From data to knowledge

Hospitals are information-intensive organizations where data is processed into mean-ingful information and knowledge at multiple levels for administrative and clinical purposes to guarantee safety of care (Graves & Corcoran 1989, Nelson 2002). The value chain of information consists of data, information, knowledge, and wisdom (DIKW), in that order, evolving in scope and meaning from simple data items to ab-stract thinking of human beings (Matney et al. 2011). It presents a hierarchy of how raw data is transformed into wisdom in practice. The DIKW framework serves as the overarching concept for informatics and information management in hospitals (Nelson 2002, Matney et al. 2011). The pyramidal structure of the DIKW framework presents a causal hierarchy that indicates the levels of each element; there are quanti-ties of data, but hardly any wisdom (Bernstein 2011). However, Choo (2006) suggests that the transformation of data into knowledge and further wisdom should actually start from signals: sensory phenomena that humans can explore, such as sounds and sights (Choo 2006). A very small number of signals are accepted and selected from vast numbers of signals occurring in practice. The selection process depends on the observer´s past experiences and beliefs about what to expect (Choo 2006).

The foundational element of the original DIKW framework is data. Data are defined as factual information (e.g., the result of measurement) used as a basis for reasoning, discussion, or calculation (Merriam-Webster n.d.a). Data can be presented in various forms; they can be numbers, words, sentences, pictures, or other formats (Matney et al. 2011). Blum (1986) defined data as “discrete entities that are described objective-ly without interpretation” (Graves & Corcoran 1989). The European Committee for Standardization (2004, p.9) excluded context in their definition of data: “Discrete, objective facts (numbers, symbols, figures) without context and interpretation.” Data are the results of patients’ examinations, vital physiological signs of the body, and other clinical information, most often processed, collected, and stored using informa-tion technology to be later used as informainforma-tion in clinical decision-making (Graves &

Corcoran 1989, Goossen et al. 1997, Moen & Mæland Knudsen 2013).

When the data is processed and put into a context, it is transformed into meaning-ful information (Graves & Corcoran 1989, European Committee for Standardization 2004, Matney et al. 2011). In other words, information represents the data, such as facts and numbers, in a certain context, which creates meaning for the information (Matney et al. 2011, Choo 2006). Thus, information is based on data processed through cognitive structuring which assigns meaning and significance to the identified facts, such as heart rate (Choo et al. 2006). On the other hand, information is considered to be the collection of linkages that connects different data objects and forms a col-lection of data (Nelson 2002). The processing of information focuses on efficiency of organization, storage, retrieval, and communication, instead of purely computational processing (Graves & Corcoran 1989).

Information becomes knowledge through belief structuring (Choo 2006). Likewise, while information is processed data, knowledge is synthesized information with iden-tified and formalized relationships between data and information (Graves & Corcor-an 1989, Matney et al. 2011). Knowledge in orgCorcor-anizations is heterogeneously gained through experience, practice, and reflection that makes it unique to each individual (Choo 2006). A classical definition of knowledge is “justified true belief,” but knowl-edge-creating theory defines it as a “dynamic process of justifying personal belief towards the truth” where the knowledge is accepted to be truth because it works at a certain time and place (Nonaka 2005, pp.376-377).

Multiple types of knowledge are identified (Choo 2006, Matney et al. 2011), and organizational knowledge is categorized as explicit, tacit or cultural knowledge (Choo 2006). Explicit knowledge is easily communicated and distributed because it is codified (i.e., made tangible). The codification defines whether knowledge is rule-based or object-based; the former applies when knowledge is codified into rules or specifica-tions, and the latter when it is codified in symbolic expressions or in tangible assets.

Tacit knowledge is personal and implicit, and is hard to formalize and communicate to others (Choo 2006, Matney et al. 2011). Cultural knowledge is defined as shared assumptions and beliefs about an organization’s targets and situation that is used to assign value and significance to new information (Choo 2006). Knowledge can be collective or individual, and knowledge management is defined as “planned and on-going management of activities and processes for leveraging knowledge to enhance competitiveness through better use and creation of individual and collective edge resources” (European Committee for Standardization 2004, p.10). As knowl-edge is a multidimensional concept, individuals mainly conduct the processing, but decision-support systems and automated expert systems may process knowledge in order to assist in decision-making (Graves & Corcoran 1989). Moreover, information systems enable effective knowledge creation in health care organizations (Kivinen &

Lammintakanen 2013).

