Background

Systems Medicine recently became a buzz word in debates on biomedical research and future health care [1, 2]. However, there is no consensus on the term’s meaning. Currently, five main positions seem to dominate the discussion: (a) Systems Medicine is the successor of Personalized Medicine [3, 4]; (b) Systems Medicine is a precursor of Personalized Medicine or P4-Medicine [3, 5]; (c) Systems Medicine is an equivalent term for Precision Medicine [4]; (d) Systems Medicine means the translation of Systems Biology into medical practice [3, 5,6,7]; and (e) Systems Medicine is an “assemblage of scientific strategies and practices that include bioinformatics approaches to human biology […]; ‘big data’ statistical analysis; and medical informatics tools” [8].

This divergence in understanding is, however, problematic as it complicates sound debates on chances, risks and limits of Systems Medicine. Moreover, it is difficult to identify and solve ethical as well as legal issues that could arise in the context of Systems Medicine. As a consequence, it is impossible to discuss questions of its matter and (future) handling. Finally, the term’s underspecification may lead to unfounded hopes of patients, for instance, regarding the possibilities of Systems Medicine approaches [6, 9]. Against this background, it becomes clear why it is still seen as one of the key challenges to define Systems Medicine [6].

The goal of this paper is to develop a sufficiently precise, formally adequate definition of Systems Medicine. Hurley [10] differentiates, among others, between stipulative, lexical and précising definitions. While a lexical definition captures the way a word is commonly used, a stipulative definition arbitrarily assigns a meaning to a certain expression, whereas a précising definition tries to reduce the vagueness of a term used in practice [11]. Hence, a précising definition is geared to the everyday usage of a term while, at the same time, it aims at standardizing, harmonizing and structuring the different occurring meanings in everyday usage with regard to formal criteria for adequate definitions.

As the term Systems Medicine is already established in everyday usage, however divergently understood, in the following we develop a précising definition. For this purpose, first, the usage of Systems Medicine in the relevant scientific literature was analyzed and, second, harmonized by appealing to formal criteria in order to develop a formally adequate definition of Systems Medicine.

Methods

The methods presented here are based on the approach developed in [16]. Differences in methods result from taking into account applicable critiques of the original account [41].

In order to identify definitions in the relevant scientific publications on Systems Medicine and provide an overall picture of the research field, we followed the approach of a systematic literature review [12, 13]. We therefore searched PubMed using the keyword “systems medicine”. We focused on PubMed as we were interested in definitions of Systems Medicine virulent in the scientific context. We did not include MeSH-terms in the search strategy as the thesaurus of the US National Library of Medicine relates “systems medicine” to the term “system analysis”, which is defined as “[t]he analysis of an activity, procedure, method, technique, or business to determine what must be accomplished and how the necessary operations may best be accomplished”.Footnote 2 Including MeSH-terms would therefore have pre-selected certain articles according to this understanding of systems analysis. Excluding MeSH-terms, on the contrary, allowed us to stay open to alternative understandings of Systems Medicine. We only searched titles and abstracts to identify those articles in which Systems Medicine is the main focus. Moreover, we assumed that those articles were more likely to contain a definition. We did not restrict the date of publication; our last search was performed on December 31, 2015. Furthermore, we included only articles written in English as our goal was to capture the international debate. Subsequently, we checked full-text availability of the articles identified. Where full-texts were not available, we contacted the respective authors.

Step 1: Description of the areas of application of systems medicine

To describe the areas of application of Systems Medicine, in a first step, all papers fitting our search strategy were assigned to deductively developed main categories based on general presuppositions regarding content alignment of the papers. These main categories are: 1) with reference to disease and 2) without reference to disease. This distinction was made in order to evaluate the current state of clinical application of Systems Medicine. Several subcategories of category 1) were developed inductively based on specific findings from the papers to ensure representing the whole spectrum of diseases dealt with in the articles. Due to the inductive approach, the classification of the diseases was mainly based on terms and references used in the articles and does not refer to a specific disease ontology. However, by categorizing the specific diseases we generally distinguish between chronic, non-cancer diseases and hematological and solid cancers as well as various (unspecified) forms of disease in order to examine the areas of application of Systems Medicine. In main category 2), those papers were included which have no reference to any disease or context of disease.

