|Home | About | Journals | Submit | Contact Us | Français|
Understanding the role of relationships in health care organizations (HCOs) offers opportunities for shaping health care delivery. When quality is treated as a property arising from the relationships within HCOs, then different contributors of quality can be investigated and more effective strategies for improvement can be developed.
Data were drawn from four large National Institutes of Health (NIH)–funded studies, and an iterative analytic strategy and a grounded theory approach were used to understand the characteristics of relationships within primary care practices. This multimethod approach amassed rich and comparable data sets in all four studies, which were all aimed at primary care practice improvement. The broad range of data included direct observation of practices during work activities and of patient-clinician interactions, in-depth interviews with physicians and other key staff members, surveys, structured checklists of office environments, and chart reviews. Analyses focused on characteristics of relationships in practices that exhibited a range of success in achieving practice improvement. Complex adaptive systems theory informed these analyses.
Trust, mindfulness, heedfulness, respectful interaction, diversity, social/task relatedness, and rich/lean communication were identified as important in practice improvement. A model of practice relationships was developed to describe how these characteristics work together and interact with reflection, sensemaking, and learning to influence practice-level quality outcomes.
Although this model of practice relationships was developed from data collected in primary care practices, which differ from other HCOs in some important ways, the ideas that quality is emergent and that relationships influence quality of care are universally important for all HCOs and all medical specialties.
Researchers and practitioners in the health care community continue to explore strategies for improving health care quality. Studying health care organizations (HCOs) as complex adaptive systems (CASs) contributes to the development of new strategies for their improvement.1 Applying CAS theory to HCOs can help one see that efforts aimed at improving health care quality should consider the role of the relationships among organizational members.
Our research group has studied primary care practices for more than 15 years. We have focused on understanding change in primary care practices,2,3 on primary care practices as “jazz groups,”4 and on building relationships in primary care practices.5 In this article we present research findings demonstrating the role of relationships in primary care practice performance and discuss the role of relationships in improving health care quality. Drawing on four data sets, we developed a model of practice relationships that identifies seven characteristics of relationships. We also discuss how these seven characteristics interact with reflection, learning, and sensemaking (unraveling surprising events) to influence quality of care.
Noting that quality of care emerges from the relationships among members of an HCO, we use CAS theory to discuss quality as an emergent property of HCOs. Although much existing research on relationships in health care is dominated by studies of only patient-physician relationships,6,7 the research that we review considers practice wide relationships, including all clinical and nonclinical roles. Finally, we offer strategies for improving relationships among members of HCOs.
CAS theory is grounded in systems thinking, which emphasizes, in part, the role of interdependencies in system outcomes. We chose a CAS perspective,1 as opposed to a complex responsive processes perspective,8 because we believe a CAS (systems) perspective provides a more suitable structure for studying relationships in HCOs. CAS theory is a more comprehensive theory and has been more widely used for studying organizations. CASs are typically thought of as being made up of agents that are diverse and that interact in nonlinear ways. CASs display emergent properties, self-organize, and co-evolve with their environment. Table 1 (right) defines these key characteristics of CASs. We pay particular attention in this article to the property of emergence.
Emergent properties are system-level properties that arise over time from the local interactions among agents. Leadership9 and strategy10 are examples of organizational-level attributes that have been studied as emergent properties, and new understandings of these phenomena have been generated by studying them this way. Using a CAS perspective enables a helpful view of health care quality as an emergent property in HCOs.
Practice improvement efforts such as continuous quality improvement, which aim to improve organizations one component or one process at a time, are often less effective than expected. We believe that this is due to a misconceptualization of quality as something that can be achieved using strategies rooted in reductionism (a perspective that quality is improved by focusing on the parts/components of a system). In contrast, a CAS perspective enables a view of health care quality as an emergent property. Emergent properties cannot be explained by separately analyzing parts of a system.11 Thus, with a CAS perspective it becomes clear that one cannot understand practice-level quality by understanding the quality of individual parts of a practice. We suggest that to improve health care quality, health care professionals must examine quality in holistic ways. Viewing quality as an emergent property provides health care professionals with an alternative way to make sense of successes and failures. For example, rather than trying to locate the individual responsible for a missed diagnosis, a practice can use the mistake as a way to think about the problem in terms of how the practice missed the diagnosis. This view also provides health care professionals with an alternative frame for designing and implementing quality improvement efforts (Table 2, page 459).
Individual components/processes are important for improving quality. Yet when one understands quality as an emergent property of HCOs, the relationships among its members become key levers for performance improvement. Efforts aimed at improving health care quality would focus on improving the relationships among the members of an HCO, rather than solely on improving individual components or individual processes of these systems. When quality is treated as a property arising from the relationships within the HCO, then different contributors of quality can be investigated and more effective strategies for improvement can be developed.
