Restoring Trust in Healthcare: Information Impact Case Study in Poland

Background: This study empirically evaluates the inuence of information on trust at the physician level, the medical profession, hospitals, and with the payer. Restoring trust in a medical setting appears to be signicantly affected due to the coronavirus pandemic. Trust improves results from medical treatment, raises perception of healthcare performance, and smoothens the overall functionality of healthcare systems. Methods: In order to study trust volatility, participants took part in a three-stage experiment designed via: (1) measured level of trust, (2) randomly dividing participants into two groups—control (i.e., re-examination of level of trust) and experimental (i.e., being exposed to a piece of certain manipulative information), and (3) checking whether observational changes were permanent. Results: Results indicate that in the experimental group the increase of trust was noticed in the payer (27.7%, p < 0.001), hospitals (10.9%, p = 0.011), and physicians (decrease of 9.2%, p = 0.036). Conclusions: The study indicated that in Poland information is likely to inuence trust in healthcare while social and interpersonal trust levels may be related to increases of trust in hospitals and in the payer versus decreases in physicians.


Background
In many countries, including Poland, a signi cant number of coronavirus infections occur in medical settings. As a result, a lowering in healthcare trust has been widely reported in terms of other illnesses and diseases not being looked at or treated-especially in a timely manner or at all [1] [2] [3] [4] [5]. Trust is crucial for the smooth functioning of complex systems, particularly in healthcare [6]. It is regarded as an effective tool for evaluating medical performance [7] [8] [9] [10] as well as of great importance to a wellfunctioning healthcare system. The study empirically evaluates the in uence of information on trust in physicians, the medical profession, hospitals, and the payer-i.e., its main components. Consequently, the notion of restoring trust in healthcare as a result of the coronavirus pandemic or any other relating circumstance is vital to a healthy and viable society.
People's trust with their physician and, generally, with a provider is vital to the care process. It can modify patient attitude and behavior which can result in better levels of treatment [11] [12] [13]. Trust can activate the placebo effect [14], increase acceptance of medical suggestions and compliance with treatment recommendations [15], diminish the risk of underusing medications in response to cost pressure [16] [17] [18], and improve motivation to seek help and use preventive care [8] [11]. Moreover, trust enhances communication between doctors and patients [19] [20] [21] as well as the perception of e cacy, selfreported health status [22] [23], well-being, and quality of life [5]. Changing attitudes and behaviors associated with patient trust also has a positive impact on healthcare providers. As such, trust in a provider may reduce the number of con icts between patients and medical staff [22], diminish the probability of complaining about medical malpractice [24], lower transaction costs (e.g., expenditures that can decrease patient anxiety via additional diagnostic testing and physician consultation) [25] [22], and increase motivation to recommend the healthcare provider to others [22]. Trust improves the perception of the performance. Studies show that patients with high trust are more likely to perceive a performance positively even if it was objectively mediocre [19] [26] [21]. As a result, low institutional trust may cause ine ciency and undermine the legitimacy of health insurance and eventually decrease solidarity [27] as well as overall success of health policy [28]. Thus, an emphasis on the importance of research factors that may affect the level of trust in healthcare become seeming fundamental.
Although there is a rich body of knowledge about factors in uencing trust in healthcare, a lack of quantitative research on how speci c information is delivered-societally-exists. Understanding how information in uences trust could have signi cant consequences, e.g., visiting a physician, the medical profession in general, hospitals, and the payer. Three points of should be considered: (1) can level of trust be in uenced or even managed regardless of real healthcare performance; (2) ability to revise the data collected, analyzed, and conclusions drawn from them; and (3) action taken to increase trust as well as criteria of resources allocation in healthcare veri ed. Trust can be considered both as interpersonal, i.e., trust in a physician, as well as social, i.e., trust in a more abstract sense such as groups of people, institutions, and health systems [8] [29] [30] [31]. [32] claimed that interpersonal trust occurs when there is a possibility of repeated testing over time, i.e., to what extent a person is trustworthy. [33] suggested that a high level of trust in a physician cannot always be explained by evaluation of objective evidence of trustworthiness since it can also arise as a response to psychic distress created by illness. This means that in healthcare trust originates from the fundamental psychological attributes of seeking care in a state of anxiety, rather than from provider characteristics or patient personalities [8]. This is consistent with other suggestions that asymmetry of information between a patient and physician [34] in accordance with the logic of professionalism [35] forces patients to trust their doctor. A meta-analysis of 47 studies showed that the correlation between trust and health outcomes is small to moderate [36].
Speci cally, trust is moderately correlated with self-rated subjective health outcomes, but there is no correlation between trust and objective as well as observer-rated outcomes [36].
Social trust can be in uenced by patient experience and the general public's view of the system [19] [32] [37]. Institutional trust in such organizations as hospitals, public payers, or insurers may be an indicator of consumer appreciation of the organization [27] and can be affected by varying degrees of interpersonal and social trust [38]. In particular, social trust may be in uenced by patient experience, general public opinion [19] [32] [37], professional institutions and legal as well as regulatory protections [32], institutional guarantees, and government regulation of medical education, protection of patient rights, and healthcare quality supervision [39]. Importantly, these issues need to be conveyed to society as understandable as well as plausibly achievable. This is signi cant since [27] pointed out that low levels of trust are due to the fact that people may not fully comprehend how the healthcare system (e.g., health insurance) works and how money in uences physicians and provider behavior. Moreover, they predict that political communication and mass media may play a central role in shaping public opinion, as "facts do matter less than the perception of the facts" [27]. In many countries, a low level of trust is directly correlated with the media, reporting on what goes wrong in healthcare [27] and why. In short, information is one of the critical factors that in uences the level of trust in a healthcare system. [27] claimed that the central query of "whether better information will indeed translate in higher institutional trust" formulates the foundation of this research. This paper explores these queries in the context of Poland and its institutional trust in its medical profession, hospitals, and payer system as well as interpersonal trust at the physician-based level. A breakdown of the study is structured as follows: Sect. 2 frames the research method, Sect. 3 illustrates the results, and Sect. 4 elucidates a discussion and conclusion on healthcare trust in terms of real performance.