The first categorization showed that papers with as well as without reference to disease can be further specified with regard to their content alignment: they are either research-related or programmatic articles, i.e. both types of papers only partially relate to a specific context of disease. Accordingly, in a second step, all papers were assigned to the inductively developed main categories A) research context and B) programmatic context based on their content alignment. Category A) includes papers which present own findings or (new) methods in the fields of basic research, clinical research or translational research and/or discuss results of other research projects in these fields. Basic research was defined as research conducted to increase fundamental knowledge and understanding of physical, chemical or functional mechanisms of life processes and diseases, thereby providing the foundation for clinical research. Clinical research was understood as patient-oriented research, conducted with human subjects (or on material of human origin) including inter alia research on mechanisms of human disease, therapeutic interventions, and the development of new technologies as well as epidemiologic and behavioral studies or health services research. Translational research fosters the multidirectional integration of basic research, clinical research, and population-based research applying basic research to human subjects and moving discoveries and knowledge into initial clinical testing. It can be described as mechanism-oriented clinical research that may include laboratory-based research aimed at clarifying mechanisms of disease, developing drugs etc. [14, 15]. Category B) contains papers which focus on the description of general potentials and challenges as well as expectations and future prospects of Systems Medicine. These include visions of a better or more comprehensive disease understanding and management as well as visions of general possibilities for translation and implementation of Systems Medicine in clinical practice.

In a third step, categories 1) and 2) were crossed with categories A) and B), in order to evaluate the current state of research on Systems Medicine in general as well as in specific contexts of disease.

In a final step, the temporal distribution of publications within the categories 1), 2), A) and B) was examined. The aim thereby was to determine possible trends over time within scientific literature regarding Systems Medicine’s areas of application.

Step 2: Reconstruction of current systems medicine definitions

In order to reconstruct the usage of defining Systems Medicine, two researchers, Christoph Schickhardt and Sebastian Schleidgen, independently screened full-texts of the papers fitting our search strategy for definitions. Included were not only explicit definitions of Systems Medicine, but also text passages containing definition-like descriptions or characterizations of Systems Medicine. In case their assessments differed, discrepancies were discussed and resolved consensually (thereby ensuring inter-coder reliability). Subsequently, as a prerequisite for developing a précising definition of Systems Medicine, a data extraction tabloid was developed which puts forward a means/ends-distinction. This decision is based on the following two assumptions: first, medicine is a field of action, in which certain goals or intentions are pursued by applying certain means or methods. Against this background, medical applications can be defined by the means they employ to reach certain ends [16]. Second, putting forward a means-ends distinction appears to be an adequate way to arrange the multiple semantic elements of Systems Medicine definitions in order to evaluate them regarding their potential for a précising definition of Systems Medicine. Christoph Schickhardt, Sebastian Schleidgen and Henrike Fleischer arranged the definitional elements found in the literature according to the extraction tabloid resulting in a list of ends and means constitutive for the use of the term of Systems Medicine. For instance, from the definition “systems medicine is an emerging discipline that aims to find novel diagnostic markers and therapeutic targets by combining omics with bioinformatics” [17] there were derived the ends “to find novel diagnostic markers” and “to find therapeutic targets” as well as the means of “combining omics with bioinformatics”.

Step 3: Development of a précising definition of systems medicine

For the purpose of developing a précising definition of Systems Medicine intending to reduce the vagueness of its usage in practice, we referred to six formal criteria for adequate definitions. These criteria are:

  1. 1.

    necessity (a definition must be necessary, i.e. there must not exist any well-established term equivalent with its definiens),Footnote 3

  2. 2.

    non-circularity (a definition must not be circular, i.e. the definiendum must not appear in the definiens),

  3. 3.

    non-redundancy (a definition must not be redundant, i.e. it must not contain any components which are implied by any other of its components),

  4. 4.

    consistency (a definition must not be inconsistent, i.e. it must not include any logical contradictions),

  5. 5.

    non-vagueness (a definition must not be vague, i.e. the definiens’ meaning must be clear), and

  6. 6.

    coherence (a definition must be coherent, i.e. its single semantic elements must reciprocally underpin each other so that their combination makes sense and is plausible) [10, 11, 16, 18].

