Organs deficit solvable?
Implementing the opt-out system in Dutch postmortem organ policy.
By: Stephan Hogenboom
Student number: 10025685
Date: 2 February 2015
Thesis Supervisor: S. He
Bachelor thesis Economics Economie en bedrijfskunde University of Amsterdam
1. Introduction
Organ transplantation is an important aspect of the Dutch health care system. However there is a considerable gap between the supply and the demand of suitable organ donors. In the Netherlands approximately 150 people die each year, because of a deficit in suitable organs. Of the people in need of an organ transplant in 2013, the number of which is
approximately 2.400, half were still on a waiting list by the end of the year. With an expected increase of the life expectancy of the Dutch population, the demand for donated organs will rise in the future as well. Resulting in a need for more organ donors (Transplantatie Stichting, 2014).
Most transplant organs come from deceased donors (in contrast to live donors who most often donate a kidney to a relative). One deceased organ donor can save up to eight lives by donating one or more organs (Transplantatie Stichting, 2014). In order to meet the number of required organ donations one can try to improve transplantation infrastructure and techniques to generate more circumstances where organ donation is possible. Another way to get more organ donations is to increase the number of registered donors (Howard, 2007). In order to meet the increasing demand for donated organs, more people have to officially become organ donors. In various other countries the adoption of another default, the opt-out system has resulted in an increase of donors. In the opt-out system, one is assumed to be donor unless one has officially withdrawn from the registration (Johnson & Goldstein, 2004). However there are also countries where the opt-out system did not increase the number of organ donors.
The first research question of this paper examines to what extent people tend to stick with the default. According to Kahneman and Tversky (1981) people systematically make suboptimal choices in many aspects of their life. Their choices are easily affected by the way they are presented to them. In the presence of a default choice people have a tendency towards doing nothing. Mechanisms behind this bias towards doing nothing are procrastination, the status quo bias and the bandwagon effect. These mechanisms will be described in the literature review. With this in mind a policy maker should set the default in such a way that it is beneficial to individuals and the society as a whole (Thaler and Sunstein, 2008).
The outcome of this first research question will also be based on the response to a survey conducted through the Internet. The surveys goal is to estimate what percentage of the non-registrants in the Netherlands is willing to donate postmortem. The relative size of this group will be used as a benchmark in order to answer the research question.
The second research question will examine, presumed that the opt-out system will increase the number of registrants, to what extend this will increase the amount of actual organ donors. In the academic realm there is still a debate going on whether the presumed consent legislation will increase organ donors. Some authors found a significant positive correlation between the opt-out system and donation rates (Abadie & Gay, 2006 : Johnson & Goldstein, 2003). Some found ambiguous results (Rithalia et al., 2011). Others found no significant correlation (Healey, 2006 : Coppens et al, 2005). An important factor that should be taken into account is the impact of family consent. In most opt-out countries relatives end up having the final say in whether or not organs should be donated.
This thesis will start with reviewing the most important theoretic and empirical literature regarding the presumed consent system in section 2. Subsequently the
methodology, the gathered data and the hypotheses will be discussed in section 3. In section 4 the empirical results will be described and finally the paper will be discussed and concluded in section 5.
2. Literature review
2.1 Theory behind the opt-out system
In classical economic theory the people are described as rational. However findings in
behavioral studies have showed that people tend to systematically behave irrational in certain situations. The choices of people are easily influenced by the way options are presented to them (Kahneman, 2012). With this in mind a policy maker can nudge a population in a
direction of his liking, by making use of a certain choice architecture. This implies that he can present options in such a way, that people are nudged to pick the option he prefers. Nudging people in a direction you think is best for them or the society as a whole is a form of
Libertarian paternalism. The libertarian part means people aren't restricted in their set of options. The paternalistic part means that people are nudged to do things that generate more utility, something they probably would not do autonomously (Thaler & Sunstein, 2008). This paper will focus on the much-debated instrument for increasing organ donations, the default.
Basically two default systems exist regarding organ policy: The opt-in system (one has to register in order to become a donor) and the opt-out system (one is assumed to be a donor unless the consent is withdrawn from the register), the latter is also known as the presumed consent legislation (Johnson & Goldstein, 2004). If a government makes citizens choose between option A and B, but sets option A as the default choice, people tend to stick with A.
The reason why people tend to stick with the government’s default will be explained in the next paragraphs.
