From Theory to Practice: Viewpoint on Economic Indicators for Trust in Digital Health

Introduction

Public trust is a cornerstone of successful health systems [1,2]. Public trust is a key enabler for the introduction of trustworthy health data spaces for sharing and using health data in research, national electronic health record systems and the introduction of artificial intelligence (AI) systems in medicine. In 2023, the World Health Organization (WHO) in the European Region stated that 77% of its Member States believe the lack of trustworthy sources to access effective medical apps is a substantial implementation barrier to digital health transformation: “Trust is paramount – trust in technology, trust in data protection and, most critically, trust in the respect and safeguarding of privacy rights pertaining to personal health data” [3]. Trust in public health systems legitimizes these activities. Public trust is predicated on the belief that those systems provide a net benefit, where benefits include individual benefit, societal benefit, and health system benefit [4]. The increasing discourse on trust and trustworthiness of digital health systems interventions (DHI) has been met with an increase in nonbinding professional guidelines and codes of conduct. These guidelines are designed to guide users toward the ethical and trustworthy use of DHIs. Adhering to such guidelines can help foster public trust which, in turn, can foster their practical adoption. Some purposefully sampled guidelines in the area of AI as a prominent example of DHIs are highlighted in Table 1. The guidelines were selected to make sure national and international as well as government and nongovernment organizations were represented but we fully acknowledge that these reflect only a subset of the guidelines that exist.

Such guidelines cover a range of trust-building principles, for example, the need to implement accountability mechanisms, security, or privacy measures, the importance of public and user-centered communication strategies, and the appropriate implementation of the regulation. One example is the Digital Trust Framework published by the World Economic Forum [6] which identifies principles along 8 dimensions: cybersecurity, safety, transparency, interoperability, auditability, redressability, fairness, and privacy. These principles aim to foster security and reliability, accountability, and oversight, as well as inclusive, ethical, and responsible data use for DHIs. The guiding principles are usually at an abstraction level that allows transferability across different contexts, necessitating careful tailoring during the implementation process. In practice, however, resources are limited. How do we determine which trust-building principles to prioritize, and which digital health policies should take precedence, especially when guidelines conflict? These guidelines are typically silent on the economic implications of trust-building, which could provide guidance on the costs and benefits of implementation. A clear need for economic guidance persists. This observation is underscored by anecdotal observations made by FG in Switzerland in the past few years during workshops and meetings discussing trust in DHIs.

After attending an expert workshop aiming at identifying a list of trust-building principles comprising a trustworthy data space, a fellow participant asked: “This list looks convincing, but I am the only IT (Information Technology) security employee in a small firm, we have limited resources, so which of the principles are cost-effective?” (Workshop on trustworthy behavior, Bern, Switzerland)

While exploring a research collaboration with an industry partner where we aimed to test the performance of a nonbinding trustworthiness codex, the industry counterpart asked: “If our company adheres to a non-binding trustworthiness codex, isn’t there a risk that we face a competitive disadvantage compared to our competitors that do not commit to these measures but still adhere to the present legislation?” (Business meeting on the performance of a nonbinding trustworthiness codex, online, Switzerland).

Discussing trustworthiness principles to conceptualize Digital Trust, workshop members from the industry suggested that what we have today is robust conceptual evidence, but we clearly lack implementation experience and evidence showing what trust-building measures are costeffective. (Workshop exploring digital trust, Zurich, Switzerland).

When it comes to the implementation of trust-building measures and the economic evaluation of their effectiveness, we stumble upon a yawning knowledge gap. Three important unanswered questions require our attention: What are the costs of implementing trustworthiness guidelines? With limited resources at hand, which measures are cost effective? What instruments exist to gauge success?

We argue that the failure to address these questions risks undermining both the implementation of trust-building guidelines and their success. What is needed is a clear understanding of trust linked to a robust economic assessment of trust building activities. Furthermore, any economic assessment of the costs and benefits should acknowledge that different stakeholders, such as investors, industry participants, policy makers, or end users, will prioritize these impacts differently.