An adequate précising definition has to satisfy all six criteria, i.e. the whole definition (its semantic components and their relations) must satisfy them. This is the kind of definition we aim at developing in the following. In order to develop such a definition, most criteria can be used to assess the single components derived in step 2 before evaluating the whole definition, i.e. to exclude the ends and means found in the literature which are not eligible for a précising definition of Systems Medicine. The criterion of non-circularity (2), for instance, can be applied to single terms like “systems approach”, which is not eligible as definitional element of Systems Medicine as the definiens must not contain any central terms of the definiendum. Other criteria like consistency (4) or coherency (6), however, can only be applied to predicative units of meaning. Thus, the criteria (2)–(6) can be applied to preselect potential components (single terms or predicative units of meaning, respectively) of a précising definition of Systems Medicine.

Accordingly, we applied the criteria to the ends and means of Systems Medicine found in the literature and assigned the remaining ends and means to categories. This does not mean to accept the ends and means identified in this step as ultimately eligible for a précising definition of Systems Medicine. Rather, in view of the criterion of coherence (6), we started an iterative process, compared the ends and means at several stages of our analysis with the goal of establishing a meaningful equilibrium between them and finally approached a précising definition of Systems Medicine.

Results

General results

We identified 185 articles containing the term “systems medicine” in title or abstract. 3 papers were not written in English, but in Chinese, German and Italian, and therefore excluded from further analysis. We screened the remaining 182 articles. 67 contained at least one definition of Systems Medicine and were thus included in steps 2 and 3 of our review (see also Fig. 1).Footnote 4

Fig. 1
figure 1

Flow diagram of data collection process according to the PRISMA statement [42]

Full size image

In our sample, Systems Medicine was first mentioned in 1992 [19], but the discourse on Systems Medicine did not substantially intensify before 2010/2011. Figure 2 shows the annual number of papers containing the term “systems medicine” in title or abstract in the period of 1992–2015.

Fig. 2
figure 2

Annual number of papers containing “Systems Medicine” (1992–2015)

Full size image

The average growth rate of literature written on Systems Medicine was about 41% per year. This is substantially higher than the annual increment of PubMed’s database in the period of 1993–2015, whose average growth rate was 5%.

Step 1: Description of the areas of application of systems medicine

The reviewed papers (N = 182) were assigned to the two main categories 1) with reference to disease and 2) without reference to specific disease as follows: 77 papers were assigned to main category 1), 105 papers to main category 2). This frequency distribution shows that 58% of the papers have no reference to specific diseases.

With regard to the specific contexts of disease mentioned in the articles, the literature review reveals that chronic, non-cancer diseases (n = 35) as well as hematological and solid cancers (n = 33) are the most frequent foci of Systems Medicine (see Table 1). The subcategory chronic, non-cancer diseases includes a variety of diseases. Pulmonary and respiratory diseases (n = 12), metabolic and nutritional disorders (n = 6), psychiatric and behavioral disorders (n = 5) are examined particularly frequently. The third subcategory various types of diseases includes papers not specifying the examined diseases as chronic or non-chronic. It consists of 9 papers. Within this subcategory, immune disorders (n = 4) are investigated most frequently. Overall, the frequency distribution of the papers shows that with regard to a single context of disease, Systems Medicine currently is most frequently focused on cancer research. With regard to content alignment, 100 papers were assigned to category A) research context, 82 articles to category B) programmatic context. Category A) includes 56 papers with reference to disease, 44 papers without reference to specific disease, which means that almost half of the research-related papers do not refer to a specific disease. Furthermore there are some disease-related articles (n = 21), which do not present or discuss findings or (new) methods of Systems Medicine and thus are classified as programmatic. Of all papers allocated to category B) programmatic context, 61 papers have no reference to disease. All articles in category B) are formulated in a very general manner meaning they are not related to any specific research methods or findings. Several papers are of a more advertising nature, others have a more basic informative intent.

