G Model

ARTICLE IN PRESS

HEAP-3368; No. of Pages 10

Health Policy xxx (2015) xxx–xxx

Contents lists available at ScienceDirect

Health Policy journal homepage: www.elsevier.com/locate/healthpol

Decision solution, data manipulation and trust: The (un-)willingness to donate organs in Germany in critical times Lars Schwettmann ∗ Martin-Luther-University Halle-Wittenberg, Department of Economics, Universitätsring 3, D-06108 Halle (Saale), Germany

a r t i c l e

i n f o

Article history: Received 6 October 2014 Received in revised form 19 January 2015 Accepted 28 January 2015 Keywords: Organ donation Trust General public Decision solution Transplantation

a b s t r a c t In 2011 and 2012 a change of rules and a data-manipulation scandal focused German public attention on organ donation. This increased citizens’ background knowledge as well as their willingness to respond to surveys. The present study is an effort to seize this research opportunity and to create evidence on which policy recommendations can be conceivably based. It uses data from two major representative surveys from 2011 to 2012 to address four central questions: Which characteristics, experiences and attitudes correlate with the written or unwritten willingness of individuals to donate (WTD) their own organs postmortem? How has the WTD changed over time? To what extent does the WTD depend on normative trust? Which factors correlate with trust? The data is analyzed through summary statistics and regression models. Several hypotheses regarding factors connected with the WTD are confirmed in the survey results. Altruistic motives, relevant knowledge and trust are decisive. The special role of trust is corroborated by the data. As current German politics prevents the introduction of post-mortem donation incentives, potential policy making proposals are restricted to institutional changes to regain trust including the implementation of an organ donor registry and the advancement of counselling talks with general practitioners. © 2015 Elsevier Ireland Ltd. All rights reserved.

1. Introduction The gap between supply and demand of human organs for transplantation has dramatic effects worldwide, thousands of deaths, human suffering while waiting for a suitable organ, and huge additional public healthcare expenditures [1,2]. Increasing the number of organ donations is paramount. Currently in most countries policy measures must comply with public opinion’s requirement that post-mortem donations have to be purely unselfish [2,3]. More radical reforms would include some form of

∗ Tel.: +49 0345 55 23443. E-mail address: [email protected]

‘do ut des’ based on non-financial or financial incentives [1,4]. A prominent example is the reciprocity rule recently introduced in Israel where those who are willing to donate have priority in case of their own need [5–9]. This rule seems to exert positive effects on organ donation rates already. Yet ethical concerns among laypeople are rendering the implementation of principles of mutual insurance and solidarity [10] into the organ donation and allocation process politically unfeasible in most countries. In Germany as in many other countries the separation of allocation from donation decisions is so deeply entrenched in law and public opinion that it will prevail in the foreseeable future. However, there is some room for political reform, which especially arises when organ donation moves into the focus of public opinion. This tends

http://dx.doi.org/10.1016/j.healthpol.2015.01.017 0168-8510/© 2015 Elsevier Ireland Ltd. All rights reserved.

Please cite this article in press as: Schwettmann L. Decision solution, data manipulation and trust: The (un-)willingness to donate organs in Germany in critical times. Health Policy (2015), http://dx.doi.org/10.1016/j.healthpol.2015.01.017

G Model HEAP-3368; No. of Pages 10

2

ARTICLE IN PRESS L. Schwettmann / Health Policy xxx (2015) xxx–xxx

to affect the willingness to donate (WTD) and the willingness to support political reforms of organ donation rules. Due to positive symbolic events such as the German case of a prominent politician serving as a living donor for his spouse, such a window of political opportunity opened in the years 2010–2012 [11]. However, a major scandal involving German physicians manipulating priority on waiting lists damaged potentially positive effects and destroyed public trust in the entire organ donation system. These processes are of interest not only for the German case but may have tentative implications for other countries that also cannot (yet) introduce more fundamental reforms. In Section 2 the German events are sketched. Section 3 describes the data and statistical methods. Section 4 presents results while Section 5 discusses the findings and concludes.

2. Recent German developments concerning organ donation 2.1. The implementation of the decision solution Due to the European Union directive of July 2010 aimed at assuring quality and safety of organ transplantation, the necessity of amending regulations of the German Transplant Act arose [12]. Also, in August 2010 the leader of the main opposition party in the German parliament, FrankWalter Steinmeier, donated a kidney to his sick wife. Hence, the shortage of donated organs received enormous media attention and politicians seemed to be prepared to rethink the present legislation [11,13]. The main public discussion revolved around the question of whether the German ‘extended consent regime’ (opting in) should be replaced by presumed consent for deceased organ donation (opting out). According to German law the declared will of the deceased potential donor is binding. If the donor has not documented it in writing, in particular by completing an organ donor card (ODC), his close relations should be asked concerning his will. If his intentions concerning post-mortem donation are unknown, next of kin could (voluntarily) opt-in to donation [4,12,14]. The system did not work well. For instance, under the old scheme 60–80% of survey respondents declared their WTD organs post-mortem, while only between 18% and 25% stated they had an ODC [15,16]. Assuming that the WTD was truthful in most cases this seems to be a procrastination case, which made asking for the donor intention of the deceased and the permission of next of kin to ex-plant unnecessarily awkward. Since the ‘nudge’ of introducing the presumed consent [17] was prevented by German public opinion, other changes of rules had to be considered. Legal policy converged on what the Germans call the ‘decision solution’ as an acceptable compromise between those who raised concerns against presumed consent as it may reduce levels of altruism, personal freedom and social solidarity, and others who highlighted the poor performance of extended consent in Germany [11,18]. The new solution became law on November 1, 2012, and supplemented the opt-in solution that remains valid in principle.

According to the decision solution adult citizens are regularly informed mainly by their health insurance company about organ donation, and asked to decide for or against becoming a potential organ donor and to document it in an ODC. The option of not making a choice still exists. In this vein, in November 2012, the Techniker Krankenkasse was the first large health insurance company to inform its members. Two aims of implementing the decision solution can be identified. First, the number of documented declarations of donor intent should be raised. Second, the number of organ donors should be increased. First surveys show indeed that the proportion of Germans possessing an ODC has increased e.g. from 22% in 2012 to 28% in 2013 [19]. 2.2. The data-manipulation scandal Parallel to the legislative initiative a second major event struck the German organ donation system. In July 2011 an anonymous phone-call to Germany’s organ procurement organization (DSO) drew attention to breaches of rules at the university clinic of Göttingen [20]. In the aftermath, investigations successively revealed that over the last 10 years doctors at five German hospitals had manipulated medical tests and falsified medical records to move their patients up the waiting list for livers [20–23]. Recently, further cases of queue-jumping in heart transplants between 2010 and 2012 have been discovered [24,25]. Starting in July 2012 the problems became widely known. A first peak of media interest was reached in September 2012, when it became apparent that data had been manipulated not only by one surgeon at two transplant centres, but also by other doctors at another hospital. Further reports of other transgressions and investigations by public prosecutors followed [22]. Support for organ donation has decreased since the second half-year 2012 [26]. The post-mortem organ donation rate of 14.7 per million inhabitants dropped to 12.8 in 2012 and 10.9 in 2013, making Germany the country with the lowest rate in ‘Eurotransplant’, the supranational association of transplant centres to which Germany belongs [26,27]. The data manipulations concerned the allocation of already donated organs, but the problems were referred to by the public at large as an ‘organ donation scandal’, thus also included donation which is institutionally separated. As a consequence of this the data-manipulation scandal attained the potential to destroy the trust of the German public in the institutions of organ allocation and donation altogether [20,27]. Finally, it is also remarkable that the public debates during that period of time have almost entirely ignored the international dispute about the definition and diagnosis of death [28], although controversially discussed issues regarding brain death had previously been picked-up in the scientific literature and covered by the media [29–31]. 2.3. The central role of trust Trust is widely regarded as the most important influence on donor rates [32,33]. To measure trustworthiness

