November 2014, VOLUME 203

Recommend & Share

November 2014, Volume 203, Number 5

FOCUS ON: Health Care Policy and Quality


Communication in Diagnostic Radiology: Meeting the Challenges of Complexity

+ Affiliations:
1Department of Radiology, Stanford University School of Medicine, 300 Pasteur Dr, Stanford, CA 94305-5105.

2Department of Operations, Business Analytics, and Information Systems, College of Business, University of Cincinnati, Cincinnati, OH.

3Center for Health Systems Excellence, Cincinnati Children's Hospital Medical Center, Cincinnati, OH.

4Department of Radiology, Cincinnati Children's Hospital Medical Center, Cincinnati, OH.

Citation: American Journal of Roentgenology. 2014;203: 957-964. 10.2214/AJR.14.12949

Next section

OBJECTIVE. As patients and information flow through the imaging process, value is added step-by-step when information is acquired, interpreted, and communicated back to the referring clinician. However, radiology information systems are often plagued with communication errors and delays. This article presents theories and recommends strategies to continuously improve communication in the complex environment of modern radiology.

CONCLUSION. Communication theories, methods, and systems that have proven their effectiveness in other environments can serve as models for radiology.

Keywords: communication, complexity, information technology, value stream

The primary business of diagnostic radiology is information. Radiology departments receive patient history and order in for mation from referring clinicians. They then extract information from the bodies of patients in the form of images, which are interpreted within the corresponding clinical context—thus creating more information—and both the imaging and interpretation data are then communicated back to the referring clinicians and patients and incorporated into clinical care.

Historically, the primary focus of radiologists as well as that of the health system's payment mechanisms has been on acquiring and interpreting imaging information; the efficient and reliable flow of that information has often been a relative afterthought. Thus, communication methods and systems that support the flow of information in radiology often fall far short of what is possible—or even of what has become routine—in many other domains of modern life.

In many industries, modern information systems function extremely reliably, accommodate and adjust for human error, and often predict and fulfill users' needs even before users are aware of those needs. In contrast, radiology workflows and information systems are often slow, fragmented, difficult to navigate, inflexible, and opaque. The problems created by this dysfunction are not trivial and range from frequent interruptions and wasted time to patient harm because of missed findings and miscommunication [13].

A major contributor to the current state of affairs is the rapidly increasing complexity of modern health care [4, 5]. In a small and simple organization, the impact of an individual's actions tends to be proximal, visible, and predictable. In such settings, informal communication mechanisms inherent in close and frequent interaction, which users may not even recognize as communication mechanisms, serve to keep care providers in sync. When questions and problems arise, they can be relatively easily addressed. In contrast, in a complex environment, many of the intuitive strategies that worked on a small scale lose their effectiveness. The impact of the information provided by the individual to an unknown individual elsewhere in the organization rather than to a known colleague often is unseen and far removed, creating problems elsewhere that are difficult to detect, correct, and prevent.

Radiology practices are becoming larger, more specialized, and increasingly integrated into other parts of the health care system [6]. Thus, the ability of a radiology department to provide safe, timely, and effective care increasingly depends on the sophistication of its communications methods.

In this article, we present a model for recognizing the value of communication methods and systems in radiology departments, illustrate how the complexities of modern health care systems make them difficult to manage, introduce theories related to communication in complex environments from the business management and communication literature, and present a regimen of strategies to continuously improve radiology communications systems and processes. Underlying our arguments is the belief that radiologists along with their clinician colleagues and technical and administrative partners should assume responsibility for the entire imaging process, from the time the clinician poses a question that could be answered with imaging to the point that the images and interpretation have been appropriately incorporated into patient care.

Fundamental Model: The Value Stream
Previous sectionNext section

In approaching the topic of how communication in radiology should function, it is reasonable to start with the question of what is the core function of diagnostic radiology. In other words, what is the primary way in which radiology provides value to the care of patients? We propose that the fundamental purpose of diagnostic radiology is to help guide clinical care and predict patient outcomes through the acquisition, interpretation, and communication of medical imaging information. In other words, the need for imaging is invoked when a patient presents with a clinical question that may be answered with imaging. Images are acquired and interpreted on the basis of that clinical question and communicated back to the referring clinician, providing a clearer picture of the patient's condition. This new information is valuable to the patient and to the referring clinician, who is acting as the patient's fiduciary, in that it guides clinical management and increases prognostic certainty. The process of creating that value can be described as a value stream, which is defined as the activities required to produce a good or service from raw material to possession by the customer [7].

