Breast cancer survivors

Breast cancer is the most common cancer globally and significantly impacts cancer-related mortality rates [1,2]. However, advances in early detection and treatment have led to a steadily increasing population of cancer survivors, presenting both opportunities and challenges for oncology care. In 2020, approximately 5% of the population in Europe were living after a cancer diagnosis, equating to about 23.7 million people out of 477.9 million in the 29 European countries studied [3].

The long-term and late effects experienced by breast cancer survivors and the treatments they undergo entail considering various factors such as the type, duration, and dosage of treatment, as well as the patient’s age during treatment. The assessment and management of these effects, such as quality of life, lymphedema, cardiotoxicity, cognitive impairment, distress, fatigue, bone health, musculoskeletal health, pain, neuropathy, infertility, sexual health, body image issues, and premature menopausal symptoms, among others, are necessary [4–6].

mHealth and PROMs

Patient-Reported Outcome Measures (PROMs) collect self-reported information directly from patients about their health, quality of life, and functionality, offering insights unfiltered by healthcare providers. Mobile health (mHealth) technologies, such as smartphone apps and wearable devices, enhance this process by enabling remote monitoring, personalized interventions, and the seamless integration of PROMs and objective measures (OM) [7]. These tools empower breast cancer survivors to track symptoms, medication, and physical activity, fostering engagement in their care journey [8]. Notably, ICHOM’s [9] standardized PROMs for breast cancer informed the Xemio app’s development and symptom domains, addressing key physical, emotional, and social well-being aspects, including pain, fatigue, and body image. Integrating these measures into mHealth platforms enhances data-driven, personalized care, with studies highlighting the potential of such interventions to improve symptom management, self-efficacy, and communication with providers [10–14].

Despite these advancements, challenges remain in sustaining adherence and engagement with mHealth tools, as shown by research emphasizing the need for optimized strategies [15,16]. Apps like the Xemio-Research App, designed to track side effects, deliver reliable health information, and support breast cancer patients during treatment and recovery, exemplify the potential of mHealth in survivorship care. Our real-world investigation into Xemio’s clinical integration aimed to assess its impact on patient behavior and well-being, addressing survivors’ informational and emotional needs while promoting a holistic, patient-centric approach to cancer care [17,18].

Objectives

This study aimed to investigate the feasibility and acceptability of the mHealth tool Xemio in a clinical setting as a support tool for breast cancer survivors. This included analyzing user adherence, system usability assessed by the System Usability Scale [19] (SUS), and patient experience through an ad-hoc questionnaire.

The secondary objective was to examine the behavior of patients in reporting side effects, by comparing self-reported effects, using the Common Terminology Criteria for Adverse Events (CTCAE) [20] provided by the Xemio-Research App (52 effects), with those reported in traditional questionnaires. Both methods comply with the ICHOM standards for PROMs in breast cancer care.

ASCAPE, OntoCR and Xemio-Research App

ASCAPE (Artificial Intelligence Supporting Cancer Patients across Europe) was a Horizon 2020 (H2020) project focused on breast and prostate cancer, two of the most prevalent malignancies [21]. Its primary goal was to use Artificial Intelligence and machine learning to improve the health and quality of life (QoL) of cancer patients through four pilot studies [22,23]. Clinical collaborators used validated questionnaires to assess various QoL aspects of these cancers. These questionnaires were integrated with data on daily activities, treatment side effects, and medical interventions, enabling Artificial Intelligence models to predict and recommend QoL improvements. The ASCAPE project collects HRQoL scores using validated instruments, including the EORTC QLQ-C30 [24], EORTC QLQ-BR23 [24], Hospital Anxiety and Depression Scale (HADS) [25] and Three-item Loneliness Scale (TILS) [26].

The Barcelona pilot of the ASCAPE project at Hospital Clínic de Barcelona (HCB) used OntoCR, an ontology-driven clinical repository [27,28] that received data from the Xemio-Research App and the clinical study questionnaires. By leveraging a dual model paradigm, OntoCR ensured seamless data transfer and semantic integrity using EN/ISO 13606-compliant extracts [18]. Xemio-Research (Fig 1) is an app developed in 2020 for breast cancer patients, providing them with proper information, enabling symptom tracking, and collecting physical activity data from its users daily (steps, time of activity, and calories). The deployment of the Xemio-Research backend was carried out on a server within the information systems area of the HCB [18].

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Xemio App Interface.

