Discussion

The Women’s Health Initiative (WHI) Strong & Healthy trial demonstrated that a remotely-delivered, “light touch,” multicomponent physical activity intervention can attenuate the age-related decline in self-reported physical activity (PA) and increase in sitting time over 9 years of follow-up in a large U.S. cohort of women, aged 66–99 years (mean 79.7 years). While all PA measures decreased and sitting increased as participants aged in both trial arms, women assigned to the Intervention arms reported statistically significant (p < .001 for all) higher MET-hours per week of walking (3.2%), aerobic (3.4%) and total (3.7%) exercise, as well as strength training sessions per week (4.4%), and fewer hours per week of sitting (−1.5%), than the Comparison arm over the course of the trial. Acknowledging that most WHISH participants were doing a mix of walking and light intensity PA, and relatively little moderate-to-vigorous activity, the difference in total exercise of 0.58 MET-hr/wk (95% CI: 0.28, 0.88) between trial arms is roughly equivalent to a difference of 14 minutes per week of walking at a slow pace of 2.5 METS. Clearly, the magnitude of PA improvements was dwarfed by the secular decline in PA measures and increase in sitting in the aging cohort (Figure 1); nonetheless, the dearth of long-term randomized controlled trial data on physical activity in geriatric populations makes these findings unique and of considerable public health value.

Identifying strategies to enhance the effectiveness of a remotely delivered PA intervention to achieve greater increases in physical activity and decreases in sedentary behavior than achieved by the WHISH PA intervention has considerable scientific merit, particularly for testing hypotheses regarding health benefits of improving PA and reducing sitting time. In keeping with the pragmatic nature of the WHISH trial, PA-I participants were not required to return PA logs or action plans; however, annual PA-I surveys collected a wealth of information which will be evaluated, including whether participants used any of the WHISH PA self-tracking tools. Many reported using the annual WHISH Activity calendars, which provided space to enter active minutes and steps, and check whether a participant did upper-body and lower-body muscle strengthening and/or balance and flexibility exercises each day of the week. On the other hand, most PA-I participants did not access the website; therefore, few tried the web PA self-tracking tool. Even fewer participants attempted to use the WHISH IVR system, and therefore only a small percent used its goal-setting feature.

The IVR system had been inspired by the success of an automated telephone-linked computer system which delivered physical activity counseling in the Community Health Advice by Telephone (CHAT) trial to motivate adults aged 55 years and older (mean 61.6) to increase PA above a target of 150 minutes per week within 6 months and maintain it for another 6 months.39 We had demonstrated the feasibility and acceptability of a WHISH IVR-prototype as a means to deliver tailored physical activity counseling, designed around the Go4Life® educational materials26,34,35 in a 10-week pilot study of 30 WHI-ES2 participants before initiating the WHISH trial.40 Unfortunately, scheduling regular IVR contacts for over 23, 000 women was not feasible within the budget constraints of the pragmatic trial, rendering that technology suboptimal for the WHISH PA-I. The system was, however, repurposed to deliver monthly outbound motivational messages, which most participants appreciated, with only 2% requesting that the calls be stopped, but the mailings continued.

PA-I surveys also asked participants about their use of each of the intervention materials. A 2019 analysis of 18 080 WHISH PA-I participants found that, except the IVR system, engagement with each of the WHISH PA-I delivery channels, eg, targeted inserts, email (opened), email (clicked links), website (logging in), and website (tracking), was associated with more hours/week of physical activity.41 Consistently across channels, PA-I subgroups with higher engagement included younger age, a subgroup which also returned PA-I surveys at a much higher rate (eTable S3 in the Supplementary Materials), and higher levels of PA and physical function.41

PA-I surveys also asked about motivational drivers of physical activity, as did a postcard that was mailed to 19 598 PA-I participants with a simple prompt, “Think about the last time you didn’t feel like being active. How did you get yourself to move or get up?” Among the 4108 (20.9%) participants who returned the postcard, women who endorsed more strategies had more hours of PA and walking.38 In addition, strategic categories that correlated with more PA included focusing on the benefits and utilizing the surrounding environment to help motivate movement.38