Table 1 Areas of application of Systems Medicine: frequency distribution (N = 182)
Full size table

Table 1 depicts the frequencies of papers assigned to categories 1) and 2), A) and B) as well as the resulting frequencies from crossing the categories:

The temporal distribution of papers indicates that there are no major scientific changes regarding thematic focus. Focusing on the articles published within the last five years, we found that the number of papers increased in all four categories approximately to the same extent. The same holds for the crossed categories.Footnote 5

Step 2: Reconstruction of current systems medicine definitions

There seems to be no consensus on the meaning of Systems Medicine. In the 67 papers containing definitions of Systems Medicine or definition-like text passages,Footnote 6 we found 98 definitions with 114 means and 132 ends, depicted in Tables 2 and 3.Footnote 7 We did not carry out a word-by-word analysis to find out the semantic overlap of all definitions and definitions-like text passages found in the literature. This would have been outside our approach which aimed to analyze all definitions and definition-like-text passages by way of building inductive categories and the applying the six formal criteria to the single elements. Furthermore, we did not seek for a consensus definition.

Table 2 Means of Systems Medicine in the literature
Full size table
Table 3 Ends of Systems Medicine in the literature
Full size table

Step 3: Development of a précising definition of systems medicine

After applying our criteria (2)–(6) for determining the adequacy of definitions (non-circularity, non-redundancy, consistency, non-vagueness, and coherence) to the ends of Systems Medicine found in the literature and excluding all ends not meeting the criteria,Footnote 8 we inductively assigned the remaining ends to the following preliminary categoriesFootnote 9:

  1. End i)

      Improving Participation

  2. End ii)

      Improving Prediction

  3. End iii)

      Improving Prevention

  4. End iv)

      Improving Stratification

  5. End v)

      Improving Treatment

  6. End vi)

      Improving Diagnostics

  7. End vii)

      Modelling

  8. End viii)

       Improving Understanding of Disease/Pathologies/Health States

  9. End ix)

      Innovative Approach to Drug Discovery

  10. End x)

      Finding Novel Therapeutic Targets

  11. End xi)

      Discovering New Diagnostic and Prognostic Biomarkers

  12. End xii)

      Re-Defining Clinical Phenotypes

  13. End xiii)

       Improving Health Care

  14. End xiv)

       Achieving better Population Health

Categories Ends ii) (Improving Prediction), iii) (Improving Prevention), v) (Improving Treatment) and vi) (Improving Diagnostics) are classical ends with regard to improving health care. Accordingly, we assigned them to category End xiii) (Improving Health Care). Categories End x) (Finding Novel Therapeutic Targets), xi) (Discovering New Diagnostic and Prognostic Biomarkers) and xii) (Re-Defining Clinical Phenotypes) were subsumed under category End iv) (Improving Stratification). With regard to category End x) (Finding Novel Therapeutic Targets), this decision followed the insight that it is a particular goal of modern efforts of stratification to base therapeutic measures on novel targets, i.e. targets identified through methods of molecular specification [20,21,22]. A similar point can be made for categories End xi) (Discovering New Diagnostic and Prognostic Biomarkers) and xii) (Re-Defining Clinical Phenotypes): it is a declared goal of activities to improve stratification (for instance through so called Personalized, Individualized or Precision Medicine approaches), to stratify diseases into smaller subgroups by using (new diagnostic and prognostic) biomarkers [16, 23]. Another goal of stratification, which is particularly associated with the molecular approach of Personalized Medicine, is to re-interpret clinical phenotypes, e.g. through a new specification and classification of tumors on a molecular basis [24]. These considerations left us with the following categories of Systems Medicine ends:

  1. End a)

      Improving Participation

  2. End b)

      Improving Stratification

  3. End c)

      Modelling

  4. End d)

      Improving Understanding of Diseases/Pathologies/Health States

  5. End e)

      Innovative Approach to Drug Discovery

  6. End f)

      Improving Health Care

  7. End g)

      Achieving better Population Health.

