© Springer International Publishing Switzerland 2015
Vimla L. Patel, Thomas G. Kannampallil and David R. Kaufman (eds.)Cognitive Informatics for BiomedicineHealth Informatics10.1007/978-3-319-17272-9_33. Theoretical Foundations for Health Communication Research and Practice
(1)
Department of Educational Psychology, University of Illinois at Urbana-Champaign, Education Building, 1310 S. 6th St., Champaign, IL, USA
(2)
Department of Health Systems Science, College of Nursing, University of Illinois at Chicago, 845 S. Damen Ave., Chicago, IL, USA
Keywords
CommunicationCollaborationCommon groundInformation processing3.1 Introduction
At first glance, communication among clinicians may seem to be the least complicated component of the health care domain, which involves providing care for a wide variety of illnesses and age groups using an array of low and high technology diagnostics and treatment, and a continually mounting base of evidence. Yet, there is mounting evidence that points to serious problems in communication within health care. Annually, 98,000 deaths are attributed to errors in health care (Kohn et al. 2000), with an estimated 60 % attributed to avoidable communication failures (Joint Commission on Accreditation of Healthcare Organizations 2005). Other research using root cause analysis revealed that approximately 70 % of sentinel events (serious negative consequences involving the unexpected occurrence or risk of death or serious injury) are caused by poor communication (Cordero 2011) and that poor information sharing and coordination is linked to patient mortality in multiple settings (Kim et al. 2010; Knaus et al. 1986; Shortell et al. 1992; Williams et al. 2007). These findings should not be surprising given the evidence that miscommunication is an important contributor to errors in other complex and high stakes domains such as aviation (Davison et al. 2003).
These failures are brought about by multiple challenges of communicating in the high stakes health care domain. A key challenge is the high cognitive demands associated with managing biological complexity in health care. This task requires interpreting and sharing uncertain and dynamic patient information, including frequent unplanned and interruptive communication among clinicians from different disciplines who must collaborate despite discipline-specific terminologies and taxonomies. These challenges are exacerbated by rapid diffusion of health information and communication technologies that can increase communication complexity. In addition, the round-the-clock nature of health care requires frequent complete transfer of responsibility for patients within disciplines, or handoffs that increase the number of clinicians who take care of a single patient. Moreover, the high cost of health care leads to multiple federal regulations that directly impact how care is delivered. For all these reasons, communication challenges may be even greater in health care than in other complex domains that involve managing engineered systems. For example, in aviation, pilots’ interaction with aircraft usually yield predictable consequences with swift feedback, while in health care the effects of clinicians’ treatment of patients is much less predictable with more variable and sometimes delayed feedback about treatment success (Durso and Drews 2010).
Given the vast differences between patient-provider communication, communication about health in the media, communication within health care organizations between caregivers and non-caregivers, and communication between caregivers, we focus our discussion on the communication between clinicians who are involved in direct patient care. Despite the volume of research related to inter- and intra-disciplinary communication in patient care, this work is often not theoretically-based. For example, a recent systematic review of hand-off communication research found that only 34 % of studies were guided by theoretical frameworks (Abraham et al. 2014; also see Patterson and Wears 2010). Therefore, we will focus on existing theoretical foundations that can inform research to address these key challenges in health care that hold strong potential for improving the quality and safety of patient care. In the next section, we describe information processing and interactive theories of communication. We then summarize some important challenges related to communication in health care contexts and argue for the importance of communication theory for addressing these challenges.
3.2 Theories of Communication Relevant to Health Care
Several theoretical approaches help identify processes involved in health communication as well as factors that influence these processes and thus the success of communication. In this section, we review approaches that have influenced research about performance in complex domains such as aviation, and increasingly in health care.
3.2.1 Information Processing
A longstanding and fruitful approach assumes that communication can be explained in part in terms of mental processes involved in information exchange by producing and understanding messages. This approach has its origins in information theory, developed by Shannon and Weaver (1949). They distinguished technical communication (how accurately information is encoded, transmitted by sending them through a communication medium or channel, and decoded by interpreting the coded message), semantic communication (how well these codes convey meaning) and effectiveness (how well the message has the intended effect). Information theory most directly addressed the technical level. The approach has been successful in many ways, for example by guiding development of communication technology (e.g., speech synthesis and recognition systems) and explaining some aspects of communication success such as reducing the impact of channel capacity, noise, and related factors on speech comprehension (Wickens and Hollands 2000). Health care as well as other domains have greatly benefited from these improvements because of its reliance on many forms of voice communication (e.g., Interactive Voice Response systems, mobile phones, and dictation systems).
