Healthcare in the Modern Era
Patient data and related Electronic Health Records (EHRs) generate volumes of clinical data at an alarming rate which will continue influencing diagnostic decision making from a cognitive load (stress) and information overload perspective. Health IT systems were touted to improve patient care. However, medical errors remain the third leading cause of death in the United States, which accounts for more than 250,000 lives annually after cancer and heart disease deaths (Makary, 2016; Sternberg, 2016). More than 95% of all eligible and critical access hospitals have adopted, implemented, and utilize certified Health IT systems through participation in the Centers for Medicare & Medicaid Services (CMS) Electronic Health Record incentive programs (CMS EHR Incentive Programs Data, 2016; Wright, 2006). Health IT Systems were touted to empower physicians, enable better communications, facilitate improvements in diagnosis, improve patient care quality, and reduce costs while helping to provide a better overall client experience (CMS EHR Incentive Programs Data, 2016; Fortune & Schulte, 2019; Wright, 2006). A review of the research reveals that sociotechnical solutions within healthcare must address macroergonomics, semantic interoperability, and other considerations to further realize the advantages touted by Health IT system vendors, i.e., All Scripts, Cerner, Epic, etc. (Fortune & Schulte, 2019; Goldstein, 2008).
Medical cognitive science emphasizes the complex nature of clinical reasoning and the significance of knowledge representation in medical decision-making. An ongoing range of cognitive processes are utilized by clinicians in constructing mental models that aptly reflect clinical scenarios and assist in making effective clinical decisions. “The use of a Health IT system, which includes an electronic health record, by providers during office visits has been equated to ‘texting while driving’ and thus raises concerns that provider’s observation, communication, problem solving, and development of trusting relationships are being impacted” (Sandoval, Palumbo, & Hart 2016). Many argue that all aspects of healthcare provider performance require cognitive processes. However, researchers claim its evident that diverse activities which comprise the office visit requires the provider to complete different cognitive tasks. Based on interviews conducted by Sandoval, Palumbo, & Hart, a prediction was made that the EHR usage increases the physician’s mental workload and thus makes the task of simultaneous data entry and engagement in patient centered care more problematic. (Sandoval, Palumbo, & Hart 2016).
Currently, clinicians do not utilize the HIT systems fully; instead, many older healthcare providers defer to notepads. Therefore, several hospital organizations have begun to realize the shortcomings prevalent within their Health IT systems. “HIT systems, such as computerized provider order entry (CPOE) with clinical decision support systems, electronic health record (EHR), and bar-code medication administration (BCMA) systems, have been hailed as possible vehicles for reducing medical errors and adverse events (Bates & Gawande, 2003), as well as a means to improve provider-provider and provider-patient communication (Kaelber & Bates, 2007)” (Lawler, Hedge, A, & Pavlovic-Veselinovic 2011). Inherent cognitive challenges are prevalent within information management activities, especially within healthcare. Distributed cognition is a branch of cognitive science that proposes that cognition and knowledge are not confined to an individual. Instead, cognition is distributed across artifacts, individuals, objects, and tools in the target environment. Therefore, Hutchins’ theory of distributed cognition can be applied to distribution processes across members of social groups (clinicians), between internal and external structures (healthcare ecosystem), and distribution through time (earlier events can transform later events). Distributed cognition lends itself well to exploring how clinicians interact with Health IT systems. Issues which distributed cognition can address includes communication and coordination, how clinician thinking extends into the environment, and performance from a broad perspective. Therefore, distributed cognition is not limited to information processing within the brain; distributed cognition focuses on information processing within sociotechnical systems, also referred to as sociotechnical informatics. (DeLanda, 2016; Hutchins & Klausen, 1996; Scott, de Keizer, & Georgiou, 2019).
