Development of a 5As-based technology-assisted weight management intervention for veterans in primary care

The prevalence of obesity in the United States (US) has increased over the past several decades [1], and recent data show that over one-third of adults in the US have obesity and over two-thirds have a body mass index (BMI) in the overweight range. Among Veterans, the prevalence of obesity is slightly higher [2, 3]. The United States Preventive Services Task Force (USPTF) recommends that all patients be screened for obesity and offered intensive lifestyle counseling since this can lead to modest weight loss and decreased risk of chronic disease [4]. Thus, Veterans Affairs (VA) Medical Centers require that all Veterans in primary care are screened for obesity by their Patient Aligned Care Teams (PACT) and referred to the VA MOVE! weight management program.

PACTs were originally launched in 2010 to implement a Patient-Centered Medical Home (PCMH) model of primary care at VA hospitals [5, 6]. PCMHs aim to improve primary care delivery through care that is patient-centered, comprehensive, increases access to services, and improves quality and safety [7, 8]. PACTs promote a whole person, team-based approach to coordinate Veterans’ medical, behavioral, and psychosocial healthcare needs [5]. The MOVE! weight management program follows evidence-based obesity treatment guidelines and has a comprehensive, multidisciplinary approach to weight management with group or individual meetings, often facilitated by dieticians or behavioral psychologists [9‐13]. The MOVE! program is effective [14‐16], but only 8% of eligible Veterans attend at least 1 MOVE! visit [17]. This underutilization of an existing intensive weight management program presents the need for other treatment options for those who do not attend, as well as increased support to promote attendance to MOVE!. Since Veterans receiving care at VAs see a primary care provider (PCP) an average of 3.6 times per year [18], primary care has the potential to be a valuable setting for both promoting healthy weight management behaviors and encouraging patients to attend MOVE!.

The 5As model has shown increasing promise in delivering a range of behavioral interventions within the primary care setting and is endorsed by the USPTF [19]. Medicare also reimburses practices for performing 5As counseling [20], and the Canadian Obesity Network endorses use of a similar model [21]. This model provides a framework for clinicians to effectively counsel patients through the completion of 5 specific groups of tasks (see Table 1). These include encouraging clinicians to “assess” current beliefs, behavior, and knowledge, “advise” lifestyle changes based on specific information about health risks, “agree” on collaborative goals based on the patient’s interest and confidence, “assist” the patient to achieve these goals by identifying barriers and creating strategies, and “arrange” for a specific follow-up plan to track progress [19, 22, 23]. Goal-setting, corresponding to the “agree” component of the 5As, is supported by the Theory of Planned Behavior [24] and other behavior change theories as a way to foster behavior change [25, 26]. Goal setting has been shown to promote both diet and physical activity changes and lead to weight loss [27, 28].

Several studies have demonstrated the feasibility and practicality of adapting the 5As model to deliver obesity interventions within the primary care setting [21, 29‐33]. Another study demonstrated that training physicians in 5As-based counseling resulted in modest patient weight loss at 12 months [34]. However, a major barrier to implementing the 5As in clinical practice is that PCPs and other healthcare team members often fail to counsel patients for many reasons including competing demands on time [35]. One way to address barriers to providing 5As-based obesity care within primary care settings is to use interactive behavior change technologies to assist with several tasks in the 5As model. These interactive technologies have the ability to assess behaviors/barriers, generate tailored advice, facilitate goal setting, and promote behavior change. Indeed, a recent systematic review demonstrated that technology-assisted weight loss interventions in the primary care setting helped patients to achieve weight loss compared to usual care [36]. In addition, they can also help to address the well-documented time constraints of clinicians. Thus, interventions using technology-assisted goal setting have the potential to overcome barriers and facilitate 5As-based weight management counseling.