In this light, a radiology department's primary product is information, which is generated and refined as it moves step-by-step through the workflow. This information value stream may be depicted in a simplistic manner as a linear chain of processing nodes connected by communication channels that pass along the information until it reaches its final destination (Fig. 1). Thus, the effectiveness of the value stream depends on the performance of both the processing nodes and the communication channels. Productivity in diagnostic radiology has dramatically increased over the past few decades, in part through the establishment of information workflows similar to those of manufacturing processes [8]. Although this has undoubtedly improved the value that radiologists provide to patients, this state of increased productivity presents new challenges that call for new solutions. Perhaps the greatest challenge is that of increased complexity.

View larger version (20K)

Fig. 1 —Highly simplified flowchart shows value stream in radiology, which begins when clinician has question that can be answered by imaging and ends when information is used in clinical care. Information is passed from one event or processing node to another via communication channels. EMR = electronic medical record.

Sources of Complexity
Previous sectionNext section

Truly linear value streams are rare in information-intensive fields such as radiology. For example, the radiology order may come from one of several clinical specialties, the examination may be performed in either the in-patient or outpatient setting, the examination may include one of several imaging modalities performed according to one of many protocols and postprocessing algorithms, the images may be interpreted by one of many types of subspecialists, and the report may be communicated by one of several communication methods. Thus, each processing node depicted in Figure 1 is actually composed of many nodes and the number of linkages between them is increased accordingly. Hence, the value stream should probably more realistically be thought of as a value network.

Even this network model is highly over-simplified in that each task is usually performed by one of many operators using a variety of equipment types, often at multiple physical locations and using numerous protocols. Without the motivation and direction to do otherwise, each operator tends to develop unique technical performance and communication styles. This may be of minor consequence in a small system when the receiver becomes familiar with the sender's style and can learn to interpret the meaning of the communication accordingly. However, in a large system, the interactions begin to resemble those depicted in Figure 2. In such a system, each sender naturally believes that he or she is communicating information to the receivers in a consistent manner because, from his or her perspective, the same communication style is used for every transaction. However, from the perspective of the receiver, incoming communication styles across senders are inconsistent and difficult to interpret and predict. Because of the difference between the perspectives of the sender and receiver, it is difficult for the sender to appreciate the burden that use of a unique communication style imposes on the receiver.

View larger version (295K)

Fig. 2A —Communications in complex environment.

A, Charts show many-to-many communications in complex environment. When each sender consistently uses his or her own unique communication style, as represented by line colors in A, receiver must cope with extensive variability across senders. From perspective of each sender (B), communication formats are consistent. From perspective of each receiver (C), communication formats are highly inconsistent.

View larger version (63K)

Fig. 2B —Communications in complex environment.

B, Charts show many-to-many communications in complex environment. When each sender consistently uses his or her own unique communication style, as represented by line colors in A, receiver must cope with extensive variability across senders. From perspective of each sender (B), communication formats are consistent. From perspective of each receiver (C), communication formats are highly inconsistent.

View larger version (93K)

Fig. 2C —Communications in complex environment.

C, Charts show many-to-many communications in complex environment. When each sender consistently uses his or her own unique communication style, as represented by line colors in A, receiver must cope with extensive variability across senders. From perspective of each sender (B), communication formats are consistent. From perspective of each receiver (C), communication formats are highly inconsistent.

Although the value stream in radiology is often modeled as a unidirectional flow of information, this model is overly simplistic. Communication is the meaningful exchange of information between individuals or groups of individuals; it is often bidirectional or multidirectional and is successful when it results in shared understanding of meaning. Communication may occur through, or be facilitated by, a variety of media, especially with modern technology. Increasing the number of available communication channels can enable better matching of the media with the message, but it also increases the complexity of the system, which, in turn, must also be managed.

These and other factors combine to create a seemingly overwhelmingly complex system that is difficult to manage and improve. Despite the complexity, however, because most of these communications and processing tasks are frequently repeated, they can be studied; improved; and, in many cases, automated.

Theories and Frameworks
Previous sectionNext section

Much of the success of modern communication systems is based on theories and frameworks in the communication and business management literature that help individuals understand, manage, and effectively operate in complex work environments. We present several that we find to be especially applicable to managing communication in modern radiology practice.

Information Theory

In 1948, Claude E. Shannon, a mathematician working at AT&T Laboratories, published a highly influential mathematic theory of communication that later became the basis for information theory [9]. The article describes communication as transmission of symbols from an information source to a destination via a noisy signal channel (Fig. 3). The model is based on the concept that a receiver cannot assume that the received message necessarily matches the sent message and that the receiver can estimate the actual content of the transmitted message on the basis of probabilities derived from experience. This estimate can then be confirmed and corrected with efficient verification and clarification strategies, much of which can now be performed nearly instantaneously by computers.

View larger version (13K)

Fig. 3 —Diagram shows general communication system by developed by Claude E. Shannon [9], originally published in 1948. (Reprinted with permission of Alcatel-Lucent USA Inc.)