Please cite this article in press as: Schwettmann L. Decision solution, data manipulation and trust: The (un-)willingness to donate organs in Germany in critical times. Health Policy (2015), http://dx.doi.org/10.1016/j.healthpol.2015.01.017

G Model HEAP-3368; No. of Pages 10

ARTICLE IN PRESS L. Schwettmann / Health Policy xxx (2015) xxx–xxx

3

Table 1 Descriptive data of dependent variables. Questions and answers offered

Coding of binary variable

[Holding an organ donor card] ‘Do you hold an organ donor card?’ Observations Yes (ODC = 0) ODC = 0 No (ODC = 1) ODC = 1

Survey 18 (April/May 2011)

Survey 21 (December 2012)

N = 1775 21.2% 78.8%

N = 1795 25.0%a 75.0%a

[Willingness to donate] Only respondents without an organ donor card: ‘Would you be willing to donate organs in the event of your death?’ Observations N = 1380 N = 1320 WTD w/o ODC = 1 14.4% 13.5% Yes, definitely WTD w/o ODC = 1 45.8% 34.7% Rather yes WTD w/o ODC = 0 31.2% 39.7% Rather no WTD w/o ODC = 0 8.6% 12.1% Definitely not [Trust in donor protection (as a paramount concern)] ‘A person carrying an organ donor card must be afraid that physicians will be more interested in his/her organs rather than saving his/her life.’ N = 1776 N = 1777 Observations I totally disagree. Trust in protection = 1 27.9% 21.9% Trust in protection = 1 25.2% 24.2% I rather disagree. I am undecided. Trust in protection = 0 34.1% 34.9% Trust in protection = 0 8.4% 11.0% I rather agree. Trust in protection = 0 4.4% 8.0% I totally agree. [Trust in compliance] only survey 2012: ‘What is your level of trust in the compliance of prevailing rules when donated organs are allocated?’ Observations N = 1.784 4.0% Trust in compliance = 1 My trust is. . .very high . . .high Trust in compliance = 1 16.1% Trust in compliance = 1 48.8% . . .medium Trust in compliance = 0 17.7% . . .low Trust in compliance = 0 13.3% . . .very low a

Percentage values are adjusted for objections in the year 2012. Unadjusted values are: ‘Yes’: 25.4%, ‘No’: 74.6%.

in healthcare, empirical studies use several aspects of two basic dimensions of trust [32,34–36]. First, ‘competence trust’ concerns the reliance on the competence of physicians and other officials. Second, ‘normative trust’ denotes the belief that they will act in the best interests of their clients (as defined by the clients’ values) rather than their own within the constraints of legal and moral legitimacy. In the organ donation and transplantation context ‘competence trust’ is related to end-of-life care, confirmation of death, and organ recovery, but also to communication and social skills. ‘Normative trust’ incorporates a reliable determination of and compliance with preferences of potential donors and their families adhering to statutory allocation rules. Corresponding behaviour is characterized by honesty, confidence, caring and respectfulness. General trust in physicians is found to have a strong impact on trust in individual physicians [36]. Offe [37] finds that trust in organizations can even weaken mistrust in specific individuals. Yet, it is widely unknown how trust at individual level affects institutional trust [35]. Furthermore, trust is a constituent aspect of healthcare quality directly correlated with patients’ satisfaction [38]. Some studies point out the strong relationship between negative experiences with or distrust of the healthcare system and distrust of the organ donation system. The latter includes the fear that doctors will prematurely declare brain death to obtain organs [33,39]. The surveys used subsequently contain two questions concerning normative trust and various potential indicators of trustworthiness relevant for two research questions: Is there a significant connection between trust and the WTD

organs? Are there links between trust and individual characteristics or attitudes of respondents? 3. Material and methods The data is part of the ‘Gesundheitsmonitor’ (Healthcare Monitor) [40] project, a regular representative postal survey financed by the Bertelsmann Foundation and the health insurer BARMER GEK on experiences with the German healthcare system. It is a characteristic of the project that some questions are included in all the surveys while others are of more transient up-to-date interest. In surveys 18 (April/May 2011) and 21 (December 2012), a total of 3573 respondents were confronted with questions regarding different issues of deceased organ donation and allocation [15,41]. Since the present study aims to investigate differences in WTD between 2011 and 2012, the main focus is on changes concerning items common to both questionnaires. 3.1. Dependent and independent variables Answers on four questions serve as dependent variables (Table 1). First, respondents indicated whether they were holding an ODC. Second, individuals without an ODC were asked to state their WTD organs post-mortem on a 4-item Likert-scale. Third, as a measure of donor trust, participants expressed on a 5-item Likert-scale the intensity of their (dis-)agreement to the statement that physicians give priority to organ procurement rather than to saving the life of the potential donor [33,42]. Fourth, though only in survey

Please cite this article in press as: Schwettmann L. Decision solution, data manipulation and trust: The (un-)willingness to donate organs in Germany in critical times. Health Policy (2015), http://dx.doi.org/10.1016/j.healthpol.2015.01.017