Physician cognition is a heterogeneous sociotechnical system with leverages intelligence, memory, and technologies which include, but not limited to electronic health records. Hutchins’ theory of distributed cognition has broadened the meaning of complex cognitive systems. Hutchins’ theory examines cognitive distribution to include members of a work team; the interplay between individual cognition processes, physical items, and artifacts; cognitive distribution is also based on time dependent events. Recently (Hazlehurst et al. 2003, 2007; Cohen et al. 2006; Patelet al. 2008), distributed cognition has been applied to Health IT to study the design and utilization consideration impacts on human activity; distributed cognition is being examined under a sociotechnical systems lens (Sittig & Singh, 2015). Human elements and inputs from work systems interact to influence cognitive load, clinician performance, and diagnostic decision making (Karsh, Holden, Alper, & Or (2006). Sociotechnical informatics includes the propagation and transformation of information across natural and engineered computational sociotechnical systems.
Like cognitive science, distributed cognition seeks to understand the organization of cognitive systems. Unlike traditional cognitive theory, distributed cognition extends it reach of what is considered cognitive beyond the individual to include interactions between materials, people, and resources. A process is not cognitive simply because it happens in the brain, neither is a process non-cognitive simply because it happens in the interactions among many brains. Therefore, it is important to understand that distributed cognition refers to a perspective on all of cognition rather than a specific variant of cognition. Distributed cognition is concerned with the range of mechanisms assumed to be participative in cognitive processes. While traditional views look for cognitive events in the manipulation of symbols inside individual actors, distributed cognition looks for broader classes of cognitive events and does not expect all such events to be encompassed within the brain of the individual. Cognitive processes involve coordination between internal and external (material or environmental) structure. (Hollan, Hutchins, & Kirsh, 2000).
From a distributed cognition perspective, arguments propose that individuals form loosely coupled systems within the healthcare ecosystem which facilitates the employment and exploitation of external structures to distribute cognition. For physicians to harness the touted possibilities of EHR, we need to better understand the emerging dynamic interactions while the task focus is no longer confined to EHR systems but reaches into the complex networked world of information and EHR-mediated interactions. As a result, we believe the theory of distributed cognition provides aids to help one better understand the interactions between individuals (physicians) and technologies (Health IT or EHR) due to its focus on the environment (health ecosystem). For example, clinicians utilize Health IT systems, which includes EHR, to obtain a better understanding of patient health history without having to rely on internal memory solely. In addition, Health IT system workflows can be configured to influence how information is transformed and propagated among teams. Inherent cognitive challenges are prevalent within information management activities, especially within healthcare. Distributed cognition can also be utilized to better understand complex sociotechnical informatics. Like assemblage theory, distributed cognition exhibits emergent behavior over time. Sarcevic and Farraro examined the efficiency of electronic health records in a fast-paced medical setting through the lens of distributed cognition to better understand information exchanges among team members and how information is documented, shared, and stored. EHRs allows for information to be mediated by expertise, knowledge, and technologies needed to render diagnostic decisions while balancing cognitive load. Assemblage theory can be utilized to better understand emergence within the broader context of healthcare (DeLanda, 2016; Hollan et al., 2000; Scott et al., 2019; Willis, 2019).
In both distributed cognition and assemblage theory, one expects to find systems which can dynamically configure themselves in coordination to accomplish various functions. From an assemblage perspective, there exists a dichotomy between material and symbolic ontological expression, matter matters. Traditional views look for cognitive events in the manipulation of symbols and inside individual actors (material), but distributed cognition seeks a broader class of cognitive events and does not expect such events to be housed within an individual. For example, an examination of memory processes within an operating room demonstrates that memory involves a rich interaction between internal process, the manipulation of objects, and the data represented among healthcare professional within the EHR system. In addition, “the physical environment of thinking provides more than simply additional memory available to the same processes that operate on internal memories”. The material world provides opportunities to reorganize distributed cognitive systems. These distributed cognitive systems make use of varied combinations of external and internal processes (DeLanda, 2016; Hollan et al., 2000; Scott et al., 2019; Willis, 2019).
