Wearable Health Tech Faces Hurdles in Oncology

Ana Maria Lopez, MD, MPH, MACP

Although use of wearable mobile health (mHealth) monitors has soared in other fields of healthcare, adoption of these electronic devices has lagged in oncology, due mostly to the need to further define how to implement these devices in this setting. Experts say wearables have great potential in oncology, but investigative studies and technological upgrades may be needed.

“There’s a lot of promise with this technology, but we need to study it and better understand where it would fit,” said Ana María López, MD, MPH, MACP, vice chair of medical oncology at Sidney Kimmel Medical College in Philadelphia, Pennsylvania.

mHealth’s practical value in other branches of medicine is clearer than in oncology. Wearable devices are attractive for their ability to monitor patient health parameters such as activity, blood pressure, sleep, heart rate, and weight in real time, making remote monitoring a possibility for physicians.¹ This functionality could lead to fewer in-person office visits, benefiting those who do not live in proximity to their care providers and who do not have their own transportation or access to public transportation. It might also improve patient—physician engagement.

For the general consumer, mHealth technology comes in the form of lightweight fitness trackers that record metrics such as steps taken and stairs climbed. The devices enable users to precisely capture objective data about their daily physical exertion. The Fitbit and Apple Watch lines of wearables are examples of commercially popular fitness monitors that unobtrusively collect data on the wearer’s health. The utility of these devices has been noted by consumers and providers alike and wearable monitors are now a large and growing segment of the healthcare industry.

Whether wearable mHealth devices can become a regular part of oncology practice hinges on their capacity to not only conveniently deliver functional and reliable data for patients and physicians alike, but also have a definitive, care-enhancing purpose.² López said that to find a place for mHealth technology, oncologists first need to clarify how wearable activity monitors add value to this setting by asking, “What’s the problem we’re trying to solve?”

“We need to keep the patient front and center [so we don’t] just become enamored with the technology,” Lopez said.

mHealth devices are seen as a potential solution to the shortcomings of existing tests used to evaluate a patient’s level of function and ability to tolerate systemic treatment, or performance status (PS), López and other experts said. Today, physicians rely on ECOG PS and Karnofsky PS (KPS) scales to guide care decisions²; these simple checklist measures can lead to subjective results because they are based on human assessment. The electronic measurement inherent in wearable technology can provide accurate data that, properly refined and interpreted, could supplement results of PS tests, although much investigation is needed before integrating these devices into modern oncology care, those interviewed for this story said.

“Our current approaches to determining PS are incredibly subjective. We ask patients about what percent of the time they’re active and what they do—that’s not very objective at all,” said George Weiner, MD, CE Block Chair of Cancer Research and director of the Holden Comprehensive Care Center at the University of Iowa in Iowa City. “Very often, they’re with family members, and it’s not uncommon for the patient to say one thing and for the family member to clarify that [what the patient said] isn’t fully accurate.”

ECOG PS and KPS scales are subjective and physician reported.² ECOG PS is a simple 6-level scale that evaluates a patient’s functionality using descriptions such as “fully active” and “completely disabled.” The KPS index has a similar, somewhat broader range of measurement intervals, but both scales have been shown to correlate with each other. Neither one is highly rigorous.

“Observing the patient for 3 minutes as they’re moving around in the examining room is highly inadequate,” Weiner said.

However, results from these tests are important to therapy choice, and inaccurate PS evaluations can lead to serious error. Because poor PS is indicative of increased risk of chemotherapy toxicity and inferior outcomes, an oncologist would be less likely to recommend a regimen that includes chemotherapy for a patient with a low PS. Therefore, imprecise PS evaluations can result in the selection of therapeutic approaches that do not best suit the patient. Moreover, because PS informs clinical trial eligibility, inaccurate PS could cause a patient to be excluded from a study that might be of benefit or lead to use of a trial therapy that the patient ultimately cannot endure.

