Women’s Health Initiative Strong and Healthy Pragmatic Physical Activity Intervention Trial for Cardiovascular Disease Prevention: Design and Baseline Characteristics

Abstract Background National guidelines promote physical activity to prevent cardiovascular disease (CVD), yet no randomized controlled trial has tested whether physical activity reduces CVD. Methods The Women’s Health Initiative (WHI) Strong and Healthy (WHISH) pragmatic trial used a randomized consent design to assign women for whom cardiovascular outcomes were available through WHI data collection (N = 18 985) or linkage to the Centers for Medicare and Medicaid Services (N30 346), to a physical activity intervention or “usual activity” comparison, stratified by ages 68–99 years (in tertiles), U.S. geographic region, and outcomes data source. Women assigned to the intervention could “opt out” after receiving initial physical activity materials. Intervention materials applied evidence-based behavioral science principles to promote current national recommendations for older Americans. The intervention was adapted to participant input regarding preferences, resources, barriers, and motivational drivers and was targeted for 3 categories of women at lower, middle, or higher levels of self-reported physical functioning and physical activity. Physical activity was assessed in both arms through annual questionnaires. The primary outcome is major cardiovascular events, specifically myocardial infarction, stroke, or CVD death; primary safety outcomes are hip fracture and non-CVD death. The trial is monitored annually by an independent Data Safety and Monitoring Board. Final analyses will be based on intention to treat in all randomized participants, regardless of intervention engagement. Results The 49 331 randomized participants had a mean baseline age of 79.7 years; 84.3% were White, 9.2% Black, 3.3% Hispanic, 1.9% Asian/Pacific Islander, 0.3% Native American, and 1% were of unknown race/ethnicity. The mean baseline RAND-36 physical function score was 71.6 (± 25.2 SD). There were no differences between Intervention (N = 24 657) and Control (N = 24 674) at baseline for age, race/ethnicity, current smoking (2.5%), use of blood pressure or lipid-lowering medications, body mass index, physical function, physical activity, or prior CVD (10.1%). Conclusion The WHISH trial is rigorously testing whether a physical activity intervention reduces major CV events in a large, diverse cohort of older women. Clinical Trials Registration Number: NCT02425345

National guidelines promote physical activity and limiting sedentary behavior to prevent cardiovascular disease (CVD) and other chronic diseases in older adults (1,2). The U.S. population aged 65 years and older is projected to reach 95 million (23% of the population) by the year 2060, with persons aged 85 years and older reaching 19 million, the majority of whom will be women (3). Decades of evidence-based interventions have shown that increasing physical activity can improve levels of physical function and cardiovascular (CV) risk factors (4-7), but these trials have mostly enrolled subjects less than 75 years old, and have mostly tested supervised, clinic-based exercise programs. There have been no randomized controlled trials with physical activity as the sole intervention that tested the hypothesis that increasing physical activity and/or reducing sedentary time reduces the incidence of major CV events. While the LIFE Trial showed that a physical activity intervention modestly preserved mobility in older adults with low physical function (8), a pragmatic centralized physical activity intervention delivered to aging adults across the functional continuum has not yet been tested. We designed the WHI Strong & Healthy (WHISH) trial as a large-scale, randomized, controlled, pragmatic trial to fill this critical evidence gap. The WHISH trial is rigorously testing whether a physical activity intervention designed to deliver national physical activity recommendations for older adults reduces major CV events in a large, geographically diverse, multiethnic cohort of older women.

Objectives and Outcomes
The primary hypothesis being tested is that a centralized, public health behavioral intervention designed to increase or maintain physical activity levels and reduce sedentary behavior will reduce the incidence of major CV clinical events, specifically myocardial infarction (MI), stroke, or CVD death, in older women. Secondary hypotheses are that the WHISH physical activity intervention will lower rates of venous thromboembolism, peripheral artery disease, and reduced physical functioning.
The primary safety aims are to evaluate whether the WHISH intervention increases risks of total clinical fracture, hip fracture, falls, or non-CVD mortality over ~8 years, compared with the usual follow-up (control). A secondary safety aim is to evaluate whether the physical activity intervention increases coronary artery revascularization (coronary artery bypass graft surgery or percutaneous coronary intervention).
The ultimate goal of the WHISH trial is to provide definitive evidence on the benefits and risks of this pragmatic physical activity intervention and, if efficacious, to disseminate an easily scalable intervention to improve the CV health of aging Americans.

