The predominance of mechanistic evidence production, review, and synthesis is not appropriate in unstable situations, such as during times of natural disruptions (e.g., COVID-19) and/or rapid technological developments (e.g., digital transformation). Attempts to cope with these new events (e.g., modeling, rapid reviews, living guidelines) can be interpreted from an organizational sociologist viewpoint as attempts to shift from mechanistic knowledge production, review, and transfer to an organic form of knowledge processing.

The organic turn in knowledge production, review, and distribution is one of the reasons for the return of the expert during times of crisis because expertise is a crucial element of organic organizations. Experts have become central not only because expert advice is readily available and trusted during times of insecurity but also because expert knowledge serves hidden functions that were lost in the decades of evidence-based healthcare.

There are several definitions of the term “expert”. In this paper, an expert is defined as a person who possesses a specialist knowledge (48). An expert uses technical skills, and is—ideally—impartial (48). An important characteristic of experts is that they “mediate between the production of knowledge and its application; they define and interpret situations; and they set priorities for action” (48).

The original idea of the evidence-based movement was to retire the experts (49, 50) and to uncover their sin (51).

In recent years, and especially regarding COVID-19-related policy consulting, many observers have noticed a shift. Despite the anti-expert sentiment that exists in the general public (52), experts and their stockpiles of knowledge are increasingly in demand (53), and there is an impression that they are emerging from “retirement” (8, 54). Advocates of evidence-based medicine are therefore increasingly asking how the role of expert consultations for public policy should be interpreted and classified (55, 56). The issue is that at the onset of new, unexpected health-threatening events, there is typically no high-quality external evidence because the necessary data is lacking (55, 56). However, there is also an urgent need for action (57). Experts have therefore filled this evidence gap.

The need for experts remains even after implementing the accelerating tools because the problem of time-consuming evidence-based medicine persists, even if it is somewhat alleviated through rapid knowledge processing. Sound knowledge production still takes time. Even fast reviewing is too slow for politicians facing instability and rapid change. Following the formal procedures, the development of clinical guidelines still requires several months (43). The main obstacle of evidence-based science lies in the core of the knowledge generation process: the production of sound knowledge in scientific studies. It is extremely difficult to speed up studies. Time constraints are relevant during stable and unstable periods because in both cases researchers must wait to measure the effects of an intervention (“impact time”). Impact time is the time from the onset of an intervention until the desired outcome emerges. As an example of the problem posed by impact time, it takes more than 15 years to plan, conduct, and publish an RCT on colorectal carcinoma screening (58). Thus, in times of rapid change where there is no time to await the end of the impact time rapid advices of experts are necessary.

The organic turn in knowledge production, review, and transfer favors experts because experts fulfill additional functions beyond providing evidence-based knowledge. These additional functions include the provision of theoretical knowledge, experiential knowledge, and knowledge regarding the context of the planned interventions (see Table 3). The probability of high-quality decisions in health policy is highest if the decisions are transparently made based on the best knowledge available in all four knowledge domains: (1) evidence-, (2) theory-, (3) experience-, and (4) context-based knowledge. Using all four knowledge domains results in knowledge-based health policy (Figure 1). Knowledge-based health policy means to make health policy decisions based on (a) the best available external evidence from systematic research and sound mathematical modeling, (b) the best available theoretical knowledge, (c) the best experiential knowledge available, and (d) the best context-specific knowledge. Experts and policymakers should combine and synthesize these four knowledge components to arrive at balanced decisions in healthcare and health policy (see Table 3). The process of synthesizing and applying knowledge should be documented for transparency. Due to their ability to combine their tacit and experiential knowledge with other knowledge components, such as evidence-based knowledge and contextual knowledge, trusted experts play a central role in the process of knowledge transfer (59–64).

To support decisionmakers in the decision-making process, experts and decision-supporting institutions should not only provide knowledge form these four knowledge domains (48, 65, 66). They should also provide knowledge synthesis, reflexive thinking and timely advice. Thus, we are able to identify seven categories of advices decision supporters can give to policymakers. Table 3 lists the seven functions that experts should engage in to maximize the quality and timeliness of policy decisions; this list is not exhaustive, but it contains the most important functions for creating sound and balanced knowledge-based health policy.

The first function of decision support (i.e., expert support) is to provide decisionmakers with the best available evidence-based knowledge by searching for, reviewing, and classifying clinical and public health research and then transferring this knowledge to decisionmakers (59–63). Knowledge is derived from observational studies and studies with quasi-experimental and experimental designs as well as from mathematical modeling (67–69).

