This paper is in the following e-collection/theme issue:

Cancer Survivorship (258) Wearable Devices and Sensors (888) Fitness Trackers and Smart Pedometers/Accelerometers (561) Novel Sensors and Data Acquisition Methods in Cardiology (196) Nutrition, Physical Activity, Healthy Lifestyle for Cancer Patients and Survivors (138) Mobile Health (mhealth) (2939) Breast Cancer (206) Postoperative Rehabilitation (42)

Published on in Vol 10 (2024)

Preprints (earlier versions) of this paper are available at https://preprints.jmir.org/preprint/51210, first published .
Heart Rate Monitoring Among Breast Cancer Survivors: Quantitative Study of Device Agreement in a Community-Based Exercise Program

Heart Rate Monitoring Among Breast Cancer Survivors: Quantitative Study of Device Agreement in a Community-Based Exercise Program

Heart Rate Monitoring Among Breast Cancer Survivors: Quantitative Study of Device Agreement in a Community-Based Exercise Program

Authors of this article:

Lindsey L Page1 Author Orcid Image ;   Jason Fanning2 Author Orcid Image ;   Connor Phipps1 Author Orcid Image ;   Ann Berger3 Author Orcid Image ;   Elizabeth Reed4 Author Orcid Image ;   Diane Ehlers1, 5 Author Orcid Image
Article Authors Cited by Tweetations Metrics

    Original Paper

    1Department of Neurological Sciences, University of Nebraska Medical Center, Omaha, NE, United States

    2Department of Health and Exercise Science, Wake Forest University, Winston-Salem, NC, United States

    3College of Nursing, University of Nebraska Medical Center, Omaha, NE, United States

    4Department of Internal Medicine, University of Nebraska Medical Center, Omaha, NE, United States

    5Department of Quantitative Health Sciences, Division of Epidemiology, Mayo Clinic Arizona, Rochester, AZ, United States

    Corresponding Author:

    Diane Ehlers, PhD

    Department of Quantitative Health Sciences

    Division of Epidemiology

    Mayo Clinic Arizona

    200 First Street SW

    Rochester, AZ, 55905

    United States

    Phone: 1 408 574 2739

    Email: ehlers.diane@mayo.edu


    Background: Exercise intensity (eg, target heart rate [HR]) is a fundamental component of exercise prescription to elicit health benefits in cancer survivors. Despite the validity of chest-worn monitors, their feasibility in community and unsupervised exercise settings may be challenging. As wearable technology continues to improve, consumer-based wearable sensors may represent an accessible alternative to traditional monitoring, offering additional advantages.

    Objective: The purpose of this study was to examine the agreement between the Polar H10 chest monitor and Fitbit Inspire HR for HR measurement in breast cancer survivors enrolled in the intervention arm of a randomized, pilot exercise trial.

    Methods: Participants included breast cancer survivors (N=14; aged 38-72 years) randomized to a 12-week aerobic exercise program. This program consisted of three 60-minute, moderate-intensity walking sessions per week, either in small groups or one-on-one, facilitated by a certified exercise physiologist and held at local community fitness centers. As originally designed, the exercise prescription included 36 supervised sessions at a fitness center. However, due to the COVID-19 pandemic, the number of supervised sessions varied depending on whether participants enrolled before or after March 2020. During each exercise session, HR (in beats per minute) was concurrently measured via a Polar H10 chest monitor and a wrist-worn Fitbit Inspire HR at 5 stages: pre-exercise rest; midpoint of warm-up; midpoint of exercise session; midpoint of cool-down; and postexercise recovery. The exercise physiologist recorded the participant’s HR from each device at the midpoint of each stage. HR agreement between the Polar H10 and Fitbit Inspire HR was assessed using Lin concordance correlation coefficient (rc) with a 95% CI. Lin rc ranges from 0 to 1.00, with 0 indicating no concordance and 1.00 indicating perfect concordance. Relative error rates were calculated to examine differences across exercise session stages.

    Results: Data were available for 200 supervised sessions across the sample (session per participant: mean 13.33, SD 13.7). By exercise session stage, agreement between the Polar H10 monitor and the Fitbit was highest during pre-exercise seated rest (rc=0.76, 95% CI 0.70-0.81) and postexercise seated recovery (rc=0.89, 95% CI 0.86-0.92), followed by the midpoint of exercise (rc=0.63, 95% CI 0.55-0.70) and cool-down (rc=0.68, 95% CI 0.60-0.74). The agreement was lowest during warm-up (rc=0.39, 95% CI 0.27-0.49). Relative error rates ranged from –3.91% to 3.09% and were greatest during warm-up (relative error rate: mean –3.91, SD 11.92%).

    Conclusions: The Fitbit overestimated HR during peak exercise intensity, posing risks for overexercising, which may not be safe for breast cancer survivors’ fitness levels. While the Fitbit Inspire HR may be used to estimate exercise HR, precautions are needed when considering participant safety and data interpretation.