G Model HEAP-3368; No. of Pages 10

4

ARTICLE IN PRESS L. Schwettmann / Health Policy xxx (2015) xxx–xxx

21 of 2012, subjects specified how strongly they trusted that rules for allocating donated organs are followed. Before 2012 the possession of an ODC was regularly interpreted by laypeople as an agreement to donation. Objection rates on donor cards were extremely low and only half of the respondents in surveys knew about this option [43,44]. In survey 2012, but not in survey 2011, participants holding an ODC were asked to detail their statements on the card and to report the date of the signature. While in 2012 about 12% reported that they objected to the donation, only 2% did so in the years before. Hence, if the two surveys are compared, and if the possession of an ODC is interpreted as an agreement to donation, it is necessary to exclude from survey 2012 eight objections dated 2012 but not those dated earlier. Response categories of the other three questions are merged into respective binary variables to make regression results comparable and to allow for further statistical analysis (see coding of variables in Table 1). With respect to donor trust, physicians caring for potential donors are institutionally separated from those caring for potential recipients. Patients can rely on being protected against premature removal of their organs so that disagreement with the statement in the question indicates trust. For trust in compliance it is understood that choosing a high or medium category indicates that trust remains after the manipulation scandal. Potential covariates are selected on the basis of existing studies. They are clustered into five groups, which are introduced stepwise into regression models to identify robust significant effects. Only covariates displaying such an effect in at least one model enter all final regressions to make results comparable. Table 2 and Tables A1 and A2 (Appendix), which also report descriptive data and drop-out rates, describe the groups of variables finally selected. To control for multicollinearity variance inflation factors (vif) are calculated. A value above 5 would suggest a problem [45]. Starting with group 1, previous work has shown that individuals with a positive WTD organs are more often young and female, have a higher level of education and a higher income [32,44,46–48]. People living in multi-person households may instruct their relatives rather than completing an ODC. In East Germany, the former socialist GDR, transplantation rates are regularly higher than in West Germany [26]. Trust in physicians is found to be higher among older, less educated patients [38], while social trust is lower in East Germany [49]. Several variables in group 2 indicate relevant knowledge. Respondents insured by the Techniker Krankenkasse received information according to the new law before the data collection started in 2012. Obesity and smoking behaviour proved to be good indicators for current health state and health behaviour. They may indicate the average number of contacts with the healthcare system, but also health-related attitudes. Group 3 contains variables reflecting general values such as altruism (blood donation) or political activity (voting at polls). Religious people are often found to be less willing to donate organs [32,39,50]. The evaluation of the healthcare system is reflected by covariates in group 4 and may display a positive influence on all

dependent variables. Finally, the data set allows for controlling for evaluations of physicians (group 5), which are expected to have an impact on normative trust in organ donation. 3.2. Statistical methods Data analysis has been carried out with SPSSTM (version 22) and STATATM (version 12). Chi-squared tests are applied to investigate differences between surveys. The main analyses are performed by estimating four regression models using the dependent binary variables described in Section 3.1. Several robustness checks are performed. For the WTD-question estimations of ordered logit models lead to almost identical results compared to the binary models in terms of directions of effects and levels of significance. Bivariate probit models with sample selection were estimated for holding an ODC and WTD, because the ODC questions serve as a filter [51]. However, the hypothesis that the resulting equations are independent cannot be rejected for any sample. Furthermore, general findings from the binary regression models on trust are in line with results of alternative multinomial logit models with three answer categories; corresponding ordered logit models do not pass the test of parallel lines indicating that the single crossing assumption is not fulfilled [52,53]. Additionally, results are also robust with respect to different combinations of response categories in the binary trust variables. The resulting indicators for trust are included in the binary logit models for holding an ODC and for WTD. To test for endogeneity of trust in these regressions, bivariate probit models are performed [54–56]. Covariates indicating evaluations of physicians (group 5) are used as instruments for the respective type of trust. In some regressions they prove to be strong instruments, while in other cases, different (further) covariates have to be incorporated to identify sufficiently strong sets of instruments. The null hypothesis of exogeneity is generally not rejected. In models with strong instruments, the trust effect considered is positive and significant. A reweighting technique, entropy balancing [57,58], is applied to account for potential collinearity especially in regressions where both trust indicators are included, and to investigate causal relationships. Here, the respective samples are divided into a treatment group (trust = 1) and a control group (trust = 0). Entropy balancing reduces covariate imbalance prior to estimating the regression model by reweighting the control group observations. Corresponding results will be reported in Section 4. Finally, in all models the binary indicator for time is interacted with other variables to investigate whether public debates about the organ donation system have influenced groups of society differently [59]. Yet, this has proven not to be the case. 4. Results 4.1. Descriptive results Table 1 reports frequencies of answers to the four questions serving as dependent variables in the regression

Please cite this article in press as: Schwettmann L. Decision solution, data manipulation and trust: The (un-)willingness to donate organs in Germany in critical times. Health Policy (2015), http://dx.doi.org/10.1016/j.healthpol.2015.01.017

G Model HEAP-3368; No. of Pages 10

ARTICLE IN PRESS L. Schwettmann / Health Policy xxx (2015) xxx–xxx

5

Table 2 Binary logit regression results (odds ratios, OR) for holding an organ donor card (ODC) or willingness to donate (WTD) without an ODC. Independent variables

Holding an ODC (0: no, 1: yes)

WTD without an ODC (0: no, 1: yes)

Surveys 2011/2012

Survey 2012

Surveys 2011/2012

Survey 2012

OR

(S.E.)

OR

(S.E.)

OR

(S.E.)

OR

(S.E.)

0.917 0.914 0.721** 0.608*** 0.549*** 1.690***

(0.190) (0.145) (0.108) (0.098) (0.085) (0.174)

0.879 0.937 0.610** 0.400*** 0.404*** 1.621***

(0.267) (0.218) (0.130) (0.096) (0.093) (0.238)

0.968 1.039 0.830 0.806 0.688** 1.580***

(0.193) (0.179) (0.123) (0.122) (0.102) (0.152)

1.022 0.972 0.739 0.561** 0.564** 1.563***

(0.298) (0.265) (0.166) (0.128) (0.127) (0.220)

1.054 1.507***

(0.148) (0.220)

1.031 1.576**

(0.211) (0.344)

1.114 1.409***

(0.134) (0.184)

0.906 1.397*

(0.159) (0.272)

0.911 1.015

(0.131) (0.146)

0.986 1.041

(0.189) (0.233)

1.230 1.521***

(0.162) (0.202)

1.027 1.332

(0.188) (0.297)

0.866 0.854 0.662***

(0.113) (0.123) (0.091)

0.712 0.581** 0.649**

(0.134) (0.128) (0.127)

1.169 1.030 1.102

(0.146) (0.145) (0.134)

1.128 1.295 1.255

(0.216) (0.288) (0.221)

1.231 1.376 1.386*** 1.174 1.402*** 1.136

(0.237) (0.358) (0.171) (0.268) (0.175) (0.141)

1.797*** – 1.191 1.272 1.182 1.130

(0.326)

(0.201) (0.353) (0.159) (0.442) (0.085) (0.176)

1.235 – 1.317 1.205 0.575*** 1.298

(0.273)

(0.228) (0.435) (0.213) (0.201)

0.933 1.166 1.266* 1.755** 0.620*** 1.437***

1.299* 1.064 –

(0.193) (0.138)

1.661** 1.120 1.750***

(0.382) (0.217) (0.241)

1.337** 0.908 –

(0.187) (0.110)

1.209 0.791 1.058

(0.249) (0.140) (0.150)