In a next step, we applied our criteria (2)–(6) (non-circularity, non-redundancy, consistency, non-vagueness, and coherence) to the means found in the literature on Systems Medicine and excluded all means not meeting the criteria.Footnote 10 Subsequently, we inductively derived ten categories from the remaining meansFootnote 11:

  1. Means i)

       Application of Systems Biology to Medical Research and Practice

  2. Means ii)

       Data Integration

  3. Means iii)

       Modelling

  4. Means iv)

       Networks

  5. Means v)

       Bioinformatics/Computer Tools/Computational Analysis

  6. Means vi)

       Addressing Complexity

  7. Means vii)

       Stratification

  8. Means viii)

       Holistic Approach

  9. Means ix)

       Understanding of Illnesses/Diseases

  10. Means x)

       Experimental Validation/Experimental Examination

A closer look revealed that some of these categories do not depict adequate means of Systems Medicine and thus have to be excluded. First, as regards category Means iv) (Networks), it has to be stated that, although terms like “network-based approach” [25, 26] or “network-level view” [25] might suggest that networks are a distinctive methodological feature of Systems Medicine, most text passages with reference to “network” show an understanding of the term in the sense of “physiological network”: networks are seen as part of healthy as well as ill physiological processes that are to be investigated rather than to be used as a means for medical research and practice. Networks are a structural characteristic of the physiological reality that Systems Medicine seeks to understand. Networks are thus object of research efforts and cannot be a means of research at the same time.

Second, regarding category Means vi) (Addressing Complexity), the text passages addressing complexity state that complexity, understood as dynamic interactions between multiple factors, is a fundamental challenge of Systems Medicine calling for a distinctive conceptual orientation. In this understanding, however, addressing complexity cannot be seen as a specific means for Systems Medicine research or practice. We thus learnt that both categories, networks and addressing complexity, are no means but rather indicate a characteristic feature of the physiological reality that Systems Medicine seeks to investigate and understand. Furthermore, for they consist in the dynamic interactions and interdependencies of several factors, networks can be considered as a phenomenon of complexity and therefore subsumed under addressing complexity. We finally subsumed addressing complexity, including networks, under the category Understanding of Illness/Diseases and considered it a specific challenge of Systems Medicine research efforts.

Third, however, regarding category Means ix) (Understanding of Illnesses/Diseases), it is obvious that understanding of illnesses/diseases is not only a means. In particular, this applies to medical approaches like Systems Medicine which not only refers to medical practice but also to biomedical research. The extent to which Systems Medicine is a biomedical research orientation, understanding of the researched phenomena, e.g. illnesses or diseases, is a primary objective. The fact that understanding of diseases was already identified as category End d) also indicates that Means ix) (Understanding of Illnesses/Diseases) rather has to be understood as an end of Systems Medicine.

Fourth, taking a look at category Means viii) (Holistic Approach), it is obvious, that “holistic approach” cannot be considered a concrete means of Systems Medicine. Usually it is used as an attribute, assigned to the perspective or, without specification, to the general approach of Systems Medicine, e.g. “a holistic approach to disease”, “a holistic view of health and disease” [27]. The precise meaning of “holistic” cannot be derived from the respective text passages. If it is used in the sense of “everything”, the term is relative in the sense that “everything” is included, which is considered necessary for the respective phenomenon to be investigated. However, in this understanding, a holistic approach is rather an ideal than a specific means of Systems Medicine.

Fifth, like category Means ix) (Understanding of Illnesses/Diseases), category Means vii) (Stratification) was already assigned to Systems Medicine’s ends (category End b)). And there are good reasons for stratification of patients and diseases not being a specific means of Systems Medicine: stratification was, for instance, identified as an essential goal of Personalized Medicine [16]. With regard to Systems Medicine, it also seems plausible to understand stratification as an intermediate end indicating the specific manner to achieve the ultimate overall goal of Systems Medicine, i.e. the improvement of healthcare provision. Therefore, we decided to exclude stratification from Systems Medicine’s means and to treat it as an end.