    Trial Registration: Clinicaltrials.gov NCT03980626; https://clinicaltrials.gov/study/NCT03980626?term=NCT03980626&rank=1

    JMIR Cancer 2024;10:e51210

    doi:10.2196/51210

    Keywords

    wearable devices; exercise prescription; validity; photoplethysmography; monitoring; wearables; devices; exercise; heart rate; breast cancer; cancer; cancer survivor; community; chest monitor; Fitbit; recovery; safety 


    Wearable sensors have gained traction in both commercial and research sectors [1], with a projected 156 million units to be purchased in 2024 [2]. These devices use microelectronic triaxial accelerometers to measure steps, energy expenditure, sleep, and time spent in different intensities of activity and photoplethysmography above the wrist to measure heart rate (HR). These data, along with options for goal setting, can be used to help individuals self-monitor and increase their daily physical activity (PA) [3]. The ease and utility of these devices have led to their adoption in health promotion research for continuous measurement of health behaviors and as behavior change tools [4].

    Traditional research-grade monitors are costly, lack consumer-friendly designs, provide little opportunity for user interaction with the device, often evaluate only 1 dimension of daily activities (eg, HR, motion, or sleep only), and have limited, real-time data transfer capacity [3,5,6]. In contrast, consumer-grade monitors continuously collect and transfer data through Bluetooth and web-based platforms to allow for data collection across months or even years [3]. Additionally, participant data can be easily monitored and accessed via web-based platforms at any point during the data collection period. The increasing number of peer-reviewed publications and National Institutes of Health–funded grant proposals, which include consumer-grade, wrist-worn monitors, emphasizes the utility of these devices in research settings [1,7].

    Many of these devices now measure HR, a key component of aerobic exercise prescription. Although electrocardiogram (ECG) is widely accepted as the gold standard for assessing HR during exercise, chest-worn monitors also have well-documented validity for measuring HR [8]. However, like ECG, they may be inconvenient or prohibitive in community-based and unsupervised exercise settings due to necessary receivers and participant discomfort. In contrast, newer devices are increasingly being designed for wear on the forearm or wrist. Commercially available wearable sensors, such as the Fitbit, represent an accessible, multifunctional alternative to HR monitoring in exercise. Appropriate exercise intensity, often expressed as a percentage of HR reserve, is a fundamental dimension of exercise prescription for achieving the health benefits of exercise [9]. For example, cancer survivors begin to reduce fatigue symptoms with a minimum dose of aerobic activity at 45% of their HR reserve, whereas benefits for other symptoms (ie, anxiety, depression, and physical function) begin at a minimum dose of 60% of their HR reserve [10]. To improve dissemination and uptake of exercise prescriptions in clinical or community-based settings, it is critical that survivors have user-friendly methods to independently monitor exercise prescription components.

    There is an increasing number of exercise oncology studies that use commercial wearable sensors to intervene in PA behaviors and reduce cancer-related symptom burden, particularly among breast cancer survivors, and evidence indicates that wearable sensors are effective, feasible, and user-friendly for breast cancer survivors in exercise interventions [11-13]. Although these studies have helped bolster the utility of wearable sensors in PA promotion research, they have failed to provide any detail on intensity or HR monitoring during their respective interventions. Many exercise interventions in cancer populations are adopting community-based, hybrid, and unsupervised designs [14]. Therefore, it is integral that researchers understand the capacity, and limitations, of commercially available wearable sensors in providing accurate measurements of HR to monitor participants’ safety and compliance with the exercise prescription.

    In the general population, the reliability of popular, commercially available activity and HR monitors has been previously examined with varying agreement between commercial products and traditional ECG monitoring [8,15-17]. Unfortunately, many of these data have been collected in controlled settings with predetermined treadmill speeds in young, healthy adult participants. The dearth of literature assessing commercially available, wrist-worn HR monitors in clinical populations during training sessions limits the utility of these devices in less controlled environments. Therefore, the purpose of this study was to examine the agreement between a commercially available, wrist-worn wearable sensor (Fitbit Inspire HR; Fitbit Inc) and a traditional chest-worn monitor (Polar H10; Polar Electro OY) for HR measurement in breast cancer survivors at different stages of exercise in a community-based program. It is hypothesized that HR monitor agreement in this study will be highest at periods of pre- and postexercise rest and lowest during the exercise session when participants were exercising at higher intensities.


    Study Design and Participants

    The Study on Physical Activity’s Relationship with Cancer and Cognition (SPARCC) was a randomized exercise trial in which 30 women diagnosed with breast cancer were randomized to a 12-week moderate-intensity aerobic exercise program (n=15) or usual care (n=15). Our study includes only those women randomized to the exercise group with valid Fitbit and Polar HR data (n=14), as exercise HR was not collected from women in the usual care control group.

