International Choice Modelling Conference, International Choice Modelling Conference 2017

Font Size: 
Preferences for donating a family member’s organs for transplantation: a discrete choice experiment
K Howard, S Jan, Germaine Wong

Last modified: 28 March 2017


Background and context

Whilst solid organ transplantation is the treatment of choice for those with organ failure, the demand for organs for transplant far exceeds supply.  Although there is widespread public support for organ donation this is not reflected in actual donor intent registrations; for example, only around 23% of Australians have registered a donation intention, despite over 75% of the population having positive views towards donation.  There is considerable evidence on the individual respondent factors associated with being an organ donor. People with higher education, higher socio-economic status and younger age tend to be more willing to donate.   In Australia, as in many western countries, particularly those with an opt-in consent system, family members have the final say as to whether potential donors’ organs are actually donated.  Qualitative research suggests that if family knows the deceased intentions, then they often abide by that decision.  However, there is considerably less research on the influence of, and preferences for, other ‘systemic’ factors, for example those factors that could act as policy levers to encourage donation. Qualitative research suggests personal, cultural and policy/systemic factors influence individuals’ decisions on whether to donate a family member’s organs however there is a paucity of quantitative data on the relative importance of various systemic factors on  preferences, or on the influence of decision maker personal characteristics on the importance of these systemic/policy factors in making decisions about donating a family member’s organs.


This study sought to quantify preferences for donating a family member’s organs and to identify whether particular groups of respondents viewed specific attributes as more, or less, important in this decision.



A discrete choice study was conducted using an online panel of Australian community respondents aged 18 and over. Following a pilot study in 127 participants, the final design consisted of 150 scenarios, blocked into 5 versions of 30 questions; the design had a d-error of 0.001 and an s-estimate of 192. Respondents were presented with the scenarios asking them to indicate for each one, whether they would agree to donate a family member’s organs under the described circumstances. Trade-offs between 8 attributes were quantified using mixed logit and latent class models: whether family member’s preferences were known; who discusses organ donation with families; how death is defined; whether families can be told anything about recipients after donation, the value of any payment to donor’s family; the value of any reimbursement of funeral expenses; whether a donor’s family receives priority for organs in the future; and whether a donor’s family receives any formal recognition of the donation.  All analyses were conducted in NLOGIT 5.0.



There were 2002 respondents with a mean age of 44.7 years (range18-87); 51.3% were female.  The mixed logit model was significant in explaining choice (LL=-13636, χ2 = 40725, p<0.000001, 52df; pseudo r2 = 0.599).  The model constant indicated a preference towards not donating a family member’s organs.  This underlying preference however, could be mitigated by other factors.  Compared to knowing that the deceased did NOT want to be a donor, knowing they did increased the likelihood of agreeing to donation significantly (OR=12.93, 95% CI: 10.77-15.52); being unsure about donation preferences of the deceased also increased the likelihood of agreeing to donation (OR 1.25, 95%CI 1.09-1.43).  Respondents were more likely to agree to donation when: death was defined through cardiac death (compared to not being told); when there was either a direct payment to family of $5000 (OR 1.18; 95%CI:1.07-1.31) or $20000  (OR 1.39 95%CI: 1.24-1.57), a funeral expense reimbursement of $5000 (OR 1.12 95%CI 1.02-1.23), or $20000 (OR 1.45 95%CI: 1.30-1.62).  Direct payments and funeral expense reimbursements of $100 and $1000 led to a significantly lower likelihood of agreeing to donation.  Family priority for receiving organs in the future increased the likelihood of agreeing (OR 1.10 95%CI: 1.01-1.18).  Who discussed donation with the family did not significantly influence likelihood of agreeing, neither did being told anything about the recipient after donation had occurred.  The following respondent sociodemographic factors all increased the likelihood of agreeing to donation:  increasing age, being a registered donor, having positive views towards organ donation; having discussed donation intentions with their family; being employed full time, having a household income of higher than $125,000 and speaking mainly English at home.  Interestingly, unlike previous qualitative research, respondents with at least university level education we less likely to agree to donate a family member’s organs


However, the mixed logit results also indicated significant heterogeneity across respondents, particularly around the constant.  Class 1 had a high underlying preference for agreeing to donate (βconstant =4.43) and had a preference for funeral expense reimbursement of $5000 but not direct payment, and compared to the base class 4, were more likely to be registered donors, have less than university education, male and spoke mainly another language at home; class 2 had a smaller, but still significant preference for agreeing to donate (βconstant =0.45), valued both direct payment and funeral expense reimbursement > $5000  and were more likely to be younger, not a registered donor, have ambivalent or negative views on organ donation and male; class 3 hd a high underlying preference for not agreeing to donate (βconstant =-2.04), favoured $20000 funeral expense reimbursement and were more likely to not be registered donors, hold negative donation views and have household incomes less than $125,000.



Our results suggest that the most important factor in deciding whether to donate a family member’s organs was knowing that they wanted to be a donor.  Other systemic/ policy factors also contributed to increasing the likelihood of agreeing to donation; sociodemographic and attitudinal characteristics of respondents also played a significant role in whether a respondent would agree to donate their family member’s organs



Conference registration is required in order to view papers.