Professional Documents
Culture Documents
research-article2018
HPPXXX10.1177/1524839918790936Health Promotion PracticeMcQueen et al. / Perceptions of Health Coaching for Behavior Change
Use of health coaching to help individuals make and Keywords: health education; health behavior change;
sustain changes in health behavior and disease man- health coaching; self-management
agement is increasing, and early evidence about its
effects is promising. However, few studies assess par-
H
ticipants’ preferences and expectations about health ealth coaching in the United States is used by
coaches, the use of the term health coach, and the health insurance plans, workplace wellness pro-
coaching relationship. To help inform the design of grams, primary care clinics, and community
future health coaching programs, we conducted quali- health centers to promote health behavior change.
tative interviews with 50 U.S. adults (25 with Medicaid, Health coaching combines health education with psy-
25 commercially insured) to assess their (1) sources of chosocial support and behavior modification tech-
health information; (2) familiarity with, preferences for, niques to meet patient-determined goals, often involving
and associations with 32 terms that could describe lifestyle behaviors such as diet, exercise, smoking ces-
“someone who uses experience and one-on-one com- sation, and stress management (Hill, Richardson, &
munication to help others change behaviors to improve Skouteris, 2015; Wolever et al., 2013). Studies of coach-
their health;” and (3) perceptions of how well different ing interventions have shown positive effects on behav-
terms apply across different health behaviors and for iors, including physical activity, nutrition, and weight
nonhealth concerns (e.g., financial, career). Results loss (Hill et al., 2015; Olsen & Nesbitt, 2010), as well as
showed wide variability in preferences for different improved medication adherence and health outcomes
coaching terms and mental models of coaching that among those with chronic diseases (Vale, 2003; Wolever
differed by insurance type. Commercially insured par- et al., 2010).
ticipants made a distinction between experts and sup-
porters as different categories of coaches. Medicaid
1
participants associated coach with sports and viewed Washington University in St. Louis, St. Louis, MO, USA
the term as less professional. All participants preferred
terms such as advisor and specialist, and all wanted
Authors’ Note: The research was conducted as part of collabora-
content experts who could also appreciate the broader tions with the Envolve Center for Health Behavior Change Research
context of their life. at Washington University School of Social Work. The authors wish
to thank all the participants and the Missouri Home State and
Louisiana Healthcare Connections health plans for their support
of this research. The authors also acknowledge contributions to
the research and reporting of results by Nikki Caito, Katie Childs,
Health Promotion Practice Rachel Garg, Jennifer Funaro, Lisa Gibson, and Karyn Quinn.
Month XXXX Vol. XX , No. (X) 1–10 Address correspondence to Amy McQueen, Washington University
DOI: 10.1177/1524839918790936 in St. Louis, 4523 Clayton Avenue, Campus Box 8005, St. Louis,
© 2018 Society for Public Health Education MO 63110, USA; e-mail: amcqueen@wustl.edu.
1
Although titles, training, and credentials of health better understand how coaching programs offered
coaches vary (Ustjanauskas, Bredice, Nuhaily, Kath, & through commercial health insurance plans and
Wells, 2016), there is some agreement about what they employers may need to be adapted to better reach and
do. The International Consortium for Health & Wellness serve Medicaid beneficiaries. Specifically, we antici-
Coaching (ICHWC) conducted a job analysis that identi- pated that Medicaid beneficiaries may have more
fied common tasks, knowledge, and skills to inform a unmet basic needs (e.g., food, housing, unpaid bills)
national standard for health coaches (ICHWC, 2017). In that act as barriers to behavior change and may extend
general, ICHWC concluded that coaches do not diag- beyond a coach’s expertise or scope of work. To explore
nose or advise, but rather elicit goals and solutions from potential differences, we interviewed Medicaid benefi-
clients. Coaches use strategies to motivate and hold ciaries and individuals with commercial health insur-
clients accountable to change behaviors in ways that are ance to gain insights about (1) the sources of health
appropriate to their life context (Wolever, Jordan, information and support currently used by members of
Lawson, & Moore, 2016). The focus of these duties can both groups, (2) their familiarity with different coach-
vary, with some coaches specializing in a particular ing-related terms and associations with our definition
chronic disease or health behavior and others address- of a health coach, and (3) preferences for particular
ing wellness holistically. Although there is some con- terms and how those preferences may vary across
ceptual overlap between health coaching and patient health behaviors and for nonhealth concerns such as
navigation, the latter has a specific focus on care coor- financial or career needs. The goal of this formative
dination within health care organizations, which can research was not to test a priori hypotheses, but to
also help increase particular health behaviors such as understand how people talked about different terms
timely adherence to screening and treatment (Ali-Faisal, and their expectations of programs we considered rel-
Colella, Medina-Jaudes, & Benz Scott, 2017). evant to “coaching.” The study was informed by schema
The promising findings from health coaching theory, which includes both relevant attributes of a
research have led to questions about the “active ingre- concept and the associations among those attributes
dients” or mechanisms that drive their effects (Hudlicka, (Fiske & Taylor, 1991). Specifically, role schemas create
2013; Willard-Grace et al., 2015) and prompted forma- cognitive scaffolding for individuals to organize their
tive research such as qualitative interviews with knowledge and expectations about individuals and
coaches to understand how they engage with clients their behaviors depending on their social position and
(Goldman, Ghorob, Eyre, & Bodenheimer, 2016) and their situational context.
