The analyses show that the particular field of study has an impact on the perception of standard of care and fairness in the health care system. This finding is in line with Ahlert, Schwettmann et al. and Ahlert, Felder et al., who conducted experiments with students on the distribution of resources in an equitable health system [34, 35]. The authors also found differences in behaviour between academic disciplines. With respect to standard of care, we showed that – contrary to our hypothesis – medical students begin their studies more optimistic than students in non-medical disciplines. Their perception increases, moreover, over the course of their studies. Zupanic et al. asked medical students about their motivations to study medicine . The main aspects were an interest in medical issues, contact with people as well as desire to help. At first sight, it seems that there is no direct connection between the motivations and their perception of standard of care, but medical freshmen may already be convinced of the system; otherwise, they might have been deterred from studying medicine in the first place. On the one hand, the growing optimism over the course of their studies may contradict some of the results found by Osenberg et al. and Zupanic et al. [21, 22]. Their studies show that German medical students anticipate the budgeting of services as an important problem of their future work, perceive deficiencies in patient care, and are concerned about those in the future. It can also be assumed that medical students perceive the problem of resource allocation in more direct and specific terms . Only recently, the professional association of medical students in Germany requested an open debate about rationing . On the other hand, there are arguments that might explain the growing optimism. Medical students are taught by professors at medical universities who do basic research and take care of serious conditions. In general, this basic research intends to deliver innovative therapies for the treatment of patients. This environment might also influence the students’ perception. It can also be assumed that medical students learn about the health systems of other countries in ways unlike non-medical students. It may be that their responses have a more relative character. An international survey conducted by Koch et al. showed that German practitioners require comprehensive reforms of the health care system but overall assess the standard of care as being good .
In contrast to the question regarding the standard of care, there was no significant difference between the freshmen`s perception of fairness, but the coefficients indicate that philosophy/religion students make up the most critical of all groups. The non-significance might result from the relatively small size of the sample.
More interesting, however, is that there is only a marginal difference between the perception of freshmen and advanced students studying philosophy/religion. Here, our assumption that the course of the philosophy/religion studies influences the perception of fairness in the health care system cannot be confirmed. Instead, our results show that the law students’ perception of fairness decreases the most. Considered individually, the perception of the law students decreases significantly (p < 0.05) during the course of study. Of course, it may be that philosophy/religion freshmen already have a relatively well-grounded, consolidated opinion. In comparison to other freshmen, the perception of philosophy/religion students does not significantly differ (see Table 3). Looking at the coefficients, however, their perception is already worse, which might explain the marginal changes during the course of study. Another reason might be that the health care system is not discussed directly in the courses of philosophy/religion students. The coefficients nevertheless show that the perception of advanced philosophy/religion students is worse compared to advanced medical and economics students. Additionally, Figure 4 illustrates how the perceptions of law and philosophy/religion students and economics and medicine students diverge. Studies have shown that medical school can often have detrimental effects like increasing cynicism and decreasing empathy on students' professional growth [39–41]. If these aspects played a role here remains open. The results however may indicate a high conflict potential between the disciplines and may also have an impact on the controversial discussion over health care reforms, prioritization and rationing in the health care sector [19, 20].
Independently of the hypotheses, we generated additional results. The type of health insurance and gender have a high significant influence on the assessment of standard of care. In our study population, men and persons covered by PHI rate the standard of care as better than persons covered by SHI. These findings are in line with those of Sawicki, who obtained similar results regarding the standard of care . In a patient survey, he found out that the insurance type has a relevant influence on the assessment and that women have a worse perception of the standard of care. A citizen survey of the European Commission analysed the socio-demographic factors that determine a poor subjective assessment of the quality of health care . No differences between men and women were found, but the authors recoded the variable ‘perceptions’ in only two categories (good/bad), which might have led to a loss of information. Furthermore, the question stressed the quality of health care, which is not exactly the focus of our inquiry.
In addition, our analysis revealed that subjective health status has an impact on the perception of the standard of care. Here, the coefficients for chronic disease, poor subjective health status, and less health-conscious lifestyle are all negative. It can be assumed that people with chronic disease have more experience with the health care system. In an international study by Schoen et al., German patients criticized deficiencies in care and poor coordination for chronic conditions in Germany [44, 45]. The study also revealed that German patients had a more negative perception compared to foreign subjects, but the negative impact that we discern also contradicts the results of previous studies. Zuchandke et al. and Lange et al., who investigated the perception of financial security in cases of long-term care need and illness, determined that people with experience of the German health care system and long-term care perceive their financial security better compared to those without such experience [46, 47]. One could therefore assume that this would also be the case for standard of care, but these perceptions are of course not directly comparable.