From an assemblage theory and distributed cognition perspective, the dynamic integration of material and symbolic expression creates the basis for a functional relationship of elements that participate together. As basic elements combine certain elements possess tendencies to more relational than other elements; these tendencies are influenced by, both, material and linguistic expressive relationships. A cognitive process is delimited by the functional relationship among the elements that participate rather than by the spatial colocation of the elements. Previously undifferentiated spaces, from a differential geometry perspective, begins to emerge through varied combinations of differentiated space via varied, and unpredictable, element interactions. Distributed cognition encapsulates a range of mechanisms that may be assumed to participate in cognitive processes. Similarly, these relationships or assemblages allow certain elements to interrelate with each other which leads to territorialization. As a result of territorialization, higher level structures begin to emerge via coding and territorialization reproduces itself. Territorialization could also be described as a form of emergent environmental shaping; environmental shaping refers to how the emergent systems exerts influence on its members (DeLanda, 2016; Hollan et al., 2000; Kane, Alavi, Labianca, & Borgatti, 2014).
Social organizations and the underlying structural capital determine how information flows through a group. Therefore, social organizations and their underlying structural capital may be viewed as a cognitive architecture. Distributed cognition includes phenomena that emerges in social interactions as well as interactions between the symbolic and material. These social interactions either reinforces territorialization or destabilizes territorialization (deterritorialization). For instance, personal identity deterritorialization may occur not only through loss of stability but also through augmentation of new skills; in short, behavior creates social encounters. Habits or repeated patterns form via an evolutionary process which affects one’s perception of the world when fragments of action or thought emerges over time. As a result, individuals emerge through habits of an expressive and material nature; expressive and material behaviors are interrelated. These emergent social identities formed through habits foster social encounters: theses emergent social encounters create networks. These emergent social encounters with expressive and material context matter. The codification of emergent networks creates social classes (underpinned by structural capital) with varying dimensionality. The spread of influence or resources within our emergent social network can be referred to as contagion. With the spread of territorialization, resource access increases as contagion increases. However, a random event can also trigger deterritorialization which disrupts tightly coupled emergent properties and deteriorates emergent relationships (DeLanda, 2016; Hollan et al., 2000; Kane et al., 2014).
From an assemblage theory perspective, two perspectives are needed to better understand patient information from a distributed cognition lens. First, the idea of socially distributed cognition to make decisions while employing information processes, planning, organizing, and sensemaking of varied information resources. The digital and physical materiality of information representations and tools coupled with personal networks of care providers, family members, friends, and healthcare professionals. As a result, cognitive processes are socially distributed across members of a group. From a healthcare perspective, distributed cognition is a broader conception that includes phenomena that emerges among healthcare team when making decisions or making sense of the data available (DeLanda, 2016; Hollan et al., 2000; Willis, 2019).
Embodied cognition is the second tenet of the distributed cognition approach which must be explored. Better stated, causal coupling is an essential fact of cognition that evolution enables us to exploit. From the perspective of distributed cognition, the organization of mind is an emergent property of interactions among internal and external resources. These two perspectives in concert form an assemblage of information practices, information resources, people technologies, and tools to provide patient-centered care. These broad perspectives
can be used to describe assemblage within healthcare ecosystems. These heterogenic perspectives can be linked together to form a functional whole balanced by socially distributed cognition networks. Both, territorialization and deterritorialization processes are necessary within healthcare ecosystems for continual improvement and long-term evolutionary survival (Kane et al., 2014; Willis, 2019).
As one can see, assemblage theory is a meta-framework which allows theories such as distributed cognition to be hung upon. Distributed cognition facilitates dynamic interaction among sociotechnical system components within the healthcare ecosystem. The sum is greater than individual sociotechnical system parts or elements. Emergence is prevalent within healthcare systems; the orchestration and choreography of functions in unexpected ways are a key characteristic of assemblage theory in action within healthcare. Evolutionary dynamic assemblages of medical capabilities (digital and physical materiality) within the ecosystem will be required to obtain true precision medicine and patient-centered care over time (DeLanda, 2016; Kane et al., 2014).