We drew data from four large National Institues of Health (NIH)-funded studies and used an iterative analytic strategy and a grounded theory approach to understanding the characteristics of relationships within primary care practices. These four studies were aimed at improving primary care practices. Methods consisted of direct observation of practices during work activities, direct observation of patient-clinician interaction, individual in-depth interviews with each clinician and other key staff members, surveys of patients and practice staff, structured checklists of the office environment, graphical representation of patient pathways during office visits, and chart reviews for clinical endpoints. This multimethod approach to observing practices amassed rich and comparable data sets in all four projects. The four studies are summarized in Table 3 (page 460), and a detailed overview can be found in Appendix 1 (available in online article).
We worked from case summaries prepared for each practice from each study. We performed a secondary analysis of project data, working iteratively from these sets of data. We developed our theory from these observations. We then looked to another set of data, the Using Learning Teams for Reflective Adaptation (ULTRA) study, to test and refine the emerging theory. After we identified a core set of relationship characteristics, we tested them in the ongoing ULTRA study.
Throughout this process we used several strategies to increase the rigor and quality of analysis. Analysis involved people with diverse roles (practice change facilitators, lead researchers, statisticians, nurses, administrators, educators, doctoral students) and from multiple research sites (Case Western Reserve University, Robert Wood Johnson Medical School, Lehigh Valley Hospital, the University of Texas at Austin, and University of Colorado). Analysis occurred in two interdependent phases—identification of characteristics and model building—both taking place during approximately a two-year period (January 2004–December 2005).
We met often to identify key relationship characteristics—that is, those that distinguish high- from low-performing practices in terms of patient outcomes—from the data and to develop a model of practice relationships. Our resulting model named mindfulness, communication, tight and loose coupling, respectful interaction, and stable patterns of interacting. We later included trust on the basis of the experience of practice facilitators. Similarly, notions of tight and loose coupling evolved during multiple discussions to social and task relatedness. We also noted the need to include heedfulness as a distinct characteristic, capturing different behaviors than those captured by mindfulness. As we worked on identifying key characteristics of practice relationships, we examined relevant literature and used it to guide both our inquiry and the refinement of the set of characteristics. After multiple discussions and returning to the data, we named the “Magnificent Seven” as follows: (1) trust, (2) mindfulness, (3) heedfulness, (4) respectful interaction, (5) diversity, (6) social and task relatedness, and (7) rich and lean communication.
We then connected these seven characteristics to the activities of reflection, sensemaking, and learning. We engaged facilitators on the ULTRA project in verifying our model by checking it against emerging data—applying the model of relationships to primary data from the ULTRA study in real time. Facilitators returned to the field to look for these characteristics, contradictory examples, and alternative characteristics. It was this ongoing, iterative process that enabled us to continually see new things in the data and to refine our model accordingly. Figure 1 (page 461) illustrates the model-development process.
We then held three additional face-to-face discussions with practice change facilitators from three research institutions. These discussions helped us assess the extent to which the characteristics identified accurately represented key characteristics of relationships in other primary care practices. A detailed time line of the model-building process can be found in Appendix 2 (available in online article).
In a majority of the cases, all seven relationship characteristics were present in practices that met the outcome goals of our studies, that is, practices that were high performers. We believe these seven characteristics to be important in both practice improvement and in the achievement of high-quality health care. We now describe each characteristic in detail. Table 4 (page 462) provides definitions of the relationship characteristic and examples of how each appears in the practices.
Trust is exhibited when one individual is willing to be vulnerable to another individual. Trust is particularly important in health care because the relationships among members of health care teams are highly collaborative and interdisciplinary. Trust can be difficult to foster; the culture of health care delivery often works against the development of trusting relationships.12 Policies and procedures in HCOs may lead to distrust. Risk of litigation and clinical documentation requirements can also erode trust. A study of trust in the context of telemedicine showed that physicians must trust each other before physicians will use telemedicine in caring for patients.13 We believe that practices with high levels of trust will be able to have difficult conversations and will be able to openly discuss and learn from successes, failures, and near failures.
Mindfulness is a social characteristic exemplified by the openness to new ideas and multiple perspectives,14 a fully engaged presence,15 a rich awareness of discriminatory detail,16 and the seeking of novelty, particularly in seemingly routine situations. Mindfulness is a purposeful cultivation of awareness. People in practices must be aware to be open to novelty. Mindfulness has been shown to be critical in the effective practice of health care.17–19 Mindful approaches are characterized by a continuous creation of new categories, openness to new information, and implicit awareness of more than one perspective.20 Mindfulness—which must be practiced because it is not innate—occurs when people question their assumptions about the nature of the world.