The second function of decision support is to provide the best available theoretical knowledge pertaining to the decision to be made. Theoretical knowledge involves using established theories or models as a guide to health policy strategies. Lewin summarized the practical usefulness of theoretical knowledge with the quote “there is nothing so practical as a good theory” (70). Theoretical knowledge is one of the central domains of experts, and experts can provide theories that guide policymakers (71, 72).

The third function of experts is the provision of experience-based knowledge relevant to the decision. Experiential knowledge denotes the proficiency and judgement that experts, advising institutions, or decisionmakers acquire through experience and practice. Expert knowledge comprises—among other forms of knowledge—experience-based tacit knowledge and experience-based intuition (61, 73–75).

The fourth function of expert advice is to provide context-based knowledge to tailor evidence-based knowledge to the situation. Context-based knowledge pertains to the physical-chemical, biological, and social situations and possible changes in those situations (18, 76, 77). One important aspect of the context is knowledge of the processes, legal frames, and rules that have to be considered in decision-making (72).

The fifth function of decision support is to synthesize evidence-based, theoretical, experiential, and contextual knowledge and to adapt it to the decisionmakers (48). Experts can provide scientific information, convince the decisionmaker to consider a single best option, or adapt scientific information to decisionmakers preferences to reduce the choices provided (78).

The sixth function of decision support is to stimulate reflexive thinking (79). Reflexive thinking involves evaluating the direct and indirect costs and the “unanticipated consequences” (80) of the decisions made. In addition, reflexive thinking means learning about contextual nuances and possible interactions between the intervention and the context (77, 81). From the organizational learning perspective (82, 83), reflexive thinking involves single-, double-, triple-, and quadruple-loop learning (82, 84–86) as well as unlearning processes (87). All these learning types can be found on the individual, collective, and organizational level. Single-loop learning is trial-and-error learning without questioning the policies, basic assumptions, and goals underlying the trial-and-error actions, while double-loop learning involves questioning these factors (82). Triple-loop learning means learning about single- and double-loop learning (88). Triple-loop learning also involves building a learning infrastructure that connects local clusters of learning (89). Quadruple-loop learning involves continuous, context-specific learning to cope with uncertainty and complexity by revising, redefining, and expanding triple-loop learning (90). Articles on the lessons learned from the COVID-19 pandemic are examples of reflexive thinking. In a recent article comparing US and South Korean pandemic strategies, researchers demonstrated that single- and double-loop learning were essential to cope successfully with the COVID-19 pandemic (91).

The seventh function of decision support is to provide the required knowledge in a timely and easy-to-understand fashion (92, 93). Individual experts or expert groups represent the most agile knowledge sources. They are flexible and accessible and they are able to combine all four knowledge components.

As Table 3 shows, the established evidence-based medicine (EbM) institutions cover the most important function of providing empirical evidence. Some institutions derive recommendations for clinical practice based on this evidence, often using the Grading of Recommendations Assessment, Development and Evaluation process (GRADE) as a framework (94). However, these institutions completely fulfill only one of the functions required to guarantee high-quality decisions. The other functions are not or only incompletely fulfilled. In a world of rapid medical change, this gap is increasingly being filled by experts. In sum, expert knowledge has four dimensions: evidence-, theory-, experience-, and context-based knowledge (Figure 1). Ideally, experts or expert groups have sufficient theoretical knowledge, are up to date with current evidence regarding the given problem, and have enough experience to use their context-specific knowledge to integrate all four aspects of knowledge and to apply them to the situation at hand. In short, the advantage of evidence-based experts and expert groups compared to evidence-based government institutions is the use of four-dimensional knowledge that can be applied to a specific situation in a timely and reflexive manner.

The third lesson indicates that integrating theory-based knowledge and evidence-based knowledge is a necessary component of the organic turn. Because this integration poses a special challenge, it is necessary to examine this task more closely.

Scholars who study the use of theory in health services research and especially in implementation research have stated that there is an underutilization of theory in health services research in general and in implementation science in particular (95); this is already a problem in normal times, but in rapidly changing times, this underutilization could contribute to disorientation in health policy with regard to the right starting-points, the right direction and the right plans.

When there is no evidence, as is the case at the onset of new situations, established theories can fill the knowledge gap and be used to advise health policy by indicating the starting-points for plans and measures. When there is an exponential growth in evidence, a lack of theories can hinder (a) the meaningful integration of existing evidence-based knowledge elements into a broader picture, (b) the explanation and prediction of complex phenomena, and (c) the guidance of evidence-based research.

Hence, the question arises as to why theories play no official role in evidence-based health policy and evidence-based healthcare; one possible reason for this is that there is a lack of quality grading for theories comparable to the quality grading of empirical studies in evidence-based healthcare. Therefore, we propose grading the quality of theories on a meta-level. With meta-level we mean assessing theories not by the quality of the content, but rather by formal criteria, like the spread and acceptance of theory in science.