The information processing approach was elaborated during the cognitive revolution in which the mind was understood metaphorically in terms of the computer, and drawing upon linguistic theories of mental structures (Miller 2003). Communication was explained in terms of the cognitive processes required to produce and understand linguistic messages, as well as the cognitive abilities and resources that constrained these processes.
Speakers (or writers) generate ideas by activating concepts in long-term memory and assembling these concepts (and associated words) into ideas (represented as propositions) that are mapped onto syntactic structures to convey the ideas through speech (Levelt 1989). Word access and propositional encoding processes involved in speech planning are shaped by our cognitive architecture, such as the capacity of working memory, which constrains how much conceptual context can be active at one time, and thus the size of the planning unit, as revealed for example in patterns of pausing when talking (Levelt 1989). To understand these messages, listeners (readers) recognize spoken (printed) words, activate the corresponding lexical codes and concepts in long-term memory, and integrate these concepts into propositions or idea units. Understanding extended discourse involves identifying relationships among these ideas, often driven by identifying referents of co-referring expressions and drawing on knowledge to identify temporal, causal, and other relationships among the ideas (Kintsch 1998). Comprehension requires more than assembling concepts into propositions: the network of propositions (the textbase) must be interpreted in terms of what we know about the concepts in order to develop a situation model, or representation of the described events and scenes (Kintsch 1998; for application to issues of comprehension and knowledge representation in the medical domain, see Arocha and Patel 1994; Patel et al. 2002). Situation models may be concrete, reflecting our perceptual experience of the described situation (Zwaan and Taylor 2006), or more abstract, capturing the essential ‘gist’ of the message (Reyna 2008). Understanding as well as producing messages is constrained by limited cognitive resources. For example, readers tend to pause at the end of clauses and sentences, presumably to wrap up conceptual integration, so that the pattern of pausing when reading mirrors the pauses when producing the message (Stine-Morrow and Miller 2009).
Individual differences in cognitive resources such as working memory capacity help explain differences in successful communication. For example, age differences in comprehension of complex messages (e.g., conceptually dense text) often reflect age-related differences in cognitive resources (Stine-Morrow and Miller 2009). Knowledge, on the other hand, can facilitate comprehension, reducing need for effortful conceptual integration and inference processes (Kintsch 1998). Age differences in comprehension are often reduced for texts that are organized to match knowledge organization (Stine-Morrow and Miller 2009).
While the information processing approach has been successful both theoretically (supporting a large body of empirical research on language understanding and production processes and how they are shaped by cognitive resources) and practically (e.g., spurring development of communication technology), it is incomplete as an account of communication in complex domains. It is based on and guided by a conduit metaphor of communication, which assumes communication depends on how precisely linguistic codes match speaker ideas and how accurately listeners decode these ideas, or how well they ‘take away’ the message (Reddy 1979). Thus, the focus is more on participants’ processes and resources than on communication medium, context, and purpose. Although this approach recognizes the importance of pragmatic as well as semantic views of language (e.g., explaining message effectiveness is an important goal of information theory), much of the work within this framework focused on how people represent message meaning, more than on the actions performed by speakers when using language; e.g., semantic rather than pragmatic effects of discourse (Austin 1960).