The figure shows a user holding a smartphone displaying the Xemio app, developed by Fundación iSYS. Reprinted under a CC BY license, with permission from Fundación iSYS, original copyright 2024.

https://doi.org/10.1371/journal.pone.0342702.g001

Ethical considerations

This study is part of the research approved by the Ethics Committee for Drug Research of the Hospital Clínic de Barcelona, Spain (HCB/ 2020/0971). All participants provided written informed consent in person during the recruitment visit, covering both primary and secondary analyses. The signed consent forms were securely stored at the hospital, with access restricted exclusively to the authorized project investigators. All data presented in this paper were anonymized. This study is a sub-study of the clinical trial registered under ClinicalTrials.gov Identifier: NCT05401643.

All data collected through the Xemio-Research app were pseudonymized and securely stored in compliance with the General Data Protection Regulation (GDPR). Access to the database was restricted to authorized investigators, and no personally identifiable information was shared externally. These safeguards were implemented to ensure confidentiality and foster patient trust, which is a critical factor for the acceptance of digital health tools.

Study design and participation

This one-year longitudinal feasibility study used a convenience sample of eligible breast cancer survivors recruited at a single academic center. The sample size was determined by the number of patients available and willing to participate during the recruitment window (December 1, 2020–December 22, 2021), with participants assigned to the Xemio intervention group. Recruitment depended on patient availability in the outpatient oncology department and was influenced by COVID-19–related constraints.

Eligibility criteria

Patients were eligible for the study if they met the following criteria: they were breast cancer survivors without hospital treatment in the previous year, they owned a smartphone, they had no cancer-related complications, and they agreed to participate in the study.

Procedures

Participation in the study was proposed to a consecutive sample of adult breast cancer survivors attending the Oncology outpatient consultations of the HCB who met the inclusion criteria. The recruitment period was 12 months, and the participation period in the study was an additional 12 months. For patients who agreed to participate in the clinical study and were assigned to the Xemio group, the App was installed. During a motivational interview, the functionality of the App was explained to them, and they were provided with a brochure containing information and a contact telephone number in case they encountered any issues. Patients were asked to make free use of the App and its contents. Throughout the study, two reminder calls to use the App were given, coinciding with the collection of electronic patient-reported outcomes (ePROs).

Outcome measures

Adherence to the application was measured using three available records: side effects, QoL questionnaires and number of daily steps (pedometer). Side effects were actively reported by the patients using the App’s symptom tracker, which included 52 effects and their intensity according to CTCAE. Daily steps were passively obtained by the mobile phone’s accelerometer. QoL questionnaires were completed at months 0, 3, 6, 9, and 12 according to ASCAPE design. Partial results from these questionnaires could prompt a call from the treating oncologist, who had access to the patients’ follow-up data. At the end of the study, a satisfaction survey was conducted (see S1 and S2 Files). All the collected data was integrated into a pseudo-anonymized dataset to assess the overall feasibility and acceptability of the Xemio-Research mHealth tool in a clinical setting.

Data analysis

The analysis focused on calculating means and standard deviations, providing a detailed summary of central tendency and variability to support the interpretation of the statistical findings.

Comparing side effect capture: CTCAE vs. QoL questionnaires.

We conducted a comparison between the spontaneously reported side effects collected using the CTCAE list and the standardized capture of symptoms through QoL questionnaires. Patients were encouraged to report side effects they experienced, choosing them from a list of 52 side effects presented in the app. The process was designed to be user-friendly and informative, ensuring that patients could accurately document their symptoms. Upon selecting a side effect, a screen appeared with a brief explanation. By clicking the next button, patients could view a list of intensities corresponding to those indicated by the CTCAE, with language adapted for patient understanding. After selecting the intensity (Fig 2), the screen displayed advice tailored to that intensity, providing immediate, relevant information and guidance based on the severity of the reported side effect. The side effect, its intensity, and the date of the report were recorded. Each side effect can have 1–4 intensity options (detailed in S2 File): 1 for mild symptoms, 2 for moderate, 3 for severe, and 4 for life-threatening symptoms.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Patient-reported side effect screens in the Xemio app.