Hypothesis development
In trust literature, the most frequently studied elements are a physician (i.e., with whom a patient has the most frequent contact), the medical profession, hospitals, payers, and the overall healthcare system [31]. Given this study is based in Poland, objects studied in this research are the same except for the healthcare system. This is important since Polish society often confuses the healthcare system with the payer, i.e., the National Health Fund (in Polish Narodowy Fundusz Zdrowia) (NFZ). In Poland, the NFZ operates as a single centralized payer which is the most visible part of its healthcare system. Confusingly, the media and the public often use the words "system" and "payer" interchangeably, thus blurring the difference between these concepts. Hence, a lack of clear separation between these objects makes it impossible to prepare the appropriate information for the intervention and subsequent interpretation of the results. Therefore, to obtain less ambiguous results from this research, only the payer (i.e., NFZ) was selected as it is a better recognized and de ned object. Trust level objects, in particular, may vary since trust in different objects may have diverse levels of susceptibility to the supplied information. In consideration of the existing literature, the degree to which information in uences trust level of a particular object, may depend on the frequency people encounter it in a particular healthcare system [21] as well as the type of trust being considered (i.e., social or interpersonal). The study considers the following three hypotheses. This hypothesis is based on two foundations. First, trust in an insurer (i.e., in this case the payer) is more amenable to change than in a physician [40] [8] [41], Second, in Poland's healthcare system patients virtually have no contact with the payer (i.e., NFZ) hence they have no direct experience with it.
Hypothesis 2 (H2): Trust in a physician is most resistant to delivered information.
According to the literature, the increase of patient trust in a physician may be associated with the improvement in receiving care promptly [42] as well as perceived physician competence and communication skills [8]. Interpersonal trust occurs when there is a possibility of repeated testing over time-to determine the extent of a person's trustworthy [32]. Additionally, there are signi cant obstacles in delivering information concerning every individual physician, hence the change of trust in this object, apart from experience, may originate mostly from the interrelationship between social and interpersonal trust claimed by [43]. Trust in hospitals and the medical profession are related to trust at the physician level [19]. [22] claimed that patients who trust their physician may worry less about the hospital due to their reliability from their physician to direct them to a suitable one, monitor their quality of service, and their clinical outcome. [38] suggested that trust in the medical profession depends to some extent on patients' previous experiences with their own doctor. Thus, trust in hospitals may be less susceptible than trust in the payer since approximately 15% of Poles have direct (i.e., personal) experiences with hospitals [44] versus 85% with their doctor. Similarly, trust in the medical profession may be more susceptible to delivered information than trust in a physician since the medical profession is more abstract than an individual physician and trust is not based directly on personal experience. However, trust in the medical profession should be considered less vulnerable than trust in the payer (i.e., NFZ) since trust in the medical profession may be more related to trust of a physician-per se [38].

