INTRODUCTION

The imbalance between primary care demand and capacity in the U.S. creates barriers to patient access and contributes to physician burnout. Many believe that enhanced inter-professional teamwork can promote more efficient primary care delivery1 4 Inter-professional teamwork has been defined as “the provision of comprehensive health services to individuals, families, and/or their communities by at least two health professionals along with patients, family caregivers and community services who work collaboratively on shared goals within and across settings to achieve care that is safe, effective, person-centered, timely, efficient and equitable.” 1 , 5 The primary care team typically includes a lead clinician, such as a physician or nurse practitioner, and other key personnel including nurses, medical assistants (MAs), care managers, practice managers, clerical staff, as well as others, when available (e.g., behavioral health and pharmacists). The utilization of teamwork enables organized care delivery that makes efficient use of patients’ and providers’ time and frees more highly trained clinicians to focus on complex aspects of patient care.6 8

Despite extensive talk about teamwork within patient-centered medical home initiatives and among policymakers, there is little practical information available to primary care providers on how to function as teams. There is a well-developed body of literature on teamwork related to industrial, military, and more recently, to hospital settings,9 but literature with regard to teamwork in primary care facilities, which differs greatly from that in a hospital setting, is in the early stages. The IOM and others have noted that there is “very little data” regarding effective teamwork in primary care, citing a need for examination of the characteristics of highly effective teams and how they are implemented.1 , 5 , 10 , 11 Researchers2 4 , 12 15 have begun to examine these topics, but much remains to be explored in terms of how primary care practices develop functional teams.

To help fill the information gap, this study examines how selected primary care practices have developed teams and how they have overcome common challenges to team-based primary care. It also explores implications for care delivery and policy.

METHODS

Identification of Participants

To identify practices using an empiric measure of high-functioning teams, we started with a list of practices recognized as PCMHs as of February 2013 via the NCQA 2011 PCMH tool,16 which included a new element on practice teams (included in the appendix) focused on eight team characteristics derived from the literature on teamwork.2 , 12 , 13 , 17 , 18

To obtain a range of practice types for our study, NCQA drew a random sample of PCMHs stratified by geographic region and by variables covering both practice size (number of physicians/independent clinicians) and ownership (e.g., physician-owned vs. hospital- or health system-owned vs. community health center (CHC), etc.). The NCQA divided practices within geographic regions into two groups: (1) those achieving 100 % of the points on the practice team element (high team score) and (2) those obtaining 0–25 % of the points on the practice teams element (lowest team score). Our goal was primarily to interview practices with high scores, but we included one low-scoring practice per region for qualitative comparison. We were not aware of the practice team element scores at the time of the interviews.

For confidentiality reasons, NCQA made the initial outreach to the practices, at which time they described the study goals and assured the practices that participation would have no bearing on their PCMH recognition and that participants would not be identified. NCQA sent a list of practices that consented to an interview to our research group, and our research assistant conducted a screening phone call to confirm practice characteristics and to identify a lead clinician who could best speak to the topic of primary care teamwork at the practice. Then, during the lead clinician interview, a second (and if possible, a third) informant involved in the operational aspects of teamwork at the practice was identified.

Given the project objectives and budget, our goal was to interview between 25 and 30 practices. After completing 63 interviews at 27 practices from across the strata, we were no longer hearing substantively new information (themes were consistently being repeated), and we concluded our outreach.

Interview Protocol

The interview protocol was partially based on a literature review, and included questions summarized in Table 1. At the beginning of each interview, we defined the term “primary care team” for respondents as “two or more people working together to provide primary care for patients.”

Table 1 Summary of Key Protocol Questions for In-depth Interviews of Practice Respondents

In-Depth Interviews

Interviews occurred between May and December 2013. Every practice completed at least two separate interviews (a lead physician and a second practice team member), and several completed three interviews. Interviews lasted 45–60 minutes and were conducted via telephone by a senior researcher and trained research assistant. The research assistant typed verbatim notes during each interview, which were reviewed by the senior interviewers.

Analysis

To avoid prematurely imposing a prescribed conceptual framework for teamwork onto the interview data, we conducted a preliminary reading and first-stage analysis that evaluated informant responses using a combination of descriptive and emergent coding.19 Three research team members reviewed every interview in its entirety and created separate lists of potential themes and codes from the transcripts. The team revisited the code lists in several meetings and came to agreement on the definitions for each through an iterative process.