The tendency to do nothing when a default is present is called the status quo bias (Samuelson & Zeckhauser, 1988). Two drivers mainly cause the status quo bias. One of the drivers is switching costs. An individual is familiar with the implications that come with his current position. But when one wants to deviate from his current position, the other options first have to be evaluated. This requires spending time and effort. The actual action to change from position A to B can also require effort and money (Samuelson & Zeckhauser, 1988). Switching costs related to organ donation can be identified as contemplation or
registering costs. Making a well considered choice about whether to donate or not takes time and effort. These contemplation costs are often perceived as high and unpleasant as
individuals dislike to be confronted with their own mortality. Choosing to be passive and therefor sticking with the default is effortless (Abadie & Gay, 2006)
The other driver behind the status quo bias is the combination of loss-aversion and the endowment-effect. Loss-aversion implies that people tend to experience losses as more intense than gains. For example: Losing €100 has twice as much emotional impact than winning €100. Loss-aversion causes people to be afraid of changes as new situations may come with a potential loss. The potential gain of the new situation is outweighed by the potential loss due to this bias (Kahneman & Tversky, 1981).
The endowment-effect is a cognitive bias that causes individuals to overvalue objects they have in their possession (Thaler et al., 1991). To test this phenomenon Kahneman & Tversky (1991) did an experiment with groups of students. They randomly distributed coffee mugs amongst the students. Half of the students got a mug the other half did not get a mug. The average student that got a mug valued the mug almost twice as high as the average student that did not get a mug. Similar experiments have shown that individuals
overestimate the value of things in their possession (Kahneman, 2012). When a default is in effect, individuals tend to use this default as reference point. Under the opt-in system people perceive their organs as part of their endowment. Where in a presumed-consent they do not. Goldstein and Johnson (2003) argue that it easier for individuals to part with their organs postmortem, if the organs are not perceived to be part of their endowment.
Another reason why people tend to stick with a default is because they are prone to indecisiveness and procrastination. People are present-biased. This implies that they have a tendency to choose options that grant immediate gratification over options that are more beneficent in the long run. The more benefits the long-term option provides, the more likely
the person is to choose this option (O'Donoghue and Rabin,2000). The problem with organ donation is that someone else receives the future benefits. This makes the act of registering an altruistic deed that has no tangible benefits for the donor. According to Thaler and
Sunstein (2008) a lot of people, who are willing to donate, procrastinate with registering up until the point they will not register at all. This is because they are not motivated enough to take action.
Caroll et all (2009) found that people with an high tendency to procrastinate are more likely to vote in an election than to register as donor. One reason that could explain this difference is that elections come with a strict deadline. Procrastination in an election will result in losing ones chance to vote. Rooij & Teppa (2014) argue that procrastination problems can possibly be resolved by forcing individuals to make a choice. For instance by implementing a mandated choice system. This system presumes no default, but requires individuals to make a choice about organ donation at some point in their life. For example in an experiment in Illinois in 2008, people who wanted to renew their drivers license had to officially register whether they wanted to donate or not before their application was accepted (Thaler & Sunstein, 2008).
Finally, the default set by the government can be perceived as signaling. The default is then interpreted as the position of the government and the general opinion or cultural norm. The opt-in system makes donating look like something extraordinary, which goes against the norm. The opt-out system will send the opposite signal, which will make donating look like the norm. People are in general sensitive to social norms and the position of their social group. Therefor they tend to behave like the group does. This cognitive bias is called the
bandwagon effect. If donating becomes the norm, only individuals who are really reluctant to donate will opt-out (Johnson & Goldstein, 2004).
2.2 The presumed legislation in practice
Various countries have implemented the opt-out system. In some countries this resulted in a larger number of organ donors per capita, in other countries the effect was negligible. Various authors argue that even if it can be presumed that switching to an opt-out will result in more registrants, this will not automatically result in more organ donors per capita (Rithalia et al. 2011: Healey, 2006). This section will discuss why the empirical findings on the presumed consent legislation are ambiguous.
The theory behind “presumed consent” is that there's no longer a need to look for evidence to find out whether an individual supported organ donation. It's presumed he
supports donation, unless he registered to opt-out. In theory this measure should increase organ donors, as people have a tendency to stick with the default. An important prerequisite for this principle to work properly is that next-of-kin involvement should no longer be a factor in whether organ procurement takes place or not. The choice of the donor should be binding.
A presumed consent legislation where next-of-kin involvement no longer plays a role is called the strong or hard version (Healey, 2006). Only a few countries have adopted the strong version of the presumed legislation, of which Austria is the most cited. In Austria the number of donors per million inhabitants rose from 4.1 donors to 10.1 within 4 years after
implementation. The donation rate increased to 27.2 per million inhabitants after Austria also improved the organ donation infrastructure and installed full time organ donation coordinators (Gnant, 1991).