Based on our prior research and experience training providers to deliver 5As [30, 32, 37], as well as other studies showing that time and competing demands could limit implementation and sustainability of 5As counseling [21, 29, 31], we sought to develop an intervention that would facilitate 5As counseling without overburdening providers/healthcare teams and also increase enrollment in MOVE!. In this paper, we describe how we used qualitative research methods and user-centered design approaches to develop and pilot test a 5As-based, technology-assisted weight management intervention to improve obesity care for Veterans within primary care.

The intervention, rooted in the 5As framework, is based on the Theory of Planned Behavior where intention to perform a new behavior predicts behavior change [24]. Studies show that interventions that produce greater changes in intention are more likely to produce behavior change [38]. It is also based on current goal setting theory where achieving small, doable goals increases a patient’s self-efficacy to set and achieve more goals [39, 40]. Our approach to intervention development was based on the ORBIT model, a systematic framework for guiding efforts to translate basic behavioral science findings into behavioral treatments for preventing and treating chronic illness, which recommends iterative design with frequent pilot testing [41]. We iteratively developed intervention components with a user-centered design approach and piloted them between July 2014 and July 2015. We utilized a multidisciplinary research team of VA clinicians, health science and public health students/professionals, software developers, dietitians, physical activity specialists, as well as extended advisors and consultants from the MOVE! weight management program and the VA National Center for Health Promotion and Disease Prevention. Here, we summarize this process in two phases: the “Development Phase” and the “Testing Phase” [see Table 2].

The Development Phase consisted of 1) a thematic analysis of previous formative research (Veteran focus groups and key informant interviews with PACT staff) [42, 43] to determine intervention components, and 2) iterative development of an online weight management tool that facilitates goal setting (later named “Goals for Eating and Moving” or “GEM tool”). The Testing Phase consisted of two rounds of concurrent usability testing [44] of the GEM tool with Veterans, and pilot testing of other components of the intervention, including health coaching materials. Round 1 of testing occurred in December 2014 and Round 2 in May 2015. The protocols for both phases were approved by the local VA Institutional Review Board and participating Veterans consented to be part of all study procedures. Additionally, all sessions were audio-recorded, with focus groups, key informant interviews, and usability testing sessions transcribed professionally.

In this paper, we describe the development of a technology-assisted 5As weight management counseling intervention. Commonly described approaches to intervention development and testing, and especially ones that incorporate behavior change and/or technology, include: intervention mapping [62‐64], adaptive intervention design [65], person-based or user-centered approach [66‐69], requirement development approach [70], behavior change wheel approach [71], participatory approach, EVOLVE model, ORBIT model, and other variations of systematic and theory-based behavioral science and mixed-methods approaches [41, 72‐77]. What is often reiterated by authors utilizing these intervention approaches, and particularly the ORBIT model is the importance of both systematic planning and iterative design, guided by a theoretical basis or framework for behavior change, needs assessments or formative studies, and then continuous and integrated user input/feedback to identify and prioritize key intervention components.

Thus, our approaches to the development of the GEM intervention are well-supported in the literature. We designed the GEM intervention to address a crucial need to improve primary care-based weight management counseling, increase attendance to intensive programs such as MOVE!, and facilitate weight loss in patients who do not attend. We chose to use the 5As behavior change model as it has already been shown to be effective in smoking cessation counseling and is now seen as a promising framework for obesity counseling within primary care [34, 78]. We also wanted to focus on goal-setting as the key approach to weight management as it is an integral part of the 5As framework as well as a key element in several behavior change theories [24]. We used rigorous formative methods, combining perspectives from both Veteran patients and PACT staff to identify, prioritize, and develop initial intervention components prior to usability testing [42, 43]. We also used rigorous user-centered approaches, conducting two rounds of usability testing and evaluation, pilot testing of health coaching protocols, and open-ended exit interviews to guide iterative development of intervention components and incorporate crucial user feedback throughout the testing phase. We believe that these methods will increase the likelihood that our intervention will facilitate seamless delivery of 5As counseling integrated within the PCMH model of primary care at VA hospitals.