1.810*** 0.708

(0.174) (0.154)

1.518*** 0.818

(0.209) (0.273)

1.567*** 0.736*

(0.145) (0.127)

1.539*** 1.133

(0.208) (0.277)

1.731*** 1.108 1.085 1.225

(0.222) (0.137) (0.191) (0.161)

1.726*** 1.198 0.976 1.480**

(0.322) (0.213) (0.247) (0.291)

1.260* 1.069 0.999 1.565***

(0.155) (0.125) (0.153) (0.183)

0.961 1.072 1.004 1.804***

(0.175) (0.190) (0.212) (0.322)

1.311** 1.548**

(0.147) (0.299)

1.243 1.839*

(0.199) (0.594)

0.909 0.590***

(0.097) (0.101)

0.926 0.590

(0.144) (0.200)

0.958 1.256** 1.235**

(0.169) (0.140) (0.129)

1.414 1.235 1.467**

(0.364) (0.194) (0.228)

0.882 0.945 1.205*

(0.134) (0.102) (0.119)

1.018 0.906 1.080

(0.240) (0.141) (0.160)

Trust in donor protection and in compliance Trust in protection Trust in compliance

5.275*** –

(0.560)

4.667*** 3.172***

(0.716) (0.650)

2.369*** –

(0.224)

1.814*** 3.609***

(0.263) (0.568)

Survey 21 in 2012 Constant

1.444*** 0.028***

(0.153) (0.008)

– 0.015***

(0.007)

0.655*** 0.342***

(0.064) (0.085)

– 0.152***

(0.057)

Number of observations Pseudo R-squared

3106 0.163

Group 1: Socio-demographic characteristics Age (ref.: 45–54) 18–24 25–34 35–44 55–64 65plus Female Level of education (ref.: low) Middle High Income (ref.: low) Middle High Household size (ref.: 1 person): 2 persons >2 persons East Germany including Berlin Group 2: Relevant knowledge, own health Techniker (health insurer) Survey in 2012 × Techniker Knows organ recipient Knew organ donor Health professional Obese Smoker (ref.: current) Past Never Informed about allocation rules (only 2012) Group 3: More general values Blood donor Non-voter (next elections) Religiosity (ref.: high): Low Middle Refused Solidarity (healthy/unhealthy) Group 4: General evaluation of healthcare system Need for reform (ref.: middle) Low High Satisfaction (ref.: high) Low Middle Low future quality of services

1563 0.222

2350 0.091

(0.263) (0.454) (0.121) (0.222)

1149 0.140

All covariates are coded as dummy variables. Robust standard errors are reported in parentheses. * 0.05 ≤ p < 0.10. ** 0.01 ≤ p < 0.05. *** p < 0.01.

models. Chi-squared tests reveal that in 2012 the proportion of respondents holding an ODC had increased (1% significance level) even if respondents who objected in 2012 were excluded. The change of rules in 2011 seems

to have had an effect. However, WTD of those without an ODC, and trust in donor protection as a paramount concern have declined. This seems to indicate an effect of the manipulation scandal.

Please cite this article in press as: Schwettmann L. Decision solution, data manipulation and trust: The (un-)willingness to donate organs in Germany in critical times. Health Policy (2015), http://dx.doi.org/10.1016/j.healthpol.2015.01.017

G Model

ARTICLE IN PRESS

HEAP-3368; No. of Pages 10

L. Schwettmann / Health Policy xxx (2015) xxx–xxx

6

Table 3 Binary logit regression results (odds ratios, OR) for trust indicators after entropy balancing. Survey

Trust variable

Dependent variable

OR

(S.E.)

2011/2012

Trust in protection

2012

Trust in protection

2012

Trust in compliance

Holding an ODC WTD without an ODC Holding an ODC WTD without an ODC Holding an ODC WTD without an ODC

5.343*** 2.381*** 4.357*** 1.940*** 2.608*** 4.899***

(0.622) (0.230) (0.714) (0.300) (0.651) (0.982)

All regression models contained the same set of covariates as the corresponding models described in Table 2. In each model the ‘trust’ variable defines the ‘treatment’ and the ‘control’ group for entropy balancing. Linearized standard errors are reported in parentheses. * 0.05 ≤ p < 0.10. ** 0.01 ≤ p < 0.05. *** p < 0.01.

4.2. Holding an organ donor card and willingness to donate The highest vif equals 2.42 (see Tables A1 and A2). This implies that multicollinearity should not be a problem. Table 2 summarizes regression results for holding an ODC and for WTD without an ODC. Many outcomes confirm our expectations, although they are not always statistically significant at the 5% level. The odds for holding an ODC rather than not, or for declaring a high WTD rather than not, are highest for middle-aged, female respondents, higher levels of education and higher income. The positive role of information can be deduced from some of the odds ratios for respondents who know an organ recipient or knew an organ donor, but also for individuals feeling informed about allocation rules. The results for survey 21 in 2012 reveal that the proportion of card owners is higher among respondents insured by the Techniker Krankenkasse. It had already been higher in survey 18 and increased further in survey 21. Results for covariates indicating relevant attitudes usually display the expected positive impact. This holds true for blood donors, past smokers, less religious respondents and individuals, who deem mandatory expressions of solidarity between healthy and unhealthy individuals as fair. Other effects are at least prima facie more surprising. Health professionals more often hold an ODC, but those without an ODC are less willing to donate. Possibly, being informed leads corresponding respondents to a more conscious decision for or against organ donation. The proportion of people having an ODC is higher in West Germany. This contrasts with the observation that transplantation rates are higher in the East. Furthermore, respondents are more likely to own an ODC if they evaluate the healthcare system negatively. Turning to the time effect, two opposing effects are observed. While the number of people having an ODC has increased between both surveys, WTD without an ODC has dropped. It can be assumed that both observations together reflect effects of the change of donation rules and the manipulation scandal. The effects of both types of trust on the two dependent variables are positive and remarkably high. This effect is not biased by other covariates. Although there are differences between mean values for some covariates (see Tables

A3–A5), the estimated coefficients for the trust indicators after the entropy balancing of samples remain essentially unchanged (Table 3). 4.3. Trust The next step is to identify covariates that are related to trust in donor protection or in compliance. Table 4 presents corresponding regression results. Several covariates display the expected influence, although, again, not all effects are statistically significant at the 5% level. The proportion of individuals with some trust is larger among females and respondents with higher levels of education. Trust is lower in East Germany, but only if trust in donor protection is concerned. Possibly, due to the manipulation scandal in 2012 trust in compliance is also very low in the West. In general, trust seems to increase with information. It is higher among those who know an organ recipient or knew an organ donor. Additionally, feeling informed about organ donation rules strengthens trust in compliance. Covariates indicating general attitudes also display the expected effects. Trust in protection is higher among blood donors and people who actively participate in politics by going to the polls. It is lower for those who refused to state their level of religiosity. Regarding attitudes towards the healthcare system, the proportion of people with low levels of trust increases if respondents think reforms are urgently needed, but decreases with satisfaction (for trust in protection) or if respondents expect the quality of medical services to decline (for trust in compliance). All coefficients of covariates concerning the evaluation of physicians (group 5) display the expected signs. They are highly significant in the first regression model in Table 4, whereas only the qualification of the physician displays a significant effect in the second model. Furthermore, trust in donor protection declined from 2011 to 2012. The second model also shows that both types of trust are correlated, the Pearson correlation coefficient equals 0.356. 5. Discussion and conclusion The implementation of the so-called decision solution and revelations of data manipulations have severe consequences for both organ donation and transplantation. Surely, talking to families would go much more smoothly