Most countries however, have adopted the weak version of the presumed legislation. In this version consent of the donor is also presumed, unless opted-out. The main difference is that next-of-kin involvement is allowed in this version. The consent of just the donor is not
enough. Before the doctor can start with organ procurement he must ask the family for their consent. Family involvement complicates the presumed consent system, as the actual decision whether to donate shifts from the donor to the relatives. A lot of organs from registered donors have been vetoed by next-of-kin in countries with a weak version of the opt-out system (Howard, 2007). According to Abadie and Gay (2006) increasing family consent has shown to be one of the most promising ways to increase organ donations. The influence of the presumed-consent legislation on family consent seems to be ambiguous. Abadie and Gay (2006) argue that the choice of the deceased is the main indicator of the family’s choice. It’s presumed that more people will become registered-donors if the opt-out system is in effect, due to people’s tendency to stick with the default. If a person is registered this can be perceived as a signal by the family. This signal increases the family's tendency to give their consent. However according to Howard (2007) families are not tricked that easily. In an informed consent country someone makes an active choice when one opts-in. Where in an opt-out country it is not clear if someone is a donor through choice or just being passive. Healey (2006) argues that families can be approached differently when there's an opt-out system. If someone is a non-donor by default the mental threshold of the next-of-kin to give their consent is high. When the deceased is donor by default the family is more likely to be persuaded (Healey, 2006). Thaler and Sunstein (2008) found data that support that the
refusal rate of families in opt-in countries seems higher than countries with an opt-out system. In the United States, an opt-in country, the refusal rate of the family of a donor who didn’t officially register, is roughly 50 percent. In Spain, an opt-out country, relatives only refuse 20% of the time in case the deceased did not opt-out. France, also an opt-out country, has a 30 percent refusal rate (Sunstein and Thaler, 2008).
The discussion about to what extent families should be involved in the decision whether to extract the organs of their deceased relatives or not is often avoided by politicians. The exclusion of the family is often seen as unethical and can easily turn a politician in a pariah (Bilgel, 2012). After observing the relatively high donation rates in Austria and the ambiguous results in other opt-out countries, Brazil tried to strictly enforce the presumed consent legislation. As a consequence family vetoes were ignored. Doctors who refused to extract organs without family consent got fined or even prosecuted. This led to public distrust in the health care politics of Brazil. As a consequence a very large percentage of the
population opted-out and the donation rates dropped considerably. Eventually Brazil had to abolish the presumed consent legislation to restore balance (Bilgel, 2012).
2.3 Methods for analyzing the effect of the opt-out system on the organ donation rates
In the existing literature three methods were used to measure the effect of implementing the op-out system on donation rates. The first method is conducting surveys to investigate how the population will respond when the opt-out system is implemented. Surveys are also used to estimate what part of the population is in favor of the presumed consent legislation. Another method is comparing a country's donation rate before and after the incorporation of the opt-out system. The last method is to compare countries with an opt-in system to
countries with an opt-out system and establish the difference between the donation rates. (Rithalia et al , 2009).
2.3.1. Conducted surveys
Throughout the years several surveys have been conducted to determine the extent of the population that wants to donate but hasn't registered because of the reasons discussed in section 2.1. The main goal of these surveys is to predict the increase of organ donation rates if the presumed consent legislation is implemented.
have a positive attitude towards organ donation. But less than half of the respondents made a decision regarding organ donation and only 28 % actually registered as organ donor. This survey established that there is a large gap between people that want to donate and people that actually sign the donation forms. Similar results have since been found in Spain, Sweden and Germany (Thaler & Stunstein, 2008 )
Johnson and Goldstein (2013) conducted a survey to predict the difference in donation rates between the opt-in system, the opt-out system and the active choice system. The active choice system implies that there is no default. Instead people are forced to make a choice regarding organ donation at some point (In America there was an experiment were people had to fill in whether they want to donate their organs before they could get a driver’s license). The respondents of the survey were independently split in 3 groups. Each group was assigned 1 of the 3 different organ policies. The outcome was that in the opt-in policy group 42 % stated to want to donate their organs. In the opt-out 82 % of the respondents stated to want to donate and in the active choice group 79 % agreed. This survey established that the opt-in system has a very negative influence on donation rates and that both an active choice system and an opt-out system can significantly increase donor registrations (Johnson & Goldstein, 2003).
2.3.2 Donation rates before and after the implementation of the opt-out system
The vast majority of the conducted surveys predict that the adoption of an opt-out system results in a large increase in organ donors. The other two methods however show ambiguous results (Rhitalia, 2009). In this section studies on donation rates before and after the adoption of the opt-out system will be discussed.
One of the biggest increases after the implementation of the opt-out system was found in Belgium. Belgium implemented the presumed consent legislation in 1986. From 1986 to 1992 the number of people who died in circumstances that were suitable for organ donations halved. But in the same time span the number organ donations doubled because of a large increase in registered donors (Michielsen, 1996). The kidney donation rate went from 18.9 per million population to 41.3 per million within 3 years after implementation (Rithalia et al, 2009).