The environment within the emergency room may serve as a great example to model to better understand how interruptions (information overload) influences decision making within an emergency room context. The emergency room scenario allows one to captures emergence within decision making while considering cognitive load and information overload effects. Oftentimes, the emergency room perform life critical tasks while the surgeon is being subjective to frequent distractions or interruptions which could lead to information overload and cognitive stress. Repeated patterns or routines form via an evolutionary process which affects one’s perception when making decisions over time. Therefore, established routines and habits present within the operating room foster an environment of emergence. From an assemblage theory perspective, the perspectives in concert form an assemblage of information practices, information resources, people technologies, and tools to provide patient-centered care. These broad perspective categories describe assemblage within healthcare ecosystems. These heterogenous linkages are linked together to form a functional whole balanced by socially distributed cognition networks. Both, territorialization and deterritorialization processes are necessary within healthcare ecosystems for continual improvement and long-term evolutionary survival (DeLanda, 2016).
Properly designed system dynamics models will allow one to better identify influences on decision making from cognitive load (cognitive stress) within healthcare while using Health IT systems. System dynamic models can be utilized to model varied aspects of assemblage present within the healthcare ecosystem within a decision-making context while under cognitive stress. System dynamic models consists of flows, sinks, sources, and stocks which can be representative of healthcare methods, processes, or system utilization. System dynamic models helps designers to analyze systems with feedbacks and interdependencies. System dynamic models help designers to explain and understand less intuitive phenomena findings; nonintuitive findings are the hallmark of system dynamic modeling. Our intuition fails because we tend to focus on direct effects and fail to think through the entire logical chain. From a decision-making perspective, processes of interest can be modeled using sources and sink while flows may be representative of directionality based on antecedents or pre-conditions. As a result, system dynamic model simulations can be used to further explain, guide, and predict action while exposing lagging cycles (Blobel, 2019; Eppstein, Horbar, Buzas, & Kauffman, 2012; Kasiri, Sharda, & Asamoah, 2012; Page, 2018).
System dynamic models allow designers to improve their capacity to think through logical chains that include positive and negative feedback present within a representative system. From a healthcare perspective, the true value of system dynamic modeling and simulation resides in the ability to help designers reason through the effects of intended actions prior to implementation to better understand potential implications downstream; to minimize cost and schedule implications, it is best to make significant changes as early in the lifecycle as possible. When a system includes both positive and negative feedbacks, it can potentially produce complexity or exhibit emergence. Due to the positive and negative feedback loops present within healthcare decision-making, the related complex adaptive personal health (pHealth) ecosystem combines varied domains represented by a huge variety of differentiated human and non-human actors which belong to varied policy domains and disciplines. The core challenge in modeling system dynamics is management of multi-dimensional knowledge domains and its representation. System dynamics models that represent flows and stock levels as mathematical functions can be calibrated to explain past values of stocks, to predict future values, and estimated effects of interventions or treatment protocols (Blobel, 2019; Eppstein et al., 2012; Kasiri et al., 2012; Page, 2018).
The varied healthcare ecosystem actors deploy differentiated methodologies, ontologies, and terminologies underpinned by experience, knowledge, and skills to accommodate business cases and business objectives; designers work to model varied aspects of decision-making and system utilization in hopes of understanding complexity within the system. It is important to note that system dynamic models can be both qualitative and quantitative in nature. New structures and order may emerge from agents as they collectively follow simple local rules that are amplified or constrained by features of the environment which changes system dynamics. System dynamic models can be utilized to search for additional improvement in healthcare. There are many potential applications for the deployment of system dynamics modeling in healthcare. System dynamic modeling helps to model the behavior of processes involved in healthcare to examine varied practical and theoretical decision-making perspectives. Quantifying the tangible and intangible benefits of using Health IT in clinical and decision-making processes is a difficult task. However, designers propose using techniques to estimate EHR benefits through modeling of system dynamics. The results of system dynamic modeling influences can also be transformed into economic value (derived from patient benefit) to estimate financial indices (Blobel, 2019; Eppstein et al., 2012; Kasiri et al., 2012; Page, 2018).