with physicians to understand the role of coaches in
their practice (Johnson, 2015) and the level of support
they receive (de la Riva et al., 2016). Fewer studies >>
Method
have interviewed potential recipients of coaching to
Sample
learn about their preferences for and expectations
about health coaches and the coaching relationship. Eligible participants were adults 18 years or older
Although several studies have interviewed patients who were able to complete a single, face-to-face inter-
after they completed coaching interventions for life- view in English. Only one person per household could
style behaviors, the studies excluded individuals who participate. For practical (e.g., limited time and person-
choose not to participate in coaching or who disen- nel resources) reasons, we recruited a convenience
gaged during the study (Andersen, Andersen, sample of two groups. Commercially insured partici-
Muurholm, & Roessler, 2014; Lyden et al., 2013; pants were recruited through a research volunteer reg-
O’Sullivan et al., 2010). Because only a minority of istry and recruitment flyers posted in community
people who are invited or eligible to participate in a locations (i.e., university buildings, coffee shops, pub-
coaching program actually start and complete one, lic libraries, and a fire station). Medicaid beneficiaries
research is needed to understand the facilitators and were recruited from Medicaid Managed Care Plans in
barriers to engaging in a coaching program. Missouri (Home State Health) and Louisiana (Louisiana
This study was conducted to inform design deci- Healthcare Connections) who provided the research
sions and best practice implementation strategies for a team with contact information (Missouri) or onsite
future health behavior change program using phone- referrals (Louisiana). Letters were mailed to Missouri
based and online health coaching to address lifestyle members, and research staff made telephone calls
behaviors (i.e., diet, exercise, smoking cessation, stress within 1 to 2 weeks to schedule interviews, which were
management) among Medicaid beneficiaries in the completed in December 2016. Louisiana participants
United States. This formative research was needed to were recruited and interviewed during a health fair at
a federally qualified health center in New Orleans in parent), and relevant synonyms of each. Terms were not
October 2016. Interviews with Missouri participants limited to health-focused applications. The list was not
were conducted in private meeting rooms at the univer- presumed to be exhaustive, and participants were encour-
sity or a local federally qualified health center. Response aged to add terms using blank cards provided to them
rates were not accurately recorded due to problems (although only one did this, adding OB/GYN).
estimating the denominators of commercially insured
participants exposed to study flyers, Home State mem- Procedure
bers who actually received a letter and phone call (60%
of phone numbers were incorrect), and Healthcare After providing verbal informed consent, each par-
Connections members who heard about the study dur- ticipant was interviewed face-to-face for 30 to 60 min-
ing the health fair. Participants did not have a relation- utes. First, open-ended questions elicited participants’
ship with research team members prior to their usual sources for health information and support. Card
interview. All participants were told the study purpose sort exercises were then used to assess the familiarity
and the role of the research team, which was independ- of participants with different terms, their reactions, and
ent of their health plan. preferences. For example, participants were asked to
sort the cards into two piles: terms that were familiar
and unfamiliar to them when applied to health. Starting
Interviewers
over, participants then sorted the same cards into two
To accommodate participants’ preferred interview new piles of terms they liked or didn’t like, then
days and times, five research team members including selected terms they preferred most. Throughout the
A. M., M. L., and T. P. were trained to conduct inter- interview and card sort exercises, participants were
views. Ages, gender, and credentials varied across inter- encouraged to reflect and elaborate on the characteris-
viewers; T. P. is a former reporter and current publications tics, qualifications, experience, services, and types of
editor in a research lab; two interviewers have graduate personal interactions they associated with each term
degrees, and two interviewers were graduate students. being discussed. Participants also were asked how well
Interviewers were trained to use the semistructured each term applied to someone who helps others with
interview guide (Supplemental Appendix A available more than one health behavior as well as nonhealth
in the online version of the article) to elicit the opinions issues, for example, career development, financial
of participants. None of the interviewers had been a planning, parenting, or civic engagement.