Randomized Consent Design
The WHISH trial uses the randomized consent design proposed by Zelen (9) (Figure 1), in which eligible participants are randomized before being contacted and before informed consent is obtained. The design is based on the intention-to-treat principle; outcomes are assessed on the entire randomized population, regardless of their subsequent level of participation. This design is appropriate when the goal is to test the impact of a public health intervention on the population at large, not just within the subgroup of the population that is willing to participate, and tests the most critical research question: "Can a centrally delivered public health physical activity intervention reduce CVD events in a large population of older women?"

Study Population and Eligibility Criteria
The WHISH trial is embedded in the WHI Extension Study (WHI-ES), which is continuing to collect health outcomes and related data on women who were originally enrolled in the WHI Clinical Trial Program or the Observational Study at 40 U.S. clinical sites from 1993 to 2005, and which is expected to continue through at least 2027. Women in the WHI-ES consented to continued participation and extended follow-up in 2004-2005 and again in 2010. Earlier phases of the WHI program have been described in detail previously (10)(11)(12)(13)(14)(15)(16)(17)(18).
Participants in the WHI-ES were eligible for WHISH if they were alive in 2015, and free of conditions that would limit their full participation, specifically known dementia, living in a nursing home, inability to walk, inability to read English, or unavailability of follow-up CVD outcomes data.
The WHISH Data Coordinating Center (DCC) used the WHI database to screen women for eligibility based on previously collected WHI data. Of the original 93 567 women who consented to ongoing follow-up in the WHI Extension Study in 2010, 10 252 were known to have died before the date of randomization into WHISH, and 8533 were excluded due to known dementia (n = 6043), living in a nursing home (n = 2650), selfreported inability to walk (n = 934), or Spanish language preference (n = 269). An additional 3311 participants had been lost to follow up due to inadequate contact information such as invalid addresses ( Figure 2).
Availability of adjudicated CVD outcomes differed for 2 categories of WHI-ES participants. The "Medical Records Cohort (MRC)" is a subcohort in which CVD outcomes continued to be verified by review of medical records and physicianadjudication. The MRC cohort includes women who had participated in the WHI Hormone Trials, or were an African American or Hispanic participant in the Diet Modification Trial or Observational Study. All MRC cohort women were eligible for WHISH if they met other eligibility criteria as detailed above. The remaining WHI-ES women are part of the "Self-reported Outcomes Cohort (SRC)" and eligible for WHISH if they met the criteria above and their follow-up hospitalization data were available through linkage to the Centers for Medicare and Medicaid Services (CMS) database. Previous validation studies showed that CMS diagnosis codes had good agreement with WHI MRC adjudicated outcomes for MI (κ = 0.74) (19), stroke (κ = 0.84) (20), and other CVD outcomes (21,22). Unavailability of CMS follow-up for major CVD events excluded 21 871 SRC participants, leaving a total of 49 331 participants eligible for randomization ( Figure 2).
The WHISH protocol is registered on clinical trials.gov (23) and was approved by Human Subjects Review committees at the Fred Hutchinson Cancer Research Center, Stanford University, and the University of California, San Diego.

Randomization and Initial Contact
Before WHISH randomization, all WHI-ES participants were mailed a newsletter describing new studies underway for eligible WHI participants. WHISH was described as a study testing whether increasing physical activity would lower the risk of heart disease and help maintain independent living. Eligible WHI-ES participants were randomized using the Zelen design in equal numbers either to the WHISH physical activity intervention or to continue usual follow-up. For practical reasons, participants were randomized in 3 waves about 1 month apart. Randomization was carried out in 24 strata formed by baseline age in tertiles (<77, 77-82, and ≥83 years), 4 WHI regions, and the method of outcomes collection (MRC or CMS Medicare hospitalization data). Among WHI-ES participants who met other eligibility criteria, 16 251 were in the MRC and 30 346 had sufficient Medicare data to be included in WHISH. An additional 2734 SRC women had outcomes adjudicated by the WHI for other reasons, yielding a total of 49 331 women randomized into the WHISH Trial.