Information theory serves as the technical basis for essentially all modern automated communication systems. It formally recognizes that for every communication there is a certain probability that the meaning of a message is incorrectly conveyed to the receiver and that corresponding error identification and corrective strategies can and should be deployed accordingly. Every communication transaction may be subject to noise, defined as any unpredictable perturbation that causes the received signal to be different from the transmitted signal. This perturbation may take many forms, from alterations in electrical currents to side conversations in a crowded room to typographic errors on a keyboard. Information theory also recognizes that the number of possible meanings of a message, although potentially very large, is ultimately finite; hence, there is value in learning from past communications to improve future communications, which can overcome miscommunication caused by noise.

This philosophy and its technical derivatives that underpin modern information systems often benefit radiologists in ways that may not be fully appreciated. For example, if a radiologist interpreting a pediatric renal ultrasound study wishes to determine a normal renal measurement for age, he or she may type “pediatric normal m” into the Google search application and the autocorrect feature will complete the term as “pediatric normal measurements,” which, if it is accepted, returns a list of web pages [10]. The first site returned by the search is the Oregon Health Sciences' Pediatric Normal Measurement page, which contains a comprehensive set of measurements of normal pediatric organ sizes on the basis of age, including renal measurements [11]. This requires just a few seconds of operator time and very little specialized operator skill, using the same tool one might use to search for local movie listings. In fact, if the user misspells pediatric as “pedmiactic” and normal as “nroml,” for example, it still provides the same result [12]. In other words, the system anticipates human error and adjusts accordingly. This type of activity occurs throughout the world billions of times per day and has essentially become a commodity in many information-intensive fields. However, this level of sophistication is only beginning to enter the domain of radiology systems.

Probability Modeling

As presented by Shannon [9], the concept of information is closely related to probability modeling in which the likelihood of an event within a set of alternative hypotheses can be estimated on the basis of available data and can be further adjusted as additional information is acquired [13]. For example, for a correction algorithm to be able to recognize that “pedmiactic” should be interpreted as “pediatric,” the algorithm must recognize that, as with any word that is entered, there is a chance that it was entered in error. On the basis of past user performance, the system compiles a list of possible alternatives to the entered word and assigns to each the probability that it is the word that the user actually meant to enter. The word that is determined to be the most likely to be correct is then passed into the search algorithm. Modern search systems are continuously searching the Internet and assign each page a rank value according to words or combinations of words (which is what enables them to be so fast). Therefore, on the basis of these prior searches, the program provides a list of web-sites that are estimated to be the most likely to correspond with the word or combination of words passed to it.

A common use of probability modeling in medical practice is that of the differential diagnosis. For example, a primary malignant sarcoma of the bony pelvis may have a variety of causes, such as chondrosarcoma, osteosarcoma, Ewing sarcoma, lymphoma, fibrosarcoma, malignant fibrous histiocytoma, and malignant vascular tumor [14]. The incidence of these tumors can be calculated on the basis of research experience, which has been shown to be 32%, 22%, 22%, 14%, 5%, 2%, and 2%, respectively. On the basis of age, sex, and clinical presentation, the likelihood of each of these lesions can be further adjusted before the corresponding images are viewed and is then modified even further by the radiologist according to the imaging characteristics and guided by his or her experience. Thus, the conditions to be included in the differential diagnosis are first enumerated and then probability modeling is applied to determine which are most likely.

Probability models are built into many of the information systems used in daily activities, including voice recognition systems, mapping and navigation applications, and online retailing, to name a few. It is one of the key principles that allow self-correcting information systems to adjust for human error and still function reliably. Probability models also underlie many aspects of medical care, although sometimes in less-recognizable ways. For example, on the basis of the American Association for the Surgery of Trauma organ injury scaling system, a solid organ injury can be placed into one of several predefined categories. These categories have prescriptive implications for treatment and predictive implications for outcomes [15]. Predictive modeling also has implications for health care operations. For example, communication with a referring clinician may be performed by several alternative avenues, including routine insertion of the report into the electronic medical record, faxed report, telephone call, video conference, or in-person physician-to-physician communication. Once the various methods of communication have been enumerated, the appropriate method of communication can be determined algorithmically with a reasonable degree of accuracy according to the clinical situation using data from experience. In this way, communication systems can be designed that can learn from past events to anticipate and automatically prepare for predictable future events.

Media Richness Theory

Since its introduction, information theory has expanded considerably and has led to several other communication theories as the number of available communication media has expanded. One of these is media richness theory, which recognizes that different communication media contain different levels of richness, defined as “the ability of information to change understanding within a time interval” [16]. As Lengel and Daft state [17], “The more learning that can be pumped through a medium, the richer the medium.” For example, in-person communication is extremely rich in that it can handle multiple information cues simultaneously, facilitate instantaneous feedback, establish a personal focus, and use natural language. The electronic transmission of a scientific report, on the other hand, is extremely lean in that although a large amount of information can be transmitted in a few seconds or less, understanding has not been substantially changed in those few seconds. Rich media tend to be less lean in that they place greater demands on individuals in terms of the time and effort required to coordinate and perform the communication. Lean media, on the other hand, tend to be less rich in that they are limited in the degree of subtlety that can be conveyed and rapidity with which clarification can occur.