health coach or involved in a health coaching program, Participants completed a brief paper survey of demo-
which further supported our participant-as-expert inter- graphic information using standard questions from
view style. The research team had few a priori expecta- national surveys and received a $25 gift card to a local
tions about term preferences and associations or grocery store (Missouri) or Walmart (Louisiana). Study
differences between groups. One exception was an early procedures and materials were approved by the institu-
assumption that Medicaid beneficiaries may be less tional review board at Washington University in St.
interested in having a health coach compared with com- Louis ID 201509027 on December 22, 2015, and were
mercially insured participants due to greater concerns approved by the Louisiana Department of Health and
about trust, intrusiveness, or time required. the participating health plans.
Phenomenology focuses on the participant’s lived experts and distinguished other groups of terms as
experience and perspectives. The constant comparative nonexperts. To construct mental models, we examined
methods of grounded theory guided the coding and the following information about each term: expert ver-
analysis processes that were both deductive and induc- sus supporter, educational background, other terms
tive (Miles, Huberman, & Saldana, 2014). A codebook viewed as closely related, services provided, associated
was developed by the first author a priori based on the characteristics of this person, relationship to client,
interview guide and initial research questions, and and specific behaviors or topics addressed. The
emergent codes were added to the codebook by the responses of Medicaid participants were examined for
coders during the iterative coding process (Supplemental the same information, but the mental model that
Appendix B available in the online version of the arti- emerged was different. All results were discussed
cle). Because of differences in the timing of recruitment among investigators, providing opportunities to chal-
efforts, we completed the coding and codebook for the lenge perceptions, explore potential negative and devi-
commercially insured sample, then applied the code- ant cases, and reduce the potential for confirmatory
book to the Medicaid sample and made few changes. bias (Esterberg, 2002; Pidgeon & Henwood, 1997).
Two staff members were trained and supervised in cod- Additional interviews were not sought once the data
ing each transcript independently. Discrepancies that became saturated per sample for our interview ques-
arose were resolved through group discussion to tions determined a priori, and no new themes or con-
achieve consensus (Padgett, 2012). During analysis, we tradictory evidence became evident. Saturation was
focused on individuals’ interpretations and meanings. determined through discussions with interviewers,
When primary coding had been completed for all tran- coders, and research leads. Illustrative quotes were
scripts, reports for each code were reviewed and sum- selected for presentation and are accompanied by the
maries were created to address deductive research code assigned by the interviewer to the interview for
questions. simple data tracking purposes. Specific codes are not
In contrast, our interpretation of participants’ cogni- meaningful per se, but indicate the interviewers ini-
tive representations or “mental models” of the 32 terms tials, sample (Medicaid Missouri, Medicaid Louisiana,
emerged from the data. Specifically, many commer- or Commercially insured), and a sequential numbering
cially insured participants grouped some terms as per interviewer.
>>
Results was frequently selected as a preferred term. Fewer
Medicaid members selected the term.
Participant Characteristics Although the mental models we developed varied by
We interviewed 50 adults: 25 commercially insured; insurance type, they had several similarities. First, all
25 Medicaid (16 from Missouri, 9 from Louisiana). The participants thought of medical providers as the pri-
ages of participants ranged from 18 to 62 years (M = 38, mary specialists who would address serious health
SD = 10.5). Most were women (72%), and either Black concerns. Everyone saw specialists as focused on one
(56%) or White (38%). Most (74%) were employed full- specific issue, and most agreed that different specialists
time or part-time, and their educational backgrounds would be needed to address different health conditions
varied from high school (20%), to some college (28%), or life concerns (e.g., financial, child care, job). Although
to bachelors or advanced degrees (38%). Ten percent of personal experience was valued, when pressed, most
participants described themselves as being in poor or participants prioritized an expert’s experience helping
fair health. Most (78%) had looked for health informa- others change behavior as more important than having
tion online in the past 12 months. Participants reported faced the same issue themselves (e.g., having a former
being active an average of 3 days a week (SD = 1.8), smoker as a quit smoking coach; Table 1).
eating an average of 2.7 servings of fruits and vegeta- Second, all participants valued family members and
bles per day (SD = 1.2), and 14% smoked some days or friends as lay supporters of their health and nonhealth
every day. Thirty percent of participants were over- needs. Participants immediately thought of people they
weight, and 40% were obese. had close relationships with when discussing terms
such as parent, friend, and partner, and thus these
were not viewed as useful terms to describe profes-
Sources for Health Information and Support
sional health coaches with whom they did not have
In both samples, the top three sources of health such relationships.