Please cite this article in press as: Schwettmann L. Decision solution, data manipulation and trust: The (un-)willingness to donate organs in Germany in critical times. Health Policy (2015), http://dx.doi.org/10.1016/j.healthpol.2015.01.017

G Model HEAP-3368; No. of Pages 10

ARTICLE IN PRESS L. Schwettmann / Health Policy xxx (2015) xxx–xxx

7

Table 4 Binary logit regression results (odds ratios, OR) for trust in donor protection as paramount concern (2011/2012) and trust in compliance (2012). Independent variables

Trust in protection

Trust in compliance

OR

(S.E.)

OR

(S.E.)

0.730* 1.051 0.857 0.926 1.178 1.333***

(0.135) (0.158) (0.113) (0.126) (0.158) (0.117)

1.172 1.734** 1.384 0.949 1.209 1.276*

(0.330) (0.437) (0.310) (0.219) (0.279) (0.185)

1.299** 1.652***

(0.147) (0.196)

1.254 1.167

(0.241) (0.238)

1.158 1.150

(0.140) (0.140)

1.183 1.059

(0.219) (0.239)

0.995 0.868 0.672***

(0.115) (0.111) (0.076)

1.512** 1.305 1.200

(0.295) (0.288) (0.220)

Techniker (health insurer) Survey in 2012 × Techniker Knows organ recipient Knew organ donor Health professional Obese Smoker (ref.: current) Past Never Informed about allocation rules (only 2012)

1.241 0.917 1.244** 1.406* 1.108 1.111

(0.246) (0.243) (0.135) (0.281) (0.125) (0.120)

0.701* – 1.264 1.959 0.834 1.015

(0.144)

0.849 0.931 –

(0.108) (0.103)

1.224 1.064 4.056***

(0.261) (0.197) (0.597)

Blood donor Non-voter (next elections) Religiosity (ref.: high): Low Middle Refused Solidarity (healthy/unhealthy)

1.234** 0.691**

(0.102) (0.113)

1.073 0.854

(0.149) (0.241)

1.144 1.000 0.603*** 1.180

(0.127) (0.106) (0.088) (0.131)

0.830 0.924 0.549*** 0.817

(0.153) (0.165) (0.120) (0.153)

1.135 0.678**

(0.111) (0.113)

1.223 0.555**

(0.197) (0.166)

0.705** 0.703*** 0.879

(0.101) (0.069) (0.080)

0.738 0.966 0.713**

(0.181) (0.156) (0.106)

4.938***

(0.750)

Age (ref.: 45–54) 18–24 25–34 35–44 55–64 65plus Female Level of education (ref.: low) Middle High Income (ref.: low) Middle High Household size (ref.: 1 person): 2 persons >2 persons East Germany including Berlin

Need for reform (ref.: middle) Low High Satisfaction (ref.: high) Low Middle Low future quality of medical services

(0.265) (0.886) (0.161) (0.179)

Trust in protection

–

Low respect (GP does not act respectfully) Thoroughness (GP examination) High qualification (physicians) Too much trust (in drugs) Survey 21 in 2012

0.566*** 1.358*** 1.342*** 0.597*** 0.642***

(0.078) (0.142) (0.142) (0.050) (0.057)

0.799 1.001 1.666*** 0.869 –

(0.169) (0.162) (0.279) (0.121)

Constant

0.757

(0.198)

0.366**

(0.153)

Number of observations Pseudo R-squared

2795 0.084

1415 0.2153

All covariates are coded as dummy variables. Robust standard errors are reported in parentheses. * 0.05 ≤ p < 0.10. ** 0.01 ≤ p < 0.05. *** p < 0.01.

if the written intent of the potential donor was more available than it is at present. The decision solution seems to increase the number of ODC, but at the same time the manipulation scandal is said to have destroyed trust. If trust is damaged, altruism, which is currently seen as the only eligible motive for donation in many countries [2], might be too weak for many potential donors or relatives

to consent to ex-plantation. Physicians may fear or even face hostile reactions during conversations with families. The data set contains survey results from the years 2011 and 2012 including answers to questions on the written and unwritten WTD organs post-mortem, but also on two aspects of normative trust. It allows investigating correlations between these responses and several individual

Please cite this article in press as: Schwettmann L. Decision solution, data manipulation and trust: The (un-)willingness to donate organs in Germany in critical times. Health Policy (2015), http://dx.doi.org/10.1016/j.healthpol.2015.01.017

G Model HEAP-3368; No. of Pages 10

8

ARTICLE IN PRESS L. Schwettmann / Health Policy xxx (2015) xxx–xxx

characteristics, experiences and attitudes and, thereby, provides answers to four major topics. First, regarding factors correlated with written or unwritten WTD, many earlier findings are confirmed. Strong correlations with typical indicators for altruism, such as blood donation, are detected. Any reform that aims to introduce incentives runs the risk of crowding-out altruists without attracting a sufficient number of donors to (over-)compensate the loss. Surprisingly, respondents with a negative attitude towards the healthcare system altogether are more likely to possess an ODC, whereas the results for those who believe that there is dire need for reform are ambivalent. Perhaps for many of these individuals, the organ donation system is just a further sign of a general weakness in the healthcare system they want to improve. Health professionals are found to be more likely to possess an ODC but more often rigorously reject donation if they have no ODC. This result seems to contradict the hypothesis of a positive impact of being better informed if a respondent works in the healthcare sector [60]. However, being informed may also lead to more conscious decisions, which could then be fostered by the decision solution. Furthermore, it is possible that health professionals are at least partly disenchanted with respect to the motives of some of their colleagues. This implication may have severe negative effects on the frequency and quality of talks to families of potential donors [61]. The second topic concerns the time effect. Data collection for survey 21 in 2012 started after the first main revelations of manipulations and also directly after the public discussion about the implementation of the decision solution. The positive impact of increasing knowledge of potential donors is corroborated by respondents insured by the Techniker Krankenkasse. The additional information they received did in fact increase their proclivity to donate. In the long run the campaign initiated as part of the new law for organ donation in Germany may be expected to increase WTD. Nevertheless, overall there are more respondents holding an ODC, but also more with a low WTD without having an ODC. Hence, there are two effects from the new law and the manipulation scandal: the probability of declaring consent on ODCs is decreasing, but the absolute number of declared consent is likely to increase. Either way, the increased number of ODC should help clinicians talk to next of kin. Third, turning to the issue of normative trust, it has a positive influence on attitudes towards organ donation. Obviously, the manipulation scandal also reduced trust concerning organ donation, although it directly affected ‘only’ the allocation side. This observation indicates that policy interventions should be aimed at regaining trust at all levels. From earlier work on trust it is known that especially the institutional level is pivotal [36,37]. For example, the implementation of an organ donor registry, which is accessible only after the confirmation of death, may have a confidence-building effect [4,62]. Since any conflicting interests of practitioners should be avoided, DSO has announced that their single aim is to guarantee that the will of the deceased is satisfied

