Feasibility of implementing a text-based symptom-monitoring program of endometrial, ovarian, and breast cancer patients during treatment

Symptom management utilizing patient-reported outcomes is an important area of focus in cancer care [1]. Cancer patients often experience symptoms related to their treatment regimens [2] and/or the disease itself [3], as well as psychosocial concerns [4, 5]. Common cancer-related symptoms, such as pain, fatigue, and depression [3, 6], may resolve after treatment completion or may persist. Past research indicates that health care providers systematically underestimate their patients’ moderate or severe symptoms compared to what patients report themselves [7]. Under-estimation, which tends to be more common than the over-estimation of patients’ symptoms [7], leads to poorer health outcomes and the under-treatment of patients [8, 9].

Routine monitoring of patients’ physical and psychological symptoms is becoming increasingly more common [5]. The use of patient-reported outcome (PRO) measures during cancer treatment has been shown to improve patients’ survival rates and result in fewer emergency department (ED) visits and hospitalizations, and less symptom burden during hospital stays [10, 11]. When patients have their symptoms addressed or at least relayed to their physicians, it has been found to improve patient satisfaction with care [12], and can help reduce patient anxiety and promote self-care [13]. In addition, monitoring patients’ health-related quality of life has been found to improve communication between patients and their health care providers [12, 14, 15].

While there are many ways to track patients’ symptoms, such as in-clinic assessments and telephone calls between visits, technological advances are making it easier to collect symptoms and adverse events using electronic devices, such as computers, tablets, or mobile phones. The use of these devices can result in faster relay times of patient information to providers [16, 17], more accurate detection of adverse events [18], and may assist in reducing the use of avoidable services like ED visits or hospitalizations [19].

Recent work has shown that health systems and providers are increasingly likely to adopt the use of electronic patient-reported outcomes (PROs) with their patients [18, 20, 21]. Successful remote assessments of PROs using electronic devices (ePROs) have used email or text-message reminders with direct links to patient forms [20], invitations through patient communication portals (e.g., MyChart) [18], and REDCap with automated emails [22, 23]. However, routine cancer care has been slower to adopt ePROs. Challenges to implementing ePROs in clinical practice have been identified, with attempts to address these barriers by involving stakeholders (physicians/staff and patients) [19, 24], and interviewing patients, caregivers, and providers to ensure the relevance of measures selected for PRO assessments [20, 25, 26].

In 2014, a systematic review was conducted on ePRO use in clinical oncology settings (n = 27) [27]. Results indicated that 30% of the ePRO systems reviewed were accessible from the home, and 37% were accessible from both the home and the clinic. Many of the assessments were conducted on computers or tablets, but few used cell phones. Most of the ePROs were designed to be completed by patients during active treatment (63%), but others were also used for follow-up care (40%). Reminders to complete the ePRO surveys were sent in 63% of the systems, with email being the most common method (53%), and 33% using phone, text, or letter reminders. Real-time alerts were used in 85% of systems to send patients’ responses to their providers. The systems reviewed collected a variety of PRO data that were reported to providers (e.g., current scores, longitudinal changes, population norms, or reference values), and varied by the needs of the health care providers in caring for their patients.

The current study reports on a text-based, symptom-monitoring program with patient navigation to assist endometrial, ovarian, and breast patients during treatment. The purpose of the program was to identify patients’ symptoms and needs in a timely manner, before symptoms or problems intensified compromising effective treatment. This paper reports on the feasibility of implementing the program in these patient populations. The a priori goals for program success were that (1) ≥ 85% of patients approached would participate in the monitoring program; and (2) adherence to the surveys during the 12-month period would be ≥ 75% for all cancer types.

Across the three cancer types, 346 patients were approached to take part in the program with 313 agreeing (90.5%). Program declines by cancer type were 9/79 (11.4%) ovarian; 5/55 (9.1%) endometrial; and 19/213 (9.8%) breast. There were no significant differences in program declines or the reasons for refusing by age, race, ethnicity, or cancer type. The major reasons for not participating were lack of interest (36.4%) or believing the program would not be useful (33.3%).

Demographic and clinical characteristics of the participating patients are provided in Table 1. The breast cancer patients were younger on average than either the ovarian or endometrial patients and were more likely to be married, have higher educational attainment, and be employed. Of note is the wide age range of patients across all three cancer types, with the oldest patients over age 80. The majority of patients were non-Hispanic white, which is indicative of the catchment area of the OSUCCC [33], which is 22% rural and includes the Ohio Appalachian region. Greater than 20% of patients across all cancer types reported incomes below $35,000 per year.