information were the Internet (e.g., WebMD, Google), Third, some terms consistently elicited strong
doctors or clinics, and family members or friends—ide- impressions or associations that may not be helpful for
ally people with health-related experience or previous engaging people in certain health coaching programs.
training (e.g., nursing). For example, trainer was always associated with per-
Thirteen of the 25 commercially insured partici- sonal trainers who focus on exercise. Geek was not a
pants reported prior exposure to coaching: seven had desirable term and was always associated with comput-
some prior experience with a health coach, two had ers and tech support. Coach and cheerleader were
experience with a personal trainer, and four were almost always associated with sports, which led to
familiar with the practice through friends or coworkers. mixed reactions regarding the appeal of the terms
Spontaneous examples included Weight Watchers among participants. Elders were viewed as senior mem-
coaches, personal trainers, disease support groups, and bers of a family or church, which also led to mixed
web- or telephone-based health coaching programs. reactions.
Only one of the 25 Medicaid participants reported Commercially insured participants made a strong
some prior experience with coaching. Most Medicaid distinction between experts and lay supporters, and
participants (n = 15) thought of health care providers neither were perceived to be emotionally or personally
when considering the definition we provided. Nine involved with the client, but rather acting in paid posi-
additional participants described helpful individuals tions to serve clients. Commercially insured partici-
such as social workers, health educators, counselors, pants also articulated when advanced educational
housing authority advocates, nutritionists or dieticians, degrees or other credentials would be needed (e.g., for
wellness professionals, and school-based staff who mental health services), and when specific training or
helped families receiving state assistance such as food experience would suffice (e.g., for proper use of exer-
stamps. cise equipment). Participants said they would want an
expert if they had a serious health concern, but were
willing to see a nonexpert for “preventive health or
How Terms Related to Our Definition of a Health
general health information” (TPC6). Several responses
Coach
suggested that lived experience was especially valued
Preferences for different terms are presented by for nonexpert sources. For example, one said, “the phy-
participant health insurance type in Figure 1. Both sician or expert can give you the medical knowledge of
groups preferred the terms specialist and advisor. For it, but to see and hear a person who’s really had it
commercially insured participants, the term coach makes you feel better” (AMC02). It’s the nonexpert who
can show you “what recovery [from surgery] looks Figure 2 illustrates commercially insured partici-
like” (TPC5). Some participants thought having an ini- pants’ mental model of terms associated with these two
tial counselor or consultant who would get to know types of care (experts vs. supporters). The vertical
you and refer you to the best person to address your organization distinguishes a consumer-focused affilia-
problem would improve participant engagement and tion versus a personal relationship. The ordering of
trust (NCC1). terms in the figure was based on participants’ collec-
Distinct characteristics and expectations were tive comments about each term. Terms that were men-
ascribed to each term. For example, participants felt tioned together or that shared similar characteristics
that an advisor provides information, resources, and are shown more closely together, even overlapping.
referrals in their domain of expertise, but not necessar- Supplemental Table S1 (available in the online version
ily solutions to a problem. A supporter was not viewed of the article) provides sample quotes showing how
as an expert, but was assumed to be a good listener and terms were related.
encouraging rather than providing advice. Neither Compared with commercially insured participants,
advisors nor supporters were associated with having a Medicaid participants classified many more of the
close personal relationship with the client like a men- terms as experts compared with the commercially
tor or ally would. Participants expected to ask ques- insured participants, but they were not specific about
tions of consultants who would provide direction and what credentials they should have. Anyone in a posi-
advice, whereas participants would delegate tasks to tion providing support to the participant was perceived
(subservient) assistants who were not expected to be a to have valuable expertise through training or experi-
source of new information or resources. ence. Although many participants described family
prior experiences with a health coach. The mental mod- healthcare utilization outcomes: A meta-analysis of randomized
els also may be affected by the indirect elicitation meth- controlled trials. Patient Education & Counseling, 100, 436-448.
ods we used. By design, our sampling approach included Andersen, L. N., Andersen, S. S., Muurholm, B., & Roessler, K. K.
many people without prior experience with a health (2014). A qualitative study of citizens’ experience of participating
in health counseling. Scandinavian Journal of Psychology, 55,
coach; however, the far greater proportion of Medicaid
558-566.
participants without prior exposure may have biased
Carpenter, K. M., Lovejoy, J. C., Lange, J. M., Hapgood, J. E., &
our comparison. Future studies may benefit by stratify-
Zbikowski, S. M. (2014). Outcomes and utilization of a low inten-
ing participants by prior coaching experience per sam- sity workplace weight loss program. Journal of Obesity, 2014, 1-7.
ple. Knowledge of members’ prior experience with doi:10.1155/2014/414987
coaching also might inform how organizations design Charmaz, K. (1995). Grounded theory. In J. A. Smith, R. Harre, &
and market their own coaching programs. L. Van Langahove (Eds.), Rethinking methods in psychology (pp.