And in Sydney the kidney donations increased from 4.7 to 31.3 per million inhabitants within 3 years after the implementation. The priority allocation rule was however implemented around the same time and is likely to be responsible for a big percentage of this increase
(Soh, 1992; Rithalia et al , 2009)
Healy (2005) found that the countries that experience the biggest growth in donation rates after the adoption of the opt-out system have also implemented other measures around the same time that are likely responsible for a portion of the growth.
2.3.3 Comparing donation rates between countries with different consent legislations
Abadie and Gay (2006) observed the donation rates of 22 countries in a period of 10 years. They compared the mean of the donation rate in countries with an opt-in system with the mean of countries that have an opt-out system. Donation rates were at average 25-30 % higher in countries with an opt-out system after controlling for other variables. In a similar study regarding 11 European countries Johnson and Goldstein (2003) found a 16.3 % higher donation rate in countries with an opt-out system.
Although these numbers seem to be very promising it is hard to distinguish what part of the difference is due to the presumed consent legislation and what part is due to other factors. Michielsen (1996) states that it's hard to generalize the effect of the presumed consent legislation when comparing between countries. When countries are compared to establish the effect of the opt-out system, variables like religion and historical and cultural background should be taken in account. Rithalia et al (2009) state that when comparing donation rates amongst countries, the presumed consent legislation cannot be used as a binary variable, as the presumed consent has been implemented and is enforced very differently in various countries. Furthermore the vast majority of the countries that adopted the opt-out system also take other measures to increase the postmortem donation rate. The adoption of the presumed consent legislation can be seen as one of the many commitments the countries have made to increase their donations rates (Healy, 2005). Healy’s statement is supported by Boyarksy et al. (2012). Studies that try to identify the causal link between presumed consent and increased donation rates, tend to oversimplify the factors involved and omit variables that also can be responsible for higher rates in the presumed consent countries.
Both Healy (2005) and Coppens et al. (2005) found a strong correlation between relevant mortality and organ donation rates. Relevant mortalities are mostly in-hospital deaths after traffic accidents or cerebral vascular diseases. Both studies compared opt-in countries with opt-out countries and found that the donation rates are on average higher in opt-out
controlled for relevant mortality rates.
Abadie & Gay (2006) argue that the outcome of comparing donation rates from opt-out countries to those of opt-in countries suffers from reverse causality. In countries where a large percentage of the population is in favor of organ donation it is more likely that the opt-out system is implemented. The higher donation rate in this particular opt-opt-out country is not solely due to the presumed consent legislation but also partly due to the preference of its population. When this method is used to estimate the effect of the opt-out system on a
sample that includes such countries, this will probably cause a positive bias in the correlation between the opt-out system and the organ donation rates.
3. Methodology
In this section the methodology is discussed. To gather more data regarding the willingness of the Dutch population to become an organ donor a survey has been conducted. In the first part of this section the structure of this survey will be discussed and the corresponding hypotheses will be formulated.
3.1 The survey and methodology
In the literature section, three methods have been discussed to predict the effect of the opt-out system on the donation rates. This paper uses the survey method to gather new
information. The main goal of the survey is to predict what part of the population that has not registered, is willing to donate. Willing non-registrants will be referred to as the group of interest (GOI) in this paper. In this paper the sum of registrants and non-registrant that are willing to donate will be referred to as, the total willingness to donate (TWD).
As discussed in section 2.1 there are various reasons why people who are willing to donate, fail to register as donors. By establishing what percentage of the population belongs to this group, the potential effect of the opt-out system on the organ donation rates can be predicted. As discussed in section 2.3.1, various previous studies have estimated that the GOI forms a considerable part of the population in other countries (Abadie & Gay, 2006: Rithalia et al, 2009: Johnson & Goldstein, 2004). The first hypothesis of this paper
corresponds with this assumption. The second hypothesis is depending on the outcome of the first hypothesis.
H1A: A large part of the Dutch population is willing to donate, but has not registered as donor H1B: Introducing the strong version of the opt-out system in the Netherlands will lead to
higher donation rates.
In this paper the effect of both versions of the presumed legislation, as discussed in section 2.2, will be examined. In the weaker version of the presumed consent, next-of-kin can effectively veto organ procurement of legal donors (Healey, 2006). The weaker version of the opt-out system is the most common one. Therefore the second part of the survey is to find out if relatives are likely to give their consent for extraction. Participants were asked if they would give their consent to extract organs from a deceased relative. And also if they think it's likely that their next-of-kin in turn will give their consent to extract the organs of the respondent, should he or she pass away. This part of the survey is to predict how many organ donations would potentially be lost if the weak version is adopted, due to a lack of family consent.