Unlike the first regression, fewer control variables have a significant impact on the perception of fairness in the health care system. In particular, there is no significant difference between women and men. As in the first regression, participants with a less subjectively health-conscious lifestyle assess the system as less fair. This is interesting, since the provision of health care in Germany should not be dependent on the lifestyle. Another interesting result is that the variable ‘serious illness among family or friends’ has a significant negative impact on the assessment of fairness. About 33% of the study participants report having at least one person with serious illness among family or friends, but they do not assess the standard of care as significantly worse. Why the perception of fairness is more negative in this group raises a new research question.
In what follows, we highlight limitations with respect to the variables used, the methodology, and the study population. The objective of our study was to determine whether the assessment of the German health care system in terms of standard of care and fairness differs between fields of study. We used two statements as proxies. Although we can describe certain tendencies in the responses, we cannot indicate, for example, which attributes make up a fair system for the participants. In particular, we do not know whether the students assess the standard of care and fairness of the system in comparison to other systems or with respect to an idea of an optimal system. Additionally, we do not know the participants' actual knowledge of the health care system and to what extent a combination of knowledge and views affect each other. However, we found a significant correlation (p < 0.01) between both variables (standard of care and fairness). Respondents might somehow have experienced a form of rationing that underpinned both an unfavourable assessment of standard of care and of fairness.
In this context, we want to stress the subjective character of our dependent variables, which can lead to heterogeneous interpretations. Unfortunately, we are not able to control for such interpersonal heterogeneity as we do not have panel data. While the heterogeneity can be reduced by transforming the ordinal variables to binary variables, this would lead to a loss of information and consequently less precise results. We therefore decided to keep the ordinal structure of our dependent variables. However, further research is needed to identify the ideas of an optimal system in terms of standard of care and fairness. A mixed methods approach consisting of combined quantitative and qualitative surveys might be appropriate. Additionally, further questions regarding a fair priority setting could be addressed here, including the identification of relevant attributes, the balancing of values like efficiency, equity, or reasonableness, and trade-off between different focuses (achieving fair outcomes or a fair process).
With respect to the dependent variable ‘standard of care’, we want to highlight that our statement focused on the standard of care of seriously ill patients. The general label ‘standard of care’ may therefore be misleading. We nevertheless decided to formulate the statement more precisely since previous studies showed that the health care of seriously ill patients with high severity is more important to society than the health care of minor conditions [48, 49].
A problem of imperfect multicollinearity may result due to the inclusion of ‘year of birth’ as a control variable. There is obviously a high correlation between age and the number of study terms (p < 0.01). We nevertheless wanted to control for possible age differences among freshmen, but we have also redone the regression analyses without considering the factor ‘age’ to test the impact on the standard errors. The significance levels did not change.
Ordered probit and ordered logit models are often used in applied econometric analyses. The models nevertheless have some limitations especially when it comes to analysing marginal effects. First of all, the assumption of a normal or logistic distribution leads to the case in which the sign of the coefficients does not necessarily represent the direction of the effect for all outcomes of the dependent ordinal variable . Moreover, assumptions such as the single index function or constant thresholds lead to further restrictions. A detailed overview of limitations based on ordered response models is provided in Boes and Winkelmann .
With respect to the subjects, they all studied in Hannover or Göttingen. It is possible that students in southern or eastern Germany have a different perception. It is also possible that advanced students have been influenced by the opinion of single professors at these universities. Since lectures are given by various professors, however, the possibility of a special influence is minor. A wider study would nevertheless also have the advantage that more students who study philosophy and religion could be integrated into the data set. Due to the small number of students, we formed one group from these two subjects, which is also a restriction. We therefore implicitly assume both groups to be identical with regard to the perception of ‘standard of care’ and ‘fairness’.
As already mentioned above, we used cross-sectional data instead of longitudinal data. We did not analyse the perception of the same study group over time. Besides the mentioned possibility of controlling for interpersonal heterogeneity, another important issue is the dropout rate. This may especially be relevant for law and economics, in which the rate is up to 50%. It would therefore be interesting to conduct the same study again and interrogate the same group over time.
Finally, we want to stress that the results of this study are not representative for the general population, since students are young and well educated. Age and the socio-economic status may have an important influence on the perception of fairness and standard of care. It would therefore be interesting to repeat the study, ask the general population about their opinion, and control for the age and socio-economic status of the subjects.