We conducted this field and qualitative study from July 19, 2024, to October 11, 2024, in collaboration with 1199 Service Employees International Union Training and Employment Fund (1199SEIU-TEF), a labor management fund of the 1199 Service Employees International Union (SEIU) United Healthcare Workers East, the largest health care union in the United States [22]. The 1199SEIU-TEF provides education and training to more than 55,000 HHAs who are employed by over 50 licensed and certified home care agencies in New York City, New York [23].

To be eligible, HHAs had to (1) be aged ≥18 years, (2) be English speaking, (3) own a smartphone (iOS or Android), and (4) be employed as an HHA in New York City. To recruit our intended sample of 25 to 30 study participants, 1199SEIU-TEF staff reached out to HHAs who had previously expressed interest in participating in research initiatives and invited them to take part in our study using standardized recruitment materials (email script and flyer). If potential participants were interested, staff referred them to the study team to be screened for eligibility. HHAs who were eligible for the study provided written consent. All aspects of this study took place in person at 1199SEIU-TEF headquarters in New York City.

Before conducting the field study, the research team met with participants to elicit their initial perspectives on activity tracking devices and the device form factor (wristband vs ring). From these conversations, we found that HHAs were enthusiastic about the potential benefits of activity trackers for increasing their health awareness. HHAs preferred wristband devices over rings as they felt that the wristband’s small screen would be easier to check discreetly during the workday than their phone and its form factor would not interfere with their daily tasks.

The field study involved giving HHAs Fitbit Charge 6 devices, which they were asked to wear for 28 days. Before providing the devices, we conducted a 1-hour onboarding session where participants were introduced to the research team; asked to respond to a few self-reported demographic questions (age, gender, race, Hispanic or Latinx ethnicity, educational level, and years of experience as an HHA); and guided through setting up the Fitbit device and mobile app, highlighting key features including activity and sleep tracking. After the participants created their Fitbit accounts, we walked them through the authorization process to allow the research team remote access to their data through the Fitbit application programming interface.

Participants were then asked to wear the Fitbit daily for 4 weeks, during which the research team contacted them weekly via their preferred method (SMS text message). Before the field study, we confirmed participants’ contact preferences (eg, SMS text message, email, or phone), and all preferred SMS text messages. Prior work has also supported this modality for communicating with HHAs [24].

SMS text message reminders prompted participants to wear the wristband, review their data, and contact the team with questions or issues. However, these reminders did not substantially affect engagement or acceptance. Most HHAs either ignored the messages or briefly acknowledged their receipt. In interviews, participants reported wearing the devices consistently because they valued the insights into their health and well-being, with SMS text messages serving primarily as a channel for troubleshooting or procedural clarifications (eg, how to access participant gift cards). Following the 28-day field study, we used the Fitbit application programming interface to collect participants’ data to create personalized visualizations showing each participant’s activities in relation to (1) the National Institutes of Health’s (NIH) recommended PA and sleep guidelines for adults and (2) their broader group of peers in the study.

Four investigators from the research team (IRS-K, MT, IY, and KC) then conducted 1-hour semistructured interviews with each participant to understand their experiences, data preferences, and insights into the wearable devices (Multimedia Appendix 1). Along with participants’ general perspectives, the investigators focused on device acceptability—as illustrated by participants’ attitudes toward the devices’ usability and perceived value. Following these discussions, we introduced the personalized visualizations of their data and prompted participants to engage in a think-aloud process [25]. Participants reflected from multiple perspectives: as individuals balancing complex personal and work lives, as members of a distributed peer group, and as part of a broader community advocating for improved working conditions. Interviews ended when thematic saturation was reached.

Participant characteristics were analyzed using descriptive statistics. We collected participants’ Fitbit data, comprising 187 JSON files containing metrics on activity levels (sedentary, light, fairly active, and very active), step count, distance, heart rate, and sleep duration and stages (wake, light, rapid eye movement, and deep), each tagged with precise date and time stamps. We focused on sleep duration and step count and assessed use feasibility, defined as participants’ ability to consistently engage with the devices, by examining the date and time stamps associated with these metrics to determine whether the devices captured participants’ data on a daily basis.