3.2.1.1 Persuasion/Risk Communication
Other theories within this tradition focus on affective as well as cognitive processes in communication, how these processes interact to influence addressees, and how speakers design messages to influence addressee beliefs and actions. For example, the elaboration-likelihood model argues that persuasive effects of messages depend on both a direct route (addressee’s deliberative processing of message meaning) and on peripheral routes (addressee’s responses to indirect, secondary aspects of the message context) (Petty and Cacioppo 1986). This research addressed social psychological problems related to attitude change, but the mechanisms presumed to underlie persuasive effects of messages were similar to information processing models of text processing. Similarly, cognitive models of risk communication focus both on how addressees represent message meaning and how these representations interact with beliefs and domain knowledge in order to influence behavior. For example, mental model theory focuses on how to design messages to target addressees’ conceptions of risk (Morgan et al. 2001). The fuzzy trace theory argues that addressees can understand risk at both verbatim and gist-based levels, which have different effects on decision making and action (Reyna 2008). A large body of research that builds on such theories investigates how to leverage technology to deliver health-related messages tailored to addressee characteristics (Kreuter and Wray 2003).
3.2.2 Communication as Interaction
Interactive theories analyze how meaning emerges from coordination of speaker and listener actions (“sense-making in the moment”), and how this activity is influenced by constraints imposed by communication media and context and by speakers’ and addressees’ cognitive resources. An important idea is that communicative success depends on the interaction of participants’ cognitive resources, with joint attention as a key resource. Therefore, interactive theories can be interpreted within the framework of distributed cognition, which analyzes human performance, including communication, as emerging from cognitive resources distributed across social contexts such as conversational partners, and external contexts such as tools (Hutchins 1995; a detailed description can be found in Chap. 2 in this volume).
3.2.2.1 Common Ground
Conversational partners communicate by coordinating cognitive effort in order to construct meaning (Clark 1996). To do this, they ‘ground’ information, or agree that the information is mutually understood and acceptable (accurate and relevant to joint goals). Grounding rests on and contributes to shared situation models (Morrow and Fischer 2013). In this view, speakers not only ensure that they are understood, but collaborate with their addressees to create meaning. For example, speakers may implicitly invite their addressees to co-construct their message by presenting an intentionally ‘underdeveloped’ contribution, encouraging their addressees to help specify the content. For example, a resident might say “I checked that patient.”, prompting a nurse to complete the contribution “The one with edema?” “Right. It’s down”. Speakers may also present a message they think is clear, but the addressee’s request for clarification reveals that it is undeveloped (Clark 1996; Coiera 2000). Thus, ongoing feedback is essential to communication. For example, stories are more understandable when speakers receive immediate feedback from their addressees (Bavelas et al. 2000).
Communication depends on joint or collaborative effort, as well as the individual effort involved in producing and understanding messages (Clark 1996; Coiera 2000). Developing common ground itself requires effort. Pre-emptive grounding involves devoting effort ahead of time to develop shared knowledge about the message domain, communication strategies, and other aspects of communication (Coiera 2000). The upfront work involved in pre-emptive grounding facilitates communication as it occurs (e.g., little cost to grounding during communication). For example, aviation communication depends heavily on shared knowledge about aviation concepts, terminology, and communication procedures that is acquired by Air Traffic Controllers and pilots as part of their professional training. This shared knowledge enables rapid and efficient communication during flight operations (Morrow and Fischer 2013). Similarly, in health care, two providers may discuss patients before a formal hand-off in order to expedite the later conversation. Just-in-time grounding, on the other hand, involves devoting effort during, rather than before, communication. In this case, partners share less knowledge about the domain and/or the conventions of communication, and so must devote more effort ‘on the fly’ during communication (Coiera 2000). For example, communication between domain experts and novices or between experts from different subdomains (e.g., physicians and nurses) often requires partners to be more explicit, devoting more effort to ground contributions.
Common ground theory helps to refine the view of how cognitive resources constrain communication. Speakers initiate contributions by getting their addressee’s attention using verbal or nonverbal (e.g., gestures) cues. They then present messages that are designed to be understood based on common ground (shared knowledge of the language, cultural context, as well as more specific concepts that are relevant based on the prior discourse and context of communication). Listeners not only understand the message (which involves activating and integrating concepts, as described above), but also signal to the speaker that they do or do not understand. In the latter case, they may implicitly (e.g., puzzled expression) or explicitly request clarification. The speaker and addressee accept the message as mutually understood, so that the contribution enters common ground. These phases typically overlap: speakers often initiate contributions by presenting messages, and listeners often signal acceptance by responding to the message with a relevant contribution (Clark 1996).