These considerations left us at the following list of Systems Medicine means categories:

  1. Means a)

       Application of Systems Biology to Medical Research and Practice

  2. Means b)

       Data Integration

  3. Means c)

       Modelling

  4. Means d)

       Bioinformatics/Computer Tools/Computational Analysis

  5. Means e)

       Experimental Validation/Experimental Examination

When relating categories Means a)-e) to the categories Ends a)-g), we discovered three incoherencies: First, with regard to the category End a) (Participation) we found almost no means outlining how to achieve this end; of the 98 definitions or definition-like text passages only two mention “patient-centered integrated care” [28], one speaks of an “patient-centered approach” [29]. Patient-centered care is widely understood as aiming at patients’ individual needs as well as personal preferences and values in order to actively include them in the process of shared decision-making [30]. However, the text passages referring to patient-centered care consider it as part of the strategy, i.e. the means of System Medicine. Stating that participation (as an end) is pursued through patient-centered care (as a means), however, does not make sense: patient-centered care is a particular form of participation and thus not a plausible means to realize participation. Against this background, we conclude that none of the mentioned means addresses the question of how to concretely improve patients’ participation or empowerment in comparison to the status quo. Furthermore, it is striking that “participation” as an end in none of the reviewed papers is mentioned sole or in combination with a semantically identical or similar term. It is only referred to as one of the “P”s in “P4-Medicine”. This indicates that “participation” in the context of Systems Medicine is used as an empty phrase, which is presumably why it has not been taken into consideration by which means the end of (improving) participation could be achieved. Therefore, we excluded “participation” as an end of Systems Medicine. It is important, however, that by excluding participation (including patient-centered care) from the ends of Systems Medicine we are not suggesting on a normative level that participation should not be pursued by Systems Medicine approaches. The exclusion is solely due to the application of formal criteria to currently existing definitions of Systems Medicine in the literature.

Second, it is striking that End g) (Achieving better Population Health) is mentioned only once [29]. Although there is much academic debate on the term “population health” [31, 32], in our understanding it designates the aim of improving public health, defined as “the art and science of preventing disease, prolonging life and promoting health through the organized efforts of society” [33] and referring to “all organized measures (whether public or private) to prevent disease, promote health, and prolong life among the population as a whole. Its activities aim to provide conditions in which people can be healthy and focus on entire populations, not on individual patients or diseases” [34]. Classical means to improve population or public health are surveillance and assessment of a population’s health and well-being, the identification of health problems and health hazards in a community as well as providing health protection services (regarding, e.g., the environment, food safety, social determinants and economic inequalities) [34]. Among the means analyzed, however, none of such or similar means are mentioned. Against this background, we excluded the end “achieving better population health” due to it being not coherent with the means.

Third, category Means c) (Modelling) was also mentioned as category End c). If, however, the application of Systems Biology to medical research and practice (Means a)) is a means of Systems Medicine, modelling can only be a means, too, for Means b)-e) (Data Integration, Modelling, Bioinformatics/Computer Tools/Computational Analysis and Experimental Validation/Experimental Examination) are strongly connected to system biology. Data integration (category Means b)) means merging different types of data or data gained from different sources. This aggregation of Data is an essential means of Systems Biology: “Modern Systems Biology tackles biocomplexity through a unique approach that integrates biological data […]” [35]. Such data integration, in turn, is strongly connected to (mathematical) modelling (category Means c)) as part of Systems Biology: “systems biology is a methodology that employs mathematical modeling and computational biology tools to integrate and analyze quantitative biological data” [36]. Against this background, we excluded modelling as an end and arrived at the following list of categories of Systems Medicine ends:

  1. End A)

      Improving Stratification

  2. End B)

      Improving Understanding of Diseases/Pathologies/Health States

  3. End C)

      Innovative Approach to Drug Discovery

  4. End D)

      Improving Health Care

Ends B) und C) can be assigned to a new category “Medical research”, leaving us at the final list of Systems Medicine Ends:

  1. End I)

      Improving Stratification

  2. End II)

      Improving Medical Research

  3. End III)

      Improving Health Care

Ends II) and III) show that Systems Medicine aims at improving the classical target areas of medicine per se. Against this background, in turn, End I) is plausibly to be understood as an interim goal that indicates the manner in which improvement shall be achieved in the areas of medical research and health care.