rather than increasing the number of organ donors [63,64]. Fourth, normative trust is a complex notion and is in turn influenced by several individual attributes. Furthermore, there are strong connections between trust and evaluations of physicians. Consequently GPs should play a more prominent role than at present when it comes to counselling patients about end-of-life decisions including organ donation [44]. This is in line with the process of ‘shared decision-making’ of physicians and patients known in the healthcare literature [65]. But GPs should then be adequately reimbursed. Experience shows that GPs in their professional roles tend to respond to incentives in entrepreneurial ways that render risks of crowding-out intrinsic motivation negligible. Differences between East and West Germany reveal further aspects that may be important for organ donation. Fewer respondents in East Germany had an ODC or confidence in donor protection, but organ donation rates are regularly higher, there. This discrepancy may partly be due to the fact that in the former GDR an opt-out system was in force so that corresponding procedures and attitudes towards donation may have survived. However, competing explanations such as a lack of civic culture or of confidence that the ODC request will be respected are also possible. Some shortcomings of the study should be acknowledged. First, several aspects of interest could not be covered due to the predetermined set of questions and the restriction that certain questions should appear in both surveys. Second, only two questions on trust and organ donation have been asked. Hall et al. [36] propose a larger set of trust-related questions to derive a general scale, and to include something akin to their approach certainly would have been desirable. Third, the sample covers only two surveys from 2011 to 2012. It would be interesting to include a longer period and to see whether responses remain robust when information is less readily at hand once the theme loses its focal place in public attention and debate. Fourth, public discussions on organ donation in Germany in the period covered by the study have widely ignored the death debate. The public understanding and acceptance of the brain-death criterion seems to be reasonably high [19]. However, especially recent reports about mistakes in the diagnosis of brain death potentially undermine competence trust [66]. To conclude, the pure altruistic model in general is put into question. Reform proposals include the reciprocity solution, presumed consent or modest payments provided to deceased donor families [1,2,62]. As opposed to these more sweeping reforms this paper strictly kept within the scope of what is presently politically feasible in Germany and under EU regulations [4]. The German case in this and other regards may be relevant beyond Germany. Conflict of interest The author declares that there is no conflict of interest. Acknowledgements I would like to thank the Bertelsmann Foundation and BARMER GEK for providing the data and permission

Please cite this article in press as: Schwettmann L. Decision solution, data manipulation and trust: The (un-)willingness to donate organs in Germany in critical times. Health Policy (2015), http://dx.doi.org/10.1016/j.healthpol.2015.01.017

G Model HEAP-3368; No. of Pages 10

ARTICLE IN PRESS L. Schwettmann / Health Policy xxx (2015) xxx–xxx

to realize these analyses. Helpful comments by participants of the Annual Meeting of the German Association of Health Economics (dggö) in Munich 2014 are gratefully acknowledged. I am also indebted to Marlies Ahlert, Thomas Kirschstein, Hartmut Kliemt and Mathias Kloss for many instructive comments. The usual disclaimer applies. Appendix A. Supplementary data

[17]

[18]

[19]

Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/ j.healthpol.2015.01.017. References [1] Beard TR, Kaserman DL, Osterkamp R. The global organ shortage: economic causes, human consequences, policy responses. Stanford, CA: Stanford University Press; 2013. [2] Beard TR, Osterkamp R. The organ crisis: a disaster of our own making. The European Journal of Health Economics 2014;15:1–5, http://dx.doi.org/10.1007/s10198-013-0530-z. [3] Chkhotua A. Incentives for organ donation: pros and cons. Transplantation Proceedings 2012;44:1793–4, http://dx.doi.org/10.1016/ j.transproceed.2012.05.038. [4] Breyer F, van den Daele W, Engelhard M, Gubernatis G, Kliemt H, Kopetzki C, et al. Organmangel. Ist der Tod auf der Warteliste unvermeidbar? Berlin: Springer; 2006. [5] Chandler J, Burkell JA, Shemie SD. Priority in organ allocation to previously registered donors: public perceptions of the fairness and effectiveness of priority systems. Progress in Transplantation 2012;22:413–22, http://dx.doi.org/10.7182/pit2012324. [6] Gruenbaum BF, Jotkowitz A. The practical, moral, and ethical considerations of the new Israeli law for the allocation of donor organs. Transplantation Proceedings 2010;42:4475–8, http://dx.doi.org/10.1016/j.transproceed.2010.09.146. [7] Guttman N, Ashkenazi T, Gesser-Edelsburg A, Seidmann V. Laypeople’s ethical concerns about a new Israeli organ transplantation prioritization policy aimed to encourage organ donor registration among the public. Journal of Health Politics, Policy and Law 2011;36:691–716, http://dx.doi.org/10.1215/03616878-1334686. [8] Lavee J, Ashkenazi T, Gutman G, Steinberg D. A new law for allocation of donor organs in Israel. Lancet 2010;375:1131–3, http://dx.doi.org/10.1016/S0140-6736(10)61137-3. [9] Lavee J, Ashkenazi T, Stoler A, Cohen J, Beyar R. Preliminary marked increase in the national organ donation rate in Israel following implementation of a new organ transplantation law. American Journal of Transplantation 2013;13:780–5, http://dx.doi.org/10.1111/ajt.12001. [10] Gubernatis G, Kliemt H. A superior approach to organ allocation and donation. Transplantation 2000;70:688–707, http://dx.doi.org/ 10.1097/00007890-200008270-00032. [11] Wainwright D, Hanser AC. The formation of organ donation policy in Germany: evidence, politics and public opinion. Universal Journal of Public Health 2014;5:137–46, http://dx.doi.org/10.13189/ ujph.2014.020501. [12] Deutsche Stiftung Organtransplantation [DSO]. Organ donation and transplantation in Germany – annual report 2012. Frankfurt: Deutsche Stiftung Organtransplantation; 2013. [13] Tuffs A. Germany is poised to change law on organ donation to allow automatic “opting in”. BMJ 2010;341:c4763, http://dx.doi.org/10.1136/bmj.c4763. [14] Li D, Hawley Z, Schnier K. Increasing organ donation via changes in the default choice or allocation rule. Journal of Health Economics 2013;32:1117–29, http://dx.doi.org/10.1016/ j.jhealeco.2013.09.007. [15] Ahlert M, Schwettmann L. Einstellungen der Bevölkerung zur Organspende. In: Böcken J, Braun B, Repschläger U, editors. Gesundheitsmonitor 2011: Bürgerorientierung im Gesundheitswesen. Gütersloh: Bertelsmann Stiftung; 2011. p. 193–213. [16] Bundeszentrale für gesundheitliche Aufklärung [BZgA]. Einstellung, Wissen und Verhalten der Allgemeinbevölkerung zur Organ- und Gewebespende. In: Zusammenfassung der wichtigsten Ergebnisse