Study design
There is signi cant di culty in designing a study that assesses the in uence of speci c information on societal trust level. Understandably, the possibility of controlling information delivery to individuals and measuring the difference of trust level before and after the delivery is challenging. To overcome this problem, a three-stage experiment in an unchanging group of respondents was applied. The experiment was conducted between September 2015 and March 2016 in two medium-sized Polish enterprises within the context of a multi-staged training program concerning quality systems. The training was conducted in permanent groupings and on all organizational levels, ensuring constant composition of the groups using demographically diverse samples (Table 1). At all stages, the level of trust in a physician, the medical profession, hospitals, and the payer were surveyed. Between the rst and second stage, the period of at least one month was used to minimize the likelihood that participants could remember their previous responses. In order to determine whether the change was permanent, the third stage was carried out at least two months after the second stage.  (7) 14 (11) 17 (14) 9 (8) 11 (9) Health status Very well 33 (14) 18 (14) 14 (11) 20 (17) 13 (11) Well 159 (64) 87 (69) 72 (58)  81 (68) 68 (57) Average 45 (18) 17 (14) 29 (24) 14 (12) 30 (25) Bad

Components of mass media information
In the study, the assumption has been adopted that information delivered by mass media is a contribution to the decision-making process, consisting of two components: (1) statisticalobjective and (2) emotional-subjective [45]. Information (i.e., statistical-objectivity) in uences the audience when it is comprehensible for the average user, adequate, knowledgeable, trustworthy [46], and presented in a structured manner [47]. As a result, the emotionalnarrative component of information has a more signi cant impact on the audience than a statistical one [48] [49]; however, demonstrating statistical data as a graphical representation can increase its impact on decision-making [50] [51].

Perception of healthcare in Poland
Surveys conducted throughout Poland indicate a wide discrepancy between the general public trust in healthcare and individuals using healthcare frequently [44] [52]. Responsibility for a lower level of trust from non-users may arise due to a highly correlative link from negative information concerning healthcare disseminated by the media [44] [52] -mainly as a side-effect of the system's rapid change and violent political competition [53] [54]. Poland's healthcare is under constant reform, trying to adjust the post-communistic system to Westernized standards, which leads to con icts between interest groups struggling to protect their current interests versus efforts to try and obtain better access to public funds. The primary result of this strife is negative media output.

Design of the information package used
The information package presented to participants was designed in a manipulative manner by presenting Poland's healthcare as superior (i.e., in a better light) compared to other countries. All the provided data were drawn from the World Health Organization and Eurostat in which particular indicators were chosen in such a way that Poland was a near top healthcare provider. The information package was not in line with the main "climate" currently being portrayed by Polish mass media. The contrast was designed by comparing the United States as the country with the highest spending rate on healthcare in the world and other wealthy Western European countries as well as with some former communist countries which were on the same economic level before their collapse in 1989. The information package consisted of two types of information. First, it targeted an emotionalnarrative by starting off with the rst 30 minutes of the movie "Sicko" directed by Michael Moore, dubbed in Polish [55]. Second, statistical-objectives were stressed to elucidate data from o cial international health statistics presenting a number of graphs mostly illustrating country-related expenditure and data concerning medical errors in American hospitals. The information package was prepared in a contradictory manner to the mainstream point of view represented in mass media.
It was hypothesized that the movie would have an impact on trust for the payer since it focuses on healthcare insurance and system-speci c aspects. It might also affect the medical profession, especially when it portrays a medical doctor making a public confession that she had one primary duty-to use her medical expertise for the nancial bene t of the insurer stating "…doctors at health insurance companies actually are responsible for the death of patients". The data regarding medical errors in the US informed participants about the inevitability of medical risk. As a result, this was supposed to increase the positivity and perception of the performance of Polish hospitals as well as show other countries as less forward-thinking. Moreover, indicators such as standardized death rates for speci c cancers or ischemic heart disease illustrated Poland as a comparable alternative to countries spending several times more on healthcare, hypothetically in uencing all three objects excluding individual physicians since provided information could not be directly linked to each of the participant's personal doctor.

