Building Teams in Primary Care: What Do Nonlicensed Allied Health Workers Want?
Summer 2011 - Volume 15 Number 3
Introduction: Nonlicensed allied health workers are becoming increasingly important in collaborative team care, yet we know little about their experiences while filling these roles. To explore their perceptions of working as health coaches in a chronic-disease collaborative team, the teamlet model, we conducted a qualitative study to understand the nature and dynamics of this emerging role.
The high-functioning health care system, based on the chronic care model, hinges on a team-based approach to care. Health care teams, though, are broadly defined and vary tremendously in how much each member interacts directly with the patient and collaborates with other members. As primary care practices consider how best to configure such teams, they have increasingly turned to nonlicensed allied health workers to play a significant role.
One method of incorporating these workers in collaborative care practice is to involve them as health coaches, such as in the teamlet model. Thomas Bodenheimer coined the term teamlet to describe the smallest, most patient-centric model that pairs a clinician with a medical assistant (MA) or a community health worker. This teamlet works to provide a variety of services for a panel of patients and to help patients and their families manage their own chronic conditions within the context of their daily lives.1,2 Figure 1 describes how health coaches enact their roles within the teamlet model both during and between clinic visits.
Although the functioning of a teamlet has been previously described,3,4 we know very little about what nonlicensed allied health workers experience when working as a health coach within this model. Involvement in such a collaborative team requires nonlicensed allied health workers to significantly expand their traditional job responsibilities.5,6 To optimize patient care and job satisfaction, it would be important for those involved in practices to understand what nonlicensed allied health workers identify as important issues to consider when they assume new roles in the collaborative care of chronically ill patients and their families.
To examine how nonlicensed allied health workers—specifically MAs and community health workers—experience the role of health coach, we designed a qualitative study and conducted in-depth individual interviews with the 11 health coaches who participated in a 1-year pilot intervention based on the teamlet model of care.7 We did not intend to assess which of the 2 types of nonlicensed allied health workers gave better health coaching but rather to gather the perceptions of both types of workers of filling the new role of health coach. The study was approved by the institutional review board of the University of California, San Francisco.
The San Francisco General Hospital Family Health Center (FHC), a family medicine teaching clinic, is the largest primary care clinic within the San Francisco Community Health Network, serving more than 10,000 active patients. The patient population is racially and ethnically diverse (39% Latino, 27% Asian, 17% white, 13% African American), with 83% uninsured or covered by Medicaid. Patients speak 29 different languages: Most common are English (42%), Spanish (25%), and Cantonese/Mandarin (8%). The FHC is the primary ambulatory training site for the 41 resident trainees in the Family and Community Medicine Residency Program at the University of California, San Francisco. Residents in our practice all have a continuity panel of patients whom they monitor throughout their 3 years of training. They have between 1 and 4 half-day clinic sessions each week of residency, and their patients see them as their primary provider for the large majority of their clinic visits.
Eleven health coaches participated in the program for the entire academic year. Before the study, they were working in the FHC as either MAs or community health workers, two groups that represent different health care disciplines with variations in training and roles in the clinic. At the FHC, MAs are trained to carry out clinical care, including measurement of vital signs, vaccination administration, point-of-care testing, and varied administrative tasks. In contrast, community health workers in our system are paraprofessionals, usually with a background in health education and community health (eg, advocacy, outreach). They are unlicensed, are required to have a high school diploma, and frequently are similar to our patient population in ethnicity, language, and socioeconomic status.8 They are clinic based and do not monitor patients outside the health center. They are trained in administrative skills, such as billing and clerical tasks and patient education, but they lack clinical training. Community health workers are also more likely than MAs to have received some training in collaborative communication skills.
We implemented the teamlet model with 13 first-year family medicine residents, 11 health coaches, and 150 patients. This implementation coincided with our participation in an improvement collaborative involving teaching clinics throughout California. The FHC Medical Director and Nurse Manager assigned all available MAs and community health workers (n = 11) to be health coaches with the first-year family medicine residents. A key component of this teamlet model intervention was to ensure a high degree of cultural and linguistic concordance among the health coach, physician, and patients. Sharing language and cultural identity was considered important for optimizing team functioning, with patients' access and engagement in their care. Table 1 demonstrates how the health coaches' race or ethnicity reflected that of our active patient panel.
