We selected participants using dissimilar case sampling (16), in which cases are purposively selected to maximize differences for a variable of interest. NCP identified MOVE! programs with high and low retention, based on the average number of visits per patient enrolled in the group program. After stratifying by geographic location (urban vs rural), and facility complexity level (a category assigned by VHA to reflect levels of patient risk, teaching and/or research, and patient volume, and assessed as high, medium, or low) (17), we selected 20 programs, 10 with high participant retention and 10 with low retention. We determined through discussion with local MOVE! staff that coordinators would be the most knowledgeable about the group program; therefore, we contacted program coordinators (n = 20) via e-mail and invited them to participate in an interview about their program. Eleven coordinators (10 dietitians and 1 registered nurse) agreed to participate (response rate, 55.0%), representing 12 MOVE! programs. (One participant was coordinator at 2 facilities, both with low retention. Because the purpose of this was to explore program differences related to retention, and these 2 programs had key differences in implementation, these programs were analyzed separately.)

Our study was reviewed and approved by the authors’ local VHA institutional review board. Coordinators were informed that the purpose of the interview was to gather information about their MOVE! program to identify factors related to patient retention, and to discuss any efforts to improve retention at their facility. The first author (S.M.L.), an experienced qualitative researcher, conducted telephone interviews in 2010 using a semistructured interview guide (Appendix). Three members of the study team developed the guide; it was then reviewed for clarity and completeness by local MOVE! program staff. Interviews lasted approximately 30 minutes and were recorded and transcribed verbatim.

Two research team members developed categories of program stages (referral, orientation, group meetings, and postprogram follow-up) based on earlier research describing patient assessment, referral, program attendance (individual, group, or intensive), and optional follow-up (18). Additionally, coordinators were asked to discuss several topics relevant to MOVE! overall; these responses were categorized as general MOVE! issues. We used descriptive coding techniques (19) to analyze content and used NVivo 8 qualitative data analysis software (QSR International, Doncaster, Victoria, Australia) to code responses in program stages. Members of the study team met throughout coding and analysis to discuss the codebook and the coding and categorization of coordinator statements. After coding, 3 members of the research team compared frequencies and responses between high- and low-retention programs to determine which characteristics differentiated them.

Referral method was not related to differences in retention (Table 1). In 10 programs (5 low- and 5 high-retention), patients were referred to MOVE! by a primary or specialty care provider; providers were reminded through a clinical alert to offer MOVE! to patients who had a body mass index (BMI) of 25 kg/m² or higher and, if the patient was interested, request a MOVE! consultation. In 5 programs (3 low- and 2 high-retention), patients were also self-referred; self-referral was the primary referral method in 2 programs, both with low retention.

Although referral method appeared unrelated to retention rate, provider support for and knowledge about the program differed between high- and low-retention programs. Coordinators at 4 low-retention programs said patients attended orientation without knowing why they were there or were often referred by providers who offered little discussion of MOVE!: “[The patients] don’t know they’ve been referred . . . they don’t know anything about it . . . I . . . think maybe the doctors just say, ‘Do you want to lose weight?’ and the patient says yes. . . . Maybe the doctors just don’t know what MOVE! is” (low-retention program). Only 1 coordinator of a high-retention program noted any such problems: “I would like to see a more consistent understanding among providers of what [MOVE!] is.”

Program reputation may also be related to retention. One coordinator from a high-retention program discussed a positive reputation: “Our program has a good reputation here. We . . . always have plenty of recruits, plenty of veterans referred to us.” On the other hand, a coordinator from a low-retention program discussed a negative reputation: “MOVE! has gotten a lot of negative attention in the past. . . . The doctors got bad feedback.”

After referral, participants in 10 programs (5 low- and 5-high-retention) attended an orientation session. Five programs held orientation sessions individually (1 low- and 4 high-retention) and 5 as a group (4 low- and 1 high-retention). One facility that allowed walk-ins without referral at their orientation session had low retention. Overall, use of orientation sessions was unrelated to differences in retention.

After orientation, patients in all programs could select self-management support by telephone, Internet, or group meetings. Coordinators said veterans were encouraged to enroll in group meetings.

Retention was unrelated to whether veterans could join mid-cycle (Table 2). Patients in 6 programs (3 low- and 3 high-retention) began attending classes immediately after orientation; in the remaining 6 programs (4 low- and 2 high-retention), patients waited until a new cycle began: “I have them wait until the beginning. . . . It’s easier to remember where they started . . . [and they develop] camaraderie with the other participants” (low-retention program).

Program retention also was unrelated to frequency of meetings; 5 low- and 5 high-retention programs offered classes once per week or every other week. The number of classes offered in a cycle also did not relate to retention; 9 programs (5 low- and 4 high-retention) offered 6 or more sessions in a cycle. On the other hand, retention rate differed by flexible scheduling; 2 high-retention programs offered more than 1 meeting time, to accommodate different schedules, whereas no low-retention programs did so.

When planning the program and determining which sessions to hold, all coordinators used handouts and activities from the MOVE! website; only 2 coordinators, both of high-retention programs, used complete lesson plans provided by NCP. Most coordinators instead supplemented MOVE! materials with presentations by guest speakers, the coordinators’ professional knowledge, and materials from other weight-management programs.