A summary of key program components is provided in Table 2.

This paper reported on the feasibility of implementing text-based symptom monitoring with patient navigation to assist ovarian, endometrial, and breast cancer patients undergoing treatment. A major focus was on developing a symptom-monitoring system that could be utilized by most patients, and did not perpetuate biases against patients who lacked electronic devices. Unique aspects of this program included being able to provide smartphones and training to patients without these devices, as well as institute alert values to trigger patient navigators to triage patients’ clinical and non-clinical care needs. Our focus was primarily on larger health systems or academic medical centers that may have resources either through research grant funds or other sources to support these programs. We also sought to utilize or build on existing resources to offset costs of this program. For example, REDCap is available to many academic health centers in the U.S., and can support these types of monitoring program economically.

Successes of this program were that greater than 90% of patients in all three clinics elected to participate in this voluntary activity and complete the text-based surveys for up to 12 months. Adherence to the monthly surveys averaged to approximately 75%, but adherence was lowest among breast patients after 5 months, coinciding in part, with the completion of active therapy. In general, the majority of patients reported value in completing the monthly surveys and having another means to communicate with their health care team. The oncologists and staff in the participating clinics provided critical feedback. They found merit in being able to monitor patients remotely between clinic visits, although the frequency and the timing of the patient assessments, instructions given to patients to contact their health care team directly with severe symptoms or concerns, and the presentation of the survey results back to the health care team will need further streamlining. In addition, “real-time” symptom reporting to the health care team was requested, and the ability to focus on select patients with customized monitoring was believed to be an important use of this technology moving forward. The provision of cell phones to patients, as well as providing navigation services, went smoothly with no difficulties. A limitation of this program, however, was that since this was a pilot quality improvement program and not a randomized intervention study, there were no control groups for comparison purposes. In addition, although we made headway in training older patients to use smartphones or other electric devices to complete the monthly surveys, we still had greater numbers of older patients who preferred to complete these assessments by telephone administration and/or to refuse to enroll in MyChart. These results are similar to those reported by other investigators [21, 23, 27].

A recent review of mobile health interventions/programs found a positive impact among application users in the area of improved symptom control, and determined that changing the patterns of communication between patients and providers is one of the most beneficial aspects of mobile health [25]. Patients in our text-based program also reported similar benefits with more than 70% of patients indicating that completing the symptom surveys helped them communicate better with their providers. In addition, greater than 85% thought that other patients would benefit from this type of symptom monitoring during treatment.

Our original intent was to develop a system that could utilize MyChart to collect patients’ symptoms and needs over time. However, a major drawback of using MyChart is that it lacks flexibility in being able to more quickly add or modify questionnaire items, unlike REDCap. In addition, lower enrollment in MyChart among older adult patients, who constitute the majority of cancer patients, and/or those without electronic devices, again excludes patients from such monitoring and perpetuates health disparities. This is particularly problematic in the state of Ohio, given the large rural and Appalachian populations with sometimes unreliable internet service. Thus, we elected to use REDCap as our mode of survey delivery and data capture. This system worked well for our program purposes, and was very efficient and easy to manage. We will continue to refine this program to discern who might benefit the most from this type of monitoring, how best to meet the needs of the oncologists and staff in treating their patients, and explore options to integrate these data into the EHR, if desired by the health care teams, or use MyChart for some program components.

The value of this and other similar programs will be determined by whether they result in cost savings in terms of fewer hospitalizations, emergency department visits, having patients with better mental health and social support, or assist patients to solve personal/economic barriers to treatment through the use of patient navigators. Not all health care systems can access all of these program components, but routine symptom monitoring using a smart phone or computer/website may be accessible to many. Monitoring can be done in a variety of different ways when patients are not in clinic. This program was just one of the ways patient’ symptoms could be assessed in “real time” using a common technology to address patient needs while undergoing therapy.

This study established the feasibility of implementing a text-based, symptom management program with navigation support for cancer patients undergoing treatment. Future reports will examine the outcomes of this program on clinic flow and metrics (emergency department visits, hospitalizations), as well as more in-depth analyses of the impacts on patient quality of life.