We selected sleep duration and step count because these metrics were intuitive for participants (ie, required little explanation, were immediately interpretable, and were clearly actionable) and have clear policy relevance, such as the No More 24 campaign [26] advocating to end 24-hour HHA shifts. Furthermore, prior research indicates that Fitbit-derived sleep duration and step counts are reasonably reliable and accurate; systematic reviews and validation studies have found that Fitbit devices provide sleep duration estimates comparable to those of research-grade accelerometers [27,28], perform well in distinguishing wake from sleep [28,29], and meet acceptable validity and reliability standards for tracking PA in free-living conditions [30].

Participants’ Fitbit data (steps and sleep duration) were processed using exploratory data analysis methods in Jupyter Notebook [31]. As mentioned above, during the exploratory data analysis process, we generated data visualizations for the reflection interviews, which also incorporated auxiliary data such as participants’ sleep and work schedules along with the NIH’s PA and sleep recommendations to create benchmarks for optimal levels.

Each participant interview lasted approximately 60 minutes, resulting in 17 hours of audio-recorded interview data. The interviews were professionally transcribed using NoScribe (noScribe.ai) [32], an open-source artificial intelligence tool run on the research team’s computers. We verified each transcript against the original recordings, correcting errors and redacting identifying information. Our analysis used grounded theory and, as such, used an inductive open coding approach [25] with 4 investigators (IRS-K, MT, IY, and KC). We established a baseline code set by jointly coding one interview and then validated our approach by coding a second interview in 2 separate pairs (IRS-K and KC as well as MT and IY), meeting to reconcile differences. The remaining 15 interviews were coded independently, with regular meetings to resolve disagreements and refine the coding scheme. Finally, we conducted affinity diagramming to synthesize these codes into high-level themes with representative quotations.

All participants gave informed consent to take part in the study, including consent for note taking, photo taking, audio recording of interviews, and collection of participants’ Fitbit data. Identifiable details in the data and features in participant images were removed to ensure participant privacy and confidentiality. Participants were compensated for their time with US $25 gift cards for each of the 3 components of data collection (preliminary discussions, field study, and final interview), for a total possible compensation of US $75 in addition to the Fitbit device, which they were allowed to keep. Cornell University’s institutional review board classified the study as expedited and reviewed and approved the project under protocol IRB0148598.

A total of 17 HHAs participated in this study. They had a mean age of 48.7 (SD 12.2) years, 88% (n=15) were women, 65% (n=11) identified as Black, 29% (n=5) identified as Hispanic or Latinx, and they had worked as HHAs for a mean of 11.6 (SD 7.5) years (Table 1).

Overall, 94% (16/17) of the HHAs wore their activity trackers for the full 28-day study period, with only 6% (1/17) wearing the device for 26 days, and 82% (14/17) indicated continued use after the study’s conclusion. As shown in Table 2, participants took a mean of 10,230 (SD 3586) daily steps during the study period, ranging from 3855 to 20,528 steps. In total, 53% (9/17) of the participants surpassed the NIH recommendations for 10,000 daily steps for healthy adults [33]. Table 2 shows that participants slept for a mean of 6.27 (SD 0.58) hours per night, ranging from 5.40 to 7.38 hours, with 18% (3/17) exceeding the NIH recommendations for 7 hours or more of daily sleep duration for healthy adults [34].

A total of 4 major themes arose. Each theme is presented below with accompanying quotations that represent key concepts.

To our knowledge, this is the first study to examine how activity tracking devices influence HHAs’ awareness of their health and well-being using empirical data from the devices and qualitative reflection interviews with participants. Our findings demonstrate that the use of activity trackers was both feasible and acceptable among HHAs. Feasibility was reflected quantitatively by participants’ ability to consistently engage with the devices, with 94% (16/17) wearing them for the full 4-week field study period. Acceptability was shown qualitatively through participants’ reflections during follow-up interviews, where they expressed positive attitudes toward the devices’ usability and perceived value for increasing health awareness. This level of engagement is noteworthy as previous research has shown lower participant adherence to consumer wearables [35]. The high engagement observed among HHAs in this study suggests that activity tracking interventions may be particularly effective for supporting the health and well-being of HHAs more broadly, although further research is necessary to validate this assumption.