Examples of the user interface for patient-reported symptom tracking in the Xemio mHealth app. (a) Reporting painful intercourse, where users can adjust the intensity level using a color-coded slider. (b) Reporting dry skin, where users select affected body areas and can indicate additional symptoms like rash or pus. (c) Reporting vomiting, where users choose intensity levels based on predefined episodic frequency categories. These features allow real-time symptom tracking, facilitating personalized patient management. Reprinted under a CC BY license, with permission from Fundación iSYS, original copyright 2020.

https://doi.org/10.1371/journal.pone.0342702.g002

Sociodemographic characteristics of the study sample

The participants in the sample had a mean age of 58.9 years. Regarding BMI, 38.3% had a normal weight, 36.2% were classified as overweight, and 25.5% were categorized as obese. In terms of employment, 40.4% were on labor incapacity, while 14.9% were unemployed. Most participants (80.6%) were undergoing treatment with aromatase inhibitors. Educational attainment varied, with 31.9% having no formal education and 29.8% holding university degrees. Economically, 40.4% resided in high-income districts, with an income ranging from €38.604 to €98.883. Additionally, 53.2% had never smoked, and 29.8% reported never consuming alcohol.

Adherence results

Adherence to the Xemio-Research intervention was analyzed as the frequency of app usage throughout the study period. Fig 3 showed a significant decline in app usage after the first three months, with the highest adherence observed during the initial trimester. Usage dropped sharply in subsequent periods and stabilized at approximately 32% during the latter half of the year.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. App usage analysis.

Visualization of patient engagement with the Xemio app over a 12-month period. (a) The number of patients actively using the app in four time intervals: 0-3, 3-6, 6-9, and 9-12 months. A decline in user engagement is observed over time. (b) The total number of app usage days accumulated by patients in each time period, showing a similar declining trend. Despite a decrease in active users, the remaining users maintained relatively high engagement levels.

https://doi.org/10.1371/journal.pone.0342702.g003

Fig 3 illustrates the usage of the app by patients at quarterly intervals throughout the year. During the study period, 49 patients used the Xemio-Research app. In the first 3 months, 47 patients (95.92%) accumulated 669 days of usage. By the 3–6 month period, usage dropped to 20 patients (40.82%) with 411 days. In the 6–9 months, 16 patients (32.65%) used the app for 219 days. Finally, in the 9–12 month period, 17 patients (34.69%) accumulated 327 days of usage. Despite the decrease in active users, the average days of use per patient increased in the later periods.

Feasibility and acceptability

At the end of the study, participants completed an evaluation using the SUS questionnaire and the ad-hoc feasibility questionnaire (S1 File). The following sections present the results.

SUS questionnaire results

Of the 49 patients who used the app, 36 completed the System Usability Scale (SUS) questionnaire. The SUS score, calculated from responses to the 10-item usability questionnaire, was 82.78 (SD = 14.21), indicating a level of usability classified as ”excellent” [25–27].

Ad-Hoc questionnaire results

A total of 33 patients completed the ad-hoc feasibility survey with four sections, designed to evaluate the perceived ease of use of each app functionality. The results from the first section of the questionnaire ad-hoc are detailed in Table 1, where 9 questions to answer using a Likert Scale ranging from 1 (Totally disagree) to 5 (Totally agree).

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Ad-hoc Xemio assessment.

https://doi.org/10.1371/journal.pone.0342702.t001

In the second part of the ad-hoc questionnaire, patients were asked to respond to the open-ended questions: “What did you like most about the Xemio-Research App?” and “What did you like least about the Xemio-Research App?” Participants provided qualitative feedback in free-text format.

1. Positive comments emphasized the app’s ability to reassure users about normal symptoms, thereby reducing fear and anxiety.
2. Patients also highlighted its user-friendly interface, high-quality content, personalized symptom tracking, and the wide range of topics covered.
3. The app’s visual appeal, ease of navigation, and accessibility to informative resources were frequently praised.

Conversely, participants identified areas for improvement, including:

1. smartphone compatibility issues,
2. difficulties in symptom tracking,
3. information overload,
4. technical problems,
5. perceived impersonality.

Some participants further explained that the perceived impersonality of the app made them feel it acted as a constant reminder of their illness.

In the third section of the ad-hoc questionnaire, patients evaluated the app’s functionalities using a scale ranging from 1 to 5, reflecting their frequency of use. Among the 31 patients who completed this section, 21 provided ratings for usage frequency. As illustrated in Fig 4, the distribution of scores (1–5) across five app functionalities: Consult Information, Symptom Tracking, Medication Logging, Physical Activity, and Event Schedule.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. Usage frequency of different app’s functionalities by individual patients.