The appropriate level of richness depends on the needs of the communication transaction. For example, the report of a normal chest radiography examination performed as screening for tuberculosis can generally be communicated via a lean medium, such as a radiology report transmitted electronically to the patient's medical record, to be reviewed at the convenience of the ordering clinician. On the other hand, if that same patient has imaging findings suspicious for active tuberculosis, the findings should be provided through a rich medium that can facilitate rapid and unequivocal communication, such as telephone or in-person review, that accommodates discussion regarding clinical context and confirmation of accurate receipt of the message by the appropriate clinician. Because a variety of communication media may be required in radiology depending on the needs of a given situation, appropriate media channels must have been already instituted before those situations arise.

In general, the use of rich communication media in situations in which lean media would suffice leads to waste; the use of lean communication media in situations in which rich media are required leads to communication errors. As communication transactions are repeated and standardized, situations that previously required rich media generally can be increasingly accommodated by lean media. For example, a “code blue” page in a hospital is a lean communication that may convey a large amount of information to its recipients within a short message—as long as those participating in the process have previously invested the resources to reach a shared understanding of the meaning of the term.

Media Synchronicity Theory

The media synchronicity theory extends the media richness theory and addresses the degree to which individuals must work together at the same time to achieve common understanding [18, 19]. It posits that communication can be broken into two fundamental processes: conveyance and convergence (Fig. 4). Conveyance is the act of transmitting information from a sender to a receiver. Convergence, in contrast, involves processes of verification, discussion, and clarification until both parties recognize that they mutually agree (or fail to agree) on the meaning of the information. For highly standardized routine unambiguous communications (such as placing an order at a fast food restaurant), conveyance dominates the transaction, with minimal convergence activities. For complex nonroutine communications that involve a large amount of new information or uncertainty or are of critical importance (such as a salary negotiation), convergence activities tend to dominate. Convergence-dominated communication activities rely on the rich back-and-forth exchange of small packets of information between parties. In contrast, for lean conveyance-dominated communication activities, there is less need for the parties to interact simultaneously; the sender might provide a large amount of information at one point in time with the receiver handling it later when he or she is ready to process it. Hence, the degree to which conveyance or convergence dominates the transaction greatly influences how lean or rich it is. The need for synchronicity has significant implications for the design of communication systems in radiology; efficiency can be improved by establishing robust conveyance channels, but safety and quality may be compromised if convergence channels are degraded.

View larger version (46K)

Fig. 4 —Illustration shows communication transactions in radiology workflow, derived from media synchronicity model [19]. Although conveyance is primary means by which information flows through value stream, it also flows into, through, and out of stream by many other channels. Organizational and information system design should accommodate these information flows with intelligent blend of formal and informal channels.

Other information flows exist outside the primary communication stream, which are also illustrated in Figure 4. For example, in addition to images, relevant information such as patient history, allergies, scheduling information, and examination-specific details are also passed through the chain by formal and informal means. When the primary workflow is disrupted, contingency workflows are deployed so that the radiology practice may continue to function. Furthermore, individuals often also rely on additional information, policies, references, consultations, and tools to help them complete a task correctly. Communication systems can enable these auxiliary information sources to be delivered to the individual performing the task just at the time that it is needed in a way that filters out distracting information (“clutter”). Radiologists' communications also provide value to referring clinicians outside the primary value stream through activities such as case consultations, imaging recommendations, and multidisciplinary conferences. To facilitate continuous improvement of the systems, individuals also need an effective means of communicating feedback to system designers and managers.

Systems Thinking

The remarkable achievements that occur in modern health care, including in radiology, are based in large part on the explosion in scientific knowledge derived from decades of societal investments in biomedical research [20]. The application of this knowledge requires high levels of skill and specialization. To develop, maintain, and apply that level of skill, organizations tend to establish divisions along specialty lines. Although specialization is necessary to achieve the organization's goals, it tends to create silos, which begin to function as autonomous units that may be far removed from the customers [21]. Whereas customers' purchasing power serves as the control that helps ensure that the organization's efforts remain in line with customers' needs, that effect is diluted as it disseminates to the organization's silos. Subunits tend to try to optimize their respective functions on the basis of their local objectives and develop unhealthy rivalries that may be at odds with, or even undermine, the organization's global objectives. Thus, for the organization to work effectively to consistently satisfy customers' needs, the subunits must be organized and managed in a way that enables them to continuously understand those needs and work cooperatively with other subunits to fulfill them. Effective communication systems are critical to maintaining that cooperation.