Consent
Participants in the WHI-ES had all previously consented to follow up and collection of outcomes data. Women randomized in WHISH to usual follow-up received no further information about the trial. Women randomized to the physical activity intervention were sent an initial mailing from the WHISH DCC that included intervention materials, a Go4Life Workout to Go: Mini Exercise Guide (24), and a telephone number to call if they preferred to opt out of receiving additional physical activity materials. After removing 870 (3.5%) women who "opted out," 53 who had undeliverable addresses and 81 who had died after randomization but before the intervention began, Stanford received names and contact information for 23 653 participants for the WHISH physical activity intervention.
Intervention participants received written information meeting HIPAA requirements in the fourth Stanford mailing (November 2015), and were provided a Stanford telephone number, postal address, and e-mail addresses to use if the participant did not consent to the use of data collected from surveys and postcards, data entered by participants into a WHISH website tracking tool, recorded through the WHISH automated telephone-based interactive voice response system, or returned as handwritten entries on annual WHISH calendars, or input received by telephone, e-mail, or other means. Intervention participants were reminded that they could request no further contact at any time in the future, thereby withdrawing passive consent, or request discontinuation of any specific component of the intervention, including but not limited to quarterly mailings, outbound monthly telephone messages, or emails.

WHISH Physical Activity Intervention
The WHISH physical activity intervention goals are based on the U.S. Department of Health and Human Services 2008 Physical Activity Guidelines for older Americans (1), as presented in the National Institute on Aging (NIA) Go4Life health education campaign (25), and the updated 2018 Physical Activity Guidelines for older Americans (2). These federal initiatives emphasize increasing or maintaining aerobic physical activity (primarily walking) and decreasing sedentary behavior (especially sitting), as well as multicomponent physical activity recommendations regarding muscle-strengthening, balance, and flexibility. Based on state-of-thescience behavioral theories, including Social Cognitive Theory (26), the Transtheoretical Model of behavior change (27), and Stages of (Readiness of) Change (28), the intervention is delivered through multiple channels including quarterly (seasonal) WHISHful Actions newsletters, with inserts targeted at 3 participant groups based on lower, middle and higher levels of self-reported physical functioning and physical activity levels; monthly outbound telephone calls with short (~1 minute) motivational messages; monthly motivational e-mails, for the approximately one-fourth of participants who provided email addresses; access to the WHISH website which includes videos of participants demonstrating exercises as well as many additional resources; and occasional personal contact with intervention staff by phone, e-mail, or regular mail as required or requested. The physical activity intervention adapts to participant input received from annual surveys or other inbound communication channels, regarding physical activity preferences, resources, barriers, and motivational drivers (29)(30)(31).
In the first year of the intervention, all physical activity participants received the Go4Life Exercise & Physical Activity: Your Everyday Guide (32), pedometers and calendars with instructions on how to track their physical activity, and resistance bands with examples of exercises using the bands for muscle-strengthening. In subsequent years, participants have received annual calendars, new pedometers and resistance bands, targeted inserts demonstrating various strength, balance, flexibility and endurance exercises, a WHISH sun visor, and chances to win gift cards by participating in special WHISH challenges, and they may request replacements of these materials as needed. Further specific details of the intervention program will be presented in a subsequent manuscript.