In this article, we offer direction on how to close this gap by improving the conceptual understanding of trust, implementing actionable trust performance indicators, and sharpening the trust discourse. While the importance of trust in DHI is commonly recognized, we argue that we still need more evidence-based policies to ensure the trustworthiness of these programs are implemented efficiently and effectively. With this viewpoint, we aim to initiate further development of economic thinking in trust building. More specifically we aim to transfer and adapt knowledge from economic research to other domains where trust building matters, but economic considerations have not yet entered the professional discourse and implementation practice of trust building activities.

Shortcomings and Remedies

Trust has been part of economic research for centuries [11]. Ample research exists investigating the relationship between trust and societal prosperity [12], the causal impact of trust on economic development [13,14], and how trust lowers transaction costs [15]. As Arrow [16] stated in 1972, “Virtually every commercial transaction has within itself an element of trust, certainly any transaction conducted over a period of time. It can be plausibly argued that much of the economic backwardness in the world can be explained by the lack of mutual confidence…”

While the existing research provides arguments for the economic incentives for trust-building, the costs and benefits of trust-building in digital health have not been well studied. In some emerging economic evaluation guidelines tailored to DHIs, trust can be a component of an impact inventory in user experiences or considered as a non–health-related impact [17]. Yet, trust as an outcome variable is not a well-established component of economic evaluation of digital health. We argue that the lack of focus on the economics of trust-building in this domain include conceptual and methodological challenges that need to be addressed.

We Need Conceptual Clarity of Trust in Digital Health Interventions as a Prerequisite for Economic Evaluation

Embedded in wider sociopolitical as well as cultural norms and values, the concept of trust is complex and influenced by the context in which it evolves. No commonly accepted definition of trust exists. We suggest that trust in a person or institution is a belief about the trustworthiness of that individual or entity. It is a belief that makes one willing to take a risk and make oneself vulnerable, to rely on that person or institution for one’s future benefit. The trustworthiness trait is defined as having characteristics that are conducive to cooperation. What traits count as trustworthy differ by the situation. Still, we can categorize them as either traits relating to competence (the ability to fulfil the tasks expected of them) and traits related to integrity, the commitment to prioritize the best interests of the other [11,18].

Trust is volatile, and while the set of trust-building principles might remain unchanged over time, the relative importance of individual principles can vary depending on the context. In particular, trust repair after health system scandals or negative shifts in societal sentiments toward the government require different approaches [19]. The trait of trustworthiness, often the main target of guidelines, should not be conflated with trust. Trustworthiness is a precondition of trust-building but not a guarantee for trust. Trust-building can be based on calculated decisions and rational thinking [20]. At the same time, a body of literature highlights the importance of emotions in building trust [21]. Combining calculated and emotional motivations to establish trust broadens the spectrum of possible measures available to policy makers and health system actors. Trust research and guidelines in the digital health domain tend to neglect the emotional motivations of trust-building.

Many guidelines contain a comprehensive portfolio of trust-building principles which reflect the overall complexity of the topic but also reflect the lack evidence about which principles work best in practice. Without considering economic constraints and trade-offs, the guidelines may be biased toward presenting too many trust-building principles that are too broad for fear of missing out, with too little guidance on how to prioritize them. As a result, the guidelines may be perceived by users as overwhelming due to the number of principles included or to their high abstractness, with too much onus placed on the implementer to figure out how to navigate the guidelines.

On the receiving end of the guidelines, we see discussions emerging around which organizational units are most suitable to take on the implementation task. The broad spectrum of trust-building principles, from user centered communication and IT security measures to adherence to the rule of law, places responsibility on different organizational units to lead the implementation of guidelines in areas as diverse as public relations, law, compliance, and IT services. Table 2 provides handpicked examples of specific trust-building activities recommended by the guidelines presented in Table 1 that we categorize about accountability for past mistakes, transparency about the present, and promises and benchmarks for the future [11].

