Digital symptom monitoring with patient-reported outcomes (PROs) in community oncology practices improves patients' symptom management and yields better clinical outcomes than usual care.
When cancer patients come for a clinic visit, clinicians face hurdles such as limited time and competing topics, and the patient may forget to bring up symptoms that prevent information from being communicated. At home, patients may be reluctant to contact the clinician's office until symptoms become severe. Communication hurdles may inhibit the ability of clinicians to detect and react early to symptoms before they worsen. An alternative approach is to use digital symptom monitoring with PROs.
“A computer system can trigger reminders to a patient by email or text to report symptoms at regular intervals. The system can then alert clinicians in real time about severe or worsening symptoms to prompt symptom management, and can generate reports showing longitudinal trajectories of symptoms that can be discussed at clinic visits,” said Ethan Basch, MD, MSc, Chief of Oncology at UNC-Chapel Hill. “This approach can enhance the connection between patients and clinicians and focus their interactions. It is designed to catch symptoms early and enable clinicians to intervene to prevent downstream complications.”
Previously, the STAR randomized controlled trial, conducted at a single tertiary cancer center, found benefits of a digital symptom monitoring system on multiple clinical outcomes. At a live presentation at the American Society of Clinical Oncology Plenary Series, Basch provided results from the multicenter PRO-TECT trial. This is a cluster randomized trial conducted at 52 U.S. community oncology practices, funded by The Patient-Centered Outcomes Research Institute. At each practice, up to 50 patients could be enrolled who had a metastatic cancer receiving systemic therapy. Practices were randomized to digital symptom monitoring with PRO surveys or to usual care.
Patients in the PRO arm were asked to complete a weekly electronic survey with 12 common symptoms. Email alerts were sent to clinical nurses in real time any time a patient reported a severe or worsening symptom, along with evidence-based symptom management pathway advice. Reports showing the longitudinal trajectory of symptoms were generated for the clinical team at visits. Patients at intervention sites were offered a choice of completing their weekly PRO surveys via web, handheld device, or an automated telephone system.
Usual care consisted of discussions between patients and clinicians at visits or interactions between visits. Copies of symptom management pathways were provided to nurses and patients.
Data on overall survival, the primary outcome based on linkage to national mortality databases, is not yet mature. Basch reported on key secondary outcomes, including effects on patients' physical function, symptom control, health-related quality of life, implementation feasibility, and satisfaction with digital monitoring. Physical function, symptom control, and health-related quality of life were assessed at baseline and at months 1, 3, 6, 9, and 12.
The study enrolled 1,191 patients (593 PRO, 598 control). “There were statistically significant differences in favor of the PRO arm at multiple time points, with some convergence towards the end of the study when there was substantial attrition of patients due to death or hospice,” said Basch. There was a similar benefit in favor of the PRO arm for symptom control and significant benefits for health-related quality of life.
In a related analysis, measurement of clinically meaningful benefits at 3 months found the PRO arm had 13.8 percent more patients with physical function benefits, 16.1 percent more patients with symptom control benefits, and 13.4 percent more patients with health-related quality of life benefits, as compared to usual care, with similar findings at the other study time points.
“On average, 91.5 percent of patients completed their weekly PRO survey without attrition over time, even up to week 52, despite the level of illness in this population,” said Basch.
About one-third of the weekly PRO self-reports triggered an email alert to nurses. Nurses took action in response to 59 percent of these alerts, most frequently responding with telephone counseling and supportive medication prescriptions. Patients overwhelmingly agreed that the questions were easy to understand, and the system was easy to use, improved discussions with their doctor and nurse, made them feel more in control of their care, and would recommend it to others.
“Nurses, similarly, had positive impressions of the system, largely noting that the patient-reported information was helpful for electronic health record documentation, improved discussions with patients, and improved efficiency and quality of care. Nurses felt that a small amount of protected time was needed to address the alerts, and future implementations should account for this need for protected nursing time to assure successful uptake,” said Basch.
He noted the PRO system was not integrated with electronic health records at practices, which would have increased efficiency and should be considered in the future. “The high number of alerts could potentially be reduced, and this is a topic of ongoing machine learning research,” Basch added.
Also, more than a year of follow-up was conducted during the COVID-19 pandemic, when competing priorities may have pulled clinical staff from reviewing patient-reported information. This could have reduced the impact of the intervention, he noted.
“Digital symptom monitoring with PROs is feasible, improves clinical outcomes, and is valued by patients and clinicians during routine cancer care,” Basch stated. “Prior research suggests that the mechanism is early detection of symptoms, which enables clinicians to intervene and prevent downstream complications. Future work should focus on barriers and facilitators to effective implementation. PROs offer a tool to contribute to high-quality and patient-centered oncology practice.”
Mark L. Fuerst is a contributing writer.