The GEM intervention has evidence-based components that have shown to be efficacious in other primary-care based weight management interventions. For instance, there have been three Practice-based Opportunities for Weight Reduction (POWER) studies, funded by the NIHLBI, which represent the most comprehensive primary care-based weight management studies to date [79‐81]. Together, the POWER studies showed that various primary care-based weight management interventions can produce clinically significant weight loss in primary care practice. The use of non-clinician support staff (e.g., health coaches, medical assistants, health educators) was an important component of all these interventions, and two incorporated technology. However, none of these studies were delivered within medical homes nor did they integrate with existing, intensive weight management programs at the respective sites. The GEM intervention is innovative because it leverages the PCMH model of primary care to improve implementation of the 5As framework through the use of technology-assisted, collaborative goal setting by medical home-based health coaches integrated within primary care teams.

Several methodological limitations should be acknowledged and addressed. For usability testing of the early prototypes of the GEM tool, different professionals (i.e. clinician vs. non-clinician) were involved in Round 1 and Round 2 of testing, which may have resulted in different outcomes related to health counseling. However, we purposely had a clinician (MJ, the PI) provide the health coaching in Round 1 to ensure that the health coaching protocol was relevant and feasible in a best-case scenario. We then had a non-clinician (KFM) trained in the protocols be the Health Coach in Round 2 in order to assess whether someone without formal clinical training could assist patients with using the tool and provide brief 5As counseling. Another limitation is that Veteran patients that previously participated in our formative research were invited back to participate in usability testing. Our intent in using this approach was to involve members of our target population in the prototype testing of our tool. Inviting back study participants from our formative research allowed us to highlight how their focus group feedback was used to inform design of our tool and intervention. Additionally, we recruited only 5 Veteran patients per round of usability testing. There is no universally accepted standard sample size for usability testing, but existing literature recommends considering quality of assessment over quantity, while keeping in mind operational limitations [51‐53]. Our iterative approach used two rounds of in-depth testing with members from our target population, combined with internal testing and development with a multidisciplinary team of clinicians, dietitians, other health professionals, software developers, and MOVE! program staff. As part of an on-going pilot RCT, we are further testing the feasibility of using non-clinician health coaches with more extensive training, as well as the acceptability of the intervention among a larger sample size of Veteran patients that have not involved in our formative or development research.

Another important limitation is that practices that do not use a PCMH model could potentially find this intervention difficult to implement, especially if they lack additional resources and programs to support weight management. However, multi-professional, team-based care is highly valued and increasingly being adopted [82]. We also acknowledge that much of the formative and development work to design the GEM intervention was conducted at a single VA site, and thus may not be directly translate to other settings and populations. However, a strength of the Manhattan VA is that it is one of the most diverse VA hospitals within the Veterans Affairs Healthcare System; compared to both national and New York state Veteran demographics, our sample population was more inclusive and representative of minority groups, specifically female, African American, and Hispanic Veterans [45]. Importantly, we utilized a multidisciplinary research team to design and test the GEM tool and intervention. Thus, although the GEM intervention was developed at a single site, we believe that the diverse training and perspectives of our team were key to developing a complex intervention that is widely applicable and has the potential to be implemented and easily tailored to other VA sites and other healthcare settings more broadly [83, 84].

We describe a systematic intervention development process that aligns with the ORBIT model of behavioral intervention development. The use of qualitative research methods and user-centered design approaches enabled us to quickly detect salient issues and make iterative changes to prototypes of the GEM tool, improve health coaching protocols, and strengthen the overall design, and will facilitate integration of the intervention into primary practice. We believe that through this rigorous process, the resulting intervention has a higher likelihood to improve delivery of weight-management care for patients within primary care settings. We are currently testing the proposed intervention within the clinic setting for feasibility and acceptability as part of a pilot randomized controlled trial (RCT). Future studies will test the efficacy of this intervention in other VA centers as well as public outpatient centers.