    Eligibility criteria for this study included the following: female participants aged 21 years or older; postmenopausal at the time of diagnosis; first, primary diagnosis of breast cancer (stage I-IIIa); within 3-24 months of completing surgery, chemotherapy, or radiation therapy; self-reported an average of <60 minutes of moderate to vigorous PA per week for the previous 6 months; having received physician’s clearance to participate in an exercise program; and randomized to the 12-week aerobic exercise program in the SPARCC study. Participants were recruited from a midwestern academic medical center, a private cancer center, and the community (eg, via flyers to community organizations, social media posts, and word of mouth). Interested individuals were scheduled for a phone appointment to confirm eligibility, absence of neurological disorders, and interest in participating in the study. Eligible women were then asked to attend an in-person or Zoom-based orientation session to receive more information about the study, decide if they would like to participate, sign the Institutional Review Board (IRB)–approved informed consent, and schedule baseline testing appointments. After baseline data collection was complete, participants were randomized in a block design to the 12-week aerobic exercise program or usual care. Participants were not instructed to change physical activity behaviors prior to beginning the exercise program.

    Ethical Considerations

    This study was approved by The University of Nebraska Medical Center IRB and is registered with the National Institutes of Health’s ClinicalTrials.gov (NCT03980626). All participants provided written informed consent prior to participation. All data presented herein were deidentified using study identification numbers and stored separately from participants. identifiers. Data were collected and managed using applications hosted by the study institution (ie, Research Electronic Data Capture [REDCap] or Box Enterprise) [18,19]. Participants did not receive payment for their participation in this research but received a Fitbit Inspire HR that was theirs to keep. All participants were offered a 3-month membership to a local fitness center.

    Exercise Protocol

    Breast cancer survivors randomized to the exercise program engaged in small group or one-on-one, moderate-intensity walking sessions facilitated by a certified exercise physiologist. These sessions were held at local community fitness centers 3 times per week for 1 hour per session. All participants completed a treadmill-based submaximal cardiopulmonary exercise test prior to randomization to establish baseline fitness and safety and inform individualized exercise prescriptions (Table 1).

    Table 1. Exercise prescription.
    WeekIntensity (rating of perceived exertion), rangeIntensity (% heart rate reserve), rangeDuration (minutes), range
    19-1145-5015-20
    29-1145-5020-25
    39-1145-5025-30
    411-1350-5525-30
    511-1350-5530-35
    611-1350-5535-45
    713-1555-6535-45
    813-1555-6540-50
    913-1555-6540-50
    1015-1765-7540-50
    1115-1765-7545-50
    1215-1765-7545-50

    As originally designed, all 36 exercise sessions were scheduled to be delivered by the exercise physiologist in the supervised, community-based setting. However, due to the COVID-19 pandemic, the exercise program was modified for some participants to include both supervised and unsupervised sessions. Participants who were in the middle of the intervention in March 2020, were transitioned to unsupervised exercise with weekly Zoom-based counseling from their trainer. Participants enrolled after March 2020, engaged in only 4 supervised exercise sessions held once per week in the research team’s exercise laboratory in weeks 1-4. All sessions in weeks 7-12 were unsupervised, home-based sessions with weekly Zoom-based exercise counseling. Across the study, 4 breast cancer survivors completed 36 supervised sessions as originally designed, 5 were in the middle of the intervention in March 2020, and 6 were enrolled after March 2020 and received 4 supervised sessions. Depending on the enrollment time (ie, before or after the COVID-19 public health restrictions), participants engaged in an average of 13.33 (SD 13.71) supervised sessions (range: 4-36). Participants who received the original intervention received their fitness center membership during the study. Those enrolled during or after the onset of the COVID-19 pandemic were offered a fitness center membership when it was safe to do so based on local IRB and public health requirements. Only supervised sessions (n=200) were included in our analysis.

    The exercise program was progressive in nature such that the volume of exercise increased across weeks from 15-20 minutes of walking in weeks 1-2 to 40-45 minutes in weeks 8-12 and from 40%-55% estimated HR reserve in week 1 to 65%-70% HR reserve in weeks 9-12 (Table 1). All sessions began with a 5-minute light-intensity walking warm-up and ended with an active cool-down including light walking and static stretches (Figure 1). The exercise program was designed to follow American College of Sports Medicine guidelines for exercise in cancer survivors [10].

    ‎
    Figure 1. Exercise session stages for heart rate measurement.

    Measures

    HR Monitors

    All participants received a Fitbit Inspire HR sensor to wear on their nondominant wrist for the duration of the study and a Polar H10 chest strap to wear during supervised exercise. The Fitbit was chosen for this study because it is one of the most popular wrist-worn activity trackers, represents approximately 20% of the commercial wearable sensor market, and has sold 63 million devices worldwide in the last decade [3]. The Fitbit Inspire HR measures HR via optical photoplethysmography, which is processed using proprietary algorithms. Briefly, this is done by shining a light on the skin, assessing the reflected light, using algorithms to determine changes in blood volume based upon reflected light, and calculating HR based on oscillations in blood volume [20,21]. The Polar H10 chest strap monitor was chosen as the comparator device because it has high validity with ECG, the gold standard for measuring exercise HR [8]. Exercise trainers fit participants with the Polar H10 monitors, placed on the distal sternum, at the start of each supervised exercise session and used Polar HR readings to adjust treadmill speed and grade to meet prescribed exercise intensity. HR was measured concurrently using the Fitbit and Polar monitors at 5 stages of the exercise session: pre-exercise seated rest; midpoint of the 5-minute warm-up; midpoint of the moderate-intensity exercise; midpoint of the 5-minute cool-down; and after a 5-minute seated recovery (Figure 1).