Returning to our list of Systems Medicine means categories, both data integration and (mathematical) modelling (Means b) and c)) point towards bioinformatics (category Means d)) as a means of Systems Biology: bioinformatics, defined as “advancing the scientific understanding of living systems through computation” [37], includes a whole arsenal of computational means and methods of system biology such as information management (of omics and clinical data), data analysis and interpretation, design of new omics experiments, quality control and pre-processing of omics data, (statistical) data analysis methods of large and complex omics-based datasets serving the goal of elucidation and analysis of biological networks [38].

Finally, experimental validation and experimental examination (category Means e)) is a necessary supplement of modelling in Systems Biology, as “systems biology requires ‘detail’ and hence accurate experimentation in vivo” [39]. The procedure of combining (mathematical) modelling with in vivo experiments is an established method of system biology tackling “biocomplexity through a unique approach that integrates biological data, mathematical modeling, and experimental verification in iterative feedback cycles” [35].

It follows that Means b)-e) are well-established means of Systems Biology and, thus, have to be subsumed under category Means a). This leaves us with the only Means of Systems Medicine:

Means Application of Systems Biology to Medical Research and Practice.

Our considerations on adequate means and ends can finally be summarized to derive the following précising definition of Systems Medicine:

Systems Medicine is an approach seeking to improve medical research (i.e. the understanding of complex processes occurring in diseases, pathologies and health states as well as innovative approaches to drug discovery) and health care (i.e. prevention, prediction, diagnosis and treatment) through stratification by means of Systems Biology (i.e. data integration, modeling, experimentation and bioinformatics).

It is apparent that neither the means, as already established in the context of Systems Biology, nor the ends of Systems Medicine, especially stratification, are for themselves new elements of medical research and practice. Novelty, nevertheless, results from the specific combination of these means and ends. Thereby, our definition meets the first criterion for adequate definitions (necessity). Furthermore, our previous analysis and iterative critical comparison of means and ends shows that criteria 2–6 (non-circularity, non-redundancy, consistency, non-vagueness, and coherence) are also met.

Discussion

43% of the papers with reference to disease focus on cancer contexts (hematological or solid cancers), 82% of these papers present research results. This suggests that development of Systems Medicine is most advanced in the area of oncology. 45% of the papers on Systems Medicine, however, have a programmatic alignment, which shows that the description of general future prospects, potentials and challenges play an important role in ongoing works on Systems Medicine. Accordingly, a variety of articles are of general and visionary nature. This, in combination with the fact that 58% of all papers have no reference to a specific disease, implies that Systems Medicine is rather in an early stage of development and not (yet) part of everyday biomedical research and practice. Moreover, almost half of the papers on Systems Medicine research do not refer to any diseases. This strengthens the impression of Systems Medicine’s visionary character: much research on Systems Medicine seems to be basic research rather than translational or clinical research. Furthermore, we found no evidence for Systems Medicine being part of health care practice. The temporal distribution of publications on Systems Medicine can be interpreted similarly as it indicates no major changes in scientific publishing with regard to thematic focus. This may be due to issues regarding methods as well as feasibility that are direct consequences of the Systems Biology approach [7].

Against this background, it is vital to avoid unrealistic expectations regarding Systems Medicine approaches. This especially holds for goals like improving patient participation which are intensely debated in the context of Systems Medicine: Although our analysis showed that improving participation is referred to in the scientific community, we found no respective proposals for adequate means by which this goal could be achieved. Similar to the debates on the meaning of Personalized or Precision Medicine [16], the vision of Systems Medicine – due to its focus on approaches of the natural sciences and bioinformatics – does not focus on wants or preferences of individual patients, but rather on physiological needs.