Outcomes Ascertainment
The primary outcome of the WHISH trial is major CV events, defined as the first event since randomization in WHISH of either MI, stroke, or CV Death. The primary safety outcomes include non-CV death, hip fracture, other clinical (non-hip) fractures, and falls. Secondary outcomes are venous thromboembolism, peripheral artery disease, physical functioning as determined by the Rand-36 10-item score (33), and coronary revascularization (coronary artery bypass graft or percutaneous coronary intervention) as the secondary safety outcome. Additional outcomes, including cancer, heart failure, and a wide array of clinical diagnoses, are ascertained on all participants through CMS linkage and annual WHI surveillance (34). Two ancillary studies to the WHISH Trial have been funded specifically to adjudicate heart failure and atrial fibrillation endpoints (R01HL130591, CB Eaton, Principal Investigator, and R01HL136390, MV Perez, Principal Investigator).
Details of the WHI outcomes ascertainment and adjudication procedures have been published (34,35) with modifications presented in the WHI-ES Protocol (36). Briefly, WHI-ES participants are followed annually, primarily by mailed questionnaires, to collect data on self-reported health events and related conditions. Mailings include a Form 33 -Medical History Update (Supplementary Appendix A) and additional questionnaires (see below), which most participants (or their proxies) return by mail. WHI staff may collect Form 33 by phone for women who have not returned it after repeat mailings.
Outcomes ascertainment procedures for WHISH depend on the outcome and whether the participant is in the MRC or in Medicare. These procedures use a combination of self-report and adjudicated events, as well as National Death Index and CMS linkages to identify outcome events (Table 1).
An overlap of 8431 participants in the MRC with CMS linkage allows a comparison of outcomes from the 2 sources. Highly concordant results reported from these sources in WHI for MI (19), stroke (20), peripheral vascular events (22), and venous thromboembolism (21) suggest that it is possible to combine adjudicated and CMS data to monitor outcomes in the WHISH trial.

Physical Functioning Assessment
Physical functioning (PF) is based on participant responses to questions in an annual WHI form (Form 151 Activities of Daily Living, Supplementary Appendix B), which is mailed with Form 33, that ask about limitations (no, not limited at all; yes, limited a little; yes, limited a lot) regarding 10 specific activities (Q 7-16) from RAND-36 (33). The PF score ranges from 0 to 100 (in 5-point increments), with 100 meaning no limitations on any activities.

Physical Activity Assessment
Physical Activity assessment is based on both a self-reported physical activity questionnaire (Form 521 Physical Activity Questionnaire, Supplementary Appendix C), which is mailed annually to all WHISH participants with Form 33 and labeled as a WHI form, to maintain blinding and reduce response bias, and through periodic accelerometry in a subsample of WHISH participants (see below). Form 521 ascertains walking, sedentary time, sleeping, details about falls, and an abbreviated CHAMPS (37,38) questionnaire. Whereas WHISH investigators are blinded regarding clinical outcomes, with the exception of Dr.Kooperberg, (see Monitoring), trial investigators regularly review the comparisons of self-reported physical activity data as a process measure for the intervention materials and adapt the intervention as appropriate.
A previous WHI ancillary study, the Objective Physical Activity and Cardiovascular Health (OPACH) Study (R01 HL105065, AZ Self-report Self-report LaCroix, PI) provided an opportunity to conduct a substudy of accelerometry within a subset of 4730 WHISH participants who had successfully worn and returned Actigraph GT3X+ accelerometers in 2012-2014, thereby providing pre-WHISH Trial data for comparison between randomized groups (39). Surviving WHISH-OPACH participants who had returned usable accelerometer data were invited to join and actively consent to participation in the "WHI Physical Activity Follow-up Study" with no reference to the WHISH trial. To achieve balanced groups of at least 1000 women in both the WHISH arms throughout the planned follow-up period, the recruitment target was set at 2300; 2349 women consented.

