The clinical problem the platform targets is substantial. Chemotherapy-induced myelosuppression, the suppression of bone marrow activity that produces white blood cells, red blood cells and platelets, is among the most frequent adverse events in breast cancer treatment, with a cumulative incidence reported as high as 80 to 90 percent among patients receiving chemotherapy. When myelosuppression strikes, infection risk rises, anemia develops, and bleeding can occur, all of which can compromise the delivery of chemotherapy, undermine its therapeutic efficacy, and degrade quality of life. Breast cancer itself remains the most commonly diagnosed malignancy among women worldwide: GLOBOCAN 2022 estimates from the International Agency for Research on Cancer attribute roughly 2.30 million new cases to the disease in 2022, some 11.5 percent of all newly diagnosed cancers, along with 666,103 deaths, or 6.8 percent of all cancer-related deaths. Projections cited in the study suggest that by 2040 more than 3 million new cases and over 1 million deaths will occur annually.

Despite this burden, the researchers found that routine management of myelosuppression still depends heavily on hospital-based follow-up and passive patient reporting, a model that produces infrequent monitoring, poor adherence, limited individualization, and inefficient use of healthcare resources. The consequences can be serious: delayed recognition of clinical deterioration, postponed interventions, and ultimately suboptimal outcomes. The scale of the gap became clear in the team’s own survey data, in which 92.81 percent of patients reported having forgotten to recheck their blood tests and 89.22 percent had experienced myelosuppression. Mobile health technologies have been increasingly adopted in patient education, nursing care, disease prevention, and treatment support, and prior interventions have shown potential to deliver individualized support, provide early warnings of adverse events, improve adherence, and reduce acute complications. Yet existing mHealth tools have struggled to sustain user engagement, enable intelligent monitoring, and deliver personalized support, challenges that are particularly acute in low- and middle-income settings where linguistic, health-system, and technological barriers complicate implementation.

To build something matched to this specific need, the team followed a user-centered design approach across three sequential stages: a needs assessment among patients and healthcare professionals, expert panel consultations to refine core modules and interaction logic, and pilot testing to assess technical functionality and initial usability. The needs assessment used two tailored questionnaires, distributed both online and on paper to reduce access barriers for participants less familiar with digital tools. Of 346 questionnaires distributed to patients, 306 valid responses were returned, a valid response rate of 88.44 percent. The respondents, all women, ranged in age from 25 to 80 years with a mean of 50.62 years, and 83.66 percent had a junior high school education or below. Among them, 88.56 percent indicated a need for the platform and 82.35 percent expressed willingness to use it. The highest-rated functional demands, each scored near the top of a five-point scale, included access to disease-related knowledge and coping strategies at any time, knowing whether complete blood count results were normal, receiving reminders before each follow-up blood test, and having a quick way to contact a doctor and return to the hospital when symptoms became severe.

Healthcare professionals were equally receptive. Of 180 questionnaires distributed, 164 valid responses came back, a 91.11 percent valid response rate, from staff with a mean age of 29.91 years and a mean of 10.62 years of work experience. A striking 93.90 percent considered the platform necessary and 85.37 percent expressed willingness to use it in clinical practice, with physicians reporting greater willingness than nurses. Their top-rated needs included promptly contacting patients who had developed severe myelosuppression to arrange hospital return, reminding patients to review their complete blood counts at appropriate time points, visualizing patients’ longitudinal blood count data, and providing targeted guidance for different grades of myelosuppression. The researchers interpret this divergence in willingness through the lens of clinical roles: physicians, as key decision-makers, require timely updates on patient conditions, while nurses, though central to patient education and follow-up care, have limited prescribing authority under China’s medical system, which may shape how they prioritize decision-support functions. Patients’ willingness, meanwhile, may reflect their awareness of myelosuppression and individual health literacy.

The design was then stress-tested by a panel of thirteen multidisciplinary experts spanning breast surgery, nursing, medical informatics, psychology, and nutrition, whose authority coefficient averaged 0.88 on a scale where values ranged from 0.75 to 1.00. Two rounds of structured meetings produced concrete changes. Experts warned that grading systems should not reserve intervention for severe cases alone, leading to an alert function that classifies routine blood test data into Grades I through IV, with each level triggering algorithm-based automatic grading and self-care management recommendations grounded in evidence-based medicine. Questionnaires were restricted to a completion window of three days before or after the preset time to prevent early or late submissions. Every alert-triggered intervention was logged in a custom patient follow-up record form for later analysis. Blood count values for neutrophils, white blood cells, hemoglobin, and platelets were rendered as line charts for easier comprehension, and the system backend began automatically setting the units for each laboratory indicator so patients knew which value to enter. Three reminders per blood test were judged excessive and cut to two, delivered at 20:00 the night before and 14:00 on the day of the re-examination, stopping automatically once the test was completed.