Heedfulness occurs when an individual pays attention to his or her specific task at hand21 as well as to the task of the larger group. In heedful practices, people watch to see how their actions influence the actions of the group, and they seek awareness about how their actions are intertwined with the actions of other members of the practice. Heedfulness is difficult to achieve because of the many competing demands placed on health care professionals. Fostering heedfulness, however, might be an effective strategy for reducing medical errors because “when heed is spread across more activities and more connections, there should be more understanding and fewer errors.”21(p. 366)
Respectful interaction is characterized by honesty, self-confidence, and appreciation of others. In relationships characterized by respectful interaction, new meanings often emerge through interaction.21 For example, in a staff meeting where practice members are interacting respectfully, it is likely that the solution to a particular problem will be created by the group, as opposed to an individual. Medical errors are an unfortunate part of the health care delivery process, but respectful interaction can enable learning from mistakes. Practices can learn from mistakes when people actively seek out and value the opinions of others (appreciation of others), freely share opinions even when these opinions may be unpopular (honesty), and willingly change their minds in response to new meaning created within the practice (self-confidence).
Primary care practices are made up of diverse people. Here we focus on cognitive diversity. Cognitive diversity is the differences in perspectives and world views of individuals (how people think). Moderate levels of diversity can help organizations operate effectively in competitive environments, process information, and learn in real time.22 Too little diversity can block creativity and innovation, and too much diversity can block communication. Diversity in a primary care practice can increase people’s capacity for making sense of the world and broaden the range of available solutions for problems.
Both social and task relatedness are important in practice relationships. Social relationships are personal in nature and are often based on friendships or family relationships that extend outside of work. Task relationships are focused on work issues. Members of a practice characterized by high task relatedness rarely discuss non-work-related topics with one another. The data from the four studies indicated that practices with relationships that were too socially oriented (conversations were dominated by personal topics) and practices with relationships that were too task oriented (conversations were dominated by work topics) tended to perform more poorly than practices with a mixture of social and task relatedness. Our findings suggest that social and task relatedness is not an “either/or” attribute. We suggest that both social and task relatedness are needed for practices to deliver high-quality health care.
We noted the following commonly used communication channels (in the order of richest to leanest) in primary care practices: (1) face-to-face, (2) telephone, (3) personal documents (for example, letters, e-mails, reminders), (4) impersonal documents (mass e-mails and impersonal memos), and (5) numeric documents (appointment schedules and budgets). When ambiguity is high, practices should use face-to-face communication channels, which allow for rapid information flow and for the clarification of meaning in real time (one-on-one conversations and small-group meetings). Less ambiguous messages can be communicated using a leaner channel (memo or e-mail). The medical record—electronic or paper—is often a major communication channel in primary care practices, and its richness/leanness varies depending on the user and the specific context in which it is being used.
How do relationships support the emergence of health care quality? We believe that for relationships to contribute to an emergence of high-quality care, practices must participate in effective reflection and learning and sensemaking. The model shown in Figure 2 (above) integrates our current understanding of how relationships support the emergence of health care quality.
Reflection is a dynamic, conscious process that occurs when individuals attempt to make sense of and/or learn from challenging situations.23 Both reflection-in-action and reflection-on-action23,24 are important in mediating the relationship between practice relationships and practice outcomes. A model previously published by our group3 demonstrates that reflection on key actions by practice members can both improve practice outcomes and change the nature of relationships among practice members. The model we present here builds on that previous research.
Sensemaking and learning are particularly useful strategies for dealing with the kinds of ambiguity that often arise in HCOs. For a practice to grow, change, or improve it must be able to make sense of and learn from its environment. Sensemaking is a social act of retrospectively unraveling a surprising flow of events.25 Qualitative differences can exist in the sense that is made from an event, and not all sensemaking is beneficial to organizations. Because sensemaking is a social activity, we believe that practice relationships are critical to the quality of the sense that is made from unexpected events.
Learning is also a social act. One way to improve health care quality is to encourage a culture of learning—learning from mistakes, learning by doing, and learning by experiencing history richly.26 Our model relies on the logic that effective learning can improve the quality of care delivered by a practice.
Improving the relationships among practice members is one way to improve sensemaking and learning. People in a practice who trust one another will be more likely to admit when they are unsure about how to solve a particular problem, increasing the likelihood that learning will occur. Clinicians relating heed-fully with others will be more likely to recognize when a front-office staff member needs more information about the patient to do his or her job well. A diverse practice is more likely to have a broader set of perspectives with which to observe and make sense of important practice issues. Effective sensemaking and learning can improve a practice’s capacity to make decisions and take actions that lead to better health care quality.