Results suggest that regardless of insurance type, 27-49). London, England: Sage.
participants may respond well to terms such as health de la Riva, E. E., Hajjar, N., Tom, L. S., Phillips, S., Dong, X., &
specialist or advisor, but that Medicaid beneficiaries Simon, M. A. (2016). Providers’ views on a community-wide
without prior exposure to coaching programs may not patient navigation program: Implications for dissemination and
respond as positively initially to the term health coach. future implementation. Health Promotion Practice, 17, 382-390.
It is possible that with greater exposure to the term Dillon, E., Panattoni, L., Meehan, A., Chuang, J., Wilson, C., &
Ming, T.-S. (2016). Using unlicensed health coaches to improve
health coach and coaching programs that Medicaid
care for insured patients with diabetes and hypertension: Patient
participants would come to understand and value the and physician perspectives on recruitment and uptake. Population
term more. However, familiarity alone is not the only Health Management, 19, 332-340.
answer; positive messaging about how such an indi- Esterberg, K. G. (2002). Qualitative research methods in social
vidual could help clients also may be needed. Dislike research. Boston, MA: McGraw-Hill.
for the term geek was because of ubiquitous familiarity Fiske, S., & Taylor, S. (1991). Social cognition (2nd ed.). St. Louis,
with geek squads who provide computer support. MO: McGraw-Hill.
Although some terms were clearly disliked by all par- Goldman, M. L., Ghorob, A., Eyre, S. L., & Bodenheimer, T. (2016).
ticipants (i.e., guru), it is unlikely that any one term How do peer coaches improve diabetes care for low-income
will evoke only positive reactions and associations patients? A qualitative analysis. Diabetes Educator, 39, 800-810.
among everyone. Further audience segmentation Goulding, C. (2005). Grounded theory, ethnography and phenom-
research could help identify specific services and per- enology. A comparative analysis of three qualitative strategies for
sonal characteristics that are most highly valued by marketing research. European Journal of Marketing, 39, 294-308.
different participants considering a health coaching Greif, S. (2010). A new frontier of research and practice:
program in the United States or abroad. Observation of coaching behaviour. Coaching Psychologist, 6,
97-105.
Previous studies have attempted to identify which
aspects of the counseling or coaching relationship con- Hill, B., Richardson, B., & Skouteris, H. (2015). Do we know how
to design effective health coaching interventions: A systematic
tribute to successful behavior change (Greif, 2010; Hill
review of the state of the literature. American Journal of Health
et al., 2015; Wolever et al., 2013). More research is Promotion, 29, e158-e168.
needed from the client perspective to understand what Hudlicka, E. (2013). Virtual training and coaching of health
factors influence initial and continued engagement behavior: Example from mindfulness meditation training. Patient
with health coaches. Many insurance companies and Education & Counseling, 92, 160-166.
employers commit significant resources to providing International Consortium for Health & Wellness Coaching. (2017).
health coaching programs, but only a small proportion Health and wellness coaching job task analysis findings. Retrieved
of members or employees enroll, limiting program from http://ichwc.org/wp-content/uploads/2015/03/JTA-ICHWC-
effectiveness and population health benefits (Carpenter, Feb-1-2017.pdf
Lovejoy, Lange, Hapgood, & Zbikowski, 2014; Dillon Johnson, J. D. (2015). Physician’s emerging roles relating to trends
et al., 2016). in health information technology. Informatics for Health and
Social Care, 40, 362-375.
Supplemental Material Lyden, J. R., Zickmund, S. L., Bhargava, T. D., Bryce, C. L., Conroy,
The Supplemental Material is available in the online version of M. B., Fischer, G. S., . . . McTigue, K. M. (2013). Implementing
the article at http://journals.sagepub.com/home/hpp. health information technology in a patient-centered manner:
Patient experiences with an online evidence-based lifestyle inter-
vention. Journal for Healthcare Quality, 35(5), 47-57.
References
Miles, M. B., Huberman, A. M., & Saldana, J. (2014). Qualitative
Ali-Faisal, S. F., Colella, T. J., Medina-Jaudes, N., & Benz Scott, data analysis: A methods sourcebook (3rd ed.). Thousand Oaks,
L. (2017). The effectiveness of patient navigation to improve CA: Sage.