Trial Monitoring
An NHLBI-appointed Data Safety and Monitoring Board (DSMB) meets annually and receives updates on the physical activity intervention, including measurements of participant engagement, and monitors self-reported and accelerometer-measured physical activity and clinical outcomes data.
In early discussions, the DSMB determined that the trial would not have formal monitoring boundaries for stopping early for futility with respect to the CVD outcomes in recognition of the potential value of increasing or maintaining physical activity on multiple health conditions. The DSMB also suggested that the trial not have monitoring boundaries for stopping early for CVD efficacy because the benefit of getting a definitive answer will likely outweigh any benefit of making a less clear result public. This decision was justified based on several considerations: (i) the need for a broader assessment of the benefits and risks of increasing physical activity in older women; (ii) a strong anticipation of CVD benefits from prior studies of physical activity (4,5); and (iii) the availability to the general public including the WHISH usual follow-up comparison group of the NIA Go4Life materials (24,25,32) which were provided to the physical activity intervention group. Thus, trial monitoring for WHISH focuses on 2 main purposes: assuring the safety of participants; and testing the primary and secondary hypotheses adequately. WHISH could be stopped if the intervention group does not appear to be more physically active or less sedentary than the comparison group, which could result in a noninformative test. In addition, the NHLBI has requested an interim monitoring for benefit and futility in Spring 2022.
The primary outcome of CVD is also monitored as a (severe) safety outcome and together with non-CV death and hip fracture is evaluated annually, that is, each of these 3 outcomes are monitored at each annual interim meeting using a Z-value of 2.42. Based on simulations, the cumulative probability that the Z-statistic for one of these outcomes will exceed 2.42 at any meeting, if in truth there is no effect, is approximately 5%, assuming that these 3 outcomes are independent. The primary safety aims, total clinical fracture, hip fracture, and falls are also monitored, whereas less severe outcomes, including coronary revascularization, falls, and other clinical fractures, are provided regularly but not formally monitored because the clinical implications for these outcomes may or may not be apparent at the time they are observed. For example, increased physical activity in the intervention group may uncover symptoms of angina or claudication, which could lead to revascularization procedures which are intended to reduce the rate of subsequent, more serious events such as MI, and hence do not carry the same adverse safety signal as other events being monitored.

Duration of Follow-up and Statistical Considerations
The WHISH trial was funded for an initial 5 years (2015-2020). Renewal funding will extend the follow-up to approximately 9 years (to September 2024). Power calculations for the renewal of the WHISH trial were based on the observed effects of the WHISH physical activity intervention, the observed event rates for the intervention and comparison arms combined, and the observed relation between physical activity and CVD from the OPACH study. The OPACH study was also carried out in WHI participants and should, thus, give us the best possible effect estimates. As of February 1, 2020, WHISH had 4 years of follow-up, the renewal would extend that follow-up to 8 years. The current difference in MET-hours/week walking, as reported by the complete cohort, is over 8.5%, and has been increasing during the study. The currently observed event rates as a function of age for the primary WHISH outcome (CHD, stroke, CVD death), and the "censoring outcome" (death from other causes) is estimated from the observed data, combining both arms. The observational OPACH study obtained a hazard ratio of 0.79 for the WHISH CVD outcome, for a 1-hour increment of light physical activity which is comparable to nonpurposeful (daily life) walking in the age range of women studied in WHISH. In OPACH, the average light physical activity was 4.76 hours/day, so an 8.5% intervention group difference is equivalent to 0.4 hours/day, corresponding to a 9% decreased rate of the primary CVD outcome. Using the age distribution of WHISH participants and the observed event and censoring rates, we obtain the power estimates shown in Figure 3. We note that with an intervention effect of 9% we have over 90% power, and with an effect of 8% we have close to 85% power after 8 years.
For each of the primary, safety, and secondary outcomes, intervention effect sizes, confidence intervals, and Z-statistics are based  on Cox-proportional hazards models, with the exceptions of falls, which are compared using Poisson regression, and physical function, for which the mean difference between the groups is compared using linear regression. All analyses are stratified by categories applied to the randomization of participants, including age on April 1, 2015 (in tertiles), WHI region, and outcomes data source (MRC vs CMS fee-for-service). Using standard risk sets in survival analysis, participants who change from fee-for-service to managed care will have their self-reports adjudicated after the switch and are moved to the MRC stratum at the time of the switch in insurance coverage. Similarly, outcomes for participants who have 2 types of data, but do not return their self-report Form 33, will be assessed using CMS data after their last returned Form 33. All analyses will be carried out using an intention-to-treat analysis. Participants randomized to the intervention who "opted out" or requested that the intervention be stopped as the trial progressed will be included in all analyses, as assigned, for monitoring and primary reporting purposes. Deaths that occurred between randomization and the start of intervention are also included in the analyses.