Healthcare IT requires continuous process improvement, but Health IT benefits are complex, hard to measure, and usually achieved over time. Healthcare IT directly influences decision making in healthcare from both a cognitive overload and macroergonomics perspective. System dynamic modeling and simulation are effective tools that allow researchers to project the potential implications of decision-making within a health ecosystem or network which uses Healthcare IT systems. It is important to note that Healthcare IT systems influences cognitive load when moderated by information overload and macroergonomic design mismatch. Using system dynamics models, we designers can better understand both intended and unintended effects within system boundaries. Therefore, we need multiple models; any single model simplifies the world and highlights limited dimensions. Outcomes are dependent on variables being explored (Blobel, 2019; Eppstein et al., 2012; Kasiri et al., 2012; Page, 2018).
Decision-makers in healthcare settings must identify the impact of EHR systems on clinical processes or diagnostic decision making to ensure benefits are significant to warrant future investment. Therefore, system dynamic modeling and simulation allows designers to map the impact of healthcare decisions on human capital, policy initiatives, strategic partnerships, technology investments, and other considerations. Multiple models may be needed to properly capture unintended effects of distributed cognition upon decision making when using Health IT systems. From a system dynamics and simulation perspective, system dynamics modeling of decision-making in the presence of cognitive overload may help researcher to better understand the unintended effect of healthcare misdiagnosis. Better understanding of the correlation between decision making and its unintended consequence of misdiagnosis may arm researchers with the initial tools to properly address factors which influence misdiagnosis errors. As a result of this research initiative, we will be able to contribute to practice by shedding light on decision making from a distributed cognitive perspective within healthcare. In addition, this research will provide us with the opportunity to contribute towards literature on decision making, distributed cognition, and the broader literature on sociotechnical systems (Blobel, 2019; Eppstein et al., 2012; Kasiri et al., 2012; Page, 2018).
References
Blobel, B. (2019). Challenges and Solutions for Designing and Managing pHealth Ecosystems. Frontiers in Medicine, 6(83). doi:10.3389/fmed.2019.00083
DeLanda, M. (2016). Assemblage theory: Edinburgh University Press.
Eppstein, M. J., Horbar, J. D., Buzas, J. S., & Kauffman, S. A. (2012). Searching the clinical fitness landscape. PloS one, 7(11), e49901.
Hollan, J., Hutchins, E., & Kirsh, D. (2000). Distributed cognition: toward a new foundation for human-computer interaction research. ACM Transactions on Computer-Human Interaction (TOCHI), 7(2), 174–196.
Hutchins, E., & Klausen, T. (1996). Distributed cognition in an airline cockpit. Cognition and communication at work, 15–34.
Kane, G. C., Alavi, M., Labianca, G., & Borgatti, S. P. (2014). What’s different about social media networks? A framework and research agenda. MIS quarterly, 38(1), 275–304.
Kasiri, N., Sharda, R., & Asamoah, D. A. (2012). Evaluating electronic health record systems: a system dynamics simulation. SIMULATION, 88(6), 639–648. doi:10.1177/0037549711416244
Page, S. E. (2018). The model thinker: What you need to know to make data work for you: Basic Books.
Scott, P., de Keizer, N., & Georgiou, A. (2019). Applied Interdisciplinary Theory in Health Informatics: A Knowledge Base for Practitioners (Vol. 263): IOS Press.
Willis, M. (2019). I’m Trying to Find my Way of Staying Organized: the Socio-Technical Assemblages of Personal Health Information Management. Computer Supported Cooperative Work (CSCW), 28(6), 1073–1102.
