Remote monitoring of patients with cancer receiving chemoradiotherapy using wearable devices, supported by AI tools, may improve symptom tracking and patient experience. However, researchers reported operational and organizational challenges, increased staff demands, and concerns that technology cannot replace the doctor-patient relationship.

These findings were reported in the ESMO Real World Data and Digital Oncology and described preliminary results from the continuous activity monitoring (CAM 2.0) study, which included 73 patients with gastrointestinal, lung, or head and neck cancers who were receiving chemoradiotherapy.

All participants used wearable devices that continuously tracked patient activities. Patients in the intervention group also interacted with an AI-powered chatbot called “Penny,” which was designed to triage common adverse events automatically and provide support through text messaging.

Speaking with Univadis Italy, part of the Medscape Professional Network. Francesco Perrone, MD, president of the Italian Association of Medical Oncology and director of the Clinical Trials Complex Unit at the Pascale Cancer Institute in Naples, Italy, said, “Conceptually and technically, these tools are an extension of ePROs [electronic patient-reported outcomes], which are also recommended by the ESMO Guidelines, but this extension is neither necessary nor mandatory.” “It only makes sense if it improves outcomes for both the patient and the clinician, without harming anyone.”

Perrone noted that the balance between automation and direct clinical observation should depend on whether technology proves useful in real-world clinical settings.

Previous findings from the CAM 1.0 study suggested that wearable monitoring tools may support nursing triage and potentially reduce emergency department (ED) visits and hospital admissions.

The primary endpoint of CAM 2.0 evaluated whether chatbot use reduced in-person triage visits compared with standard monitoring. The secondary endpoints included treatment interruptions, ED visits, hospital admissions, and staff workload. Although these outcomes remain under evaluation, researchers highlighted several early observations.

First, the analysis showed that introducing AI may increase the workload of the care team. The authors noted that nurses took on new responsibilities, including helping patients use the chatbot and verifying the accuracy of the reported alerts. They also emphasized that digital health tools may redistribute the workload of the healthcare staff rather than reduce it. The inclusion of staff workload as a secondary endpoint in the CAM 2.0 study reflected these concerns.

The study also highlighted the digital divide — differences in technology access and use, driven both by patients’ motivation and by limitations of the platform itself — which may interfere with treatment adherence and continuity of care.

Speaking with Univadis Italy, Laura Marziali, oncology patient advocate, oncology activist, and president of C’è Tempo OdV, an Italian volunteer organization, said, “I believe the risk of digital divide in oncology is quite real. If the tool is not designed according to principles of universal accessibility, we risk creating first-class patients who are technologically advanced and closely monitored and second-class patients who are older or socially vulnerable.”

Marziali added that AI should not become “another barrier and another cause of discrimination.”

The CAM 2.0 study raised another concern about the use of chatbots and AI in oncology care: Overall quality of life was lower in the chatbot group than in the standard care group. This finding suggests that digital support cannot fully replace the relational aspects of care. The authors acknowledged that clinical reassurance and empathy cannot be replicated through automated responses and emphasized that human interaction remains central to effective symptom management and to the success of each patient’s treatment journey.

“AI should not replace human interaction but rather free up more time for it. The chatbot can manage administrative tasks, streamline the first steps, and manage routine symptom-related processes such as recording data or providing standard responses to known adverse effects. However, it can never validate trauma, fear, or any other human emotion linked to the experience of illness,” Marziali said. She added that the chatbot becomes useful only when it enables doctors or nurses to begin direct patient interactions already informed about the clinical situation, allowing them to bypass technical steps and devote consultation time to empathetic listening, patient’s personal history, symptom history, medical history, and treatment pathways. “Without human contact serving as a source of reassurance, patients may perceive the chatbot as a form of digital abandonment.”

“As the saying goes, the final reckoning comes at the end, and in this area, we are only at the beginning,” Perrone noted. He emphasized that much work remains and said that future progress should be guided by contextual analysis and intellectual honesty rather than by uncritical technological idolization.

This story was translated from Univadis Italy.