Technically, the platform was first developed in a Windows 10 environment using Apache 2.4, MySQL 8, and PHP 7.2, then deployed on a CentOS 8 server with Nginx 1.24.0 and PHP 7.3 for stability and compatibility. The back end runs on PHP, the front end on HTML, CSS, and JavaScript, with SQL handling data queries under a front-end/back-end separation architecture designed to enhance modularity, maintainability, and scalability. Users register by scanning a QR code, after which administrators assign role-based access permissions. The patient side comprises five modules: Health Education, Health Interaction, Message Center, Health Records, and User Center, supporting education, communication, and the recording of assessment scales, laboratory results, health status, and medication use. The clinician side adds routine management, permission control, intelligent alerts, chemotherapy tracking, and SMS logging. A Health Management module tracks daily indicators including temperature, oral mucosa status, skin integrity, and bowel function, while the Alert module maps uploaded blood counts to WHO grading thresholds, with Grade 0 remaining unflagged and Grades I, II, III, and IV triggering green, blue, yellow, and red alerts respectively, each paired with stratified management guidance.

The pilot, conducted from February 1 to March 8, 2024, in the Breast Surgery Department of a tertiary Grade A hospital in Guizhou Province, enrolled eight patients beginning their first chemotherapy cycle, with a mean age of 43.9 years. Usability was measured with the validated Chinese version of the mHealth App Usability Questionnaire, a 21-item instrument covering system ease of use and satisfaction, information architecture, and perceived usefulness, each rated on a seven-point scale where lower scores mean higher usability. The domain scores were 3.72 for ease of use and satisfaction, 3.90 for information architecture, and 3.27 for perceived usefulness, yielding an overall 3.63. Because lower is better, perceived usefulness scored best while information architecture scored worst, suggesting that information organization and interface layout warrant further refinement. Server logs showed 64 SMS reminders sent, three complete blood count records uploaded, and a mean of seven user sessions per participant, ranging from 3 to 14. Across the three uploaded records, the system generated 11 indicator-level alerts: five green Grade I, two blue Grade II, and four yellow Grade III, with manual review confirming 11 out of 11 concordance with the prespecified WHO grading thresholds.

The authors position BCI-MP against earlier digital symptom systems, noting that many conventional alert tools focus on early prevention with predefined indicators; one tablet-based system for patients with hematologic malignancies generated alerts only for febrile neutropenia, while an interactive symptom-triggered application lacked CIM-specific functionality. By contrast, BCI-MP uses routine blood test data to drive algorithm-based grading and alert generation while integrating automated reminders and individualized recommendations, linking objective laboratory data with remote monitoring and stratified guidance. The team is candid about the limitations: the pilot sample was small, the absence of a control group precludes conclusions about clinical effectiveness, routine blood count results were not automatically synchronized and required patient-initiated uploading, which explains why only three records were captured and highlights adherence to monitoring tasks as an implementation challenge, and privacy and data security within the third-party WeChat ecosystem were not formally evaluated. Heavy clinical workloads may also limit timely professional responses outside working hours. The authors call for larger controlled samples, longer follow-up, and mixed-methods evaluation of usability, adherence, effectiveness, feasibility, and data security before broader clinical integration.

Subject of Research: Development and pilot testing of a WeChat-embedded mobile health platform for managing chemotherapy-induced myelosuppression in breast cancer patients

Article Title: Designing and optimizing an intelligent health informatics management platform for patients undergoing chemotherapy for breast cancer: from needs assessment to pilot testing

Article References: Yu, X., Zhang, L., Yang, L., Chen, S., & Wang, X. (2026). Designing and optimizing an intelligent health informatics management platform for patients undergoing chemotherapy for breast cancer: from needs assessment to pilot testing. Frontiers in Oncology, 16, Article 1865670. https://doi.org/10.3389/fonc.2026.1865670

Image Credits: AI Generated

DOI: 10.3389/fonc.2026.1865670

Keywords: breast cancer, chemotherapy-induced myelosuppression, mHealth, WeChat, remote monitoring, mobile health, complete blood count, user-centered design, health informatics, patient self-management, automated alerts, pilot study