    Demographic and Clinical Information

    Participant demographics (ie, age, race, education, income, employment status, marital status, and comorbid conditions) were self-reported and collected via REDCap hosted by the study institution. Clinical information on breast cancer diagnosis and treatment were obtained via electronic medical records. BMI was calculated from height and weight measured via the Seca 703 scale and stadiometer (Seca Corp) by the study staff at a baseline testing visit.

    Data Analysis

    HR from the Polar monitor was operationalized as the criterion measure and used to assess absolute and relative paired differences between monitors [8]. Agreement between HR measurements was assessed using Lin concordance correlation coefficient (rc) with 95% CIs. This test measures the degree to which the paired observations fall on the identity line and defines statistical agreement as rc≥0.85 [22]. The agreement was also represented visually across stages using Bland-Altman plots with upper and lower limits set using 95% CIs [23,24]. Absolute paired differences were calculated by subtracting the Fitbit-measured HR from the Polar-measured HR at each stage of the exercise session. Relative paired differences were calculated as relative error rate (RER) across exercise session stages [25,26], as follows:

    RER = (Polar HR measurement – Fitbit HR measurement) × 100/Polar HR

    Negative resultant RERs are indicative of an overestimation of HR by the Fitbit, and positive RERs are indicative of an underestimation of HR by the Fitbit, as compared to the Polar monitor. Data were analyzed using SPSS (version 27; IBM Corp) and RStudio (version 1.3.1093; R Core Team).


    Participant Characteristics

    Participants (mean age 63.1, SD 8.7 years) were White women with a history of early-stage breast cancer; on average, overweight; and physically inactive (Table 2). Additionally, more than 1 quarter of participants had a history of clinically diagnosed anxiety or depression at the time of enrollment. Breast cancer survivors in this study were enrolled approximately 17 months after their diagnosis. Participants’ breast cancer treatments included surgery, radiation, and chemotherapy; however, most women in this study did not receive chemotherapy. One participant randomized to the exercise program did not have valid Polar data for supervised sessions and was, therefore, excluded from the analysis (N=14).

    Table 2. Participant characteristics.
    CharacteristicsValues
    Demographics

    		Age (years), mean (SD)63.07 (8.66)

    		Non-Hispanic White, n (%)13 (93)

    		Bachelor’s degree, n (%)8 (57)

    		Income >US $40,000 per year, n (%)8 (57)

    		Employed full-time, n (%)8 (57)

    		Married, n (%)12 (86)

    		Comorbiditiesa, mean (SD)2.38 (2.10)

    		β-Blocker medication use, n (%)0 (0)

    		Antihypertensive medication use, n (%)1 (7)

    		Diagnosed with depression, n (%)4 (29)

    		Diagnosed with anxiety, n (%)4 (29)

    		BMI (kg/m2), mean (SD)29.14 (4.71)
    Clinical features

    		Cancer stage, n (%)


    		I12 (86)


    		II2 (14)

    		Time since diagnosis (months), mean (SD)16.57 (7.97)

    		Chemotherapy, n (%)3 (21)

    		Radiation, n (%)9 (64)

    		Hormonal therapy, n (%)9 (64)

    		Months of hormonal therapy, mean (SD)13.44 (5.64)

    aComorbid conditions include diagnosed arthritis, osteoporosis, asthma, chronic obstructive pulmonary disease, angina, heart failure, previous myocardial infarction, vascular disease, diabetes, tremors, gastrointestinal disease, visual impairment, hearing impairment, degenerative disk disease, anxiety, and depression.

    HR Monitor Agreement

    Agreement between the Fitbit and Polar HR monitors was highest during seated rest at postexercise (rc=0.89, 95% CI 0.86-0.92) and pre-exercise (rc=0.76, 95% CI 0.70-0.81). This was followed by the midpoint of the moderate-intensity exercise session (rc=0.63, 95% CI 0.55-0.70) and cool-down (rc=0.68, 95% CI 0.60-0.74). The warm-up was associated with the lowest level of agreement between monitors (0.39, 95% CI 0.27-0.49). RERs ranged from –3.91% to 3.09% and were most pronounced during warm-up (RER: mean –3.91%, SD 11.92%). When inaccurate, the Fitbit overestimated HR during most stages of the exercise session (RER range: –3.91% to –0.52%), except at the midpoint of moderate-intensity exercise, where HR was underestimated (RER 3.09%). RERs and concordance coefficients are provided in Table 3, and Bland-Altman plots are provided in Figure 2.