H1C: Next-of-kin vetoes will lead to significant lower donation rates under the opt-out system in the Netherlands
H1D: Introducing the weak version of the opt-out system in the Netherlands will lead to higher donation rates
The third part of the survey is to establish if the willingness to register among the non-registered participants will increase if the priority allocation rule were to be implemented. Subjects were asked if they would register if that would give them priority over those who did not, when in need of an organ donor.
H1E: Introducing the priority allocation rule in the Netherlands will increase donation rates.
The fourth part of the survey is to establish what part of the sample has faith in the Dutch government. Changing the default is considered a libertarian paternalistic tool. If the people don't trust their government, it is likely such a tool will have less effect. Johnson and Goldstein (2003) claim that a government can influence peoples choices by signaling what she thinks is best for them.
H1F: Donation rates and trust in the government in the Netherlands are correlated
In order to create control variables respondents were asked information about their age, gender, education and religion. These questions were asked to test whether the sample is externally valid. And whether there’s a correlation between this variables and the
willingness to donate.
3.2 Data collection
The data was collected from a survey made for this paper. The survey was constructed using a free survey tool found on the website http://www.thesistools.com. Afterwards it was spread through the Internet. The collection period was from 20th November until 1st December 2014. 144 anonymous respondents filled in the survey. Because of an error two versions were generated. One version did not contain two questions. This resulted in only 48 observations regarding the priority allocation rule.
3.3 Data description
In this section the sample is described and compared with the average Dutch population to check if the sample is externally valid. The sample was also compared to a similar Dutch survey on this subject with a sample size of 18.783 (Mossevelde , 2014).
The most notable result from the survey is that 44.4 % of the respondents are registered organ donors. In the survey conducted by Mossevelde (2014) 49,68% was registered and in the survey conducted by van Rooij and Teppa 50% was registered. All three samples have relatively more registered donors than the Dutch population. The
registration rate of the Dutch population above the 18 years old is approximately 26 % (CBS, 2014; donor register, 2014). An explanation for the big difference between the registration rate of the surveys and the Dutch population is presumably self-selection bias. The survey was posted on the Internet and because of this there was no control over the sample selection. Respondents on the Internet independently decide whether to participate or not. The people that are interested in the topic of the survey are more likely to respond
(Bethlehem, 2010: Armstrong & Overton, 1977).
Another reason for the higher donation rate may be due to the education level of the sample. Gimbel (2003) found a significant correlation between higher education and donation rates. The sample of the survey, compared with the Dutch population in the table below, shows an above average educational level. The sample may therefore give a positively biased representation of the Dutch population's willingness to donate.
respondents are aged 18 – 25, 26 respondents are aged 26 – 45 and 49 respondents are above 45. The male: female distribution is 39.6 : 60.4%. This may have some influence on the external validity of the sample. The sample has relatively more non-religious people, less Christians and Muslims than the Dutch population. According to previous studies in other countries Christianity and especially Catholicism seem to have a positive influence on donation rates. The religions Judaism and Islam seem to have a slightly negative effect on organ donation rates (Rithalia et al, 2009). An article from the Central Bureau of Statistics (2012) showed that Dutch non-religious people are more likely to donate than religious people. 69 Percent of the non-religious population stated to be willing to donate.
Approximately 50 percent of the Christians and 27 percent of the Muslims are willing to
donate. The effect of Christianity in the Netherlands seems to have the opposite of the effects found in previous studies in other countries.
# of ob. Percentage
Total Respondents
144
Non-Donors
80 56%
Not-registered but willing
to donate (GOI) 46 31.94%
Total willingness to donate 110 76.39%
Females registered 47 54.02%
Males registered 17 29,8 %
Total willingness females 70 80.46%
Total willingness males 40 70.18%
Non-donors affected by
priority allocation rule 19 46,3%
Family consent No family consent
Donor 61 3
Family consent No family consent
GOI 38 8
No trust in government Trust in government
Non-Donor 37 42
Donor 12 52
No trust in government Trust in Government Not register, belongs to
GOI 18 16
Not registered, does not
belong to GOI 19 26
46 of the 80 registrants stated to be willing to donate, which is 57,5%. The group of non-registrants that are willing to donate represents 31.94 % of the sample. 48 respondents returned the question about the priority allocation rule. 19 of the 41 non-donors stated that they would register if the priority allocation were incorporated. In the sample women showed
a higher tendency to register than men (54.02% and 29,8). Women also displayed a larger willingness to donate than man 80.46% vs. 70.18%. People who stated to trust the
government are more likely to have registered than people who don't.