Although this model of practice relationships was developed from data collected in primary care practices, which differ from other HCOs in some important ways, the ideas that quality is emergent and that relationships influence quality of care are universally important for all HCOs and all medical specialties. All HCOs are made up of people who learn and make sense of their experiences. The importance of relationships has been recognized by the Accreditation Council for Graduate Medical Education, whose general competencies place significant emphasis on communication and interpersonal skills, systems-based practice, and professionalism.27 Similarly, the recent review of the United States Licensing Examination has brought forth recommendations to change the examination to include measurement of a broader array of competencies that will include domains related to our findings.28
Table 5 (page 465) provides vignettes of two primary care practices, one practice with good relationships and one practice with relationships that need improvement. Tables 4 and and55 can be used to help practices assess their relationships and to guide strategies for improving practice relationships.
Periodic assessment of practice relationships can be used to indicate progress in achieving good relationships.5 Practices need to evaluate trust as a precondition for using relationships to improve health care quality. Lack of trust may be most apparent when staff are hesitant to speak up and offer perspectives. Willingness to speak up and become vulnerable, particularly in discussing an error in care, might not happen overnight. Rather, it will likely require steady attention of practice leadership to create and nurture a practice culture that values candor and new ideas.
To enhance diversity, practices should avoid the tendency to hire people just because they “fit in” and should take advantage of the range of experience that may be available. A balance of social and task relatedness can be apparent in conversations that take place, and a healthy work environment will find staff talking about both job-related tasks and their social lives outside the office. Similarly, practice leadership should ensure that modes of communication are appropriate to the message—a posted memo may be most appropriate for communicating routine administrative detail but would be inappropriate for communication around plans to downsize.
Attending to mindfulness, heedfulness, and respectful interaction present more of a challenge. Encouraging interactions among people in practices who don’t normally interact can help foster relationships that are mindful, heedful, and respectful. Physician leaders, in particular, need to make an effort to seek input from practice staff with whom they don’t routinely interact.
Finally, practice leaders must understand that learning and sensemaking are influenced by patterns of relating that occur in their practice. Practice management needs to provide time and space for reflection. It is difficult in the hectic daily routine of health care to avoid the pitfall of believing that one is too busy to take time to understand what is going on around one and how one’s efforts to improve are playing out in real time. If we believe that health care practices are CASs and that quality of care is emergent, then strategies that focus on relationships, and the time to use them, become important in efforts to improve health care.
We need to reduce our tendency to train health care professionals in isolation from one another and consider ways of integrating training programs so that health care professionals can come to a better understanding of their interdependence. The analysis presented in this article suggests a need for training efforts designed to enhance the ability of health care professionals to work together to achieve the goals of the practice. Training used to help businesses improve interpersonal communications may also be applicable in efforts to improve relationships in HCOs.
Evidence confirming the model of practice relationships is limited. Research is needed to test the interdependencies among the constructs in the model. For example, is trust required for progress in the other characteristics? Previous research in ICUs, operating rooms, and nursing homes has shown the importance of relationships for improved performance. Additional research, however, in multiple health care settings that examines specific characteristics of relationships is needed. Research is also needed that helps us understand the ways in which promotion of these seven relationship characteristics can improve health care quality. More broadly, HCOs’ participation in such research will depend on their willingness to examine the behavioral aspects of health care delivery.
The authors gratefully thank the agencies that funded their investigations of primary care offices: the National Cancer Institute (R01 CA60862, 2R01 CA60862, and 3R01 CA60862), the Agency for Healthcare Research and Quality (R01 HS08776), and the National Heart, Lung, and Blood Institute (R01 HL70800).
The authors appreciate support provided by the Herb Kelleher Center for Entrepreneurship Growth and Renewal and IC2 Institute, the University of Texas at Austin. Support was also provided by a Research Center Grant from the American Academy of Family Physicians. Dr. Stange’s time is supported in part by a Clinical Research Professorship from the American Cancer Society. The authors express special thanks to the practices that participated in these studies and the multiple research facilitators who generated the data for the practice summaries used in this analysis. The authors also express special thanks to Diane Rhodes for editorial assistance. This work was presented in part at the North American Primary Care Research Group Conference, Tucson, October 17, 2006.
Holly J. Lanham, Department of Information, Risk and Operations Management, McCombs School of Business, University of Texas, Austin, Texas.
Reuben R. McDaniel, Jr., Health Care Management.
Benjamin F. Crabtree, Department of Family Medicine, Robert Wood Johnson Medical School, New Brunswick, New Jersey.
William L. Miller, Department of Family Medicine, Lehigh Valley Hospital and Health Network, Allentown, Pennsylvania.
Kurt C. Stange, Departments of Family Medicine, Epidemiology, and Biostatics and Sociology, Case Western Reserve University, Cleveland.
Alfred F. Tallia, Department of Family Medicine, Robert Wood Johnson Medical School.
Paul A. Nutting, Department of Family Medicine, University of Colorado Health Sciences Center, Denver.