    Table 3. Heart rate (beats per minute) monitor differences according to the stage of the exercise.
    ActivityHeart rate, mean (SD)Fitbit differences from Polar H10, mean (SD)

    		Polar H10FitbitPaired differencePercent difference (RERa)Agreement (rc)b
    Pre-exercise rest78.00 (7.68)78.61 (8.74)–1.34 (5.53)–1.84 (7.51)0.76
    Warm-up94.78 (7.91)96.26 (9.17)–2.96 (10.27)–3.91 (11.92)0.39
    Exercise117.04 (8.10)115.15 (9.3)3.77 (6.98)3.09 (5.64)0.63
    Cool-down103.56 (6.68)104.24 (7.90)–0.72 (5.83)–0.77 (5.74)0.68
    Postexercise rest85.86 (7.42)85.99 (7.71)–0.38 (3.53)–0.52 (4.04)0.89

    aRelative error rate.

    bLin concordance correlation coefficient.

    ‎
    Figure 2. Bland-Altman plots. Bland-Altman plots depict the average heart rate (Polar H10 and Fitbit) by the relative difference between the two measures for each session by stage in the exercise protocol. Points on the plots indicate individual sessions, solid lines indicate the mean difference across the sample, and dashed lines indicate upper and lower bounds for each stage.

    Principal Results

    This study was the first to examine the agreement between the wrist-worn Fitbit Inspire HR and chest-worn Polar H10 HR monitors in cancer survivors and during an exercise intervention. Major findings from this study indicate that the Fitbit monitor did not produce statistically accurate measures of HR during most exercise stages, especially the warm-up stage; however, agreement during seated rest (pre- and postexercise), midpoint of exercise, and cool-down were associated with moderate agreement between devices. Data also suggest that the Fitbit underestimated HR only during the primary, aerobic portion of the exercise session (ie, the midpoint of the exercise session), which may have serious implications for safety and compliance monitoring in exercise programs. This study extends the current literature on consumer-grade, wrist-worn HR monitors and provides data to inform future studies hoping to use consume-grade sensors to monitor safety, exercise program compliance, and longitudinal behavioral patterns in cancer survivors [8,25,27]. This is particularly important as exercise interventions become less centralized and hybrid and unsupervised approaches increase in prevalence [28-30].

    Comparison With Prior Work

    The wrist-worn, Fitbit Inspire HR monitor accurately measured HR only during seated rest postexercise compared to the chest-worn Polar H10 monitor. Although pre-exercise seated rest, midpoint of moderate-intensity exercise, and cool-down also exhibited high levels of agreement, they did not reach statistical agreement as defined by Lin concordance correlation coefficient [22]. Results from previous studies have found that Fitbit devices are most accurate in measuring HR during low-intensity activities where the wrist is moving in a repetitive fashion [17]. Nevertheless, in contrast to these previous studies, warm-up represented the period of the poorest agreement. It is also unclear why the Fitbit accurately measured HR during post- but not pre-exercise seated rest; however, both pre-and postexercise seated rest reflected the highest levels of agreement with the Polar monitor, consistent with previous findings [8,16]

    Although HR was highest during the midpoint of the exercise session and cool-down, these stages represented higher levels of agreement when compared to the warm-up stage. This may, in part, be due to the slower speeds at which breast cancer survivors were walking in this study, as compared to healthy, young, or middle-aged adults in other studies [8,16,17,31]. While previous research found that lower treadmill speeds showed the highest agreement, overall speeds in those studies ranged from 2 to 9 miles per hour [16,17]. In comparison, breast cancer survivors in this study did not reach speeds greater than 3.5 miles per hour. It is possible that speeds in this study were more similar to light-intensity walking in previous studies, which would align more closely with the results from this study [16,17]. This does not, however, explain the poor agreement during the warm-up and cool-down stages.

    Of note, the RER at the midpoint of the exercise session indicates that the Fitbit monitor underestimated HR as compared to the Polar chest strap. Previous studies have reported similar trends in Fitbit data as compared to traditional ECG monitoring [8,16,17,31]. One study found the Fitbit Ionic to be comparable to other wrist-worn monitors and statistically accurate at rest [17], while another found that the Fitbit Blaze provided the least accurate optically measured HR [8]. In contrast to the findings presented here, these studies also found that higher-intensity activity led to decreased accuracy in HR measurement [8,17]. However, differences in Fitbit accuracy at peak exercise intensity between previous studies and data presented in this study may be due to the lower absolute intensity in this study, as both previous studies were conducted among athletes [8,17]. Given the generally lower intensity of exercise prescribed to cancer survivors, participants may not reach an exercise intensity high enough for devices to decrease in accuracy during steady-state exercise. This should, theoretically, reinforce the utility of Fitbits in cancer survivor populations.