[20] [21] [22]

[23]

[24]

[25]

[26]

[27]

[28]

[29] [30]

[31]

[32] [33]

[34]

[35]

[36]

9

der Repräsentativbefragung 2010. Köln: Bundeszentrale für gesundheitliche Aufklärung; 2010. Available online: http://www. organspende-info.de/sites/all/files/files/files/RepBefragung Bericht final.pdf [15.09.14]. Thaler RH, Sunstein CR. Nudge: improving decisions about health, wealth, and happiness. New Haven, CT: Yale University Press; 2008, http://dx.doi.org/10.1007/s10602-008-9056-2. Steinmeier FW. Organ donation is true solidarity. European Journal of Cardio-Thoracic Surgery 2012;41:240–1, http://dx.doi.org/10.1093/ ejcts/ezr157. Bundeszentrale für gesundheitliche Aufklärung [BZgA]. Wissen, Einstellung und Verhalten der Allgemeinbevölkerung zur Organ- und Gewebespende. In: Zusammenfassung der wichtigsten Ergebnisse der Repräsentativbefragung 2013. Köln: Bundeszentrale für gesundheitliche Aufklärung; 2014. Available online: http://www. organspende-info.de/sites/all/files/files/Organ-%20und%20 Gewebespende 2013 Zusammenfassung.pdf [15.09.14]. Pondrom S. Trust is everything. American Journal of Transplantation 2013;13:1115–6, http://dx.doi.org/10.1111/ajt.12277. Shaw D. Lessons from the German organ donation scandal. The Intensive Care Society 2013;14:200–1. Siegmund-Schultze N. Organ donation: getting out of a tense situation. Deutsches Ärzteblatt 2013;110 [A-2118/B-1872/C-1826]. Available online: http://ec.europa.eu/health/blood tissues organs/ docs/ev 20131007 art21 en.pdf [15.09.14]. Stafford N. Surgeon is accused of manipulating data to move his patients up organ waiting list. BMJ 2012;345:e5039, http://dx.doi.org/10.1136/bmj.e5039. Deutsche Welle. Berlin doctor accused of queue-jumping in heart transplants; 2014. Available online: http://www.dw.de/berlindoctor-accused-of-queue-jumping-in-heart-transplants/ a-17872414 [15.09.14]. Bundesärztekammer. Kommissionsbericht der Prüfungs- und Überwachungskommission: Überprüfung des Herztransplantationsprogramms des Deutschen Herzzentrums Berlin am 1. April, 15. April, 16. April und 8. Juli 2014; 2014. Available online: http://www.bundesaerztekammer.de/downloads/TPG Bericht PK UeK-Herzzentrum 30092014.pdf [11.01.15]. Deutsche Stiftung Organtransplantation [DSO]. Organ donation and transplantation in Germany – annual report 2013. Frankfurt: Deutsche Stiftung Organtransplantation; 2014. Eurotransplant. Annual report 2013. Leiden: Eurotransplant International Foundation; 2014. Available online: http://www. eurotransplant.org/cms/mediaobject.php?file=AR20135.pdf [15.09.14]. Manzei A. Der Tod als Konvention. Die (neue) Kontroverse um Hirntod und Organtransplantation. In: Anderheiden M, Eckart WU, editors. Handbuch Sterben und Menschenwürde. Berlin: De Gruyter; 2012. p. 137–74. Birnbacher D. Der Hirntod – eine pragmatische Verteidigung. Jahrbuch für Recht und Ethik 2007;15:459–77. Müller S. Revival der Hirntod-Debatte [Revival of the brain death debate]: Funktionelle Bildgebung für die HirntodDiagnostik. Ethik in der Medizin 2010;22:5–17, http://dx.doi.org/ 10.1007/s00481-009-0044-5. Sahm S. Ist die Organspende noch zu retten?; 2010. Available online: http://www.faz.net/aktuell/feuilleton/geisteswissenschaften/ hirntod-ist-die-organspende-noch-zu-retten-1605259.html [11.01.15]. Childress JF, Liverman CT. Organ donation: opportunities for action. Washington, DC: The National Academies Press; 2006. Morgan S, Harrison TR, Afifi WA, Long SD, Stephenson MT. In their own words: the reasons why people will (not) sign an organ donor card. Health Communication 2008;23:23–33, http://dx.doi.org/10.1080/10410230701805158. Boulware LE, Cooper LA, Rattner LE, LaVeist TA, Powe NR. Race and trust in the health care system. Public Health Reports 2003;118:358–65, http://dx.doi.org/10.1016/ S0033-3549(04)50262-5. Calnan M, Rowe R. Researching trust relations in health care: conceptual and methodological challenges. Journal of Health Organisation and Management 2006;20:349–58, http://dx.doi.org/ 10.1108/14777260610701759. Hall MA, Camacho F, Dugan E, Balkrishnan R. Trust in the medical profession: conceptual and measurement issues. Health Services Research 2002;37:1419–39, http://dx.doi.org/10.1111/ 1475-6773.01070.

