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.
The number of older patients with gastrointestinal cancer is increasing due to an aging global population. Minimizing reliance on an in-clinic patient performance status test to determine a patient’s prognosis and course of treatment can improve resource utilization. Further, current performance status measurements cannot capture patients' constant changes. These measurements also rely on self-reports, which are subjective and subject to bias. Real-time monitoring of patients' activities may allow for a more accurate assessment of patients’ performance status while minimizing resource utilization.
This study investigates the validity of consumer-based activity trackers for monitoring the performance status of patients with gastrointestinal cancer.
A total of 27 consenting patients (63% male, median age 58 years) wore a consumer-based activity tracker 7 days before chemotherapy and 14 days after receiving their first treatment. The provider assessed patients using the Eastern Cooperative Oncology Group Performance Status (ECOG-PS) scale and Memorial Symptom Assessment Scale-Short Form (MSAS-SF) before and after chemotherapy visits. The statistical correlations between ECOG-PS and MSAS-SF scores and patients’ daily step counts were assessed.
The daily step counts yielded the highest correlation with the patients' ECOG-PS scores after chemotherapy (
This study demonstrates the feasibility of using inexpensive, consumer-based activity trackers for the remote monitoring of performance status in the gastrointestinal cancer population. The findings need to be validated in a larger population for generalizability.
The number of gastrointestinal cancer cases is predicted to increase due to the aging population [
Although patients spend most of their time between cancer treatments at home, tests such as the ECOG-PS and MSAS-SF are conducted at clinic visits and do not provide a daily view of patients' performance status [
Incorporating patient-generated health data can reduce bias and improve the accuracy of the patients' performance status tests. Electronic mobile activity trackers provide new methods for collecting and monitoring patients' daily activities and function in real settings. The feasibility of commercially available activity trackers has already been demonstrated for patients with other types of cancer [
The development phase of this study began in February 2019. The Memorial Sloan Kettering Cancer Center Institutional Review Board authorized the conduct of this study in August 2019. Medical professionals were recruited from Memorial Sloan Kettering Cancer Center in New York, and the resulting team included oncologists, an oncology nurse specialist, oncology rehabilitation physicians, and a customer relationship management expert.
Patients were eligible to participate in the study if they were aged ≥18 years, had gastrointestinal cancer, and started a new line of chemotherapy. Patients were excluded if they were using assistive devices such as a walker or a cane or were receiving concomitant radiation and chemotherapy. Additionally, patients needed to be enrolled in the study for at least seven days before starting the new chemotherapy line to allow for a proper baseline activity assessment. All patients gave their written consent to participate in the study.
Each participant was given a Misfit Shine AT fitness tracker (Misfit) after institutional review board approval and written informed consent. This particular model was selected after assessing various consumer-based activity trackers based on the following four characteristics:
No feedback provided. Patients should not receive any feedback regarding their step count, nor any positive or negative reinforcement in response to a high or low number of steps [
Long battery life. Patients should not need to remove the device to recharge it, which would potentially result in forgetting to put it back on again [
Waterproof. The activity tracker should be waterproof to allow patients to continue wearing it while showering.
Ability to act as an independent device. The activity tracker should be able to act as an independent device and not require synchronization with a cell phone.
The Misfit Shine exhibits all these characteristics and best fit our needs for this study. Misfit Shine has been validated for clinical use in prior studies [
A Misfit account was created for each patient and patients were instructed to wear the Misfit Shine on their nondominant wrist. The number of daily steps was recorded automatically in the app via Wi-Fi. Clinicians had access to the patients’ data on the administrator web page. An unidentified code was applied to each patient for security. It is important to mention that patients did not have access to their accounts in order to prevent them from reviewing their step count. Step count data were collected for each patient for 7 days prechemotherapy and 14 days postchemotherapy. A day with a step count >100 was referred to as a “full day of data collection,” a day with a step count <100 was referred to as a “partial day of data collection,” and a day with no step count recording was referred to as “no data collection.” Only patients with at least three full days of data collection during both the prechemotherapy and postchemotherapy periods were included in the final study.
A research study assistant collected patient data in two phases. C1D1 (cycle 1, day 1) indicates that the data were collected before the first cycle of chemotherapy (“prechemotherapy”). C2D1 (cycle 2, day 1) indicates that the data were collected after the first cycle of chemotherapy and before the second cycle (“postchemotherapy”). Patient data were collected 7 days before C1D1 and 14 days after C1D1. There was no intervention involved in the activity monitoring, and the data were collected after the completion of each cycle, not in real-time.
Data on the presence and severity of symptoms were collected at baseline and at C2D1 by administering the MSAS-SF [
A total of 41 patients consented to the study, but one patient dropped out of the study because they decided to receive treatment at another institution. Only 27 patients (68%) had adequate activity tracker data, as shown in
Completeness of patient data collection.
Demographic and clinical characteristics at baseline visit.