    Despite this, underestimation of exercise HR may be problematic in programs using Fitbit to monitor intensity during exercise in cancer survivors for several reasons. First, participants may be asked to increase the intensity of a session to achieve the prescribed HR range. If the Fitbit monitor underestimates HR, as it did in this study, participants who reach the prescribed HR as measured by Fitbit may be exercising at an intensity higher than that prescribed, leading to concerns regarding participant safety—particularly if the session is unsupervised. In a previous analysis of exercise prescription adherence in this sample, data indicated that participants only met prescribed intensity during supervised sessions 57.5% of the time when assessed via Fitbit HR, as compared to 92.2% when measured via Polar. However, adherence to the prescribed intensity via Fitbit was 83.2% during unsupervised sessions after the onset of the COVID-19 pandemic [32]. These data, when taken together with the findings of this study, suggest that participants may have been exercising above their prescribed HR range during unsupervised sessions. Although many breast cancer survivors may comfortably exercise at higher intensities, reliance upon consumer-grade wearable sensors only may introduce safety concerns not previously observed in more traditional, controlled exercise trials. Additionally, future studies that use Fitbit to measure the dose of exercise required to effect specific outcomes (eg, cancer-related fatigue and cognitive performance) may underestimate the required intensity of activity to elicit an effect. Although these devices may have utility in exercise oncology, it is critical that researchers and practitioners are aware of limitations that may increase the risk of adverse events or decrease methodological rigor in quantifying compliance with exercise prescriptions.

    Limitations

    Although this study is one of the first to investigate HR agreement between the Fitbit Inspire HR and Polar H10 chest monitor during community-based exercise in breast cancer survivors, it is not without limitations. First, this study was performed on a small sample of breast cancer survivors. Our sample was primarily comprised of White, educated women with early-stage breast cancer, which may not be representative of the larger breast cancer population. For example, women in this study were also not on any β-Blockers, and only 1 participant reported antihypertensive medication use. Although this improves the internal validity of this study due to the lack of HR suppression, it is likely not representative of the broader breast cancer survivor population in the United States [33]. Future studies would be strengthened by the inclusion of a larger, more diverse sample of breast cancer survivors with more supervised exercise sessions.

    Additionally, exercise physiologists were available to help and provide feedback on using the devices during the exercise sessions, making it unclear to what extent user error would influence Fitbit’s accuracy in unsupervised exercise settings. Fitbit HR measurements in this study were also compared to Polar chest strap monitors, rather than the gold standard ECG. This may have introduced systematic error in evaluating agreement. Finally, the total number of sessions observed in this study was fewer than originally planned (ie, 200 observed vs 540 planned) due to COVID-19 required adaptations. It is unclear whether additional observations would have changed or stabilized results relative to device agreement.

    Conclusions

    Overall, Fitbit devices with HR monitoring capabilities may be useful for participant monitoring in exercise oncology studies. Researchers should use caution when using these devices, however, as they likely do not provide accurate HR measurement during critical stages of exercise sessions. This study showed that Fitbit monitors were only statistically accurate during seated rest and likely underestimated HR during steady-state exercise. As a result, Fitbit HR measurements are likely best for estimating exercise intensity rather than evaluating compliance with exercise prescriptions. Because of their ease and potential utility in behavioral PA interventions, future studies should further examine the agreement between wrist-worn wearable sensors and a gold-standard measurement of HR, such as ECG, in a larger, more representative sample of breast cancer survivors. Additionally, studies should analyze agreement by relative HR intensity to determine whether Fitbit may be more appropriate for specific exercise prescriptions.

    Acknowledgments

    Funding for this study was provided by the University of Nebraska Medical Center Fred and Pamela Buffett Cancer Center Support Grant (P30CA036727) as well as the Nebraska Tobacco Settlement Biomedical Research Development Fund (NTSBRDF). The funders were not involved in the study design, implementation, or dissemination activities. The authors would like to acknowledge Tiffany Dudley, BS, American Collect of Sports Medicine (ACSM) Certified Exercise Physiologist (ACSM-EP), and Joseph Scharfenkamp, MS, ACSM Clinical Exercise Physiologist (ACSM-CEP) for the implementation of the exercise intervention and their flexibility in pivoting from in-person to remote delivery in March 2020.

    Data Availability

    The data sets generated during this study are available from the corresponding author on reasonable request.

    Authors' Contributions

    LLP carried out conceptualization, data curation, formal analysis, investigation, methodology, project administration, and writing of the original draft. JF was in charge of data curation, methodology, and software, as well as reviewing and editing of the manuscript. CP contributed to the formal analysis and visualization as well as reviewing and editing of the manuscript; AMB and ECR carried out the methodology and contributed to the reviewing and editing of the manuscript. ECR also provided resources for the study. DKE was in charge of the conceptualization, funding acquisition, methodology, project administration, resources, and supervision, as well as reviewing and editing of the manuscript.

    Conflicts of Interest

    None declared.