All FHC nursing staff, including MAs and community health workers, participated in health-coach training. The training9 helped workers develop an understanding of cardiovascular risk factors, including diabetes, and acquire knowledge about and skills for developing collaborative partnerships with patients and families to come up with action plans for healthy behavior changes and make medication adherence easier. Training required active participation through role-plays to develop skills for negotiation of behavior-change action plans, medication reconciliation, and patient-centered communication. Health coaches were trained to identify low literacy as it related to medication adherence and learned strategies to help patients track their medications. They were trained to constantly assess for health literacy by asking patients what they want to learn about and checking for understanding, having been instructed in "Closing the Loop" strategies.9
Health coaches took part in 6 training sessions before taking on their new role with patients. Ongoing training involved live observations, mentoring, and case discussions to further build patient-communication skills. Total training time ranged from 14 to 16 hours, and competency was determined through direct observation by the trainers.7 The health coaches also met regularly throughout the year with teamlet model faculty members to provide input about their experience, which helped shape the model during its implementation.
During the intervention year, the physician–health coach teamlets cared for 150 patients who had cardiovascular risk factors. The patients were part of the continuity practices of the family medicine residents. The health coaches participated in 16 to 18 clinics and were involved with 5 to 10 patients. The frequency that health coaches had contact with patients, whether by telephone or in person, varied according to patients' needs and interests. Some patients did not engage after 1 contact and others had as many as 20 over the year. The role of these nonlicensed allied health workers as health coaches was seen as separate from their regular job duties and, at times, required them to move geographically to another clinical team within the FHC.
The study instrument, a semistructured interview, was developed by the research team. Questions for the qualitative study instrument for staff who participated in the chronic-care clinics were articulated around four major areas of inquiry:
Table 2 displays the complete health-coach interview instrument.
In-depth individual interviews with the 11 health coaches participating in the teamlet model of care were conducted during a 2-month period. The same investigator (ST), who was not involved in the teamlet project or in employee evaluation, conducted all interviews in private, away from the participants' workplace. Interviews were developed to elicit personal experiences working as a health coach. All 11 health coaches in the study were approached by the research investigator regarding participation in the individual interviews. They were informed that their responses would remain anonymous, be reported only in aggregate, and not influence their employment. All agreed to be interviewed. After participants signed consent forms, interviews lasting 45 to 60 minutes and carried out in English were audiotaped. The audiotapes were transcribed, and any identifying information was removed from the transcriptions. The audiotapes were then destroyed. Data analysis was performed concurrently with data collection.
Grounded theory methodology10 was used to understand the perceptions of the health coaches. The constant comparative method of grounded theory involved continual comparison of the data and the themes that emerged. Multistaged coding was used and began with open coding of raw data to develop key ideas. Axial coding then organized these categories into patterns, and finally selective coding was used to develop theoretical formulation that linked key variables to themes. Three investigators (ST, LW, and GS) read one transcript to identify and to develop initial descriptive codes. All 11 transcripts were then reread individually by each investigator using this initial list of codes. ATLAS.ti scientific software (version 5.2; ATLAS.ti GmbH, Berlin, Germany) was used to code the interviews using these descriptive codes and identified patterns. The 3 investigators discussed initial coding, and then interpretative codes were created to better describe meanings generated from the data. Discrepancies in analysis were resolved through discussion and reaching consensus among all 3 investigators.
Qualitative analysis of the interviews revealed four themes. Table 3 displays sample quotes from the health coaches for each of the themes.
Responsibilities and Roles Varied from Theoretical Expectations of Model
Health coaches described some variation between what was defined conceptually in the teamlet model3 and what occurred in its real-world application:
However, not all of the health coaches consistently assumed the responsibilities, and variations depended on individual health coaches' perceptions of available time, interest in the particular responsibility, and patient need.
Beyond duties and responsibilities, the health coaches identified three primary roles that they assumed:
Communication and Relationships Were Key to Health-Coach Experience
Communication and relationships were essential to how the health coaches valued their participation in the teamlets. They identified a triad of important relationships that developed among the three members of the team (patient, clinician, and health coach).
Patient–Health Coach Relationship
The aspects of communication between health coaches and patients that the health coaches valued included:
What they described indicated that a special and strong emotional patient–health coach bond often existed in which patients believed that someone cared for them (ie, "being there for them"). They expressed that this bond was enhanced by virtue of the extended patient contact (time) and availability of the health coach (focal person). Participants described a high level of trust between patients and health coaches because the patients saw the health coaches as their equals and not as authority figures like their physicians.
Health coaches reported that they often gathered patient information that might not be routinely revealed to the clinician and that affected treatment plans. For example, health coaches described how patients would tell them of their involvement in gangs, not taking the right medication, and feeling depressed. As one health coach said, "The relationship between the coach and the patient … is the key [to] this program work[ing] or not."