Dependent variables and experimentation
Given the lack of scale, measuring for the level of trust and testing for their reliability in terms of societal awareness, Poland inclusive of other central and eastern European countries [56] scales that were developed and tested in the US were adopted ( Table 2). Trusting a physician and the medical profession, in general, was measured using ve element scales developed by [57]. To estimate the level of trust in the payer (i.e., NFZ) and hospitals, four and three element scales developed by [58] were applied. The questionnaires were translated into Polish. To ensure authenticity and accuracy of the translations, they were translated back to English by a secondary translator to check whether the meanings remained the same. For each question, a Likert scale was used, and respondents were asked to choose an answer from the following range: 1-Strongly disagree, 2-Disagree, 3-Neutral, 4-Agree, and 5-Strongly agree. A doctor would never mislead you about anything.
All in all, you trust doctors completely.6.
Trust in health care payers 2,3 Health care payers are good at what they do.
When needed healthcare payers will pay for you to see any specialist.
When questioned about what treatments are covered healthcare payers are honest with their answers.
Healthcare payers will pay for everything they are supposed to, including treatment that is expensive.  [58]; 3 the Polish translation of the word "payers" was changed into the singular form "płatnik", i.e., NFZ, in every item since there is only one payer in the Polish healthcare system In the rst stage of the experiment, for the entirety of the participants, the level of trust in a physician, the medical profession, hospitals, and the payer were measured. In the second stage, participants were randomly divided into experimental and control groups of equal size and characteristics. In the control group, re-examination of the level of trust was surveyed, while in the experimental one (i.e., before the survey) participants were shown the information package. In the third stage, the survey was performed again to observe if any change in trust was observable. The manipulation-based check in this experiment is based on observation as well as whether participants believed the intervention. It was performed by a number of research assistants, who monitored participants and occasionally intervened when they were distracted. The research assistants also asked questions to monitor participant attention and continually veri ed if all participants, in a similar manner, had understood the information.

Results
Statistical analysis was performed using Statistica version 13 software with the p-value of 0.05 (i.e., a 95% level of signi cance). Considering the experiment consisted of three stages in which two of them were performed in two groups, in total, a comparative examination of ve tests had to be carried out. To validate the signi cance of differences in the mean value of the various stages the Tukey post hoc test was applied. This test provides sounder, more conservative results than the comparison of pairs via the use of analysis of variance. Results from the test are illustrated in Fig. 1.
Lack of statistically signi cant differences in the level of trust between the rst stage of the study and the control group in the second stage showed that between the two surveys no factors had in uenced the initial level of trust. As a result, changes observed in the experiment were consequential to the information delivered during the study. The results obtained from the Tukey post hoc test indicated substantial increases of trust in the experimental group in the second stage. Comparatively, the control group can be observed by way of the payer (i.e., an increase of 27.7%, p < 0.001) followed by hospitals (i.e., an increase of 10.9%, p = 0.011), and, surprisingly, also in physicians (i.e., a decrease of 9.2%, p = 0.036). It is worth noting that trust in the medical profession and hospitals were presented a very similar level (i.e., p > 0.05) during entirety of the experiment except for the experimental group after the delivery of the information package where trust in hospitals increased (Fig. 1). Moreover, in the experimental group, trust in the payer reached a similar level to the medical profession (i.e., p > 0.05) but trust in physicians decreased and converged with the increased level of trust in hospitals.

