor is healthy and the donation could make them ill (altruistic), whereas the recipient is ill and the donation will make them healthier. Therefore, the recipient has to decide on an action that is beneficial to them but detrimental to the donor. This concept has been found to result in concern for the donor, distress and anticipated guilt (McGregor et al., 2010). There is evidence that an individual’s emotional state can influence the direction of message framing effects in relation to health behaviours (e.g. Gerend and Maner, 2011). Specifically, fear makes people risk averse and anger risk seeking (Lerner and Keltner, 2001), thus fear of what may happen to the donor makes recipients risk averse. Future work needs to examine the emotions of the recipient (and the donor) regarding their decisions in order to make predictions about framing effects in this context.

Within this study, the majority of participants within each group did not know of someone on a transplant waiting list nor did they know someone who had had a transplant (see Table 1). This suggests that the topic of living organ donation was not a particularly relevant subject matter to the majority of participants. Prior experience and relevance have been shown to influence behaviour and perceptions by allowing the person to process the information presented more in-depth and consequently leave them susceptible to framing effects (Detweiler et al., 1999). However, given the rarity of both procedures, most people will not have any prior experience. Therefore, the findings here have direct relevance to the area of organ donation, as they show that even in the absence of prior experience there is still a framing effect.

A consistent finding is the main effect of NHS organ donor registration. Participants who were on the organ donor register, stated that they would be more likely to talk to doctors about donating and considered the overall risk to the donor to be lower, when compared to those not on the NHS organ donor register. This perhaps suggests that a more positive attitude towards organ donation in general can lead to more positive perceptions of living donation. Therefore, the continued promotion of deceased organ donation could be encouraged as it may subsequently promote favourable attitudes towards living organ donation. However, an alternative explanation could be that individuals who are more favourable towards living donation are already registered.

A more detailed measure of attitude towards living organ donation would be beneficial to future research in this area. It has previously been found that attitude can moderate framing effects within the context of another novel, high risk behaviour, namely the use of male hormonal contraception (O’Conner et al., 2005). As predicted, a loss frame resulted in higher intention to use male hormonal contraception, but this relationship was moderated by a positive attitude towards the behaviour. The moderating impact of attitude within the context of living organ donation is yet to be assessed.

One of the main limitations with this study is with regards the number of participants within each of the groups and the unequal gender balance. A further large-scale study, involving a more equal proportion of males to females, would increase the power of the tests and allow gender effects to be analysed. Currently, the study described here is underpowered with only 87 participants (a minimum of 20 in each organ × frame combination). However, it is important to note that effect size of framing effects in the health domain are very small. In a recent meta-analysis, conducted within the domain of health prevention behaviours, a small effect size for a gain frame advantage was found (r = .03) (O’Keefe and Jensen, 2007). A similar result was found in a further meta-analysis with health detection behaviours (r = -.04) (O’Keefe and Jensen, 2009). The results of this small-scale, pilot study should, therefore, not be overlooked but rather should encourage further investigation. Furthermore, these results replicate those of Apanovitch et al., (2003) and Ferguson and Gallagher (2007), who both showed gain framed advantages for low risk but no effect for frame on high risk health behaviours. Therefore, while the effects are small they are replicable and as such start to throw some light on the limits of framing effects. That is, for high risk behaviours framing effects are not observed and it may well be the case that once the environmental saliency is great the influence of this type of intervention is negligible.

In addition, further possible mediators and moderators of the relationship between frame and intentions could be measured and analyzed in future studies. Other studies involving low risk behaviours (e.g. ecological meat consumption, sun screen use) have found a gain frame advantage, but this relationship was moderated by high levels of self-efficacy and body consciousness respectively (Hevey et al., 2010; Werrij et al., 2011). Both variables could reasonably be considered important factors when considering donating or receiving an organ. In a high risk example, namely mammography screening for breast cancer, the loss frame advantage was found to be moderated by participants’ perceived susceptibility to the disease (Gallagher et al., 2011). Modifiable characteristics of those in receipt of the message have the potential to influence resulting intentions in relation to living organ donation and should be investigated further. Within the context of MMR vaccination, beliefs regarding the efficacy of the outcome were found to mediate a loss frame advantage (Abhyankar, et al., 2008). With respect to living organ donation, whether or not this surgical procedure is believed to help the recipient may also serve to explain any framing effects on intentions to donate and again would benefit from further investigation. People may be more willing to help where they know their donation will make a difference (cost to the recipient of not being helped is high), but more so when the cost to the potential donor is low. When the cost to the donor is high they may, under these circumstances, not help but redefine the situation as one that does not require help. When the cost of helping the recipient is low then donors will be unlikely to help (see Dovidio et al., 2006).

Living donation is often medically regarded as the best form of treatment for patients with organ failure and therefore its effective promotion is vital. It is important that every potential donor and recipient are fully aware of the risks and benefits of living donation and are not dissuaded from investigating the procedure simply because of how the information is presented.