Our précising definition also shows that the vision of Systems Medicine, first, aims at improving all classical target areas of health care (i.e. prevention, prediction, diagnosis and treatment) through stratification. Second, it aims at improving biomedical research through a specific research orientation, namely understanding of complex processes occurring in health and diseases by means of Systems Biology. It is highly plausible to consider Systems Medicine a comprehensive approach addressing both biomedical research and health care and aiming at advancements in both research and health care. Progress in health care relevant stratification is attainable only through the discovery, understanding and development of new biomarkers as well as new targeted therapies.

Regarding its focus on stratification, Systems Medicine indeed can be understood as some kind of successor of Personalized or Precision Medicine: Systems Medicine addresses the same, or at least similar goals as Personalized or Precision Medicine. It differs, however, regarding its means: while Personalized Medicine focused on utilizing biological information and biomarkers on the molecular level in order to improve stratification of health care [16, 40], Systems Medicine focuses on means of Systems Biology.

Admittedly, this conclusion is not without difficulties, given that Systems Biology, as it is practiced today, is a relatively young discipline differently defined and understood. However, Systems Biology is now fully established in the life sciences community and evolved into a worldwide recognized discipline [35]. As opposed to Systems Medicine there is a general consensus on its key elements amongst the scientific community. Our understanding of Systems Biology is based on the following descriptions reflecting these key elements: Systems Biology is based on the principles of systems theory. Bridging biological sciences, applied mathematics as well as computational and engineering sciences, it is focused on a systemic understanding of biological processes. It employs mathematical modeling and computational biology tools to integrate and analyze quantitative biological data [35, 36].

The frequent statement that Systems Medicine is the application of Systems Biology in medical research and practice thus clearly points in the right direction and is confirmed by our findings – as long as Systems Biology and its typical means are specified. It is important, however, to keep in mind that stratification in the context of Systems Medicine is not to be understood as an end in itself, but rather as an intermediate goal referring to the ultimate end of improving medical research and health care.

Conclusions

Our literature review revealed 67 papers containing 98 definitions of Systems Medicine, which mention 114 means and 132 ends of Systems Medicine. By applying six criteria for adequate definitions to these means and ends (non-circularity, non-redundancy, consistency, non-vagueness, and coherence), we derived the following précising definition:

Systems Medicine is an approach seeking to improve medical research (i.e. the understanding of complex processes occurring in diseases, pathologies and health states as well as innovative approaches to drug discovery) and health care (i.e. prevention, prediction, diagnosis and treatment) through stratification by means of Systems Biology (i.e. data integration, modeling, experimentation and bioinformatics).

Our study also revealed the visionary character of Systems Medicine: most papers presenting research results are to be located in the context of basic research. These insights allow for a realistic identification of actual ethical as well as legal issues arising in the context of Systems Medicine and, in consequence, for a realistic debate of questions concerning its matter and (future) handling.

It is important that our definition is not a “consensus definition”: our literature analysis reveals that there is no consensus on the meaning of “Systems Medicine”. Against this background, our approach of developing a definition of Systems Medicine combines two different approaches of determining the meaning of a term: one being not normative at all by confining itself to reconstructing the meaning(s) attributed to a term by the members of a community when using the term, and one being highly normative by postulating the “right” meaning of a term, independently from how it is used in practice. It has a reconstructive aspect in starting from definitions or definition-like passages as used or proposed by members of the scientific community. On the other hand, it also has a normative aspect when requiring all (elements of) definitions to meet six formal criteria of an adequate definition. The fact that we only worked on what people had stated and that we did not include content elements by our own might be seen as a “democratic” or consensus oriented procedure. However, we excluded all definitional elements which did not meet our six criteria. Hence, consensus was neither a goal nor a criterion.

Limitations

The definition presented above is to be understood as formally adequate, insofar it satisfies our six criteria for adequate definitions. On the other hand, as regards content, we do not claim that it is the only correct definition of Systems Medicine.