The intersection between digital health and trust is somewhat consistent regarding accountability for past mistakes and transparency about the present. Benchmarks for the future are less well-articulated across specific trust building activities [7-9], or their scope is limited.

The choice of who within an organization is responsible for the management and implementation of specific guidelines has implications for how financial resources get allocated and how the guidelines get interpreted. A joint effort across different units may be preferable as different trust-building principles require collaboration between different organizational units. Research suggests that a collaborative approach involving a range of actors across the organization enhances trust-building [22].

In response to these requirements, we need implementation evidence. Systematically collected evidence, case studies, and experience reports can inform how resources are allocated for different trust-building principles and identify successful trust-building mechanisms in a given context. Given the context specificity of trust, this evidence will not be a blueprint that can be copied and pasted across contexts. Instead, the evidence can help disseminate experience from comparable settings and is an essential step toward better informed resource allocation.

We Need Actionable Trust Performance Indicators to Collect Meaningful Evidence

To gather better evidence to inform resource allocation and make judgments about cost effectiveness, we need appropriate methods to collect data about trust-building from the implementation to the evaluation phases. The standard methodology to gather evidence about trust in the context of health care and medicine is survey instruments. Unfortunately, studies have repeatedly shown that the existing instruments have weaknesses. Goudge and Gilson [23] called for advancements in the quantification of trust; Ozawa and Sripad [24] showed in their review of 45 trust measures in health systems that half of the measures used qualitative methods in the design process, and only 33% were pilot tested; Aboueid et al [25] found in a meta-analysis of 26 publications on trust in health professionals and health systems, 10 studies did not mention the dimensions explored and of those that did, some did not define them. Taylor and colleagues [1] call for conceptual clarity and methodological creativity after reviewing 50 years of trust research. Considering the stagnation in methodological development in the field and the increasing inclusion of guidelines regarding trust in digital health, we see a need for refined methods to address questions of cost effectiveness and resource allocation.

An understanding of the resources invested, and outcomes gained is essential. Since we cannot pin a price tag on trust, we need to agree on sensible indicators that we understand as evidence informed. Table 3 gives an overview of currently applied indicators regarding trust in DHIs that have been developed by some of the essential stakeholders in the field, such as the World Bank or WHO. These indicators are the result of a systematic search on indicators to measure digital health system maturity conducted by Maaß et al [26]. The review identified that trust toward digital health applications, AI, the government, and health care providers plays an essential role among assessment tools for digital health system maturity. However, the indicators rarely reported their methodology and data source, thereby limiting their applicability. We require well designed indicators with transparent methodology and a clear definition on what is meant by “trust.”

Furthermore, most of the guidelines we identified and discussed in Tables 1 and 2 are not published or sponsored by governments or international institutions, raising the question who’s interest it actually is to understand the economics of trust-building in digital health? Without appropriate performance indicators that show whether trust-building principles are successful, we cannot evaluate trust building mechanisms or make decisions about prioritization. Indicators are considered instrumental in driving health system improvements and for running comparative evaluations across the health system [38,39]. DHI investors and developers have good economic incentives to adopt trust performance indicators as it allows them to make better informed decisions about resource allocation for trust building and allows them to better monitor whether trust building actions are successful. Indicator sets are based on robust conceptual frameworks “that set out the rationale and design principles for an indicator set” [40] with the intention to provide standardized benchmarks for the systematic evaluation and monitoring of quality, performance, and outcomes. Addressing company executives, Deloitte argues that trust itself should be a key performance indicator for companies, as drops in consumer trust can considerably erode the market value of the firm [22]. Van der Schee et al [41] suggest public trust in the health system is an important performance indicator for assessing national health systems. Several other examples of indicators in specific areas of the health system exist, such as in digital health care, which include indicators relating to trust at least at the margins [27,28,30,31,33,35-37,42-46]. Discussion is currently centered around identifying useful performance indicators for trust in DHIs and determining the optimal methods for evidence collection. The crucial task is to navigate conceptual complexity and to accommodate context specificity, ensuring the identification of indicators that are suitable for a diverse range of DHIs while aligning with trust-building guidelines.