Teamwork Approach That Includes the Patient
Health coaches spoke about wanting to create a well-functioning health care team that included the physician, patient, and themselves:
Multiple Factors Influenced Health Coaches' Experience of the Model of Care
Overall Experience of Health Coaches
The positive attributes of the teamlet model pilot included the following:
The negative attributes of the teamlet model pilot included the following:
Prior Roles of Nonlicensed Allied Health Workers
Whether a health coach had been trained as an MA or as a community health worker had considerable influence on how satisfied they were with being a health coach. MAs reported less of a benefit for themselves from being a health coach compared with their usual roles and clinical duties. New responsibilities in patient education or in navigation of the health care system were least satisfying. Some described the feeling of being "stuck in the room" and not being available to manage patient flow outside of the room for other patients. There was a perceived loss of autonomy in addressing patient needs because the health-coach role was seen as more formulaic. Participants trained as community health workers, however, uniformly enjoyed the one-on-one patient interaction and wanted to have more dedicated time for health coaching. All six of the community health workers were interested in continuing this role in the future. Only one of the five medical assistants could foresee continuing as health coaches in the future. The most common reason the other four gave for not wanting to continue health coaching was a lack of interest in providing one-on-one, in-depth patient education and motivation. These MAs preferred managing patient flow and carrying out discrete clinical tasks to help patients move efficiently through their clinic appointments.
Systems and Personnel Obstacles to Model Effectiveness
Two obstacles were identified that have significant implications to practice improvement.
Systems-level obstacles included not having enough protected time to schedule appointments, complete treatment authorization requests, or conduct follow-up phone calls. Health coaches consequently had to complete calls to patients outside of work hours. They felt pressure to respond to patient clinical demands, particularly when there was insufficient availability in primary care clinicians' follow-up appointments and insufficient time during the visit to do all health-coaching tasks (ie, medication reconciliation, creation of an action plan, patient education, and closing the loop). They also felt that there was insufficient protected time to review their patient panel with physicians.
Health-coach obstacles included a lack of knowledge for use in answering patient questions and a feeling of insecurity when compared with physicians. Finally, the influence of prior training and experience for MAs significantly reduced their motivation to carry out health-coaching tasks.
Our study had several limitations that affect the generalizability of its findings. It explored the experiences of nonlicensed allied health workers trained as health coaches working in a pilot intervention based on the teamlet model. The nature of that intervention presented certain limiting factors: a small number of health coaches, health coaches working in only one clinic that was an academic training center, and a pilot that was conducted for only one year.
The pilot intervention focused only on one form of collaborative team care—the teamlet model. It was not powered to be statistically significant, which limited our ability to examine whether the teamlet model improved the primary care of patients. This also did not permit exploration of any differences in patients' outcomes related to whether their health coaches were MAs or community health workers. A separate analysis of clinical and process outcomes of this pilot intervention did reveal trends toward improvement.7 Other sites using MAs in diabetes care have also shown significant improvement in clinical outcomes such as control of blood sugar levels, blood pressure, and cholesterol level.11
Although the experiences of the physicians and patients who participated in this pilot intervention based on the teamlet model were recorded, their analyses will be forthcoming in a future study. The pilot implementation did not test for optimal panel sizes or case loads for health coaches. Rather, its focus was exploring the acceptability of a very new role for these allied health workers. Regarding training, the health coaches were observed during role-play sessions and directly observed with patients in the clinic to assess competency. However, no formalized competency testing was conducted in either knowledge or communication skills. Potential variation in competency could have affected health coaches' experiences of the teamlet model.
In addition, health coaches' experience of the teamlet model was limited to their work within an academic training center, which may affect generalizability to other practice settings. Some of benefits might be less significant in nonacademic settings or settings involving nonpublic hospitals. For example, the cultural concordance may be especially important for a culturally diverse patient population. The focus on health literacy, communication, navigation, and health education may be more effective when used with a less literate and/or immigrant population. Perhaps most important, the value that the health coaches attributed to their role in eliciting and conveying information that the patient might not feel comfortable discussing with a physician may be most powerful in settings where most physicians are less accessible to their primary care patients (eg, academic settings, community practices with part-time clinicians).
Generalizability of the findings must consider these significant limitations. However, given the burgeoning nature of the field, this case-study experience can provide direction for future research into the selection, training, roles, and impact of nonlicensed allied health workers assuming roles as health coaches in primary care redesign and in the patient-centered medical home. For example, randomized, controlled trials are currently being conducted that take into consideration the limitations identified in this study (eg, clinical outcome, competency training, nonacademic practice settings, clinician and patient satisfaction with the teamlet model, optimal panel size and case-load ratios).