Interactive approaches emphasize that communication depends on partners’ collaborative or joint effort. For example, Air Traffic Controllers communicate by radio with many aircraft in the same air space in order to manage the flow of traffic. They may try to reduce their own effort by presenting one long, rapidly delivered message to a pilot rather than breaking it into several shorter messages that require more radio time and complicates the task of talking to multiple pilots on a single radio line. However, this strategy may increase the addressed pilot’s effort involved in understanding and accepting the long message (by ‘reading back’ or repeating key concepts from the message in order to demonstrate understanding and help establish common ground). It also increases the likelihood that the pilot misunderstands and requests clarification or that the pilot does not explicitly accept the message at all (responding with minimal or no acknowledgement). The controller in turn must spend more radio time in order to ‘close the communication loop’ by clarifying their message and seeking confirmation that the pilot understood. The upshot is increased collaborative effort (Morrow et al. 1994). In health care, outgoing ICU nurses who hand off patients to incoming nurses may overestimate common ground when the incoming nurses are already familiar with the patients, so that they present overly abbreviated reports. This strategy minimizes their own effort involved in hand-offs, but at the expense of the incoming nurse who is likely to misunderstand and to request clarification, resulting in increased collaborative effort (Carroll et al. 2012). In short, successful communication depends on partners’ ability to coordinate contributions through shared attention and collaborative effort. Such communication problems in turn contribute to adverse events that reduce patient safety (The Joint Commission 2005).
3.2.2.2 Common Ground and HCI in Health Care
Common ground theory (and distributed cognition theories more generally) is important for identifying factors that influence communication in health care settings, which depends heavily on technology and is often distributed over space (synchronous remote) and time (asynchronous remote), as well as occurring face-to-face. Next, we consider how collaborative effort and communication success depend on resources related to communication media, participants (e.g., cognition), and health care tasks. We also consider the role of technology as an external resource that shapes communication.
Communication Media
Media differ in terms of the opportunities they afford for, and constraints they impose on, establishing common ground (Clark and Brennan 1991; Monk 2008). In face-to-face communication, partners are co-present and typically see and hear each other as well as the referent situation. They can use nonverbal (gesture and facial expression) as well as verbal resources in order to coordinate attention on linguistic information as well as the nonlinguistic context when presenting and accepting messages. Communication can be efficient, with less need for elaborate verbal description compared to other media (Convertino et al. 2008; Gergle et al. 2004). Turn-taking is rapid, with messages received almost as they are produced (contemporality), the possibility of signaling that a message is understood as it is presented (simultaneity), and the order of contributions easily determined (sequentiality) (Clark and Brennan 1991). Face-to-face communication is especially suited for coordinating to accomplish joint tasks such as performing surgery. Nonverbal cues are also critical for conveying emotion, which is important for provider-patient communication. For example, provider communication behaviors such as leaning toward the patient and using facial expressions that convey concern predict patient satisfaction (Ambady et al. 2002).
Face-to-face communication also has drawbacks. Communication at work is often complex, with partners having to keep track of interacting topics or conversational threads. This complexity requires partners to easily access concepts from prior discourse (reviewability) and revise contributions in light of the evolving discourse (revisability) (Clark and Brennan 1991). The transient nature of speech complicates comprehension of complex messages because of listeners’ working memory limits. More generally, the time pressure of face-to-face communication may preclude the deliberation needed to craft complex messages, as well as to understand them.
Other media may impose more constraints on communication, providing fewer resources for grounding. Synchronous communication with partners at different locations (e.g., telephone), like face-to-face communication, allows rapid turn-taking, contemporality, and simultaneity that support grounding (e.g., immediately indicating and repairing comprehension problems), but eliminates visual cues (unless using videophone), which can increase collaborative effort involved in accepting contributions (Clark and Brennan 1991). Synchronous remote communication has become pervasive with mobile phone technology. When partners are not visually co-present, this medium may increase overall workload because speakers cannot modulate their communication as required by listener context. A good example comes from driver distraction research. Driver-passenger conversation is less likely to disrupt driving (e.g., lane control; likelihood of seeing highway exit) compared to cell phone conversation, in part because driver and passenger are co-present and the passenger can modulate their talk as required by the situation (Drews et al. 2008). Texting, like online chat, is similar to synchronous remote communication because proficient texters use compressed language that allows rapid turn-taking, although this medium does not allow contemporality. Texting also eliminates auditory (e.g., speech prosody) cues for grounding and affective messages, which may be remediated in part by innovative use of punctuation and other symbols that convey affect. Texting also provides a record of the message, which supports reviewability.