4. Results.
In this section the data gathered with the survey is analyzed in 4 different parts. In the first part the effect of the opt-out system is estimated if it would be implemented in the
Netherlands as the strong version. In the second part the effect of the opt-out system in the Netherlands is estimated if incorporated as the weak version. In the third part the potential effects of introducing the priority allocation rule in the Netherlands are discussed. Finally the correlation between the Dutch people’s willingness to donate and their faith in the
government is estimated.
4.1 Post-stratification
The sample contained a relatively large number of registered donors (44%) compared to the Dutch population (26%). This difference gives reason to assume self-selection bias.
Weighting methods are used to get a more representative sample of the population in case of self-selection bias. One of the most common methods used is post-stratification.
Post-stratification compares certain variables from the sample to well-known valued variables in the population. These variables are called auxiliary variables (Bethlehem, 2010).
Characteristics of respondents that are more present in the population than in the sample are assigned a certain extra weight in order to make the sample more balanced. In this case people who are not registered donors are used as auxiliary variable. Approximately 74% of the Dutch population is non-registrant (CBS, 2014: Donor register, 2014). The weighting factor is determined by the following formula: wk = (Nh / N) / (nh / n). Where Nh is the frequency of the variable in the population. N is the total population. Nh is the frequency of the variable in the sample and n is the total sample size. The weighting adjustment for non-donors is 0.74 / 0.556 = 1.332. The weighting adjustment for non-donors is 0.26 / 0.44 = 0.5909.
4.2 Multiplicative weighting
Another method to adjust non-response or self-selection bias is multiplicative weighting. This adjusting method is used when the cross sectional data of two variables are unknown at population level. Values that appear more frequently in the population than in the sample are
assigned more weight and vice versa (Bethlehem, 2008). As males and non-registrants are less represented in the sample than in the population, they are assigned an extra weight in comparison with females and registrants. At the start every cross variable is assigned a weight factor of 1.
Female Male Weight factor Sum product Population
Non-donor 0.278 0.278 1 0.556 0.74
Donor 0.326 0.118 1 0.444 0.26
Weight factor 1 1
Sum product 0.604 0.396 1
Population 0.5 0.5 1
The aim is that the sum product becomes approximately equal to the population estimate. The sum product for non-donors is for example 1 x 1 x 0.278 + 1 x 1 x 0.278 = 0.556 at the start. The weights are adjusted until the sum-product of non-donors looks approximately like this: w1(female)*w2(non-donor)* 0.278 + w3(male)*w4(non-donor) = population = 0.74.
Female Male Weight factor Sum product Population
Non-donor 0.278 0.278 1.3 0.74087 0.74
Donor 0.326 0.118 0.6 0.25746 0.26
Weight factor 0.9 1.15
Sum product 0.501 0.497 1
Population 0.5 0.5 1
The product of the weighting factors will become:
Female Male Weight factor
Non-donor 1.17 1.495 1.3
Donor 0.54 0.69 0.6
Weight factor 0.9 1.15
The weights given to non-registrant females =1.17 & males = 1.495. The weights given to registrant females = 0.54 & males = 0.69.
4.3 Adjustment by assuming direct relation between willingness to donate and participating in this survey
The survey only established three levels of willingness to donate .The highest level are the registered donors. People who are not registered, but are willing to, are considered a lower level. They are willing to donate, but apparently not willing enough to actually register. Lastly there are people who did not register and are not willing to do so. A crude assumption to establish a lower bound for the GOI is to assume willingness to donate and response rates are directly related. People in the 1st and 2nd group would more likely respond than people in group 3. The population level of group 1 is approximately 26%. People within this group presumably are .44/ .26 = 1.69 times more likely to participate in this survey than group 3. People in group two are therefore between 1.69 - 1 times more likely to respond than people in group 3. Presumed that people from both group 1 and 2 are 1.69 times as likely to respond to this survey as people from group 3, the lower bound of the group of interest (willing non-registrants) can be estimated to be around 31.94% / 1.69 = 18.89%.
By using three different weighting methods, the percentage of non-registrants in the
population who are willing to donate in is estimated to be between 18.89% and 42.56%. The total willingness to donate is estimated to be between 44.89% and 68.56%.
4.4 Testing statistics
In this section some of the variables are tested with the independent t-test and other variables are tested with the Mann-Whitney U-test.
Firstly was tested if the GOI was large enough to assume a difference in registered donors if the strong version were to be implemented. The independent t-test showed that there was enough evidence to assume a difference in the number of registrants depending on which legislation (n=144, df=142, t=8.19, p = 0,00). Secondly was tested if the difference was still significant when controlled for family consent. The t-test concluded there was enough evidence to assume this is the case. (n=144, df = 142, t = 7.19, p = 0,00)
Only 24 percent of the respondents that don’t trust the government registered, against 55,4 % of the respondents who do trust the government. In order to test the difference
between these two groups the Mann-Whitney U-test was used. With the test-statistics being (n=144, z= -3.611, p =0,00) enough evidence was found to assume that the means of the
groups are not equal. The Mann-Whitney U-test also showed that there's enough evidence to assume the total willingness between the two groups is not equal (n=144, z=-2.488,
p=0,0128) .