To determine the conceptual framework on teamwork that best fit the themes and codes we had identified in this first phase, we used the list of available frameworks from our prior literature review1 , 2 , 9 , 20 and settled on the Ghorob and Bodenheimer framework,2 , 12 which identifies five key elements of team building: 1) Defined Goals, such as specific measurable operational objectives; 2) Systems, including the physical arrangement of team members and clinical systems (e.g., standing orders for common situations); 3) Division of Labor, including clear task definition and role assignment; 4) Training of team members on their functions; and 5) Communication, including communication structures (e.g., instant messaging and informal face-to-face gatherings of the clinical team known as “huddles”21 , 22) and processes (e.g., feedback, conflict resolution).

After matching codes/themes to each of the five elements (in some cases codes/themes supported more than one element), the second stage of our analysis involved two team members collaboratively applying the codes to the data in Atlas.ti.23

Participants

The 63 in-depth interviews included 60 primary care practice participants in 27 practices from 17 states as well as three national experts on teamwork who were also experienced clinicians. Practices comprised a range of sizes, ownership types, and EHR use. The practice respondents included 22 physicians, 3 nurse practitioners/physician assistants, 7 RNs/LPNs, seven MAs, 12 administrative/front desk staff, and 9 practice managers. Practice and respondent characteristics are provided in Table 2.

Table 2 Respondent and Practice Characteristics

RESULTS

Primary care teams had certain common characteristics. Several practices with clearly defined roles and practice support used a core clinical team ratio of one lead clinician to two medical assistants to three exam rooms. Larger practices divided themselves into smaller functional team units. If the practice had nurses, these ratios varied. In most practices, MAs were paired with the same clinician on a day-to-day basis.

In addition to a core clinical team, some respondents had full-time or part-time care managers whose roles included ongoing between-visit communication and coordination of care for patients with complex health needs.24 A few practices, particularly community health centers and those in larger systems, included other team members (e.g., behavioral health workers, pharmacists) who helped with chronic care management, medication reconciliation, education, and counseling. Almost every respondent described the patient as a team member, noting that efforts to function more effectively were focused on patients. Patient roles were particularly focused on carrying out care plans, self-management support, and personal decision-making.

Table 3 summarizes participants’ descriptions of roles and responsibilities and how they changed as teams became more structured. In small to medium-sized practices, where RNs may be too costly, some MAs are increasingly gathering information from patients with the assistance of physician-developed EHR templates and performing routine clinical tasks such as vaccinations and specimen collection (Table 4). Physicians in the four practices that relied most on templates noted that they had to spend time creating templates and protocols to ensure that MA roles were safe for patients and consistent with evidence-based care.

Table 3 Primary Care Team Member Roles and Responsibilities Have Changed with Increased Teamwork
Table 4 Optimizing the Role of Medical Assistants on the Primary Care Team

Methods for overcoming common challenges in moving to team-based primary care are presented in Table 5. The following illustrative examples are organized within the five primary care teamwork elements.2 , 12

Table 5 How Practices Overcame Challenges to Teamwork

Goals

Physician and staff buy-in for the alignment of practice members around common goals in patient care and workflow, as well as staff involvement in developing work flow processes, was aided by data demonstrating the results of team-based care. A Colorado family physician used quality reports generated from the practice’s EHR data by an independent practice association (IPA): “Doctors love data, but only when they own it… and when the tools you are using for clinical care as part of a team have improved because of the providers’ input, you have them, hook, line and sinker.”

Getting provider and patient buy-in for creating and using patient care plans (individualized plans of care that include treatment goals and how the team will work with the patient toward those goals25) was one of the most significant goal-setting challenges. A physician noted, “It’s been hard to get people on the team to understand the purpose behind the care plan,” and that a “lack of staff infrastructure” posed implementation challenges. Similarly, participants stated that patients often did not understand the purpose of the care plan and did not comply with it.

Scores on NCQA team element items indicate that low-scoring practices did less well with training about self-management support (which typically includes care plans). None of the low-scoring practices raised the topic of care plans during their interviews, while respondents from high-scoring practices were more likely to mention care plan challenges (suggesting a greater awareness of the care plan concept). The one solution noted for care plan implementation challenges involved nurse care managers engaging patients and their physicians in designing, revisiting as needed, and executing the plan.