In health care, individual workers are generally highly skilled experts who tend to understand in detail the processes under their direct control, optimizing their local performance as best they can—usually with admirable intentions. However, they cannot appreciate all of the complexities of the larger organization or understand how certain details of their work that may be insignificant to them might actually be highly significant to a recipient of their work—a recipient who often is far removed in terms of distance, time, and organizational affiliation [22]. Central authorities and system managers, on the other hand, generally better understand the larger goals of the organization but cannot understand the details of individual processes as well as the local experts do or comprehend the subtle interactions between individuals and between subunits. Hence, there is a natural tension between the desire for flexibility and autonomy at the local level and the desire for centralized control by higher authorities and system designers. Systems thinking seeks to broker this tension by focusing on the relationships between the parts of a system, how the system works over time, and how the parts operate within the larger environments [23]. Thus, according to this philosophy, leaders in an organization should respect and support the individual's ability to learn and adapt locally, and individuals should respect the need to optimize the organization as a whole rather than only within their own local domains.

Appropriate Standardization

Standardization is the fundamental basis of successful performance and communication in a complex organization [24]. For two individuals to arrive at shared meaning, the individuals must share a common language, including a lexicon and rules governing the use of the lexicon [25]. Standards may incorporate evidence-based best practices. Alternatively, when there is little inherent value of one practice over another (such as which side of the road to drive on), mutually observed conventions enable individuals to interact with one another in a predictable manner. However, standards require an investment to be established, disseminated, and enforced and tend to diminish innovation and individual adaptation [26]. Therefore, deciding on the degree of standardization is a constant balancing act. For frequently repeated simple interactions, communication standards generally are the most appropriate option. For infrequent or more sophisticated interactions, however, there may not be sufficient volume to recoup the investment in, or offset the risk of, standardization, making other tactics preferable [27]. Once well-conceived and refined standards are developed and observed, they decrease overall costs on the system, foster increased consistency and effort-free performance, and free individual operators to focus on higher-level cognitive functions and interactions.

A Regimen for Continuously Improving Communication in Radiology
Previous sectionNext section

On the basis of our experience in applying these theories and frameworks in radiology, we recommend a handful of strategies to overcome the communications deficiencies that a complex system can create.

Establish a Shared Vision

Whether or not the participants are aware of it, an implicit mutual obligation exists between skilled professionals and organizational leaders: individual professionals do their best to do excellent work and system leaders make it as easy as possible for those individuals to do excellent work. However, in a complex organization filled with many subsystems, it can be difficult to define and align the priorities of the organization. Thus, a key function of leaders is to foster the establishment of a shared vision, which guides decision-making at all levels of the organization [28]. Frontline workers and system leaders then work together to make that vision become reality. Furthermore, individuals in the organization should work to develop shared mental models, which are defined as “deeply ingrained assumptions, generalizations, or even pictures of images that influence how we understand the world and how we take action” [29]. In other words, individuals in the organization continuously seek to “get on the same page,” both in terms of high-level vision and day-to-day work. This enables the organization to standardize in a way that decreases unnecessary complexity without sacrificing specialized expertise.

Let People Do What People Do Best and Machines Do What Machines Do Best

As a rule, highly trained individuals are best at critical thinking, problem solving, and interacting with other individuals. Modern electronic systems process, calculate, synthesize, organize, search, transfer, and display data and information with extreme efficiency and reliability. Given advances in information technology over the past few decades, many functions that used to require human intervention now can be automated. Organizations should implement technology in a way that enables people to apply their expertise and skill to higher-level functions, such as drawing meaning from information, making appropriate judgments, and cultivating interpersonal relationships. Systems should automate repetitive tasks while supporting skilled individual performance and level-appropriate communication. The workflow should be made to be as lean as possible by automating routine information flows and by removing unnecessary elements of individual tasks. Any process in which an individual is required to routinely perform tedious tasks that can be automated should be regarded as a failure. In short, to the extent possible, systems should assist individual workers in performing their important work instead of forcing workers to constantly compensate for the idiosyncrasies and inadequacies of dysfunctional systems.

Start by Streamlining the Primary Workflow

In improving communications in radiology, a leader's first priority should be to improve communication systems that support the primary information workflow, such as the management and integration of work lists, PACS, voice recognition, and electronic medical record systems. Images and other key information should be instantaneously available to users when they are needed and in the appropriate format and order. Work lists should display examinations in a way that automatically incorporates patient acuity and examination urgency [30]. Continuous efforts should be made to eliminate every unnecessary click, keystroke, and bit of ambiguity and visual clutter. In this era, humans should almost never have to serve as the routine conduit from one electronic information system to another.