Additionally, the Fitbit Charge 6’s wristband form factor and small screen provided participants with a discreet, functional user experience. Beyond use and feasibility, HHAs not only found that the devices offered them more awareness of key aspects of health—PA and sleep—but even in a short time, they reflected that they were empowered to make positive behavior changes because of wearing the devices. Finally, we found that the devices may have health benefits not only for HHAs but also for the patients for whom they care.

Our study builds on scientific literature in a few key ways. First, while many studies have focused on the acceptance and impact of wearables among older adults [13,14], few have addressed their use among family caregivers and facility-based health care workers [15-19], and even fewer have examined these applications with HHAs [36].

For example, several studies conducted in Finland and Norway have examined the use of wearables among family caregivers and facility-based home care nurses, nursing assistants, and occupational therapists [17-19]. However, these studies often relied on technologies such as accelerometers [17,18] and electrode-based monitors [19], which can be difficult to set up and may not offer real-time or easily interpretable health feedback for lay users, unlike consumer-friendly devices such as Fitbit trackers.

To date, there has been one prior case study in which Fitbits were deployed among caregivers at the SEIU 775 Benefits Group in Washington state [36]. The study found that using Fitbit devices, combined with health coaching, resulted in improvements in health behaviors over a 4-month period. These improvements included increased PA (measured via step counts and active minutes), weight loss, and reduced blood pressure. The findings were based on multiple blood pressure readings, self-reported weight loss, and recorded activity data from the devices. However, this was a pilot program conducted by SEIU 775 Benefits Group in collaboration with Fitbit, Inc (a Google subsidiary), resulting in a self-authored “case study” rather than a peer-reviewed journal article. While the case study presents promising findings, its details and depth are limited due to its intended audience. Our study builds on this preliminary work by outlining a detailed and robust study methodology focused on Fitbit devices for improving HHAs’ awareness of their health and well-being. Additionally, our study incorporates both wearable data and qualitative reflections, aimed at providing a deeper understanding of how HHAs navigate barriers to improving their health and how tracking devices may help support these efforts.

Prior research has shown that wearable devices support health monitoring and facilitate behavior change, particularly in PA and sleep [13,16], and our study extends the literature to HHAs. In doing so, it highlights how these devices have the potential to improve the health of a vulnerable population and shape workplace dynamics and patient care within home care settings. Importantly, we found that wearables can not only track behavior but also offer a simple and feasible way to readily influence workplace routines, patient interactions, and broader occupational health strategies in home care settings. This finding is notable as previous research has shown that influencing behavior change is often challenging and may reflect HHAs’ perception of being in control of their own health [37].

In recent years, there has been a growing call for better understanding and improving the health of HHAs [4,7,9,10]. However, empirical studies that aim to improve their health behaviors and health are scarce. Our study is a first step toward this through harnessing technology and offering HHAs real-time access to their own health data. While we found striking benefits, it is also important to note the challenges. For example, while these devices can offer HHAs personalized feedback and enhance user motivation and self-efficacy [13,16], we also found that external and structural barriers in the environments in which the HHAs live and work can limit the extent to which HHAs can act on their health feedback, making it difficult to prioritize meaningful and sustainable behavioral health changes. For example, unpredictable work schedules, multiple jobs, and familial obligations all left them with limited time and resources to obtain sufficient sleep.

Nevertheless, HHAs found the intervention valuable, using wearable devices to self-monitor, reflect on current health habits, and make changes to their behaviors. Some participants adjusted their sleep routines, such as modifying bedtimes and reducing screen time before sleep, whereas others increased their PA in response to real-time device feedback. Consistent with social cognitive theory and the theory of planned behavior, these self-directed changes suggest that wearables can increase health awareness and motivation to change, representing a promising approach to promote health behavior change among HHAs [11,12].