Division of Labor

To overcome the challenge of physician reluctance to delegate, the most common strategy was to introduce task delegation incrementally, starting with items that physicians could agree were safe to hand off to MAs (e.g., standing orders for urinalysis for UTI symptoms and pneumococcal vaccination). Next, some physicians developed templates within their EHR to guide staff when gathering information from patients with common complaints. Third, having outside practice coaches or facilitators teach teamwork and verify nurse competency reassured physicians about delegating tasks. As MA/RN interaction with patients increased, some physicians gained comfort by talking with patients about exactly what messages they had received.

To address the commonly noted challenges of role definition, many practices used part of their all-staff meetings to allow staff members to describe their daily tasks so that others could understand the respective roles and identify where tasks could be streamlined. Team training and practice coaches were helpful in this area.

Offloading of routine tasks to MAs and LPNs resulted in increased job satisfaction for physicians in several practices, who could instead focus on patients’ more complex and personal needs. MAs reported improved job satisfaction from feeling more involved in patient care and becoming more than “just the vital signs person.”

Communication

Maintaining real-time and structured communication was a common challenge, particularly for practices with low scores on the NCQA practice team element. Of the 27 practices in this study, 23 mentioned huddles as key to maintaining structured communication within their teams. A physician leader said that “regular huddles are one of the first things all practices should do.” Learning to huddle typically entailed heavy involvement of a leader (physician or PCMH champion) and observation of huddling by other teams. In some practices, getting huddles “down to a science” took over a year.

Huddles generally were used to clarify the “game plan” of what the team would be doing for patients scheduled to come to the office that day. They typically took place at the beginning of morning or afternoon office hours and lasted 5 to 10 minutes (but allowing teams flexibility on when they huddled was helpful), and usually included the physician and MAs or the physician and MA and nurse. In some cases, a front-desk team member was part of the huddle, but usually the MA or nurse relayed information relevant to front-desk staff after the huddle. They often began by taking stock of which staff members were present in the office, followed by a quick discussion of each scheduled patient, the special issues or tasks for particular patients, and, when such tasks were not already part of predetermined team member roles, identification of who would carry them out.

Some practices used a paper or electronic “huddle sheet”, created with a customized EMR template and populated with data from the EHR and from the pre-visit planning work that the MA or nurse had completed (typically) in the two weeks prior to the scheduled visit. In practices where part-time or shared staff (e.g., nutritionists, nurse care managers, and behavioral health specialists) were not always on-site, a standardized process for notifying them of relevant information from the huddle would be accomplished via the EHR.

Several other techniques also helped to minimize communication breakdowns. Chief among these was co-location of the MD/RN/MA in order to enable in-person communication. Some respondents mentioned creating a “safe” culture9 in which people could feel comfortable providing feedback without threat of reprisal, and in some instances this was facilitated by creating a “go-to” person for staff to express concerns or ask questions. Use of the EHR, including instant messaging features and task assignment into different “buckets” or “task lists” for team members, also facilitated communication.

Feedback from many respondents, as well as a wide body of literature, suggests that continuity of care between clinicians and patients is a prerequisite to trust and effective communication.26 Physician participants generally felt that teamwork enhanced interpersonal continuity with patients, as the necessary data to inform a patient encounter had been collected before the visit, freeing them to spend more of the visit discussing issues most important to the patient. To maximize patients’ familiarity with team members, most practices restricted the core clinical team to two or three people who had ongoing contact with the patient, and some handed patients a card with the names, titles and roles, and contact information of the key team members.

Systems

Systems-based challenges, both within and external to the practice, were the themes most frequently mentioned by respondents, and typically required system-based solutions. For example, a common problem (particularly among practices with low practice team element scores) was the tendency to “slide back toward pre-teamwork behaviors.” To address this, many high-scoring practices used a combination of daily huddles and checks built into their electronic system (tracking of task completion in the EHR) to ensure that problems were quickly corrected.

One example of the way that systems can be used to support teamwork involved diabetes management and the volume of related performance reporting in the face of “only 15-minute visits.” In response, a few practices instituted a diabetes clinic day for each provider. As a nurse in such a practice noted, “If a patient’s A1c is greater than 9, they are given a 30-minute appointment.” In these instances, the health educator and care coordinator are present, and the MA who works at the diabetes clinic also has more extensive training in diabetes. The nurse noted, “The providers and patients are happy, and we have some examples of patients lowering their A1c.”