Implement Systems and Processes to Streamline Other Information Flows

Systems and processes should also support information flows beyond that of the primary workflow. For example, as part of confirming patient identity and correct examination and site, technologists can routinely ask patients for a brief history and compare it with the one provided. This information can then be passed along through the information stream to the speech recognition system, which can be made available for radiologist review and automatically integrated into the structured radiology report [31]. Information systems should also aid the correct performance of work—systems that may not immediately be recognized as communication systems per se. These include clinical decision support, automated incorporation of radiology tools, and easy access to departmental policies.

As existing communication systems are abandoned or modified and replaced with newer ones, it is critical that important communication channels that exist outside the primary workflow, as depicted in Figure 4, are preserved and enhanced. Specifically, system designers must recognize that existing informal communication methods may have been developed locally that may be disrupted by a formal communication system to the detriment of the overall goal of patient care. Although such communication channels may be informal, they can still play a critical role in patient care. For example, the implementation of a new PACS may facilitate rapid report turnaround time, which decreases the need for clinicians to visit the reading room, which in turn weakens interpersonal relationships between radiologists and clinicians, which in turn erodes an important informal communication media through which rich information formerly passed on a regular basis. Although the new communication media may be leaner, which may have a positive impact in routine cases (the majority of cases), it may compromise the ability for clarification and verification and have a significant negative impact in more complex or ambiguous cases (an important minority of cases), diminishing the clinicians' trust in diagnostic radiology and limiting radiologists' ability to fulfill their fundamental purpose: to help guide care. If new mechanisms are not established to replace those critical communication channels, the technologic advance that was designed to improve radiologists' ability to improve patient care may actually compromise patient care despite the best intentions.

Individuals Work With Internal and External Customers and Suppliers

Although no single individual can understand the system in its entirety, individuals and organizational subunits should at least work with and understand the needs of immediate predecessors and recipients of their work. Individuals should view the direct recipients of their work as customers, whether internal or external, and should take their needs into consideration when developing workflows and communication standards [32]. Customers also should make their needs clear to their suppliers and provide constructive feedback accordingly. Through the internal customers, organizational subunits become linked to the external customers. For example, by providing better service to referring clinicians, radiology departments better serve their patients. The organization should provide the means to facilitate this interaction and the tools to make improvements as appropriate. As individuals accommodate the needs of their internal and external customers, they create a system that can learn, adapt, and better overcome the sense of inability to change that often pervades large complex organizations [33].

Simplify and Organize the Information

Information should be as well organized and accessible as possible. For example, simple naming and numbering conventions that facilitate clear communication should be devised and adhered to whenever feasible. This tends to be done well in some aspects in radiology, such as in identifying patients (medical record numbers) and examinations (accession numbers). Many other naming conventions are currently less well developed or less widely adopted, such as common procedure names, protocol names, or sequence names. To achieve highly reliable system performance, standard naming conventions or name mapping will eventually need to be in place for every piece of equipment, every protocol, every structured report, and other set elements in a radiology practice [34]. Such basic organizational efforts serve as a backbone for effective management of a complex system, facilitating automation and freeing individuals from doing the tedious and distracting work of sifting through disorganized information.

Anticipate Future Situations on the Basis of Experience

Systems should be designed to anticipate at least the most common situations that may arise at each processing node. Defaults should be set to handle the most common situations while allowing sufficient flexibility and options to accommodate others. For example, by anticipating that clinicians often struggle to determine the exact examination name when ordering an imaging examination, synonyms can be provided that are linked to the primary order. Pick lists in structured radiology report fields can include common pathology in descending order of frequency. The need for references, tools, and communication channels can be anticipated and can be made readily available to operators at the moment they are needed. Information that is needed at the next step should automatically flow from one processing node to another; for example, organ measurements included in DICOM structured reports should automatically be passed into the radiologist's report [35].

Link It Together

The performance of the system is dramatically improved as reliable communication linkages are established. Communication begins to flow extremely quickly and reliably through the system. The correct course of action more frequently happens by default. When problems arise, they are addressed quickly and permanently; thus, the system becomes self-correcting. What can be standardized is standardized, and for what cannot be standardized, a commonly accepted approach to efficiently arrive at the best answer is established. Such systems seem to “just work” with relatively minimal input from individual users, much as we have come to expect from most products and services we encounter on a daily basis.