Demographics and characteristics | Values | |
|
||
|
Male | 17 (63) |
|
Female | 10 (37) |
Age (years), median (range) | 58 (37-83) | |
|
||
|
Married | 18 (66) |
|
Single | 9 (34) |
|
||
|
College graduate or higher | 15 (55) |
|
Lower than college degree | 12 (45) |
|
||
|
Adjuvant | 8 (30) |
|
Neoadjuvant | 6 (22) |
|
Metastatic | 13 (48) |
|
||
|
Ever | 10 (27) |
|
Never | 17 (63) |
|
||
|
0 | 17 (63) |
|
1 | 10 (37) |
|
>1 | 0 (0) |
|
||
|
Global distress index subscale | 0.63 (0.37) |
|
Physical symptom subscale | 0.69 (0.52) |
|
Psychological symptom subscale | 1.28 (0.75) |
|
Total Memorial Symptom Assessment Scale | 0.63 (0.37) |
The daily mean, median, maximum, and minimum activity level pre- and postchemotherapy.
Treatment phases and days | Mean activity level | Median activity level | Maximum activity level | Minimum activity level | ||
|
|
|||||
|
Day 1 | 0.713a | 0.724a | 0.269 | 0.572a | |
|
Day 2 | 0.729a | 0.779a | 0.350 | 0.540a | |
|
Day 3 | 0.857a | 0.782a | 0.773a | 0.776a | |
|
Day 4 | 0.820a | 0.720a | 0.875a | 0.895a | |
|
Day 5 | 0.842a | 0.785a | 0.558a | 0.798a | |
|
Day 6 | 0.720a | 0.724a | 0.372 | 0.632a | |
|
Day 7 | 0.729a | 0.692a | 0.585a | 0.584a | |
|
||||||
|
Day 1 | 0.193 | 0.096 | 0.296 | 0.091 | |
|
Day 2 | 0.625a | 0.556a | 0.596a | 0.563b | |
|
Day 3 | 0.676a | 0.591a | 0.709a | 0.406b | |
|
Day 4 | 0.858a | 0.793a | 0.923a | 0.361 | |
|
Day 5 | 0.894a | 0.900a | 0.805a | 0.569b | |
|
Day 6 | 0.913a | 0.929a | 0.817a | 0.550b | |
|
Day 7 | 0.885a | 0.897a | 0.835a | 0.439b | |
|
Day 8 | 0.716a | 0.738a | 0.630a | 0.566b | |
|
Day 9 | 0.549a | 0.591a | 0.421b | 0.584b | |
|
Day 10 | 0.862a | 0.892a | 0.738a | 0.609b | |
|
Day 11 | 0.829a | 0.807a | 0.822a | 0.442b | |
|
Day 12 | 0.863a | 0.870a | 0.744a | 0.563b | |
|
Day 13 | 0.858a | 0.862a | 0.776a | 0.507b | |
|
Day 14 | 0.725a | 0.764a | 0.642a | 0.351 |
a
b
The correlation of mean, median, maximum, and minimum activity levels pre- and postchemotherapy.
Activity levels | Prechemotherapy mean activity level | Prechemotherapy median activity level | Prechemotherapy maximum activity level | Prechemotherapy minimum activity level |
Postchemotherapy mean activity level | 0.839a | 0.763a | 0.794a | 0.644a |
Postchemotherapy median activity level | 0.823a | 0.774a | 0.762a | 0.695a |
Postchemotherapy maximum activity level | 0.792a | 0.648a | 0.838a | 0.509b |
Postchemotherapy minimum activity level | 0.501b | 0.605b | 0.293 | 0.412b |
a
b
The overall average number of steps per day for all patients was 6290 before chemotherapy and 6325 after chemotherapy. The average step count prechemotherapy for patients with an ECOG-PS of 1 was 7023 steps per day, while patients with an ECOG-PS of 2 had an average step count of 5405 steps per day (
The step count per day by ECOG-PS score. C1D1 indicates that the data were collected before chemotherapy, and C2D1 indicates that the data were collected after chemotherapy. ECOG-PS: Eastern Cooperative Oncology Group Performance Status.
The overall median number of steps walked by patients pre- and postchemotherapy was 4983 and 5480, respectively. Overall, the step count decreases after chemotherapy; the median difference between pre- and postchemotherapy for the cohort was 497 steps, and the IQR decreased from 5916 steps to 5119 steps postchemotherapy. Prechemotherapy, patients younger than 60 years of age walked more than patients older than 60 years of age (median number of steps 5618 versus 4738,
The daily step count of each patient before the first cycle of chemotherapy (labeled as C1D1) and after the first cycle of chemotherapy (labeled as C2D1).
We calculated the volatility of step counts pre- and postchemotherapy to illustrate the degree of behavior change, as follows:
where
where
Step count volatility for each patient.
The annualized variance of the step counts of patients before (labeled as C1D1) and after (C2D1) the first cycle of chemotherapy.