    1. Wright SP, Hall BTS, Collier SR, Sandberg K. How consumer physical activity monitors could transform human physiology research. Am J Physiol Regul Integr Comp Physiol. Dec 01, 2017;312(3):R358-R367. [FREE Full text] [CrossRef] [Medline]
    2. Wrist-worn wearable shipments worldwide 2019-2024. Statista. URL: https://www.statista.com/statistics/296565/wearables-worldwide-shipments/ [accessed 2024-09-02]
    3. Feehan LM, Geldman J, Sayre EC, Park C, Ezzat AM, Yoo JY, et al. Accuracy of Fitbit devices: systematic review and narrative syntheses of quantitative data. JMIR mHealth uHealth. Aug 09, 2018;6(8):e10527. [FREE Full text] [CrossRef] [Medline]
    4. Mercer K, Li M, Giangregorio L, Burns C, Grindrod K. Behavior change techniques present in wearable activity trackers: a critical analysis. JMIR mHealth uHealth. 2016;4(2):e40. [FREE Full text] [CrossRef] [Medline]
    5. Imboden MT, Nelson MB, Kaminsky LA, Montoye AH. Comparison of four Fitbit and Jawbone activity monitors with a research-grade ActiGraph accelerometer for estimating physical activity and energy expenditure. Br J Sports Med. Jul 2018;52(13):844-850. [CrossRef] [Medline]
    6. Dominick GM, Winfree KN, Pohlig RT, Papas MA. Physical activity assessment between consumer- and research-grade accelerometers: a comparative study in free-living conditions. JMIR Mhealth Uhealth. Sep 19, 2016;4(3):e110. [FREE Full text] [CrossRef] [Medline]
    7. NIH Reorter. URL: https://reporter.nih.gov/ [accessed 2024-11-02]
    8. Gillinov S, Etiwy M, Wang R, Blackburn G, Phelan D, Gillinov AM, et al. Variable accuracy of wearable heart rate monitors during aerobic exercise. Med Sci Sports Exerc. Aug 2017;49(8):1697-1703. [CrossRef] [Medline]
    9. Garber CE, Blissmer B, Deschenes MR, Franklin BA, Lamonte MJ, Lee I, et al. Quantity and quality of exercise for developing and maintaining cardiorespiratory, musculoskeletal, and neuromotor fitness in apparently healthy adults. Med Sci Sports Exerc. 2011;43(7):1334-1359. [CrossRef] [Medline]
    10. Campbell KL, Winters-Stone KM, Wiskemann J, May AM, Schwartz AL, Courneya KS, et al. Exercise guidelines for cancer survivors: consensus statement from international multidisciplinary roundtable. Med Sci Sports Exerc. Nov 2019;51(11):2375-2390. [FREE Full text] [CrossRef] [Medline]
    11. Hartman SJ, Nelson SH, Weiner LS. Patterns of Fitbit Use and Activity LevPatterns of Fitbit use and activity levels throughout a physical activity intervention: exploratory analysis from a randomized controlled trialels Throughout a Physical Activity Intervention: Exploratory Analysis from a Randomized Controlled Trial. JMIR Mhealth Uhealth. Feb 05, 2018;6(2):e29. [FREE Full text] [CrossRef] [Medline]
    12. Hardcastle SJ, Jiménez-Castuera R, Maxwell-Smith C, Bulsara MK, Hince D. Fitbit wear-time and patterns of activity in cancer survivors throughout a physical activity intervention and follow-up: Exploratory analysis from a randomised controlled trial. PLoS One. Oct 19, 2020;15(10):e0240967. [FREE Full text] [CrossRef] [Medline]
    13. Ringeval M, Wagner G, Denford J, Paré G, Kitsiou S. Fitbit-based interventions for healthy lifestyle outcomes: systematic review and meta-analysis. J Med Internet Res. Oct 12, 2020;22(10):e23954. [FREE Full text] [CrossRef] [Medline]
    14. Bluethmann SM, Vernon SW, Gabriel KP, Murphy CC, Bartholomew LK. Taking the next step: a systematic review and meta-analysis of physical activity and behavior change interventions in recent post-treatment breast cancer survivors. Breast Cancer Res Treat. Jan 2015;149(2):331-342. [FREE Full text] [CrossRef] [Medline]
    15. Reddy RK, Pooni R, Zaharieva DP, Senf B, El Youssef J, Dassau E, et al. Accuracy of wrist-worn activity monitors during common daily physical activities and types of structured exercise: evaluation study. JMIR mHealth uHealth. Dec 10, 2018;6(12):e10338. [FREE Full text] [CrossRef] [Medline]
    16. Wang R, Blackburn G, Desai M, Phelan D, Gillinov L, Houghtaling P, et al. Accuracy of wrist-worn heart rate monitors. JAMA Cardiol. Jan 01, 2017;2(1):104-106. [CrossRef] [Medline]
    17. Pasadyn SR, Soudan M, Gillinov M, Houghtaling P, Phelan D, Gillinov N, et al. Accuracy of commercially available heart rate monitors in athletes: a prospective study. Cardiovasc Diagn Ther. Aug 2019;9(4):379-385. [FREE Full text] [CrossRef] [Medline]
    18. Harris PA, Taylor R, Thielke R, Payne J, Gonzalez N, Conde JG. Research electronic data capture (REDCap)--a metadata-driven methodology and workflow process for providing translational research informatics support. J Biomed Inform. Apr 2009;42(2):377-381. [FREE Full text] [CrossRef] [Medline]