All programs offered sessions on diet and nutrition and psychology; for 11 programs, diet and nutrition topics made up more than half of program content. Eight programs (6 low- and 2 high-retention) used lecture primarily; 4 programs (1 low- and 3 high-retention) were interactive, using group discussion and activities. Whereas 9 programs discussed physical activity in class, high-retention programs were more likely to incorporate physical activities into class sessions; 2 low- and 4 high-retention programs included physical activities, such as walking, chair exercises, and balance/stretching in class.

All coordinators identified retention as a major issue in the group program. Availability of resources, such as space, was a barrier for both high- and low-retention programs: “Our classroom right now, if I would have everyone show up . . . it would be extremely crowded. I kind of bank on some people not showing up” (high-retention program). “[The veterans] don’t like where we are. It’s a small area. . . . There’s nowhere else. . . . They haven’t allotted us space” (low-retention program).

Another barrier to improving retention was patient mix. Coordinators of 2 high- and 5 low-retention programs noted poorer attendance and participation among women: “I notice that some women don’t like to be in a room with 12 men. That makes them uncomfortable” (high-retention program). “The women have different issues than the men. . . . They would probably feel more comfortable with a group of their own” (low-retention program).

Two programs (1 low- and 1 high-retention) created a separate MOVE! program for women through collaboration with the Women’s Health Service: “We didn’t have a lot of women participating in the group classes, and when we did, their attendance was sporadic. . . . [Participants] really enjoy just having a women’s-only class and not having to deal with any pressures [with discussing certain issues] in front of a man” (low-retention program).

Despite resource limitations, coordinators discussed strategies for improving retention. Coordinators from high-retention programs were more likely to discuss active efforts to improve retention (5 high- vs 1 low-retention), including participant contracts addressing active participation and attendance and team competitions and challenges. Coordinators of low-retention programs, on the other hand, often discussed poor retention as inevitable: “I know that my retention rate is going to drop off by about 60% to 80% after the first week.” “I tell [patients] in their first meeting . . . ‘Look around, because half of you won’t make it to class 3.’ They just don’t come.”

At the end of the program cycle, participants were presented with follow-up options. One option that all programs allowed is re-attending classes. Re-attendance was subtly discouraged in 3 programs (2 low- and 1 high-retention) because of space constraints or observations of poor engagement by returning participants: “After they’ve been through [MOVE!] a time or 2, it’s like a social event for them. They come and visit and they’re not very active with trying to lose weight” (high-retention program).

Ten programs (5 low- and 5 high-retention) had a support group for past participants. Support groups met weekly or monthly; topics were determined by what was timely (eg, dieting during the holidays) or of interest to participants. Four programs (2 low- and 2 high-retention) also used TeleMOVE! for follow-up. TeleMOVE!, a new home-based program, uses a telephone-connected device to collect and transmit information between the participant and their VHA facility (11): “Some of them, now that we have that program, are choosing to follow up that way. It’s a little more convenient for them, especially if they have a long travel distance” (high-retention program).

Some general infrastructure issues (program location, staff/leadership support) were related to program retention. Eight programs (3 low- and 5 high-retention) were located in the facility’s Nutrition and Food Services; 4 (all low-retention) were located in the facility’s Primary Care department. Additionally, 7 programs (5 low- and 2 high-retention) lacked a physician champion or had one in name only. Only 2 low-retention programs had a physician champion who was involved in teaching classes. Although 2 high-retention programs also lacked a physician champion, coordinators at these facilities felt they had strong support from other sources, such as facility leadership: “Our primary care service line manager is very supportive. He’s all gung-ho for [MOVE!]” (high-retention program). “[Leadership has] been supportive of innovation. . . . I’ve had a lot of autonomy to manage the program” (high-retention program).

Although the job description specified the skills and attributes of the physician champion, coordinators felt the role of the physician champion was unclear, which was a barrier to greater involvement. Coordinators reported they had sufficient staff members to conduct classes but lacked personnel for other MOVE! activities, such as scheduling patients, informing facility staff about the program, and planning sessions: “I’m a one-man band here. I do all of the scheduling, all the canceling . . . the support groups, and there’s getting to be too many patients” (high-retention program). “[I would like] all those things that [MOVE!] is supposed to have . . . it’s supposed to be a team effort. . . . It just doesn’t happen. . . . I would like it to be the way they advertise it — a team” (low-retention program).

When coordinators were asked what patient characteristics they felt influenced retention, they listed age, health status, and personal investment or motivation. Driving distance was identified as a factor by coordinators of 6 programs (2 low- and 4 high-retention). These coordinators dealt with driving distance in 1 of 2 ways. One way, used by 2 high- and 2 low-retention programs, was to enroll patients in TeleMOVE!. Another method, used by 2 high- and 1 low-retention programs, was to offer MOVE! at affiliated community-based outpatient clinics (CBOCs): “[Our medical center is] not very centrally located . . . and so depending on . . . how far they have to travel, that’s probably the biggest [issue]. . . . We have the CBOCs involved, so they can travel to a CBOC and see the same thing” (high-retention program).