Clinical Prediction Models of Severe Cancer-Related Fatigue in Cervical Cancer Patients and Effectiveness of Mindfulness-Based Stress Reduction: A Randomized Controlled Trial

Objective: To build a clinical prediction models (CPMs) for severe CRF of cervical cancer (CC) patients, and conduct online 8-week mindfulness-based stress reduction (MBSR).

Design: Single-centre, cross-sectional, randomized controlled trial.

Setting: Gynecological clinic of a tertiary hospital in Shenyang. Population 284 patients, and 100 patients with severe CRF included in randomized controlled trial.

Methods: Patients were divided into construction group and validation group to verify the accuracy of CPMs. All patients participated in online 8-week MBSR and were followed up for 3 and 6 months. Main outcome measures The primary outcome were the accuracy of the CPMs and effectiveness of MBSR on CRF at 6 month. Secondary outcomes included the influencing factors of CRF, effectiveness of MBSR on sense of coherence, coping style, and perceived social support.

Results: In the CPMs: the C-Index was 0.921; the sensitivity was 0.821; the specificity was 0.900; the accuracy was 0.857; the AUC was 0.916; the area under the ROC curve was greater than 0.8; the calibration curve fitted well (P<0.05); and the net benefit of applying the model to most thresholds was good in DCA. At 6 months, CRF was from 37.70 to 31.25; SOC was from 55.07 to 64.17; ways to face was from 19.20 to 20.95; and ways to yield was from 12.10 to 8.90 in MBSR group (P<0.01).

Conclusion: The CPMs could be used as a prediction and evaluation tool for severe CRF, and the online 8-week MBSR can effectively improve the CRF of CC patients.