Survey measures of trust levels can indicate changing trust levels over time [24]. Such measures are useful for assessing the level of trust before and after implementing trust-building principles. Similarly, qualitative user interviews and broader citizen engagement can provide rich information about perceptions and experiences regarding the trustworthiness of a DHI and the effectiveness of its associated trust-building measures [47]. The use of routinely collected data generated from digital health technologies needs to be harnessed and leveraged for better decision-making around trust-building and should include measures of the user experience at different levels of the system. The evidence should inform both the design phase of trustworthy DHIs and for continual assessment of their safety, effectiveness, and quality assurance [48]. This potpourri of methods allows us to think creatively about meaningful performance indicators for various contexts, actors, and DHIs. A comparable example of a newly developed indicator set is the Digital Public Health Maturity Index [49]. This index provides initial indications of possible architecture for an actionable and stakeholder-supported set of performance indicators in an area that faces the same kinds of problems as trust. Developing a trust index and performance indicators with appropriate methods to collect evidence must be integral to future implementation research on trust-building in digital health.

We Need to Sharpen the Trust Discourse

The increased use of the word “trust” by politicians, industry, and researchers alike raises the question of whether the use of the word trust is always justified. The inflationary use of trust in the public sphere, as seen in commercials, labelling, news articles, and media campaigns, possibly undermines trust’s value in settings where trust discussions are essential [50]. If we are not cautious, we may see the trust debate in the digital health sphere drift into symbolic actions and “trust washing,” similar to experiences in “green washing” and “ethics washing” [51]. Trust degenerates into a hollow marketing term when health system actors invest resources merely to chase trends without meaningfully engaging with internal trust-building procedures and actions leading to overall trustworthiness.

Trust is vital for adopting DHIs, but it is not a panacea. Health system actors can act to the best of their knowledge, and policy makers can design evidence-informed policy, including incorporating trustworthiness guidelines in their policy designs, but that does not mean the public or users will trust us [52]. Trust cannot be enforced or demanded. The public and the users decide how to assess trustworthiness and where to place their trust [53]. That insight implies that investments in trust-building in the context of DHIs carry the risk of failure.

Even if trust building is successful, it may not result in material benefit. For example, if several alternative health system actors adhere to trustworthiness guidelines and are all judged equally trustworthy, users will likely make choices not based on trust but on convenience, price, accessibility, or other adoption enablers. Trust is not the only reason we engage with DHIs. We might depend on a DHI and, consequently, engage with it because we have no other choice. We might assess the risk of misplaced trust and its associated consequences to be low and, therefore, not really care about trust. We might carryover the trust we have for more comprehensive health system to new DHIs, allowing us to spontaneously engage with the new and unfamiliar, similar to the concept of spontaneous sociability [12]. Such considerations are essential and require those aiming to build trust to assess the context carefully and to understand user motivations.

Realistic expectations about the capabilities of trust-building are critical in deciding which trust-building measures to adopt and how best to efficiently deploy the necessary resources.

Conclusions

Meaningful guidance for trust-building in digital health must go beyond proposing principles to build trust. We argue that the costs and benefits of trust-building measures are currently missing to a large extent. Health system actors and the research community need trust performance indicators to appropriately collect evidence about resources invested and trust gained in return. We need to be mindful of the present inflated use of the term “trust” in the public sphere when designing new trust building efforts, and to acknowledge that trust cannot be guaranteed. To successfully navigate the complexity of trust-building, we need more engagement in the economics of trust-building in digital health. Economics of trust-building needs to be a deeply integrated and structured part of DHI policy making and health system governance at large to develop sustainable trust building processes while bearing in mind resource scarcity. Furthermore, stakeholders across the health system from regulators, to health organizations, to end users need to have guided access to relevant and comprehensive information about these guidelines in order for meaningful progress to be made.