There are different gradings for empirical evidence (96). However, to our knowledge, no clear-cut, standard, quality-oriented ranking of theories exists that differentiates between “low quality” and “high quality” theories similar to the hierarchical levels in evidence-based medicine. The quality of theories is an important topic in the areas of artificial intelligence, machine learning, and big data (97–99), but it is less discussed in health services research, though there are a few exceptions to this, such as in the realm of theories of health behavior changes (100, 101).

In general, there is no consensus in the literature on how to define and assess the quality of a theory. The measure of quality depends on the paradigm used, such as positivist, post-positivist, critical theory, or constructivist paradigm (97, 102, 103). To overcome this problem, we investigated the quality of theories from two formal perspectives. The first perspective is the post-positivist perspective, which has the advantage of being compatible with the positivist approach of evidence-based medicine but also allows the inclusion of qualitative results and probabilistic hypotheses (102–105). The second perspective is the constructivist perspective, which has the advantage of considering the process of the social construction of knowledge in a scientific community (106–109). Based on these two perspectives, we propose the use of two criteria to grade the quality of a theory: (1) acceptance of the theory in the scientific community and (2) empirical confirmation of the theory or parts of the theory. From a constructivist point of view, the acceptance of a theory in the scientific community, which can be measured approximately for example by how often the theory is cited in the literature (98) or by how often the theory is used in the scientific community (95, 109), is an indicator of the intersubjective quality and usefulness of the theory. From the post-positivist point of view, the quality of a theory depends on the successful testing of the theory or parts of the theory (e.g., single hypothesis) in empirical, evidence-based studies.

To apply these two criteria to theories in the health sciences, we propose the grading of theories as shown in the “scientific knowledge triangle” in Figure 2. The formal, meta-analytical hierarchy of theories on the left side of the triangle complements the evidence-based knowledge hierarchy that comprises the right side of the triangle, thereby creating the scientific knowledge triangle.

We used the grades of evidence listed by Shekelle et al. (110) to grade evidence-based knowledge. In the following section, we focus on the left side of this scientific knowledge triangle.

The base of the scientific knowledge triangle is the opinions of experts, which are not evidence-based and do not have a theoretical basis or justification. Expert opinions alone are classified as having little or no recommendation strength. Two sides emanate from this “zero base”: the evidence-based side on the right and the theory-based side on the left.

Based on the identified criteria for ranking the formal quality of a theory, the first level of theoretical knowledge is achieved if there is a systematic, internally consistent theory, model, framework, or concept that has not yet been accepted by a relevant part of the scientific community and has not been empirically confirmed (level 3).

The next formal quality level of a theory, level 2b, is reached if the theoretical knowledge is based on a systematic, internally consistent theory, model, framework, or concept that has been accepted by a relevant part of the scientific community but has not been empirically confirmed; an example of this is the sociological systems theory (18, 111).

Because of the preference for empirical evidence over consensus, the next formal quality level of theoretical knowledge (level 2a) is attained when there is a systematic, internally consistent theory, model, framework, or concept that has not been accepted by a relevant part of the scientific community but has been empirically confirmed (level 2a in Figure 2). An example of this is the use of social learning theory (112, 113) within implementation science, where this theory is not really broadly accepted or used (95).

The highest level of theoretical knowledge (level 1) is attained when a systematic, internally consistent theory, model, framework, or concept has been accepted by a relevant part of the scientific community and has been empirically confirmed (level 1 in Figure 2); examples of this are the theory of planned behavior in the science of behavior change (100, 114–116) and in management science (98).

The evidence-based approach would benefit from the addition of a theory-driven approach. In the various fields of medicine, public health, psychology, and the social sciences, the overarching goal should be to implement both forms of knowledge generation to better explain and understand reality. Combining evidence-based and theory-based knowledge creates a body of scientific understanding that can inform health policy in a sound and balanced way.

The relationship between the two types of knowledge is reciprocal. Theory can inform evidence-based knowledge and vice versa. For example, one of the most important functions of theories is to guide the direction and process of empirical research (orientation function), such as by planning experiments in a theory-based manner from the outset (117) or by using logic models (118). Another path of cross-fertilization runs from evidence-based knowledge to theory-based knowledge, which is also known as the inductive approach. A classic form of empirical theory-building is given in some forms of qualitative research (119). The third path of cross-fertilization is the classic approach to build a theory by testing parts of the theory with empirical research. If a trial confirms a hypothesis, the theory-based knowledge is further supported by empirical work with reliable evidence. However, if the RCT does not confirm an important hypothesis of the theory, this represents a classic example of the falsification of a hypothesis (120) that puts the quality of the theory used into question.