Just as physicians may over- or underestimate patients’ physical activity levels and abilities, patients also may misjudge these measures. PS reports are frequently limited by bias, such as recall bias, which refers to patients’ inability to completely and correctly GEORGE WEINER, MD remember their symptoms and estimate their activity levels in the time that lapsed after their previous doctor’s appointment.³

“Many studies have demonstrated that patients will usually overestimate their physical activity,” Gillian Gresham, PhD, a postdoctoral student at Cedars-Sinai Medical Center (CSMC) in Los Angeles, California, said.

A 2018 study found substantial discord between the physical activity levels electronically measured by CamNtech Ltd’s ActiHeart device and those reported by individuals through completion of the International Physical Activity Questionnaire-Short Form (IPAQ-SF). Investigators used ActiHeart, a heart rate monitor and accelerometer, to chart the physical activity of 108 students in the Tlokwe local municipality of South Africa for 7 consecutive days. Although the majority of students (57%) responded that they were highly active during IPAQ-SF review, electronically measured physical activity indicated that 93% of the participants were inactive.⁴ This study involved a group of 15-year-old students rather than a cohort of patients with cancer, but it nevertheless highlights the potentially misleading nature of subjective reports of daily physical exertion.

The objective data that wearable activity monitors amass about the duration, intensity, and frequency of physical activity can supplement ECOG PS and KPS assessments, offsetting subjectivity and bridging the informational gaps caused when patients miss appointments. “When patients see their provider, it’s only for a small amount of time; you’re only really getting an hour, maybe even less, to try to make assessments of their functionality and activity. You don’t really get the full picture,” Gresham said. “If a patient misses a clinic visit, then you don’t have anything.”

Early Exploration

Gresham was the lead author of a CSMC study, with results that demonstrated the feasibility of using wearable activity monitors to assess PS in patients with cancer. The investigators used the Fitbit Charge HR to measure the daily activity of 37 patients with stage IV or unresectable advanced stage III cancer.3 The majority of patients had received diagnoses of gastrointestinal cancer (n = 27) and had stage IV disease (n = 34). Patients of varying ECOG PS and KPS ratings agreed to wear the Fitbit for 3 consecutive clinic visits over a 2-week period. Investigators assessed participants’ ECOG PS and KPS scores and determined associations between metrics (steps, distance, and stair climbing) and PS, clinical outcomes (adverse events [AEs], hospitalizations, and survival), and patient-reported outcomes (PROs) during this time.

The highest correlations observed were between average daily steps and PS scores. Each 1000-steps/ day increase was associated with reduced odds of AEs (OR, 0.34; 95% CI, 0.13-0.94), hospitalizations (OR, 0.21; 95% CI, 0.56-0.79), and hazard for death (HR, 0.48; 95% CI, 0.280.83). On average, patients walked 3700 steps, or 1.7 miles, a day; climbed 3 flights of stairs daily; and slept 8 hours/night as measured by their wearable device.

The study results showed that wearable activity monitors can supplement clinical evaluation of performance status, the authors wrote.³

Gresham and López agree that further investigation is needed to identify the groups of patients with cancer for whom the technology would be most appropriate and effective and, similarly, under what circumstances it would afford the greatest value to patients and healthcare providers. Making these determinations is another challenge to implementation of mHealth devices in oncology.

“Not a lot of studies have looked at how these devices can be used as tools to obtain objective information about patient functionality,” Gresham said.

López added that “monitoring patients can seem like a good idea, but if we don’t really know what to do with the data, initial optimism can fall through.”