The median physical and psychological scores prechemotherapy were 0.53 (IQR 0.26-1.06) and 1.26 (IQR 0.66-1.86), respectively. The median GDI and TMSAS scores were 1.12 (IQR 0.64-1.56) and 0.66 (IQR 0.30-0.88), respectively. Patients’ symptom burden changed after chemotherapy. Patients had a median improvement of 0.18 for their GDI score, 0.09 on the TMSAS, and 0.26 for the psychological score, while the physical score did not change. In addition, 59% (16/27) had an improvement in their GDI and 62% (17/27) had an improvement in their TMSAS during the postchemotherapy phase. The rate of improvement for the cohort was 46% and 65% for the physical and psychological domains, respectively.
Those with an improvement in their GDI, TMSAS, and physical scores took more daily steps before and after chemotherapy compared to those with no improvement in these scores. The three patients who experienced an improvement in their GDI, physical, and TMSAS scores walked 6205, 5769, and 6239 steps before chemotherapy and 5788, 6216, and 5788 steps daily after chemotherapy. However, those with no improvement in their GDI, physical, and TMSAS scores walked 5032, 5436, and 5148 steps per day before chemotherapy, and 3934, 4562, and 4816 steps per day after chemotherapy. All MSAS scores of patients before and after chemotherapy are shown in
Patients' MSAS scores before and after the first day of the first cycle of chemotherapy (labelled as C1D1 and C2D1, respectively). GDI: global distress index; MSAS-SF: Memorial Symptom Assessment Scale-Short Form; PHYS: physical symptom subscale; PSYCH: psychological symptom subscale; TMSAS: total MSAS.
Only 13 of 40 patients did not have adequate data. There were 8 patients without adequate data prechemotherapy, 3 patients without adequate data postchemotherapy, and 2 patients with inadequate data both pre- and postchemotherapy. Of the collective 280 prechemotherapy days of the study cohort, there were 195 days (69%) with a “full day of data collection,” 2 days (1%) with a “partial day of data collection,” and 83 days (30%) with “no data.” Of the collective 560 postchemotherapy days of the study cohort, there were 405 days (72%) with a “full day of data collection,” 21 days (3%) with a “partial day of data collection,” and 134 days (25%) with “no data.” During the 7-day prechemotherapy phase, on average, patients had 5 days with a “full day of data collection.” During the 14-day postchemotherapy phase, patients had an average of 10 days with a “full day of data collection.” Patients with adequate activity tracker data were younger compared to those with inadequate data (median age 58 years versus 60 years,
Out of 27 participants, only one participant had discomfort when wearing the activity tracker prechemotherapy; however, this person found it comfortable to wear the device postchemotherapy. In addition, two patients found it uncomfortable to wear the activity tracker postchemotherapy. The patient who experienced discomfort when wearing the activity tracker prechemotherapy had an increase in the number of steps taken postchemotherapy. In contrast, for the patients who had trouble with the device postchemotherapy, the number of steps decreased. The median satisfaction score pre- and postchemotherapy remained the same at 80.
This study investigated the feasibility of employing consumer-based activity trackers to monitor patients with gastrointestinal cancer undergoing chemotherapy. As shown in
The study results indicate statically significant correlations between the number of steps patients take daily and two common performance status tests (ECOG-PS and MSAS-SF), which is consistent with earlier research findings [
We developed a steps volatility chart as a remote activity monitoring tool, as shown in
We believe patients' step counts, coupled with ECOG and MSAS scores, can help clinicians better understand patients' conditions. Activity tracker data provide a dynamic view of patients and could decrease the bias in patients' assessment tests. Although our study was limited by patient sample size, the number of monitored days, and our patients' performance status, we studied our patients in an uncontrolled environment outside clinical settings. In doing so, we illustrated the functionality of using wearable activity trackers to collect data in real life. The patients in this study tend to be healthier, with lower ECOG-PS scores, than the broader cancer population. Although this may limit the generalizability of our findings to a broader population, our results are in line with other studies on patients with severe conditions [
In conclusion, the remote monitoring of patients' physical activity could decrease the cost of health care and provide a higher quality of health care to a broader population. Remote monitoring could revolutionize how we treat patients and help to provide health care for patients who live in remote areas without direct access to health care clinics or at times when doctors cannot see their patients in person. As a next step, we will collect data from a larger sample of patients with cancer with a broader range of ECOG-PS scores and find an approach that will encourage patients to use wearable activity trackers more regularly.
cycle 1, day 1
cycle 2, day 1
Eastern Cooperative Oncology Group Performance Status
global distress index
Memorial Symptom Assessment Scale-Short Form
physical symptom subscale
psychological symptom subscale
total MSAS
The project was supported, in part, by the Beatriz and Samuel Seaver Foundation, the Memorial Sloan Kettering Cancer and Aging Program, the National Institutes of Health/National Cancer Institute Cancer Center Support Grant P30 CA008748, the Alliance for Clinical Trials in Oncology Scholar Award, and the National Science Foundation under grant CNS-1750679. We appreciate the involvement of our patients and their caregivers in this study.
None declared.