Grounding can be even more challenging for asynchronous remote communication such as email, which lacks contemporality, simultaneity, and sometimes sequentiality. More effort is needed to produce messages (typing vs speaking), which influences individual and collaborative effort involved in grounding. Asynchronous voice communication (such as exchanging pre-recorded messages) is often less effective than synchronous communication (telephone conversation). For example, introducing an EHR system in an Emergency Department may increase the use of EHR-based emails between nurses and physicians about patient treatment plans, which reduces face-to-face communication that helps to clarify and elaborate shared treatment plans. On the other hand, email supports message reviewability and revisability and affords time for deliberation, which may result in more comprehensive and understandable messages (Olson and Olson 2007).
Communication Media and Tasks
The effects of media-related constraints depend on the tasks that people communicate about (Zigurs and Buckland 1998). For example, remote communication may be more appropriate for tasks that hinge on message reviewability and revisability, such as integrating multiple sources of complex information in order to diagnose an illness or troubleshoot a problem. Face-to-face or synchronous remote media may be more effective for tasks requiring frequent interaction to accomplish goals or to resolve conflicting goals (negotiation, persuasion). For example, an unstable ICU patient receiving provisional treatment that requires close monitoring would be better served by a face-to-face nurse handoff rather than an asynchronous handoff based on a phone message from the outgoing nurse.
Participant Resources
While information processing theories focus on speaker and listener cognitive resources (e.g., attention, working memory) needed to produce and understand messages, interactive theories emphasize that resources are also essential for grounding contributions to build a shared situation model, so that communication success depends on collaborative effort. For example, speakers with fewer cognitive resources, either because of long-term effects such as aging or short-term effects such as fatigue or distraction, may take more short-cuts when producing messages (resulting in more elliptical or vague messages) potentially complicating message comprehension and grounding. Conversely, listeners with fewer resources are less likely to explicitly acknowledge contributions, providing less evidence for comprehension and undermining grounding. Older adults with fewer cognitive resources may be less adept at tailoring message presentation to listeners based on common ground (Horton and Spieler 2007). On the other hand, shared knowledge about language, the discourse topic, and other aspects of communication can reduce effort and support grounding. This knowledge arises from partner familiarity and membership in a variety of linguistic/cultural communities and includes social norms and conversational conventions (Clark 1996). Partners who share knowledge about the discourse topic more quickly establish co-reference (Isaacs and Clark 1987) and more effectively perform joint tasks by coordinating attention to key information (Richardson et al. 2007), suggesting knowledge reduces collaborative effort. Shared knowledge may also support retrieval of previously mentioned information (Ericsson and Kintsch 1995). These benefits may explain why experts benefit more than novices from collaboration (compared to working alone) when recalling and acting on domain-relevant information (Meade et al. 2009).
Technology and Communication
Technology shapes communication in many ways, especially by expanding the repertoire of media options that create new communication opportunities and constraints. In complex environments such as hospitals, where work is typically done by multiple distributed and interacting teams, technology provides many options for remote communication, both synchronous (chat, texting, videophone, electronic status boards) and asynchronous (voicemail, email, clinical messaging in EHRs).
Technology, broadly considered as “cognitive artifacts,” includes paper-based tools such as notes or whiteboards (Hutchins 1995; Nemeth et al. 2004). Distributed cognition theories analyze how artifacts, a pervasive part of work environments, reduce the need for mental computation, memory search, and other effortful cognitive processes involved in producing and understanding messages. They can support grounding by providing easily shared external referents that reduce need for explicit description (Monk 2008). They also support reviewability in face-to-face communication (Gergle et al. 2004) and visibility in synchronous remote communication such as teleconferencing because speakers can gesture to guide attention to information on the tool (Monk 2008; Whittaker et al. 1993). They especially benefit older adults by reducing demands of speech production and comprehension on cognitive resources (Morrow et al. 2003).