Managing Data

Another source of complexity involves sorting and storing the vast volumes of data gathered by mHealth devices. This information first must be transferred from an individual’s wearable device to a secure database before it can be accessed. In a 2018 study, investigators used an Apple Watch to collect PROs from 296 adults with lymphoma, myeloma, brain, pancreatic, breast, and ovarian cancer who had a life expectancy of >6 months.⁵ Participants wore the device an average of 9.8 hours/ day and took 4590 steps/day over 3 months. The study generated 6 million kB of data, an amount that might easily exceed the capacity of a medical institution’s computer systems, investigators said.¹

Defining the type of patient information to be aggregated, and how often, will be crucial to developing a data-driven workf low for institutions that intend to use mHealth technology in the future. “What is the timing of the data? Are you getting data in every day? What’s the quantity? Is it something that your team can manage?” said Susan K. Peterson, PhD, MPH, a professor in the Department of Behavioral Science, Division of Cancer Prevention and Population Sciences, at The University of Texas MD Anderson Cancer Center in Houston.⁶

Equally important is establishing a process for data cleanup, if required, prior to interpretation.1 For example, physical activity data may be duplicated if the wearer uses other fitness monitors. There is also the chance that more data might be collected than are needed or desired, based on the health parameters that an institution intends to track. Alternatively, patient data could be saved in incompatible file formats. Such cleanup, whether in the form of information filtering or conversion, would be necessary in each instance.

To provide adequate storage space for the information gathered, centers with computer systems that cannot accommodate large amounts of data might consider cloud-based alternatives.¹

Gresham said larger, randomized studies that examine how wearable technology can improve patient PS assessments will be the next step in deducing whether data from activity monitors can augment ECOG PS or KPS evaluations or, potentially, replace these standard tools.

“I suspect that, eventually, we’ll end up doing some controlled studies comparing reported PS to PS determined by wearables, and we’ll find huge differences,” Weiner said.

Table. Age and Education are Significant Determinants of Mobile Health Acceptance

In a recent study of mHealth technology acceptance among different socioeconomic strata, investigators in Philadelphia interviewed 151 patients with cancer. The survey specifically assessed the patients’ current and desired utilization of technology for healthcare services. Responses suggested that age and education levels are potential barriers to acceptance of wearable technology by patients under treatment for cancer. Patients >70 years were “significantly less likely” to use mobile applications. Conversely, those with a college-level education or higher were much more likely to utilize mHealth (Table).⁷

References

1. Liao Y, Thompson C, Peterson S, Mandrola J, Beg MS. The future of wearable technologies and remote monitoring in health care. Am Soc Clin Oncol Educ Book. 2019;39:115-121. doi: 10.1200/EDBK_238919.
2. Kelly CM, Shahrokni A. Moving beyond Karnofsky and ECOG performance status assessments with new technologies. J Oncol. 2016;2016.6186543. doi: 10.1155/2016/6186543.
3. Gresham G, Hendifar AE, Spiegel B, et al. Wearable activity monitors to assess performance status and predict clinical outcomes in advanced cancer patients. NPJ Digit Med. 2018;1:27. doi: 10.1038/s41746-018-0032-6.
4. Monyeki MA, Moss SJ, Kemper HCG, Twisk JWR. Self-reported physical activity is not a valid method for measuring physical activity in 15-year-old south African boys and girls. Children (Basel). 2018;5(6). doi: 10.3390/children5060071.
5. Thompson CA, Yost KJ, Bartz A, et al. Patient-reported outcomes, emoji, and activity measured on the Apple Watch in cancer patients. J Clin Oncol. 2018;36(suppl 15; abstr 6501). ascopubs.org/doi/abs/10.1200/JCO.2018.36.15_suppl.6501.
6. Peterson SK. Walk this way: wearable devices and remote monitoring for patients with cancer. Presented at: 2019 American Society of Clinical Oncology Annual Meeting; May 31-June 4, 2019; Chicago, IL. https://meetinglibrary.asco.org/record/168015/video.
7. Potdar R, Thomas A, DiMeglio M, et al. Acceptability of mobile health technology among cancer patients. J Clin Oncol. 2019;37(suppl; abstr e18139). doi: 10.1200/JCO.2019.37.15_suppl.e18139.