Please cite this article in press as: Schwettmann L. Decision solution, data manipulation and trust: The (un-)willingness to donate organs in Germany in critical times. Health Policy (2015), http://dx.doi.org/10.1016/j.healthpol.2015.01.017

G Model HEAP-3368; No. of Pages 10

10

ARTICLE IN PRESS L. Schwettmann / Health Policy xxx (2015) xxx–xxx

[37] Offe C. Wie können wir unseren Mitbürgern vertrauen? In: Hartmann M, Offe C, editors. Vertrauen: Die Grundlage des sozialen Zusammenhalts. Frankfurt: Campus; 2001. p. 241–94. [38] Calnan M, Rowe R. Trust matters in health care. McGraw-Hill: Open University Press; 2008. [39] Irving MJ, Tong A, Jan S, Cass A, Rose J, Chadban S, et al. Factors that influence the decision to be an organ donor: a systematic review of the qualitative literature. Nephrology Dialysis Transplantation 2012;27:2526–33, http://dx.doi.org/10.1093/ndt/gfr683. [40] Gesundheitsmonitor [Healthcare Monitor]. Bertelsmann Stiftung and BARMER GEK. Project homepage [15.09.14]: http://www. bertelsmann-stiftung.de/cps/rde/xchg/SID-232C0BD7-30F368B0/ bst engl/hs.xsl/7097.htm [41] Ahlert M, Schwettmann L. Einstellungen zur Organtransplantation und Spendebereitschaft. In: Böcken J, Braun B, Repschläger U, editors. Gesundheitsmonitor 2013: Bürgerorientierung im Gesundheitswesen. Gütersloh: Bertelsmann Stiftung; 2013. p. 63–87. [42] Ahlert M, Gubernatis G, Klein R. Common sense in organ allocation. Analyse und Kritik 2001;23:221–44. [43] BARMER GEK. Im Fokus: Organspende Ergebnisse einer Repräsentativbefragung; 2013. Available online: http://presse.barmer-gek.de/ barmer/web/Portale/Presseportal/Subportal/Presseinformationen/ Archiv/2013/130225-Organspende/PDF-Umfrage-Organspende2013.pdf [15.09.14]. [44] Bundeszentrale für gesundheitliche Aufklärung [BZgA]. Einstellung, Wissen und Verhalten der Allgemeinbevölkerung zur Organ- und Gewebespende. In: Zusammenfassung der wichtigsten Ergebnisse der Repräsentativbefragung 2012. Köln: Bundeszentrale für gesundheitliche Aufklärung; 2013. Available online: http://www. organspende-info.de/sites/all/files/files/Bericht-Studie-Organ-% 20und%20Gewebspende-2012%281%29.pdf [15.09.14]. [45] Sheskin DJ. Handbook of parametric and nonparametric statistical procedures, 5th ed. Boca Raton, FL: Chapman & Hall; 2011. [46] Decker O, Winter M, Brähler E, Beutel M. Between commodification and altruism: gender imbalance and attitudes towards organ donation. A representative survey of the German community. Journal of Gender Studies 2008;17:251–5, http://dx.doi.org/10.1080/09589230802204290. [47] Mohs A, Hübner G. Organ donation: the role of gender in the attitude–behavior relation. Journal of Applied Social Psychology 2013;43:E64–70, http://dx.doi.org/10.1111/jasp.12042. [48] Mossialos E, Costa-Font J, Rudisill C. Does organ donation legislation affect individuals’ willingness to donate their own or their relative’s organs? Evidence from European Union survey data. BMC Health Services Research 2008;8:48, http://dx.doi.org/10.1186/1472-6963-8-48. [49] Heineck G, Süssmuth B. A different look at Lenin’s legacy: social capital and risk taking in the two Germanies. Journal of Comparative Economics 2013;41:789–803, http://dx.doi.org/10.1016/j.jce.2013.02.005. [50] Boulware LE, Ratner LE, Cooper LA, Sosa JA, LaVeist TA, Powe NR. Understanding disparities in donor behavior: race and gender differences in willingness to donate blood and cadaveric organs. Medical Care 2002;40:85–95, http://dx.doi.org/10.1097/ 00005650-200202000-00003.

[51] Miranda A, Rabe-Hesketh S. Maximum likelihood estimation of endogenous switching and sample selection models for binary, ordinal, and count variables. Stata Journal 2006;6:285–308. [52] Janssen J, Laatz W. Statistische Datenanalyse mit SPSS, 8th ed. Berlin/Heidelberg: Springer; 2013. [53] Winkelmann R, Boes S. Analysis of microdata, 2nd ed. Berlin: Springer; 2009. [54] Baum CF, Schaffer ME, Stillman S. Enhanced routines for instrumental variables/generalized method of moments estimation and testing. Stata Journal 2007;7:465–506. [55] Cameron AC, Trivedi PK. Microeconometrics using stata, Rev. ed. College Station, TX: Stata Press; 2010. [56] Chiburis RC, Das J, Lokshin M. A practical comparison of the bivariate probit and the linear IV estimators. Economics Letters 2012;117:762–6, http://dx.doi.org/10.1016/j.econlet.2012.08.037. [57] Hainmueller J. Entropy balancing for causal effects: a multivariate reweighting method to produce balanced samples in observational studies. Political Analysis 2012;20:25–46, http://dx.doi.org/10.1093/pan/mpr025. [58] Hainmueller J, Xu Y. Ebalance: a stata package for entropy balancing. Journal of Statistical Software 2013;54:1–18, http://dx.doi.org/10.2139/ssrn.1943090. [59] Buis ML. Stata tip 87: interpretation of interactions in nonlinear models. Stata Journal 2010;10:305–8. [60] Radunz S, Hertel S, Schmid KW, Heuer M, Stommel P, Frühauf NR, et al. Attitude of health care professionals to organ donation: two surveys among the staff of a German university hospital. Transplantation Proceedings 2010;42:126–9, http://dx.doi.org/10.1016/j.transproceed.2009.12.034. [61] Berndt C. Ärzte engagieren sich immer weniger für Transplantation. Sueddeutsche.de; 2014. Available online: http://www. sueddeutsche.de/gesundheit/rueckgang-der-organspendezahlenaerzte-engagieren-sich-immer-weniger-fuer-transplantationen1.1867695 [15.09.14]. [62] Breyer F. Der andere Skandal: Der Mangel an Spenderorganen und mögliche Auswege. In: Haarhoff H, editor. Organversagen: Die Krise der Transplantationsmedizin in Deutschland. Frankfurt: ReferenzVerlag; 2014. p. 27–56. [63] Berndt C. Mediziner beklagen massiven Vertrauensschwund in Organspende. Süddeutsche.de; 2014. Available online: http://www. sueddeutsche.de/gesundheit/transplantationen-medizinerbeklagen-massiven-vertrauensschwund-in-organspende1.2203530 [11.01.15]. [64] Rady MY, McGregor JL, Verheijde JL. Mass media campaigns and organ donation: managing conflicting messages and interests. Medicine, Health Care, and Philosophy 2012;15:229–41, http://dx.doi.org/10.1007/s11019-011-9359-3. [65] Kon AA. The shared decision-making continuum. Journal of the American Medical Association 2010;304:903–4, http://dx.doi.org/10.1001/jama.2010.1208. [66] Berndt C. Es sollte jemand machen, der viel vom Gehirn versteht. Sueddeutsche.de; 2014. Available online: http://www. sueddeutsche.de/gesundheit/strengere-regeln-fuer-hirntoddiagnose-es-sollte-jemand-machen-der-viel-vom-gehirn-versteht1.1906767 [11.01.15].

Please cite this article in press as: Schwettmann L. Decision solution, data manipulation and trust: The (un-)willingness to donate organs in Germany in critical times. Health Policy (2015), http://dx.doi.org/10.1016/j.healthpol.2015.01.017