External system challenges to teamwork included insufficient staff and current fee-for-service incentives that emphasize office visits over population management and care coordination. To partially address these challenges, a Maryland practice negotiated with the larger health system to which it belonged to switch to a compensation model in which physicians were made responsible for their own MA salaries, enabling them to allocate funds to structure their teams for maximum efficiency and productivity.

Some practice respondents and one national expert noted fear of “running afoul of regulations” as hampering role delegation. A provider echoed the sentiments of clinicians in two large organizations when she noted that their ability to use standing orders was limited because their compliance department feared that the EHR could give the appearance that the MA was ordering the service. An RN in a California practice overcame this issue by requiring the provider to authorize the standing orders in advance of patient visits.

Training

Respondents mentioned challenges to training such as “lack of time to attend training” and the lack of awareness of training resources. Most practices had not undergone formal teamwork training. Some had a physician leader who had read about teamwork and then championed change within the practice. In Colorado, where HealthTeamWorks27 is active, practices used practice coaches28 and the military practice used PC TeamSTEPPS.20 Other practices, however, could not identify formal teamwork training. Many had participated in PCMH learning collaboratives and/or received some guidance from their larger parent organizations or IPAs, but commented that these often lacked teamwork training. Respondents from three practices also noted a need for more formal team training around patient self-management and care plans.

DISCUSSION

This paper provides practical details on how practices have developed solutions to address common challenges to teamwork. It includes new approaches as well as confirmation of previously described tools.3 , 4 , 9 , 12 , 24 , 27 32 These included using data to show the team how patient care improves with teamwork, incremental delegation by physicians,3 , 4 engaging staff in workflow redesign, creating a safe culture for feedback and questions,9 using outside coaches or practice facilitators,27 , 28 huddles,21 , 22 using EHR templates to guide data collection by MAs and nurses, and tracking task completion to help prevent sliding back to pre-teamwork behaviors.

When task redistribution maintained patient care safety, job satisfaction and interpersonal continuity of care improved both for physicians and for MAs and nurses who appreciated greater involvement in patient education and decisions about team care processes. Respondents generally believed that teamwork helped to avoid neglecting important patient care issues.

External system challenges (e.g., fee-for-service payment, resources, regulations)3 , 33 were most problematic because they were out of respondents’ control. Others have suggested that when teams lack formal authority to alter such external systems, they are particularly reliant on effective clinical leadership,34 so offering training to potential leaders in how to change care processes may help practices enhance performance and meet patients’ needs in a manner consistent with scientific evidence until such external challenges are resolved.

Limitations

We cannot generalize from this study to all primary care practices. The practice team element of the 2011 NCQA PCMH tool has not yet been widely validated, but its key elements are based on teamwork literature and feedback from providers, and it is the only practice-based sampling frame at present that provides some empiric indicator of teamwork functioning in primary care practices on a national scale.

This study’s findings have implications for practices working toward PCMH recognition as well as those already recognized as PCMHs, in addition to those simply interested in enhancing their teamwork. The provision of venues and resources for practice teams to learn from one another across practice sites, both in person and online via Web-based learning, has the potential to improve teamwork by allowing practices to share lessons learned and to model behaviors/roles for one another. Areas particularly ripe for such shared learning include effective role delegation, structured communication strategies (e.g., huddles), ways to avoid sliding back into pre-teamwork behaviors, and team roles for efficient population management. Such shared learning, particularly if facilitated by persons knowledgeable about on-the-ground use of EHRs in primary care, has the potential to help practices in improving the quality of their teamwork.

From a policy perspective, practices expressed a need for more staffing resources to enhance teamwork in care plans and patient self-management, funding for practice coaches, and assistance in better understanding the implications of regulations (e.g., scope of practice) with respect to teamwork.

As recognized PCMHs, the practices we included likely have better resources and may be more highly motivated than typical practices in the U.S., but their awareness and use of resources or curricula to support primary care teamwork was scant. We hope that the present study, in addition to current work in the field,35 will add to the existing body of research in helping practices devise ways to function as more effective teams. From participants’ comments, it is clear that the demands of primary care make it challenging for staff to leave the practice for lengthy periods of time to engage in formal classes, so teaching and tools that can be brought to the practice, or accessed as needed from within the practice, may be most feasible.