Stacked bar chart illustrating the frequency with which participants used five key features of the Xemio app: (a) Consult Information, (b) Symptom Tracking, (c) Medication Tracking, (d) Physical Activity, and (e) Event Schedule. Scores range from 1 (lowest usage) to 5 (highest usage), represented by different colours.

https://doi.org/10.1371/journal.pone.0342702.g004

Higher scores (4 and 5) are predominantly concentrated in the Consult Information and Symptom Tracking categories, reflecting more frequent use of these functionalities. Conversely, the other categories show a broader spread of scores, indicating greater variability in participant engagement with these features. In the fourth and final section of the ad-hoc questionnaire, patients were asked whether they would recommend the app. Of the 31 respondents, 27 (87.1%) answered affirmatively.

Comparing side effect capture: CTCAE vs. QoL questionnaires

To evaluate the consistency and reliability of symptom reporting across different methods, we analyzed data from both patient-driven and standardized approaches. Symptoms recorded through the non-mandatory self-reporting feature of the Xemio-Research App were compared with those systematically collected via validated questionnaires (ASCAPE design).

Side effects were documented for patients who reported at least one symptom during the study year, using the CTCAE-based application list. These were then compared with side effects recorded through standard questionnaires. Data was available for 39 of the 52 side effects included in the CTCAE-Xemio system, with varying intensity levels. The median intensity was classified as mild (≤1.4), moderate (1.5–2.4), severe (2.5–3.4), or life-threatening (>3.4).

The results are summarized in two tables, one with the CTCAE-reported effects, their median intensity, and frequency, comparing them to equivalent questionnaire items, and another that examines questionnaire responses and reported intensities, contrasting them with data from the app’s symptom tracker to highlight similarities and differences between the two methods.

Table 2 provides an overview of side effects reported at least one time per patient, ranked by frequency, and compares them with data from the EORTC QLQ-C30, EORTC QLQ-BR23, and HADS questionnaires administered at T = 12 of the ASCAPE study. It identifies both exact matches and similar symptoms, offering a detailed comparison that underscores the similarities and distinctions between data collected via the app and traditional questionnaires.

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Correspondence of CTCAE-Xemio Side Effects and QoL Questionnaire Items. All questionnaires’ results are from T = 12. ^a In the EORTC QLQ-C30 there is a generic section dedicated to pain. ^b In the EORTC QLQ-C30 Nausea/vomiting.

https://doi.org/10.1371/journal.pone.0342702.t002

The app provides a more detailed recording of side effects compared to standard questionnaires. CTCAE descriptions often differ substantially from those in validated QoL tools. For example, pain in the QLQ-C30 is broken down into more specific categories such as joint pain, bone pain, and headache, enabling the app to capture key distinctions. Side effects like joint pain (13 patients), tingling (10 patients), muscle weakness (9 patients), and nail problems (9 patients) are frequently reported in the app but are not specifically addressed in traditional questionnaires.

The intensity of symptoms reported in the Xemio-Research list of possible effects was generally rated by patients as high, with 69.23% (27/39) of symptoms rated as more than mild. Specific symptoms had mean intensity scores indicating moderate to severe levels of discomfort: dyspareunia (2.6), joint pain (2.3), vaginal dryness (2.2), and insomnia (2.0). Anxiety and fatigue had median scores of 1.8 and 1.7, respectively, indicating mild to moderate intensity. Other symptoms, such as tingling, muscle weakness, muscle pain, and memory loss, also had median scores of 1.7. Hot flushes had a median intensity score of 1.5, indicating a generally mild intensity.

Table 3 shows the median score for the EORTC QLQ-C30, EORTC QLQ-BR23, TILS, and HADS questionnaires, and the median intensity for the CTCAE. Some side effects reported in the CTCAE-Xemio app have a direct correspondence in the questionnaires, while others are more specific in the app but generalized in the questionnaire. For example, the questionnaire uses Pain as a broad category, whereas the app distinguishes between Headache and Abdominal pain. The “Type match” column indicates whether the correspondence is exact or similar.

Download:

• PNG
  larger image
• TIFF
  original image

Table 3. Median results obtained by questionnaire dimension (T = 12).

https://doi.org/10.1371/journal.pone.0342702.t003

Adherence to Xemio-Research Intervention

Adherence to the Xemio-Research App declined significantly over the study period. Initially, 95.92% of participants used the app in the first three months, but this dropped to 40.82% in the 3–6 month period, 32.65% in the 6–9 month period, and 34.69% in the 9–12 month period. Other studies [10] show that the average duration of mHealth interventions for breast cancer is 12 weeks. Similarly, German reimbursement for digital health applications (DiGA) is set at 30–90 days, showing some alignment with the behavior of users of the Xemio-Research App. Despite the drop in users, those who continued using the app maintained high engagement. This trend is common in digital health interventions [15], where initial enthusiasm wanes, highlighting the need for strategies to sustain long-term engagement, such as: develop gamification features allowing users to participate in daily or weekly challenges that positively reinforce their commitment and motivation and incorporate more effective tracking and real-time monitoring, which could be key to improving adherence and the effectiveness of digital health interventions.