However, it is important to note that the benefits of technology-based tools can be overestimated. For example, using video-conferencing or other technology to distribute information to remote team members does not in itself ensure that people work together to effectively ground and act on this information. More generally, such technology may tempt us to distribute large amounts of information to as many people as possible. This strategy may undermine communication and reduce safety if information distribution is not guided by strategies to manage joint attention to the information that is most relevant within an evolving mental model organized around shared goals.
There is much interest in designing health information technology that flexibly supports a range of communication and task goals in different workflow situations. An important challenge for communication theory and research is to understand and predict strengths and limitations of different communication media and tools in complex environments such as the ICU or primary care clinics. This requires integrating theories that identify how participant-related and media-related resources interact with tasks to influence work, with macro analysis of work processes in organizations. Macro-level theories focus on how system-level outcomes such as safety and efficiency emerge from the interaction of system levels, e.g., individuals interacting with devices in the context of teams, management, organizational policies and practices (Carayon et al. 2013; Kaufman et al. 2014). Such an approach would analyze communication in organizations, with technology-related factors at different system levels interacting to influence communication between dyads, teams, etc.
Another communication issue that becomes increasingly important as clinicians routinely collaborate with technology, rather than using technology to collaborate with each other, is how common ground is established between people and technology. People tend to take a ‘social stance’ toward technology and treat it as a communication partner (Reeves and Nass 1996). In addition, technology must be able to reciprocate by building up and acting on common ground with their human partners. This requires technology to update and reason from a model of the user’s context during communication (Coiera 2000). A simple example comes from the literature on automated tutoring systems, which are most effective when providing prompts that scaffold student learning based on a model of the student’s current state of knowledge about the target domain.
3.3 Key Challenges for Health Care Communication
3.3.1 The High Cognitive Demands Associated with Biological Complexity
Health care is an information intensive domain that includes knowledge of normal and abnormal physiology, pharmacology, multiple treatment options and health specialties, health system and organizational infrastructure, a high volume of new clinical evidence, and longitudinal information about patients, families and their communities. Delivering care to patients and communicating about care is therefore a highly complex endeavor characterized by both uncertain responses to treatment interventions and changing patient conditions (Glouberman and Mintzberg 2001). For acutely ill patients, each change in their condition requires the clinicians to reorganize and reinterpret multiple sources of data (e.g., lab values, physical exam, vital signs and patient’s subjective responses) to inform their next decision (Coiera and Tombs 1998; Collins et al. 2007; Edwards et al. 2009; Grundgeiger and Sanderson 2009; Tucker and Spear 2006), which further increases cognitive demands. Thus the potential for information overload and its associated safety implications is very high (Beasley et al. 2011).
The cognitive complexity of health care both reflects, and in turn contributes to, the need to perform multiple, interleaved tasks that result in pervasive interruptions (Coiera and Tombs 1998; Collins et al. 2007). Interruptions increase clinicians’ cognitive load because of the need to recall the interrupted tasks, introducing risk for confusion and error (Tucker and Spear 2006) and for forgetting critical tasks (Collins et al. 2007). Interruptions also increase the complexity of communication, making it more vulnerable to error during hand-offs (Behara et al. 2005), when caring for patients in ICUs (Grundgeiger and Sanderson 2009), and in many other clinical tasks. Cognitive load can be exacerbated by poorly designed information displays and electronic interfaces that fragment, rather than integrate, information needed to perform multiple tasks, which can increase the frequency and consequences of interruption. A multi-site study of EHRs found that health care workers waste much time sifting through multiple sources of the information to get a true picture of a patient’s situation, which is needed for effective communication (Stead and Lin 2009). Similarly, a study of intensive care nurses found that information needed to perform many of the common nursing tasks were inaccessible, difficult to see, and/or located in multiple displays (Koch et al. 2012). In sum, problems associated with cognitive demands in health care have important implications for communication and lead to ineffective decision making, important tasks left undone and high potential for error. These implications are more easily seen when communication is analyzed as emerging from, as well as contributing to, clinical workflow (Kaufman et al. 2014).
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
