In an effort to ensure quality, nursing homes are among the most highly regulated and rule-driven health care organizations. With such attention to quality, why do industry assessments1 continue to report poor quality? Traditional models for organizing care are derived from longstanding Newtonian beliefs about the world and have led to a reliance on regulation and rules as a primary means of control.2 Complexity science, however, offers a new way to understand organizational dynamics3 and thus provides insights about why nursing homes may struggle to provide high quality care and suggests new targets for improvement. In particular, complexity science reveals that the capacity of a human system to change and improve is shaped by the quality of relationships among its members.4 Consider the following scenario. A certified nurse assistant (Katie CNA) discusses her assignment with the nurse. Katie CNA learns that she should first get Mrs. Jones up because this resident must be upright to eat due to choking risk; second bathe and dress Mrs. Smith who is incontinent and at risk for decubitus; and last care for Mrs. Lee because she is not incontinent and doesn’t mind sleeping later. Before beginning her assignment, however, Katie CNA hears from Mary CNA that Mrs. Lee’s daughter is on her way into the building. The nurses and CNAs all know that this daughter complains and is demanding when she finds her mother in bed past 7:30 am; these complaints have been discussed in recent unit meetings. Using this new information, Katie CNA adjusts her plans and takes care of Mrs. Lee first instead of following the nurse’s instructions. Remembering what she learned from the nurse, however, Katie CNA asks Mary CNA to get Mrs. Jones and Mrs. Smith out of bed right away. Mary CNA agrees but only after she secures Katie CNA’s promise to help with her heavy care residents later in the day. Meanwhile the nurse notices the daughter’s arrival, observes the adjustments made by Katie CNA and delays the daughter at the nurse’s station to allow Katie CNA time to prepare Mrs. Lee. This scenario describes a process of self-organization.5 Each individual began with a set of plans for her work. However, each quickly adjusted her actions based on new information that came through interaction with each other and observation of events. The outcome was good and appropriate care for those involved. Self-organization is a property of all social systems—referred to in complexity science as complex adaptive systems.3 Self-organization occurs when people are free to interact, exchange information, and adjust behavior to meet the immediate demands of the environment.5 In this scenario, three critical system parameters6 for effective self-organization were apparent in this small complex adaptive system. These system parameters, which are imbedded in the relationships among the members of the system, are: Appropriate information flow. Katie CNA received a clear assignment from the nurse and was open to additional new information from Mary CNA that made her aware of the need to adjust her plans. In addition, a unit meeting provided information to all staff members and allowed for discussion of the daughter’s history of complaints. Good connections among its members. Mary CNA had open communication with Katie CNA such that Katie trusted the information received. A unit meeting about Mrs. Lee’s daughter and her complaints is more evidence of good connection. In addition, the nurse was connected to the care setting such that she observed emerging events and thus was able to adjust her own behavior in response. Finally, Mary CNA trusted that Katie CNA would reciprocate and assist Mary with the heavy care residents. This trust was another sign of good connections among staff. There was sufficient cognitive diversity. Cognitive diversity refers the variety of people in a system in terms of characteristics such as roles or positions, education levels and background, or cultural orientations. Cognitive diversity means that the system will have more “new” information available to it. For example, people holding similar education, roles, and positions, are not as likely to learn something new from each other as would people who hold different education, roles and positions. In this scenario, the nurse and CNA, being cognitively diverse, interacted in a way that ensured a better understanding of events. Specifically, the nurse, using clinical knowledge, discussed information with Katie CNA in such a way that Katie understood that the care instructions were not meaningless rules to be followed. Because she learned the clinical basis for the care assignment, Katie CNA did not delay the care of Mrs. Jones and Smith when other demands arose; instead, she asked Mary CNA to substitute for her. Even this simplistic scenario suggests that a fairly complex set of behaviors were required for effective care. One would need a fairly long list of rules to specify the behaviors necessary to achieve this outcome, which occurred spontaneously and relatively effortlessly through self-organization. Such self-organization was possible only because of the nature of the relationships among the people on this unit. Currently, the authors are conducting a series of case studies in nursing homes in part to examine relationship patterns necessary for self-organization. To date, we have completed three, six-month case studies and found that in general, the relationship patterns in these nursing homes do not create a level of interaction needed to support truly effective self-organization. Some dominant patterns appeared across these three cases that are likely to exist in many other nursing homes as well. These dominant patterns, described below, suggest a restriction in information flow, poor connections among staff, and lack of interaction among people with cognitive diversity. Thus, the system parameters described above were restricted rather than enhanced. In these nursing homes, RNs were in supervisory or administrative roles and interacted with other nursing staff mainly over staffing problems and staff conflict issues. RNs engaged in interdisciplinary care planning but had minimal interaction with front line, LPN and CNA, staff. Sometimes RNs asked LPNs and CNAs for information, but little information was returned to them. LPNs were tied to medication and treatment carts, covering up to 30 or more residents. Thus, they interacted little with anyone else in their rush to medicate and treat residents on time. CNAs mainly interacted with other CNAs. If teamwork existed, it was in small cliques with conflict and lack of cooperation between cliques. CNAs were rewarded for completing individual assignments not for teamwork. There was heavy reliance on rules and rule enforcement. This meant that managers frequently interacted with staff to correct them on what they were doing wrong. In one example, a manager was observed correcting a CNA for placing a shampoo bottle in a laundry basket but the manager never knew, and thus did not acknowledge, that the CNA used her own money to buy the shampoo for a resident who had none. Not surprisingly, staff members in these nursing homes felt unappreciated. In sum, little attention was paid to the nature and quality of staff relationships. Something better is possible however! By recognizing that self-organization is a naturally occurring characteristic of all work units, managers have some new tools for facilitating a higher quality of care. Key to this capacity is fostering relationship patterns that appropriately alter the system parameters of information flow, connection, and cognitive diversity. Prior research studies have demonstrated a link between better resident outcomes, lower staff turnover and specific management practices that facilitate the system parameters. Some evidence-based management practices are: Reduce reliance on rules and on rule enforcement.4 Facilitate open communication where staff say what they believe without fear.4 In combination with communication openness, create a climate of clear expectations, appropriate rewards and attention to staff concerns.7 Facilitate participation in decision making (PDM) by nurses.4, 8 Facilitate CNA PDM along with high interaction between RNs and CNAs.9 Facilitate leadership behaviors that are relationship-oriented.4 Encourage staff not to wait for managers but to take initiative to raise issues and take advantage of chance encounters with managers to problem solve.10 Use every means for connecting staff, including formal and informal meetings, planned and unplanned encounters in the hallways.10 Foster frequent conversations among RNs, LPNs, and CNAs.11 Understanding that relationships are truly the basis for high quality care, we need to quickly learn to do things differently in long-term care. Over-reliance on rules is not working. This article suggests some effective management practices that bring people together, improve information flow, and increase cognitive diversity. Importantly good teamwork must be accompanied by good information flow between teams and levels in the organization. Let’s talk more not less.