Interpreting the decline in adherence

Beyond the typical waning of engagement observed in digital health tools, several factors may have contributed to the observed decline. First, digital literacy and device heterogeneity (reported by participants as compatibility and usability issues) can create friction that accumulates over time, particularly in older survivors. Second, socioeconomic constraints (e.g., unemployment or labor incapacity) and treatment-related fatigue may reduce the cognitive and emotional bandwidth available for sustained self-tracking. Third, psychological factors such as anxiety or information overload may lead some users to avoid repeated symptom logging if it amplifies illness salience. These mechanisms are consistent with prior reviews on engagement with mHealth in cancer care and eHealth adherence, which highlight usability, burden, and motivational fit as key drivers of sustained use [7,8,11,15,16].

Implications for clinical adoption and sustainability

To enhance long-term use in survivorship care, interventions should (i) simplify key flows for symptom logging and reduce optional complexity; (ii) calibrate just-in-time prompts to patient context to avoid alert fatigue; (iii) offer lightweight motivational features (e.g., streaks or progress check-ins) tied to clinically meaningful behaviors rather than points alone; and (iv) integrate closed-loop feedback (e.g., clinician-acknowledged summaries or tailored advice) that reinforces perceived clinical value. Real-time CTCAE tracking may be most sustainable when embedded in routine follow-up workflows with minimal patient effort and visible clinical payoff.

Usability and participant feedback

The SUS results indicated excellent usability, with an average score of 82.78, suggesting that participants found the app user-friendly and well-integrated. The ad-hoc feasibility questionnaire reinforced these findings, with participants rating the app’s ease of use and symptom-tracking features highly. The mean score for ease of use was 4.35, and 87.1% of respondents would recommend the app, reflecting high satisfaction levels. However, there were some areas for improvement. Functionalities preference varied, with symptom monitoring and consulting information being the most frequently used features.

Symptoms and quality of life reporting

In the validated questionnaires, participants reported a good overall quality of life, with high levels of physical and social functioning. Moderate challenges were observed in areas such as fatigue, insomnia, pain, and systemic therapy side effects, while emotional well-being was generally favourable, with normal levels of depression and mild anxiety. However, notable concerns included sexual functioning and cognitive issues, reflecting the broader impact of treatment.

The comparison between self-reported symptoms in the app and those captured via validated questionnaires reveals important differences. While tools like EORTC QLQ-C30, BR23, and HADS effectively capture patient-reported outcomes across broad domains, they may overlook specific symptoms. The CTCAE complements these tools by allowing precise, real-time tracking of specific symptoms, such as joint pain or tingling, that are not captured in traditional instruments. This integration can enhance the understanding of patient experiences and improve symptom management by bridging the gap between broad assessments and detailed symptom tracking.

Compared with traditional PROMs, the CTCAE-based tracking in the Xemio app provided finer granularity and timeliness. A concise comparison of strengths and limitations is summarized in Table 4. This approach facilitates clinical integration by capturing clinically relevant toxicities as they occur, offering oncologists more actionable data than retrospective questionnaires.

Download:

• PNG
  larger image
• TIFF
  original image

Table 4. Comparative strengths and limitations of PROMs and CTCAE-based symptom tracking.

https://doi.org/10.1371/journal.pone.0342702.t004

These differences may also stem from psychological variations in how patients focus their attention. Traditional questionnaires operate through a “pull mechanism”, systematically prompting patients to recall or recognize symptoms they might not spontaneously report. In contrast, the Xemio-Research App employs a “push mechanism”, allowing patients to independently log symptoms in real time, influenced by their momentary focus and perception of symptom severity [31,32]. Combining these approaches provides a more holistic perspective, enabling comprehensive symptom reviews while capturing patient-prioritized outcomes. This dual strategy broadens the scope of symptom tracking, addressing both immediate and systemic concerns, and supports a nuanced understanding of patient well-being and treatment impact [33].