These linkages serve to minimize waste and error in the information value stream in radiology. For example, the radiology work flow begins when a clinical scenario prompts a clinician to ask a question. The ordering system guides the clinician in ordering the correct examination, decreasing the cognitive burden of determining the correct order to place and decreasing order errors. This is automatically linked to the correct protocol, which in turn is automatically linked to the correct settings in the scanner, which in turn generates images that are automatically sent to the PACS and automatically displayed with an optimized hanging protocol. Image meta-data are sent to a distributed work list that identifies the urgency of the study. The opening of a study automatically launches a standardized structured report and any pertinent electronic tools and references, with simultaneous display of patient history and other relevant clinical information. The completion of the report automatically triggers notification of the referring clinician, establishing a telephone conversation if necessary, that is automatically documented in the patient's record. Images and reports are automatically available to both referring clinicians and patients according to the institution's policy. When established correctly, the entire process can occur in as little as a few minutes and is repeated continuously, day and night. The efficiency and reliability of the system free radiologists to do more meaningful work, such as engage in richer interactions with referring clinicians, trainees, other radiologists, and patients.

Previous sectionNext section

The acquisition, interpretation, and communication of information form the raison d'être of diagnostic radiology. Thus, communication systems are a critical aspect of the provision of medical care, despite the fact that they are often underappreciated in many health care settings. Their design, implementation, maintenance, and improvement deserve the same level of commitment and skill as is applied to the provision of clinical care.

The concepts presented in this article are already in routine use in many high-performing industries, and many of them have been implemented in forward-thinking radiology departments. Radiologists should recognize that the realization of such high-functioning systems is not merely theoretic. In fact, not only is it possible, it is necessary to achieve the types of improvements in safety and effectiveness demanded by our patients and referring clinicians in our complex modern environments [36].

To achieve this type of performance, those who design, participate in, and pay for such systems must recognize that the value provided to referring clinicians and to patients depends on how accurately and reliably the information flows through the process to the point of care. Payers can greatly facilitate the design of such systems by structuring financial incentives accordingly. Radiology practices, hospitals, and other care providers must work together to establish communication standards. Vendors must accelerate their efforts in developing such systems, and those who purchase such systems must be willing to pay for the functionality that the vendors develop.

As diagnostic imaging continues to increase in complexity, there is a clear need for leaders to take responsibility for the entire radiology process, from beginning to end. We believe that the intelligent development and implementation of robust communication systems will dramatically enhance quality, safety, and efficiency in radiology in the coming decades.