Further systematic exploration of PA-I participants who particularly benefited from this resource-efficient light-touch intervention versus those who did not is warranted in each age group, given the heterogeneous composition of the older population. For example, the noticeable increase in walking in 2020, coinciding with the COVID-19 pandemic, and the steep increase in sitting time in women in both trial arms were particularly observable in participants who had reported the highest Pre-WHISH PF scores (Figure 3) and PA levels (eFigure S1 in the Supplementary Materials). Meanwhile, the WHISH PA-I seemed to have had virtually no effect on women in the lowest pre-WHISH PF and PA tertiles (Figure 3 and eFigure S1 in the Supplementary Materials), except for the modest increase in reported muscle strengthening sessions/week, which was also increased in the Intervention versus Comparison arms in all three age strata (Figure 2).

The design and conduct of the WHISH pragmatic trial differ from most previous PA trials first and foremost by not having actively recruited individuals who were interested, ready and willing to be randomized to a PA intervention or excluding individuals who were already meeting aerobic physical activity goals. The randomized consent design25 deployed in the WHISH trial aimed to simulate real-world implementation of a program that could be disseminated and implemented at a national level, ie, be generalizable to older women. Women assigned to the Comparison group, as well as Intervention women who opted out of receiving the PA-I materials, had access to the Go4LifeR® materials26,34 and national websites35,36 that PA-I participants were encouraged to use and may have been fully aware of national PA recommendations. Nonetheless, the “light touch” WHISH PA intervention resulted in significant, albeit quite modest, self-reported behavior changes in all PA measures and sitting time, particularly in women in younger age strata (66–76 and 77–82). Further investigation of scalable, resource-efficient methods for improving PA in the oldest old, ie, ≥ 85 years of age, seems particularly warranted as this group is one of the fastest growing segments of the U.S. population.42 Additionally, and speculatively, targeting adults aged 65–79 years, or even younger, might help them maintain PA behaviors into their oldest old years. While the “oldest old” are overrepresented in the WHISH subgroups with low initial PF and PA levels, the younger women with low PF and low PA levels are clearly another group to target for investigation of effective, cost-efficient strategies to increase PA within the community setting.

Strengths of the WHISH trial include the large number of women aged 66–99 years and the long-term “light touch” intervention, which resembles many booster strategies to promote maintenance of physical activity,14 and the many motivational principles that were promoted. As WHI participants, WHISH participants have been followed since 1993–1998 during which their physical activity, diet, and other lifestyle habits, as well as health outcomes, have been well documented. They currently reside in every state in the United States and are reasonably representative of the racial and ethnic composition of U.S. women of this age.43 They will also continue to be followed for several more years for physical activity and health outcomes.

A major weakness is that physical activity and sedentary behaviors are based on self-report, and the Intervention participants were not blinded to their randomization assignment, as is true for almost all lifestyle intervention studies. An intervention weakness is that we were not able to tailor it to the individual needs of the participants. Initially, only 25% of PA-I participants even used email, and although this increased to about a third as the trial continued, largely due to deaths of the oldest participants, the WHISH PA-I cohort was not, in general, tech savvy. Future cohorts of older adults already familiar with email, text, Zoom, and other communication channels, including IVR systems, will likely engage more readily with these channels, including IVR systems, making it easier to tailor remotely-delivered PA interventions, even in a pragmatic trial design.

Whether the modest changes in PA and sitting translate into clinically meaningful health benefits is currently being investigated and will be the focus of additional analyses. We will also analyze the vast amount of data collected from the PA-I surveys and explore factors associated with greater PA changes with aging. Considering the diversity of the WHISH cohort, we hope to gain insights into how to customize remotely-delivered PA interventions to different segments of this growing and diverse population for whom maintaining the ability to move is critical to enhancing the healthspan of older adults and preserving their independence. Building on the pragmatic WHISH physical activity intervention to augment behavioral changes, especially in the oldest age groups, is an important direction for population health.