Baseline Characteristics of Randomized Participants
WHISH Trial Baseline characteristics for women assigned to the Intervention (N = 24 657) versus Control (N = 24 674) were well balanced by randomized intervention group ( Table 2). The average age of women was 80 years with 9% Black or African American (n = 4514), 3.3% Hispanic/Latina (n = 1628), 2% Asian (n = 923), and 84% non-Hispanic White. Average body mass index based on the most recent WHI clinic visit (no later than 2005) was 28 kg/m 2 , with 30% of all WHISH participants in the obese range, ≥30 kg/m 2 . Use of blood pressure medications at WHI baseline was common (61%), 44% were taking lipid medication, 17% self-reported a history of having diabetes requiring treatment, and 10.2% had history of CVD. At WHISH baseline (2015), about 20% of women reported being limited in walking one block a little or a lot and 10% had a previous history of CVD defined as MI, coronary artery bypass graft/percutaneous coronary intervention, or stroke. Baseline characteristics varied by baseline age tertiles for most variables including race/ethnic distribution (Table 3).

Discussion
The WHISH trial is the first randomized, controlled trial to test whether an intervention designed specifically to increase physical activity and reduce sedentary behavior, independent of diet, weight loss, or other interventions, will reduce major CVD outcomes. As it is specifically testing this hypothesis in older women, additional aims focus on determining whether the intervention benefits physical function and other issues of particular importance to older adults, including maintaining independence and mobility, and its safety with respect to non-CVD mortality and falls and fractures. By recognizing the opportunity to utilize a randomized consent design (9) in the large multiethnic and well-characterized WHI cohort of older women, the WHISH trial meets the challenge posed by NHLBI leadership to design highly efficient pragmatic clinical trials, embedded within ongoing epidemiologic cohort studies, to test hypotheses of major public health significance (40,41). Utilization of materials available through the National Institute of Aging Go4Life program and website (24,25,32) to deliver an intervention based on accepted physical activity recommendations for all older Americans (1,2,42) adds to the efficiency and potential scalability of the WHISH trial program being tested.
The consistent evidence to date that physical activity reduces CVD mortality and incidence, including coronary heart disease and stroke, is based on a large body of observational data studying these endpoints, as well as randomized controlled trials focusing on CV risk factors rather than clinical outcomes (4,5). It is, of course, recognized that physical activity has many health benefits, with sufficient rigorous evidence irrespective of CVD reduction to recommend physical activity as part of a broader public health framework (1,2). However, because CVD deaths in women aged 80 years and older account for two thirds of all CVD deaths in U.S. women, experimental evidence that physical activity reduces CVD in this population would have extraordinary public health significance.
The randomized consent design (9) in the WHISH trial removes a serious bias of the majority of previous physical activity trials which attract and recruit individuals whose willingness to increase activity may be directly related to an inherent physiology that yields greater benefit than what the general population is likely to experience. Many such trials have excluded individuals who are already active but might benefit from even further physical activity change    or reduction in time spent sedentary. While some portion of WHI-ES participants were initially attracted to the WHI hormone therapy or diet modification trials, and were likely more health-oriented than the general population of women their ages, use of the randomized consent design basically ensures that none of the WHISH participants joined because of their physical activity habits. Another novel design element that increased the efficiency of this pragmatic trial was the "opt out," passive consent process which resulted in >96% of women randomly assigned to the intervention receiving the intervention.
The WHISH Trial accelerometry substudy took advantage of the fact that a large number of WHISH participants had participated in the OPACH study and had a first accelerometer wear in 2012-2014 before WHISH randomization. This is another example of a highly efficient study design leading to a truly rare sequence of 4 longitudinal accelerometer measures to evaluate trajectories of physical activity over time by WHISH randomization group.
In conclusion, the pragmatic WHISH trial is an innovative physical activity trial with the potential to provide strong evidence of physical activity as a sole modality to reduce CVD in older adults. If this easily scalable intervention proves to be safe and effective in improving cardiovascular health and other outcomes of importance to older women, the WHISH physical activity intervention can be widely disseminated, as well as adapted to meet the needs of many additional segments of the U.S. population.

Supplementary Material
Supplementary data are available at The Journals of Gerontology, Series A: Biological Sciences and Medical Sciences online.