39,58 percent of the respondents was male of which 29,58 % were registered. 54 % of the females were registered donors. The Mann-Whitney U-test was used to test whether males are as likely to register as females. The test concluded that there is enough evidence to assume that the means are not equal, with a confidence interval of 5% (n=144, z= -2.484. p = 0,044). The Mann-Whitney U-test concluded that there’s insufficient evidence (n=144, z=-1.416, p=0,1567) to assume that there’s a difference in willingness to donate between males and females with a confidence interval of 5 and 10%.
5. Discussion and conclusion
This paper focuses on two research questions: Will the adoption of the opt-out system in the Netherlands increase the number of organ donation registrants? And if presumed that the opt-out system will increase the number of registrant, will this result in more people actually donating their organs?
Contrary to classical economic theory, people are prone to cognitive biases that cause people to behave irrational on many occasions. One of these biases is that people have a tendency towards doing nothing when a default is present.
The literature showed various examples of the mechanisms behind this bias. First, people are procrastinators and would rather do things that grant instant gratification instead of fulfilling their moral obligations. The second mechanism is the status quo bias. This implies that when people are confronted with multiple options they tend to stick with the position that they currently hold. The two drivers behind this bias are switching costs and the combination of loss-aversion and the endowment-effect. Their current position is perceived as familiar, in contrary to the new position, which is perceived as risky. Familiarizing oneself with new options can cost a certain amount of effort and time. This is not the case if one passively sticks with the current position. The higher these costs are the less likely one is to switch to another option. Loss-aversion implies that people have a tendency to value an object higher if it is in their possession than if it is not. When they assess their options they tend to weigh the potential losses heavier than the potential gains, this makes them reluctant to change their position in general. Finally people are present-biased. This implies that they tend to rather do things that grant instant gratification than actions that grant more benefits in the
long run. The default is often perceived as the position of the government. People who are sensitive to cultural norms are more prone to accepting the default.
During the research process of this thesis the following obstacle was encountered: There appears to be a lack of literature available that criticizes the theory behind the status quo bias and procrastinating. It is necessary to collect more data on these topics in order to get a more complete understanding of the effects of a default.
The results from the survey indicated that between 18.89% and 42.56% of the Dutch population are non-registrants but willing to donate. This big interval is due to the self-selection bias of the sample. Both the surveys of Rooij & Teppa (2014) and Mosselvede (2014) also suffered from a disproportional large amount of registrants: 50% and 49%, where the Dutch percentage is only 26%. It can be assumed that surveys that research willingness to donate show biased results in general. Nonetheless even if the population of willing non-registrants were to be in the lower bound of the interval, the opt-out system would increase the number of registrants considerably. Other result from the survey showed that people who have trust in the government are more likely to register.
Based on the discussed literature and the results of the survey it is safe to say that a large part of the non-registrants in the Netherlands is willing to donate. Due to the
discussed biases, a large number of those willing to donate end up not registering. According to Rooij and Teppa (2014) procrastination is the biggest factor among these biases.
Changing the legislation will most likely increase the number of registrants. Implementing the mandated choice or posing some sort of a deadline on registering can be considered as an alternative solution to this problem.
If it is presumed that the opt-out system would increase the amount of registrants, it is not sure if this will result in higher donation rates. Healey (2006) distinguished two versions of the presumed consent legislation. In one the next-of-kin gets the final say whether organs can be procured and in the other version next-of-kin are by passed. The first version
complicates the system as the decision remains in the hands of the relatives. Most countries have adopted this version of the presumed consent. Therefor it's likely the Netherlands will adopt this version, should they decide to change the default system. At the moment the Netherlands has a strong version of the informed-consent legislation. This implies that if one has registered, next-of-kin involvement is a small factor when it comes to the final decision. The second version is more straightforward, as organs of every deceased person who has not opted-out can be procured. If the presumed consent legislation were to be implemented as the weak version this will cause a shift from individual decisions towards the next-of-kin.
The empirical results found on the presumed consent legislation are ambiguous. Rithalia et al (2011), Healey (2006) and Coppens et al. (2005) found no significant correlation between the opt-out system and higher donation rates. Abadie & Gay (2006) and Johnson & Goldstein (2003) found a significant correlation between the opt-out system and higher
donation rates. They found that the opt-out system would on average increase donation rates between 16 to 30 percent.