Although structural barriers such as power asymmetries among workers, patients, and agencies; low wages; and isolated work environments limit HHAs’ ability to improve their health and well-being, wearable data could be integrated into existing peer support programs that aim to reduce worker isolation and strengthen community ties [38,39]. For example, in our prior studies, we have found that HHAs are more likely to engage in PA if they know peers or their clients are engaged as well. Our findings in this study may help bolster and extend these efforts, which have largely relied on HHAs’ self-reported experiences. Aggregated wearable data could more concretely illustrate the challenges that HHAs face, enrich qualitative accounts, and guide the design of peer support programs by grounding discussions and activities in workers’ collective experiences to identify strategies for navigating or mitigating these structural constraints.

Notably, the findings also suggest that PA may be more modifiable than sleep for the HHAs and have clearer positive implications for their patients. That is, several participants found that wearing their activity trackers reminded them of the importance of PA on the job and motivated them to spend more time walking with their patients to increase both their step counts. This is an important observation as previous research has identified challenges in improving PA among older adults receiving home care [40] and researchers have suggested that activities such as walking are valuable for promoting PA outside of structured exercise programs [41].

In future studies, this relationship warrants investigation. Key questions include the following: does increasing HHAs’ step count in turn increase their patients’ PA and mobility and reduce the risk of falls? Does increasing time spent together moving foster closer relationships between HHAs and their patients through shared activity? Our work has previously found that higher levels of mutuality between HHAs and patients result in higher job satisfaction among HHAs [42], and this too could be formally examined after introducing technology. Another important area for exploration is the affordability of wearable devices for a workforce with limited disposable income (eg, the Fitbit Charge 6 costs approximately US $150 as of this writing). If these devices are shown to significantly impact worker and patient outcomes, they may be justified for inclusion in future workforce wellness programs. Alternatively, existing devices such as smartphones could offer a more cost-effective means of collecting data on PA and sleep. However, unlike wearable devices, smartphones are not continuously worn on the body, which may limit the accuracy of the data they capture. Regardless, the feasibility of using smartphones should be further examined.

Beyond improving personal health, these devices may help further center workers’ experiences by introducing an additional layer of granular quantitative data to complement workers’ qualitative narratives. This collection of data sources may help bolster ongoing advocacy efforts, such as supporting improving wages or better working conditions. For example, collective data can strengthen labor union and worker advocacy campaigns such as the push for Fair Pay for Home Care [43], which advocates for livable wages and job stability for HHAs. Similarly, the No More 24 campaign [26], which seeks to end the harmful practice of requiring HHAs to work 24-hour shifts, where HHAs work for 24 hours but are only paid for 13 hours under the premise that they receive at least 5 hours of uninterrupted sleep and other breaks, is directly relevant to HHAs’ poor sleep levels. Our research suggests that tracking HHAs’ activities could generate data-driven evidence on sleep patterns and quality, which, when combined with worker-centered reflections that contextualize these data, could provide compelling arguments for policymakers and bolster organizational advocacy efforts. Future research should explore how to responsibly manage and share HHAs’ collective data in ways that center their needs and experiences to improve their health and well-being.

We acknowledge that our study has several limitations. We conducted a small-scale, short-term empirical mixed methods study in a single geographic location, New York City. Further research is needed to understand how our findings might generalize to larger samples and other locales, including rural and non-US settings, particularly those where driving is a more frequent mode of transportation than walking or public transit. Participants were recruited through their affiliation with a labor management sector of a health care union and may have been more motivated to engage with the wearable devices than nonunionized HHAs, which could confound our findings and limit generalizability to the broader HHA workforce.

Our study did not ask separate questions regarding sex and gender or race and ethnicity, which may obscure participants’ intersectional experiences. Future work should aim to include these perspectives. Additionally, researchers should investigate the results of a longitudinal study to understand how participants’ experiences might change over time and how they objectively impact the patients they serve. Our study also has limitations inherent to qualitative interviews, including potential participant response bias [44]. Further research might explore techniques such as journaling and remotely distributed surveys to decrease the impact of researchers’ positionality.

Our findings suggest that personal tracking devices offer a feasible and acceptable approach to supporting positive health-related changes in this historically overlooked workforce. These devices have the potential to improve both the health behaviors and well-being of HHAs and, potentially, of the patients for whom they care. These concepts should be formally tested in future clinical trials.