    19. Harris PA, Taylor R, Minor BL, Elliott V, Fernandez M, O'Neal L, et al. REDCap Consortium. The REDCap consortium: building an international community of software platform partners. J Biomed Inform. Jul 2019;95:103208. [FREE Full text] [CrossRef] [Medline]
    20. Haghayegh S, Khoshnevis S, Smolensky MH, Diller KR. Accuracy of PurePulse photoplethysmography technology of Fitbit Charge 2 for assessment of heart rate during sleep. Chronobiol Int. Jul 16, 2019;36(7):927-933. [CrossRef] [Medline]
    21. Allen J. Photoplethysmography and its application in clinical physiological measurement. Physiol Meas. Mar 2007;28(3):R1-39. [CrossRef] [Medline]
    22. Lin LI. A concordance correlation coefficient to evaluate reproducibility. Biometrics. Mar 1989;45(1):255-268. [Medline]
    23. Martin Bland J, Altman D. Statistical methods for assessing agreement between two methods of clinical measurement. The Lancet. Feb 1986;327(8476):307-310. [CrossRef]
    24. Bland JM, Altman DG. Measuring agreement in method comparison studies. Stat Methods Med Res. Jun 02, 1999;8(2):135-160. [CrossRef] [Medline]
    25. Thomson EA, Nuss K, Comstock A, Reinwald S, Blake S, Pimentel RE, et al. Heart rate measures from the Apple Watch, Fitbit Charge HR 2, and electrocardiogram across different exercise intensities. J Sports Sci. Jun 18, 2019;37(12):1411-1419. [CrossRef] [Medline]
    26. Shcherbina A, Mattsson CM, Waggott D, Salisbury H, Christle JW, Hastie T, et al. Accuracy in wrist-worn, sensor-based measurements of heart rate and energy expenditure in a diverse cohort. J Pers Med. May 24, 2017;7(2):3. [FREE Full text] [CrossRef] [Medline]
    27. Coughlin SS, Caplan LS, Stone R. Use of consumer wearable devices to promote physical activity among breast, prostate, and colorectal cancer survivors: a review of health intervention studies. J Cancer Surviv. Jun 2020;14(3):386-392. [CrossRef] [Medline]
    28. Phillips SM, Penedo FJ, Collins LM, Solk P, Siddique J, Song J, et al. Optimization of a technology-supported physical activity promotion intervention for breast cancer survivors: Results from Fit2Thrive. Cancer. Mar 01, 2022;128(5):1122-1132. [FREE Full text] [CrossRef] [Medline]
    29. Alfano CM, Leach CR, Smith TG, Miller KD, Alcaraz KI, Cannady RS, et al. Equitably improving outcomes for cancer survivors and supporting caregivers: a blueprint for care delivery, research, education, and policy. CA Cancer J Clin. Jan 2019;69(1):35-49. [FREE Full text] [CrossRef] [Medline]
    30. Stout NL, Brown JC, Schwartz AL, Marshall TF, Campbell AM, Nekhlyudov L, et al. An exercise oncology clinical pathway: screening and referral for personalized interventions. Cancer. Jun 15, 2020;126(12):2750-2758. [FREE Full text] [CrossRef] [Medline]
    31. Muggeridge DJ, Hickson K, Davies AV, Giggins OM, Megson IL, Gorely T, et al. Measurement of heart rate using the Polar OH1 and Fitbit Charge 3 wearable devices in healthy adults during light, moderate, vigorous, and sprint-based exercise: validation study. JMIR Mhealth Uhealth. Mar 25, 2021;9(3):e25313. [CrossRef] [Medline]
    32. Page LL, Berger A, Reed E, Ehlers DK. Feasibility of an aerobic exercise intervention before and during the COVID pandemic. Ann Behav Med. Apr 2021;55(Supplement_1):S229. [CrossRef]
    33. Williams GR, Deal AM, Lund JL, Chang YK, Muss HB, Pergolotti M, et al. Patient-reported comorbidity and survival in older adults with cancer. Oncologist. Apr 2018;23(4):433-439. [FREE Full text] [CrossRef] [Medline]

    ‎
    ECG: electrocardiogram
    HR: heart rate
    IRB: Institutional review board
    PA: physical activity
    REDCap: Research Electronic Data Capture
    RER: relative error rate
    SPARCC: Study on Physical Activity’s Relationship with Cancer and Cognition

    Edited by T de Azevedo Cardoso; submitted 24.07.23; peer-reviewed by T Lund-Jacobsen, J Borrowman, M Fraser; comments to author 29.11.23; revised version received 03.04.24; accepted 08.05.24; published 20.06.24.

    Copyright

    ©Lindsey L Page, Jason Fanning, Connor Phipps, Ann Berger, Elizabeth Reed, Diane Ehlers. Originally published in JMIR Cancer (https://cancer.jmir.org), 20.06.2024.

    This is an open-access article distributed under the terms of the Creative Commons Attribution License (https://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work, first published in JMIR Cancer, is properly cited. The complete bibliographic information, a link to the original publication on https://cancer.jmir.org/, as well as this copyright and license information must be included.