When knowledge is used appropriately in managing human problems, it is called wisdom (Nelson 2002). In the DIKW framework, the level of critical thinking and rea-soning increases at each level (Matney et al. 2011). The broader definition of wisdom is stated as “forms of deliberation which combine knowledge, reflection and life ex-perience with social, emotional and ethical capabilities” (Edmondson & Pearce 2007, p.233). Further, wisdom involves good intentions with the aim to improve well-being of people (Edmondson & Pearce 2007, p.238). Clinical wisdom takes place in the judi-cious and appropriate implementation of knowledge in practice (McKie, Baguley et al. 2012). In the processing of wisdom, expert systems utilizing artificial intelligence are utilized. Figure 1 shows the DIKW framework as presented in Nelson (2002, p.13) combined with the levels and types of automated systems (Nelson 2002, p.12), and the signals and definitions used by Choo (2006, p.132). The elements of the framework are connected to the types of information systems used in information management in hospitals, and form a basis for the practice.

Figure 1. From signals to wisdom continuum adapted from Nelson (2002, pp. 13-14) and Choo (2006, p.132).

The level of information systems used to process signals, data, and information, or to apply knowledge or wisdom, is connected with the DIKW framework (Nelson 2002).

The Figure 1 demonstrates the relationships of information, decision support, and expert systems with the data, information, and knowledge as presented by Nelson (2002). Information systems are used to process data and information, whereas deci-sion support uses a knowledge base and a rule set to formulate recommendations for users. Sophisticated expert systems, that utilize artificial intelligence (AI) techniques, are the computer applications developed to solve complex problems using human-lev-el performance in cognitive tasks (Nhuman-lev-elson 2002, Ramesh et al. 2004). Clinicians need to acquire, analyze, and apply the large amount of data, information, and knowledge to improve patients’ health. Decision support and expert systems utilizing AI assist in processing all necessary subjective and objective data and enable the effective appli-cation of scientific research in clinical decision-making (Ramesh et al. 2004). However, these systems are dependent on the quality of input (e.g., data) stored in the system.

The attributes of good data, information, and knowledge presented in Table 1 provide a framework for critical evaluation of data quality (Graves & Corcoran 1989, Nelson 2002). Technology enables the transformation of various types of data into knowledge through more automated processes. Further, the access to accurate information and knowledge are core requisites for safe patient care (Moen & Mæland Knudsen 2013).

Table 1. Attributes of good data, information, and knowledge adapted from Graves & Corco-ran (1989) and Nelson (2002, p.15).

Data Information Knowledge

Descriptive* Accurate* Accurate*

Measurable* Timely* Relevant*

Relevant* Quality*

Quality*

Accessible Free from bias Comprehensive Clear

Appropriate Precise Quantifiable Verifiable

* Graves and Corcoran (1989)

Information technology is used to process and to store the data, whereas humans communicate with information. Communication between humans involves the inter-action among individuals using symbolic inter-actions or language (Shin et al. 2011), and is the transfer of messages such as information and understanding from sender to receiver by any channel (Chandler & Munday 2016). Communication also refers to a verbal or written message that is exchanged between individuals through a common system of symbols, signs, or behaviors (Merriam-Webster n.d.b). Human communi-cation is progressively affected by information, but emotions and attitudes also have a meaningful effect on it. Health communication concerns health-related issues and the factors that impact them. It might occur at different levels; for example, at the na-tional level, health communication concerns health promotion campaigns and public health plans (Shin et al. 2011). Clinical communication refers to health communication taking place in a clinical setting. It might be professional-to-professional or between professional and client/patient. In health care, communication is widely based on the clinical data (i.e., all data related to the patient episode, including but not limited to medication data, allergy data, referral or invitation to care, summary of care for the discharge, and patient-related guidelines or care instructions). Several systematic re-views have shown that failures in clinical communication (e.g., delayed, misplaced, or forgotten information) can result in adverse patient outcomes (Nagpal et al. 2010a, Braaf et al. 2011, Vermeir et al. 2015, Hohenstein et al. 2016, Kattel et al. 2016). Thus, managing information in a safe way is a prerequisite for effective communication in patient care; in particular, timeliness and content of written communication need to be improved (Vermeir et al. 2015). The understanding of differences between elements of the DIKW framework is needed when developing information management pro-cesses, as organizations should be able to transform signals and data into knowledge and to use it wisely in the practice.