Whether more registrants will results in higher donation rates, will depend on to what degree the government will enforce the presumed consent legislation. If the opt-out system would produce more registrants, the strong version will most likely increase donation rates, as relative’s consent is overruled. But this strong version will probably cause public unrest, as overruling relative’s right to decide in family matters is considered controversial. Austria on the other hand, is an example where this version resulted in significantly higher donation rates. In Brazil the strong version resulted in national distrust in the government and the healthcare system. As a result the system had to be abolished.
This thesis added newfound data to identify the percentage of the non-registrants in the Netherlands that are willing to donate organs. Through an online survey an estimate was made on how large the group of potential additional organ donors would be. This number is important as it can be used as a benchmark to find out if a change in the organ legislation can contribute to solving the organ deficit in the Netherlands in the future.
References
Armstrong J.S., T.S. Overton (1977 ). Estimating Nonresponse Bias in Mail Surveys. Journal of Marketing Research, 14, pp. 396-402.
Abadie, A. , S. Gay(2006). The impact of presumed consent legislation on cadaveric organ donation: A cross-country study. Journal of Health Economics, (25)4, pp.599-620
Bethlehem. J. (2008) Wegen als correctie voor non-response, (http://www.cbs.nl/NR/rdonlyres/3BFBFEC4-FA4B-4A9D- 8034-DD553F813BC1/0/200805x37pub.pdf), consulted 24 January 2015
Bethlehem J. (2010). Selection Bias in Web Surveys. International statistical review. 78 (2) pp 161-188
Becker. G.S., J.J. Elias. (2007) .Introducing incentives in the market for live and cadaveric organ donations. Journal of Economic Perspectives, Pp. 3-24
Bilgel F. (2010) The impact of presumed consent laws and institutions on deceased organ donation. European Journal of Health Economics (13) pp. 29-38
Byrne. M.M., P. Thompson. (2001). A positive analysis of financial incentives for cadaveric organ donation. Journal of Health Economics, pp. 69–83
CBS (2015) (http://statline.cbs.nl/) Consulted 20 december 2014
Coppen R., D. Friele, Richard L,. Marquet and S. K. M. Gevers. (2005). Opting-out systems: no guarantee for higher donation rates. Transplant International pp. 112-120
Danynag L, Z. H. (2013). Increasing Organ Donation via changes in the default choice or allocation rule. Journal of Health Economics (32), 1117-1129.
Donor register (2014), (https://www.donorregister.nl/organisatie/cijfers/registraties/) . consulted 24 January 2015 Howard D.H., (2007) Producing Organ Donors. Journal of Economic Perspectives 21(3) pp 25-36
Kahneman. D. (2012) Thinking, Fast and Slow. Penguin Group
Kahneman. D, A. Tversky. (1979).Prospect theory: An analysis of decision under risk. Econometrica 47 pp. 263–291 Kahneman. D,A. Tversky.(1981).The Framing of Decisions and the Psychology of Choice. Science 211(4481) pp.453-458 Kahneman. D., J. L. Knetsch, R. Thaler. (1991).Anomalies: The Endowment Effect, Loss Aversion, and Status Quo Bias.
The Journal of Economic Perspectives, 5(1), pp. 193-206
Healy. K.(2005). The Political Economy of Presumed Consent. DePaul Law Review, 55:1017–1043. Johnson, E, J.; Goldstein, D. (2003). Do Defaults save lives? Science, (302)5649, p1338-1339.
Michielsen P.(1996) Presumed consent to organ donation: 10 years' experience in Belgium. Journal of the royal society of medicine. (89) pp. 663-666
Ministrie van ocw (2014) (http://www.trendsinbeeld.minocw.nl/grafieken/3_1_2_31.php). consulted 20 december 2014 Mossevelde van, E. (2014). Onderzoek Orgaandonatie. (http://opiniepanel.eenvandaag.nl) consulted 3 January 2015
O'Donoghue, T, M. Rabin. (2000) Choice and Procrastination. Quarterly Journal of Economics. 116(1) pp. 121-160 Rithalia A., c. McDaid, S. Suekarran, L. Myers, A. Snowden. (2009). Impact of presumed consent for organ donation on
donatian rates: a systematic review. British Medical Journal pp.
Rooij van. M., F. Teppa. (2014). Personal traits and individual choices: Taking action in economic and non-economic decisions. Journal of Economic Behavior & Organization. 100, pp. 33–43
Schwindt R, A. Vining (1998) Proposal for a mutual insurance pool for transplant organs. Journal of Healthcare politics, policy and law. pp 725-741
Thaler R.H., C.R. Sunstein (2003a). Libertarian Paternalism. The American Economic Review, (93)2, pp. 175-179 Thaler R.H., C.R. Sunstein (2003b). Libertarian Paternalism is not an oxymoron . Chicago university law review 70(4) pp. 1159-1202
