Development and Validation of a Predictive Model for Anticipatory Grief in Family Caregivers of Cancer Patients: Based on LASSO‐Logistic Regression Model

BACKGROUND: Anticipatory grief is a significant emotional challenge for family caregivers of cancer patients, yet its early identification remains limited by subjective assessments and a lack of predictive tools. This study aimed to develop and validate a predictive model for anticipatory grief among family caregivers of cancer patients in China. METHODS: A multicenter cross-sectional study was conducted from February to October 2023, involving 642 family caregivers of lung and breast cancer patients from two tertiary hospitals in Liaoning Province, China. Latent Profile Analysis (LPA) classified caregivers into anticipatory grief risk categories based on the Anticipatory Grief Scale. LASSO-logistic regression was used to identify predictors and construct a predictive model, which was validated using discrimination (AUC), calibration (Hosmer-Lemeshow test), and clinical utility (Decision Curve Analysis). A web-based nomogram was developed for practical application. RESULTS: The mean anticipatory grief score was 72.44 ± 18.49, with LPA identifying three profiles: low (54.52%), moderate (30.53%), and high (14.95%) anticipatory grief. Seven predictors were identified: caregiver education level, monthly income, physical condition, caregiving duration, and patient cancer type, employment status, and time since diagnosis. The model showed good discrimination (AUC: 0.769 training, 0.671 validation), calibration (P = 0.095 training, P = 0.801 validation), and clinical utility (net benefit at 34%-62% threshold). The web-based tool is accessible at https://nomogrameofag.shinyapps.io/dynnomapp/. CONCLUSIONS: This study developed a predictive model for anticipatory grief, identifying key risk factors and providing a practical tool for healthcare providers to identify high-risk caregivers. The findings support targeted interventions to enhance caregiver well-being and patient care quality, though future research should expand cancer types and incorporate qualitative insights for broader applicability.