The use of nonlicensed allied health workers as part of a collaborative team with physicians to improve the primary care of patients shows great promise, as it offers increased accessibility, continuity, and advocacy in the care of the chronically ill. The health coaches we interviewed felt that they had a significant impact on patients' health and enhanced patient–physician communication through the cultural and language skills that they brought to their role. Interestingly, they believed they had fostered trust and democratized the hierarchy in the clinical relationship, which led to important revelations by patients that otherwise would not have been uncovered.
Importantly, participants indicated that for nonlicensed allied health workers to feel successful and satisfied, they must be motivated to have the new role of health coach as part of their jobs. For some, this may require them to relinquish some of the skills, responsibilities, and job satisfaction they worked hard to achieve in their previous training and certifications. MAs and other health care staff may be trained and function very differently from physicians and other health professionals—they may work best while multitasking and maintaining flexibility to move quickly from one task to another.12,13 One of the issues that perhaps led to the MAs in our study reporting lower enthusiasm in carrying out the duties of health coaches was their more clinically oriented training. MAs are trained to carry out concrete clinical tasks such as measuring blood pressures, conducting a focused patient history, and managing patient flow. The duties of our MAs in their roles as health coaches consisted of extended one-on-one time spent with patients, providing education and motivational talks, which interfered with what they saw as their primary responsibilities—managing patient flow and conducting time-limited clinical tasks.
Health coaching may or may not suit all of the staff who might be recruited for the expanded role. It may place some of them in greater emotional proximity with patients, requiring interactions that go beyond what they would find desirable in patient care. Although all participants endorsed the value of the model, they emphasized that finding the right person for the job of health coach is paramount to its long-term success. In the words of one health coach, "Get people that genuinely want to be a health coach."
The operational challenges of implementing the teamlet model using nonlicensed allied health professionals will be intimately linked to the structure of the medical office where it is being implemented. Given that our study was not structured to determine the optimal ratio of health coaches to clinicians, we remain cautious in drawing generalizable recommendations. However, the experiences of these health coaches suggest that this model of team care will not survive without clear administrative and financial support. At a basic level, these health coaches wanted physical space to meet with patients and telephones for making calls between visits. Also, they recommended ongoing training and supervision, the opportunity to learn with and receive feedback from physicians, and better standardization of health-coach roles.
Most clearly, they reported that without the time to accomplish tasks already assigned, particularly for the MAs, the added responsibilities of health coaching often felt unmanageable and led to greater dissatisfaction. This highlights the importance of having a sufficient number of staff to accomplish all of the tasks essential for patient care—both the traditional ones (measuring vital signs, rooming patients, giving immunizations, and drawing blood) and those of health coaching (using protected time to do phone follow-up with patients, meeting with physicians to discuss patients' care). The health-coach role must be fully integrated into staff members' identity, job description, performance expectations, and daily activities. As one health coach said, "The other responsibilities in the clinic … were overshadowing the things that I do for the teamlet."
Even though we found less engagement from MAs in the teamlet model during this pilot intervention, that does not preclude MAs from functioning in the role of health coaches. However, our findings suggest that if the principals of a reconfigured practice want to use MAs as health coaches, they would do better by hiring more of them to accomplish both sets of tasks rather than asking one person to simply expand his or her role.7 Another variable to consider is the size of the patient caseload for health coaches, because practices would have to allow adequate time for panel management and between-session contacts. More studies are needed to quantify the optimal caseload and the time needed for experienced health coaches to integrate health coaching into MAs' work. At that point, cost and economic sustainability would have to be considered alongside the clear demonstration that such a model improves patient outcomes and helps address the crises of patient access and physician job satisfaction currently facing the field of primary care.
For those primary care practices considering health coaches as key members of collaborative teams, the experiences of a cohort of well-trained, nonlicensed allied health workers suggests three key requirements for success:
The author(s) have no conflicts of interest to disclose.
We appreciate the valuable feedback provided by Thomas Bodenheimer, MD, in the preparation of this manuscript.
Katharine O'Moore-Klopf, ELS, of KOK Edit provided editorial assistance.
1. Laing BY, Ward L, Yeh T, Chen E, Bodenheimer T. Introducing the "teamlet": initiating a primary care innovation at San Francisco General Hospital. Perm J 2008 Spring;12(2):4–9.