Discussion And Conclusions
The statistical analysis demonstrated that in the third stage, the level of trust of all tested objects returned to initial values. However, in none of the studied objects speci c to the second stage, the experimental group showed a signi cantly different level of trust from the participants of the control group in either the second and third stage. The analyzed experimentation demonstrated that the level of trust in physicians, the payer, and hospitals appeared to be sensitive to submitted information while trust in the medical profession did not. As a result, the research may prove that information, not in line with the mainstream opinion (i.e., projected from the media) could have a signi cant impact. The study, to some extent, also con rmed H1 and H3 that trust in the payer is more vulnerable than trust in hospitals. Furthermore, attention should focalize on the requirements that decrease the level of trust in a physician, as indicated in Fig. 1. Compared to hospitals, the change is relatively large and in the opposite direction. Taking a closer look at H2, this result was not expected since rst, there was no particular information aimed at in uencing trust in a physician and second, the information package was designed to show Poland's healthcare in a more favorable light. Hence, the present research has attested that social trust in the payer and hospitals may be strongly related to trust in a physician as many researchers claim [43] [19] [38] but this relationship may also be reversely directed.
Due to the unexpected results concerning the change of trust in physicians, per se, one week after the last stage of the experiment, a meeting was organized with the experimental group to discuss the results. Some participants suggested (i.e., and some agreed with the suggestion) that after they saw the information, they felt less vulnerable and dependent on their physician then before. Earlier accounts had the majority of them emphasizing their physician was working against the de ciencies of the system. After re ecting on the information package after the results were computed they were more convinced that other elements of "the healthcare chain" work correctly, they felt more secure, and they did not need to trust so much in their physician.
The ndings from the study are similar to [8] conclusion in which "the greater the sense of vulnerability the higher the potential for trust" and to some extent to [22] suggestion that patients who trust their physician may worry less about the performance of other healthcare components. When people perceive healthcare performance as mediocre, they rely more on their physician (i.e., reciprocal with a higher level of trust). But when they realize that other healthcare components work properly, their sense of vulnerability decreases and consequently, they lose trust in physicians who supposedly compensate for the de ciencies of other healthcare components. To some extent, this reasoning might be supported by the fact that under the in uence of information, i.e., trust level in a physician, meets trust level in hospitals and approaches trust in the medical profession as well as by the payer. This phenomenon may be interpreted in a way that participants decrease their trust in a physician because they realize their physicians are not bearing exceptional efforts to organize treatment for patients, but operate in an interdependent environment and are equally important as other components of the system.
Limitations to this study invites further research to examine intervention level, i.e., it is likely it would not affect all objects in the same manner-distorting the results. Still, as the changes in the payer, physicians, and hospitals reported, these objects were assumed susceptible to delivered information. On the contrary, the lack of change in trust in the medical profession may suggest that the object is vulnerable to delivered information but that the intervention may not have been adequate (i.e., or adequately scaled).
Scales used in the study were not tested on the general society Poland-wide, therefore cultural or organizational differences in healthcare between Poland and the US might have affected the outcome.
The main conclusions point towards the ndings that information can signi cantly change people's trust in some components of healthcare regardless of their real performance. This allows for a number of additional inferences to be made. First, trust in healthcare may strongly correlate with the atmosphere in mass media. Second, patient trust in a physician (i.e., at the interpersonal trust level) and social trust in the payer as well as in hospitals may be interrelated also in opposite directions. This means that the increase of trust in hospitals and the payer may correlate with the decrease of trust in a physician. Third, the assessment of healthcare performance [59] based on trust surveys might be misleading, since any change in trust level may not necessarily translate into an immediate modi cation (i.e., need) of healthcare functionality. Fourth, delivering designed information is likely to in uence the perception of healthcare performance. Fifth, the change in trust level may not be durable. Finally, considering previous studies [40] [39], future change in trust is more likely to occur under the in uence of information rather than after a genuine change in healthcare performance indicating a long-term conclusiveness even within the bounds of the study's limitations. In retrospect of the current coronavirus pandemic, these conclusions could be applied to restore trust in healthcare during and after it reaches its end. As such, restoring trust in a medical setting is a contemporary concern that countries alike are and will need to deal with in an ever so changing global health response.

Con ict of interest statement
No con ict of interest has been declared by the authors.

Funding
Collected data was partially funded from previous research conducted during the COST Action CA15222 "European Network for cost containment and improved quality of care" and National Science Centre, Poland (grant number: 2015/17/B/HS4/02747).

Authors Contributions
Conceptualization, Data curation, Methodology, Validation, Formal analysis and Writing-Original Draft preparation, A.G.G. and R.L.; Investigation, Resources, Writing-Review and Editing, A.G.G., R.L., and G.T.C.; Project administration, A.G.G. All authors have read and approved the manuscript.