Background: Randomized controlled trials (RCTs) are the gold standard for evaluating interventions in oncology, but reporting can be subject to "spin"-presenting results in ways that mislead readers about true efficacy.
Objective: This study aimed to investigate whether large language models (LLMs) could provide a standardized approach to detect spin, particularly in the conclusions, where it most commonly occurs.
Methods: We randomly sampled 250 two-arm, single-primary end point oncology RCTs from 7 major medical journals published between 2005 and 2023. Two authors independently annotated trials as positive or negative based on whether they met their primary end point. Three commercial LLMs (GPT-3.5 Turbo, GPT-4o, and GPT-o1) were tasked with classifying trials as positive or negative when provided with (1) conclusions only; (2) methods and conclusions; (3) methods, results, and conclusions; or (4) title and full abstract. LLM performance was evaluated against human annotations. Afterward, trials incorrectly classified as positive when the model was provided only with the conclusions but correctly classified as negative when provided with the whole abstract were analyzed for patterns that may indicate the presence of spin. Model performance was assessed using accuracy, precision, recall, and F1-score calculated from confusion matrices.
Results: Of the 250 trials, 146 (58.4%) were positive, and 104 (41.6%) were negative. The GPT-o1 model demonstrated the highest performance across all conditions, with F1-scores of 0.932 (conclusions only; 95% CI 0.90-0.96), 0.96 (methods and conclusions; 95% CI 0.93-0.98), 0.98 (methods, results, and conclusions; 95% CI 0.96-0.99), and 0.97 (title and abstract; 95% CI 0.95-0.99). Analysis of trials incorrectly classified as positive when the model was provided only with the conclusions revealed shared patterns, including absence of primary end point results, emphasis on subgroup improvements, or unclear distinction between primary and secondary end points. These patterns were almost never found in trials correctly classified as negative.
Conclusions: LLMs can effectively detect potential spin in oncology RCT reporting by identifying discrepancies between how trials are presented in the conclusions vs the full abstracts. This approach could serve as a supplementary tool for improving transparency in scientific reporting, although further development is needed to address more complex trial designs beyond those examined in this feasibility study.
Background: Hyperthermic intraperitoneal chemotherapy (HIPEC) has been integrated into the management of gastric cancer (GC) as a combined approach for addressing peritoneal metastasis, serving both prophylactic and therapeutic roles following GC surgery. The pivotal decision regarding HIPEC administration typically arises intraoperatively, creating a complex clinical scenario where family caregivers must act as surrogate decision-makers under substantial time constraints. This decision-making process proves particularly challenging due to limited understanding of the procedure's risk-benefit profile and long-term outcomes among nonmedical surrogates, challenges often exacerbated by the acute stress of the surgical environment.
Objective: This qualitative study aims to explore how family caregivers of patients with GC navigate the HIPEC decision-making process, specifically examining the facilitators, challenges, and the role of information acquisition that shape the shared decision-making mode.
Methods: This study adopted a qualitative approach using semistructured interviews; 15 family caregivers of patients with GC in a major tertiary hospital in Guangxi Province were selected as research objects through a purposive sampling method. Participants were asked to comment on their experience of surrogate decision-making for the HIPEC process. The Colaizzi 7-step method was used to analyze and summarize the themes.
Results: The mean age of the 15 participants was 39.8 (SD 13.29, range 20-68) years, and all patients were on average aged 56.7 (SD 10.78, range 36-74) years. The relationship to the patient was distributed as follows: 33% (5/15) spouses, 60% (9/15) children, and 6% (1/15) other relatives. Four major themes emerged from the data analysis: (1) shared decision-making participation mode (doctor-led passive decision-making and doctor-family shared decision-making); (2) decision-information sources (decision-making information came from medical-care personnel, decision-making information came from the internet, and decision-making information came from acquaintances); (3) challenges in the decision-making process (financial burden and anticipated therapeutic efficacy); and (4) facilitator in the decision-making process (positive health beliefs and cultural dimensions of perceived responsibility: a Confucian perspective).
