The recruitment strategy was designed to enroll 300 physically inactive and overweight or obese veterans who could safely participate in self-directed PA of at least moderate intensity on most days of the week. We sought to enroll veterans for whom PCPs could recommend unsupervised moderate-intensity PA as part of routine care without requiring additional safety testing.
To be included in the sampling frame for the study, the veteran: (1) had one or more primary care clinic visits at the VAPHS UD in the 12 months before enrollment, (2) was at least 18 years of age at the time of enrollment, and (3) was overweight or obese with a body mass index (BMI) ≥25 kg/m2 calculated based on the most recent height and weight recorded in the medical record. Because we sought a representative sample of the VAPHS veteran population, exclusion criteria were limited to factors that would limit safety, feasibility, or potential to benefit from the intervention.
Excluded from the sampling frame were veterans with existing International Classification of Diseases, ninth edition (ICD-9) codes for psychoses (codes 290 to 299), alcohol or drug dependence (codes 303 and 304), mental retardation (codes 317 to 319), unstable angina (code 411.1), pulmonary hypertension (codes 416.0, 416.8), spinal cord injuries (codes 806 and 952), and long-term oxygen therapy (code V46.2). Because the intervention advised brisk walking, veterans in wheelchairs (code V53.8) or who required assistive walking devices (for example, canes and walkers), and those who could not walk at least 120 yards unassisted were excluded from the study. Because this was a print-based intervention and available only in English, with materials distributed to veterans via US postal mail, participants were required to read English at a 7th grade level or above. We also excluded those who reported they already participated in moderate intensity activities for at least 60 minutes per week, lived in an institutional setting, planned to move out of the VAPHS service area in the next 12 months, were employees of VAPHS, expressed unwillingness to adhere to the study protocol, or were participating in other clinical studies that would be expected to have an impact on PA.
We established an electronic sampling frame using clinical data recorded in the Veterans Health Information Systems and Technology Architecture (VistA), the information system built around an electronic health record that is used throughout the Veterans Health Administration. VistA provides an integrated inpatient and outpatient electronic record that supports VA clinical and administrative functions. We abstracted from VistA the records of Veterans receiving primary care at VAPHS UD after 1 June 2009. Subject-specific data were used to identify potentially eligible veterans according to the criteria described above.
To identify PCPs to participate in the screening process, we engaged the Medical Directors of the Primary Care Clinics to introduce the principal investigator (PI) and research staff to PCPs during weekly meetings. At that time, a very brief overview of the study was provided and PCPs were told that our goal was to recruit veterans for whom they would be comfortable recommending unsupervised moderate-intensity PA, comparable to brisk walking, as part of routine care and without the need for further safety testing. The PCPs also were shown screening form I (SF-I; see Additional file 1), which they were asked to complete during the veteran’s clinical visit as described below. PCPs participating in the screening process were those who were informed about the research study, were instructed in completing the screening form, and agreed to screen their patients.
Research staff identified the sampling frame electronically from the VistA records with a database query for age, calculated BMI, primary care history, presence of exclusionary ICD-9 codes, and PCP willingness to participate. This was performed at the beginning of each week to identify specific appointments for the following 5 days. Those patients who visited VAPHS for the first time during the recruitment period or had changes to their BMI or ICD-9 codes were included in subsequent database queries.
Research staff generated the patient-specific SF-I the day before each potentially eligible veteran’s scheduled appointment. The SF-I was prepopulated automatically with the veteran’s preassigned study ID number and the name of the PCP. Research staff attached the SF-I and a one-page flier describing the study to a routinely provided list of the veteran’s medications that PCPs were to review during the appointment. Clinic staff put the forms in a receptacle next to the exam room to be retrieved by the PCP when entering the room. The SF-I was printed on neon-green paper to make it noticeable to clinic staff and PCPs.
The second part of the recruitment process involved PCP screening using the SF-I. The PCP was asked to respond to the question: 'Given your knowledge of this patient, can this Veteran SAFELY do progressive, unsupervised physical activity at home (regardless of his/her interest in doing so)? Participants may be asked to work up gradually to 30 minutes of moderate physical activity (e.g., brisk walking, swimming) on most days of the week’. Response categories were 'no/not sure’ or 'yes’, further specifying with or without restrictions (Additional file 1). Defining the restrictions was left to the discretion of the PCP, to streamline the form and encourage PCPs to complete it.
Next, the PCP gave the one-page flier to the veteran and asked permission for research study staff to contact him/her directly to discuss possible enrollment in VA-STRIDE ('yes/no’). Lastly, the PCP signed and returned the form to the patient to give it to the clinic clerical staff at checkout, regardless of the veteran’s eligibility or willingness to be contacted. A unit clerk then detached the SF-I form from the clinical paperwork and filed the form in a secure location for later retrieval by research staff. In the event they were too busy to screen a veteran, PCPs were given the option to return a blank SF-I form to the research staff. When this occurred, the unscreened veteran was returned to the sampling frame for consideration at a subsequent primary care appointment. If a participating PCP demonstrated low rates of recruitment or incorrect completion of SF-I, details of the screening approach were reinforced by the PI or research staff during a subsequent face-to-face discussion. To acclimate PCPs to this process, we capped the number of SF-Is generated to five patients per PCP per day in the initial recruitment period. As PCPs became familiar with the study screening process, the screening cap was lifted. Screening was halted occasionally to accommodate staff availability. When screenings were limited or halted, veterans not approached were retained in the sampling frame for consideration at a subsequent primary care visit. The veteran was also returned to the sampling frame if he or she was not seen in clinic (that is, cancellation or 'no show’), the SF-I was not returned (that is, the patient left the clinic without turning in the form), or the SF-I was returned blank (that is, the PCP did not complete the screening).
During the third part of the recruitment process, veterans who gave their permission were then contacted via telephone by research staff to administer screening form II (SF-II; see Additional file 2) to confirm information pertaining to study exclusion criteria and the veteran’s ability to safely participate in the intervention. Waivers of Health Insurance Portability and Accountability Act (HIPAA) authorization and documented informed consent for screening and completion of SF-II were granted by the VAPHS IRB to minimize disruptions to the general functioning of the primary care clinic and eliminate response burden for those patients who would not be eligible for the study.
A brief overview of the study was provided to patients prior to administration of the SF-II to inform them that participants in both study groups would have a series of 14 mailings to their homes, including 9 PA questionnaires mailed to the intervention group. Participants in the intervention group would receive follow-up feedback reports based on responses to the questionnaires and generated by the expert system. Length of participation would be 48 weeks (defined as 12 months), and monetary compensation for time and travel would be $25 for each of the three measurements visits (baseline, 6 months, and 12 months), and $15 for the one orientation counseling visit following the baseline visit.
After the SF-II was completed, veterans who remained eligible and had continued interest in the study were scheduled for a baseline appointment. Research staff obtained signed informed consent, followed by baseline measurements at the first study visit. Veterans were randomized to either the intervention or attention-control group and scheduled for an orientation counseling session pertinent to their assignment. The veterans were informed of their study group assignment only at the counseling session.
Measurements were completed at baseline, 6, and 12 months for the primary outcome (that is, amount of PA per week) and secondary outcomes (that is, fitness for PA; health-related quality of life; weight, body composition, blood pressure, inflammation, glucose metabolism, and serum lipids; and costs).