Previous sectionNext section
1. Fitzgerald R. Error in radiology. Clin Radiol 2001; 56:938–946 [Google Scholar]
2. Gandhi TK. Fumbled handoffs: one dropped ball after another. Ann Intern Med 2005; 142:352–358 [Google Scholar]
3. Whang JS, Baker SR, Patel R, Luk L, Castro A 3rd. The causes of medical malpractice suits against radiologists in the United States. Radiology 2013; 266:548–554 [Google Scholar]
4. Plsek PE, Greenhalgh T. Complexity science: the challenge of complexity in health care. BMJ 2001; 323:625–628 [Google Scholar]
5. Plsek PE. Complexity and the adoption of innovation in health care. In: Proceedings of accelerating quality improvement in health care: strategies to speed the diffusion of evidence-based innovations. Washington, DC: National Institute for Health Care Management Foundation website. Published 2003. Accessed April 1, 2014 [Google Scholar]
6. Liebscher L, Sherry C, Breslau J, et al. The general radiologist in the 21st century. J Am Coll Radiol 2012; 9:554–559 [Google Scholar]
7. Rother M, Shook J, Womack J, Jones D. Learning to see: value stream mapping to add value and eliminate MUDA. Cambridge, MA: Lean Enterprise Institute, 2009 [Google Scholar]
8. Bhargavan M, Kaye AH, Forman HP, Sunshine JH. Workload of radiologists in United States in 2006–2007 and trends since 1991–1992. Radiology 2009; 252:458–467 [Google Scholar]
9. Shannon CE. A mathematical theory of communication. Bell Syst Tech J 1948; 27:379–423, 623–656 [Google Scholar]
11. Oregon Health and Science University. Pediatric normal measurements. Oregon Health and Science University website. Accessed April 1, 2014 [Google Scholar]
13. Bolstad WM. Logic, probability, and uncertainty. In: Bolstad WM. Introduction to Bayesian statistics, 2nd ed. Hoboken, NJ: Wiley, 2007:55–76 [Google Scholar]
14. Ritchie D. Bony pelvis. In: Davies AM, Sundaram M, James SJ, eds. Imaging of bone tumors and tumor-like lesions. Berlin, Germany: Springer, 2009:603–620 [Google Scholar]
15. Moore EE, Moore FA. American Association for the Surgery of Trauma Organ Injury Scaling: 50th anniversary review article of the Journal of Trauma. J Trauma 2010; 69:1600–1601 [Google Scholar]
16. Daft RL, Lengel RH. Organizational information requirements, media richness, and structural design. Manage Sci 1986; 32:554–571 [Google Scholar]
17. Lengel RH, Daft RL. The selection of communication media as an executive skill. Acad Manage Executive 1988; 2:225–232 [Google Scholar]
18. Dennis AR, Valacich JS. Rethinking media richness: towards a theory of media synchronicity. Proceedings of the 32nd Hawaii International Conference on System Sciences. Maui, HI: Hawaii International Conference on System Sciences. 1999 [Google Scholar]
19. Dennis AR, Fuller RM, Valacich JS. Media, tasks, and communication processes: a theory of media synchronicity. Manage Inf Syst Q 2008; 32:575–600 [Google Scholar]
20. Bastian H, Glasziou P, Chalmers I. Seventy-five trials and eleven systematic reviews a day: how will we ever keep up? PLoS Med 2010; 7:e1000326 [Google Scholar]
21. Deming WE. Introduction to a system. In: Deming WE. The new economics. Cambridge, MA: MIT Press, 1994:49–91 [Google Scholar]
22. Senge PM. A shift of mind. In: Senge PM. The fifth discipline. New York, NY: Doubleday, 2006:68–91 [Google Scholar]
23. Rubenstein-Montano B, Liebowitz J, Buchwalter J, McCaw D, Newman B, Rebeck K; Knowledge Management Methodology Team. A systems thinking framework for knowledge management. Decis Support Syst 2001; 31:5–16 [Google Scholar]
24. Liker J. Principle 6: standardized tasks are the foundation for continuous improvement and employee empowerment. In: Liker J. The Toyota way. New York, NY: McGraw-Hill, 2004:140–148 [Google Scholar]
25. Fromkin V, Rodman R, Hyams N. Morphology: the words of language. In: Fromkin V, Rodman R, Hyams N. An introduction to language. Boston, MA: Wadsworth, 2013:33–75 [Google Scholar]
26. Gilson LL, Mathieu JE, Shalley CE, Ruddy TM. Creativity and standardization: complementary or conflicting drivers of team effectiveness? Acad Manage J 2005; 48:521–531 [Google Scholar]
27. Bohmer RM. Medicine's service challenge: blending custom and standard care. Health Care Manage Rev 2005; 30:322–330 [Google Scholar]
28. Senge PM. Shared vision. In: Senge PM. The fifth discipline. New York, NY: Doubleday, 1999:215 [Google Scholar]
29. Senge PM. Give me a lever long enough…and single-handed I can move the world. In: Senge PM. The fifth discipline. New York, NY: Doubleday, 5 [Google Scholar]
30. Halsted MJ, Froehle CM. Design, implementation, and assessment of a radiology workflow management system. AJR 2008; 191:321–327 [Abstract] [Google Scholar]
31. Hawkins CM, Anton CG, Bankes WM, et al. Improving the availability of clinical history accompanying radiographic examinations in a large pediatric radiology department. AJR 2014; 202:790–796 [Abstract] [Google Scholar]
32. Lee HL, Billington C. Managing supply chain inventory: pitfalls and opportunities. MIT Sloan Manage Rev 1992; 33:65–73 [Google Scholar]
33. Ashforth B. The organizationally induced helplessness syndrome: a preliminary model. Can J Adm Sci 1990; 7:30–36 [Google Scholar]
34. Radiological Society of North America. RSNA RadLex Playbook. Radiological Society of North America website. Published November 1, 2011. Accessed April 1, 2014 [Google Scholar]
35. Noumeir R. Benefits of the DICOM structured report. J Digit Imaging 2006; 19:295–306 [Google Scholar]
36. Brennan TA, Gawande A, Thomas E, Studdert D. Accidental deaths, saved lives, and improved quality. N Engl J Med 2005; 353:1405–1409 [Google Scholar]
Address correspondence to D. B. Larson ().

Recommended Articles

Communication in Diagnostic Radiology: Meeting the Challenges of Complexity

Full Access, , ,
American Journal of Roentgenology. 2016;206:573-579. 10.2214/AJR.15.15117
Abstract | Full Text | PDF (723 KB) | PDF Plus (779 KB) 
Full Access,
American Journal of Roentgenology. 2017;209:992-1000. 10.2214/AJR.17.18381
Abstract | Full Text | PDF (926 KB) | PDF Plus (1000 KB) 
Full Access,
American Journal of Roentgenology. 2015;204:W126-W131. 10.2214/AJR.14.13116
Citation | Full Text | PDF (959 KB) | PDF Plus (847 KB) 
Full Access, , ,
American Journal of Roentgenology. 2013;201:611-617. 10.2214/AJR.12.10375
Abstract | Full Text | PDF (526 KB) | PDF Plus (635 KB) 
Full Access,
American Journal of Roentgenology. 2015;205:W470-W477. 10.2214/AJR.15.14807
Abstract | Full Text | PDF (847 KB) | PDF Plus (654 KB) 
Full Access, , , , ,
American Journal of Roentgenology. 2017;208:739-749. 10.2214/AJR.16.16963
Abstract | Full Text | PDF (1062 KB) | PDF Plus (1203 KB)