Conclusions: HIPEC decision-making by family caregivers of patients with GC was primarily passive decision-making, and many obstacles and facilitators were encountered in the process. Medical staff should share information and encourage and guide family caregivers to participate in the decision-making process through decision assistance or decision support.
Background: Digital patient-reported outcome (ePRO) tools have the potential to enhance supportive care in oncology and support timely and accurate identification of patients' needs.
Objective: This study aimed to develop, implement, and evaluate a user‑friendly, web‑based digital screening tool at a German Comprehensive Cancer Center (CCC) that systematically and efficiently assesses cancer patients' supportive care needs and enables direct referral to appropriate supportive services through seamless integration with the hospital information system.
Methods: An interdisciplinary team collaborated with the Information Technology (IT) department and the company CANKADO, an ePRO provider, to create a 14-item digital questionnaire. The tool incorporated validated instruments, such as the Distress Thermometer, the Nutrition Risk Screening (NRS), and a short form of the Integrated Palliative care Outcome Scale (IPOS), aligned with German Cancer Society certification criteria. Patients accessed the questionnaire via Quick Response (QR) codes. Screening results were automatically transferred to the HIS, where supportive care requests (SCRs) were generated automatically if indicated.
Results: Between June 2024 and May 2025, a total of 8,855 QR codes were generated. 4,909 of the questionnaires were complete and valid for analysis. This information produced 3,324 SCRs. Digital screenings resulted in a SCR in 22.4 % of cases for psycho-oncology, 18.7 % for nutrition, and 27.6 % for palliative care. The digital screening maintained or slightly improved screening rates compared to prior methods.
Conclusions: The implementation of a digital supportive care screening was feasible and effective within the CCC setting. Future efforts focus on overcoming barriers for patients with limited digital access or capabilities to ensure delivery of equitable supportive care delivery.
Background: Individuals undergoing cancer treatment often face a high symptom burden that impairs quality of life. Exercise and mind-body therapies have been shown to reduce symptoms but are underused. We developed a digital exercise and mind-body therapy program that effectively reduces symptoms while overcoming in-person delivery barriers. Understanding patient experiences can inform treatment mechanisms and guide digital health interventions in cancer care.
Objective: This study aimed to explore patient experiences with Integrative Medicine at Home (IM@Home), a 12-week live digital program delivering exercise and mind-body therapies tailored to the needs of individuals undergoing cancer treatment.
Methods: This qualitative study was embedded in a randomized clinical basket trial (NCT05053230) evaluating the effects of IM@Home versus enhanced usual care on symptoms and acute health care utilization in adults with solid tumors undergoing active treatment and experiencing moderate or greater fatigue. Using maximum variation sampling, 20 participants were selected for semistructured interviews. Interviews explored participants' experiences with the program, its impact on outcomes, unmet needs, and suggestions for improvement. Transcripts were analyzed using a combined inductive and deductive thematic analysis.
Results: Twenty participants (mean age 63, SD 9.6 years; 18/20, 90% female) were interviewed. Five major themes emerged: (1) IM@Home alleviated symptom burden and supported symptom self-management; (2) IM@Home facilitated social support and information exchange; (3) IM@Home offered a flexible, tailored program in a group setting; (4) IM@Home facilitated accessible, cost-effective support; and (5) recommendations for program enhancement. IM@Home was perceived as an accessible, flexible, and supportive program that promoted physical and emotional well-being during treatment.
Conclusions: IM@Home demonstrates a promising model for delivering integrative supportive care during cancer treatment. Findings highlight patient-valued features such as real-time guidance, tailored content, and community support. These insights can inform future implementation, integration into clinical care, and efforts to enhance digital mind-body interventions in oncology.
Trial registration: ClinicalTrials.gov NCT05053230; https://www.clinicaltrials.gov/study/NCT05053230.
International registered report identifier (irrid): RR2-10.1038/s41746-024-01387-z.
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：