Development and internal validation of a risk stratification model for high-intensity psychosocial service use among women living with and beyond breast cancer
- 1Department of Economics and Business Administration, Ariel University, P.O.B. 3, Kiryat HaMada, Ariel, 40700 Israel
- 2Department of Nursing, School of Health Sciences, Ariel University, Ariel, Israel
- 3Department of Health Systems Management, School of Health Sciences, Ariel University, Ariel, Israel
Abstract
Purpose: Breast cancer survivorship is frequently accompanied by persistent psychosocial distress and heterogeneous supportive care needs. Although distress screening is widely recommended, limited evidence exists on pragmatic tools to identify survivors most likely to require high-intensity psychosocial engagement. Because breast cancer survivorship is a major women’s health issue, tools that support more equitable and timely identification of women with greater supportive care needs may have direct relevance for women-centered service planning. We aimed to develop and internally validate a risk stratification model for high-intensity psychosocial service use and assess the incremental value of a distress index.
Methods: We conducted a cross-sectional risk stratification study among 411 women living with or after breast cancer who completed an anonymous online self-report questionnaire disseminated through community-based and professional recruitment channels in Israel. High-intensity service use was defined as the upper tertile of cumulative engagement. A base logistic regression model included demographic and clinical predictors; an expanded model added a composite distress index. Discrimination and calibration were evaluated using stratified 5-fold cross-validation and bootstrap internal validation.
Results: The base model showed moderate discrimination (cross-validated AUC = 0.675; bootstrap median AUC = 0.706, 95% CI 0.687–0.715). Adding the distress index improved discrimination (cross-validated AUC = 0.748; bootstrap median AUC = 0.778, 95% CI 0.763–0.785) and reduced model error, with the Brier score improving from 0.196 to 0.180. Lower income and shorter time since diagnosis were independently associated with high psychosocial service use.
Conclusions: A pragmatic risk stratification model demonstrated moderate ability to identify survivors likely to engage intensively with psychosocial services. Incorporating a brief distress index meaningfully enhanced model discrimination and classification performance.
Implications for cancer survivors: Risk-based targeting increased positive predictive value to 0.60–0.69 in high-risk strata, suggesting that proactive outreach based on predicted risk could more efficiently concentrate psychosocial resources among survivors most likely to require sustained support. These findings may also inform women’s health programs seeking to improve supportive care access for women facing psychosocial or socioeconomic vulnerability after breast cancer.
Supplementary Information: The online version contains supplementary material available at https://doi.org/10.1186/s12905-026-04726-2.
Related
- Development and Internal Validation of a Preoperative Risk Stratification Model for Difficult Laparoscopic Cholecystectomy in Acute Cholecystitis
- Development and Prospective Internal-Validation Protocol of the SpinePROTECT Score: A Digital Biopsychosocial Risk Stratification Framework for Spine Surgery.
- Factors associated with postoperative ileus after laparoscopic radical resection for colorectal cancer and development and internal validation of a postoperative risk-stratification nomogram
- Development and Internal Validation of the MELD–ERT Score for Mortality Risk Stratification After Liver Transplantation
- Development and internal validation of a simple risk stratification score for unplanned emergency department revisits within 24 hours after discharge: A retrospective observational study
BibTeX
@article{gonen2026development,
title = {{Development and internal validation of a risk stratification model for high-intensity psychosocial service use among women living with and beyond breast cancer}},
author = {Limor Dina Gonen and Sharon Barak and Riki Tesler and Avi Zigdon},
journal = {BMC Women's Health},
year = {2026},
month = jul,
doi = {10.1186/s12905-026-04726-2},
url = {https://doi.org/10.1186/s12905-026-04726-2},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Development and internal validation of a risk stratification model for high-intensity psychosocial service use among women living with and beyond breast cancer
AU - Limor Dina Gonen
AU - Sharon Barak
AU - Riki Tesler
AU - Avi Zigdon
PY - 2026
DA - 2026/07/30
JO - BMC Women's Health
DO - 10.1186/s12905-026-04726-2
UR - https://doi.org/10.1186/s12905-026-04726-2
AB - Purpose: Breast cancer survivorship is frequently accompanied by persistent psychosocial distress and heterogeneous supportive care needs. Although distress screening is widely recommended, limited evidence exists on pragmatic tools to identify survivors most likely to require high-intensity psychosocial engagement. Because breast cancer survivorship is a major women’s health issue, tools that support more equitable and timely identification of women with greater supportive care needs may have direct relevance for women-centered service planning. We aimed to develop and internally validate a risk stratification model for high-intensity psychosocial service use and assess the incremental value of a distress index. Methods: We conducted a cross-sectional risk stratification study among 411 women living with or after breast cancer who completed an anonymous online self-report questionnaire disseminated through community-based and professional recruitment channels in Israel. High-intensity service use was defined as the upper tertile of cumulative engagement. A base logistic regression model included demographic and clinical predictors; an expanded model added a composite distress index. Discrimination and calibration were evaluated using stratified 5-fold cross-validation and bootstrap internal validation. Results: The base model showed moderate discrimination (cross-validated AUC = 0.675; bootstrap median AUC = 0.706, 95% CI 0.687–0.715). Adding the distress index improved discrimination (cross-validated AUC = 0.748; bootstrap median AUC = 0.778, 95% CI 0.763–0.785) and reduced model error, with the Brier score improving from 0.196 to 0.180. Lower income and shorter time since diagnosis were independently associated with high psychosocial service use. Conclusions: A pragmatic risk stratification model demonstrated moderate ability to identify survivors likely to engage intensively with psychosocial services. Incorporating a brief distress index meaningfully enhanced model discrimination and classification performance. Implications for cancer survivors: Risk-based targeting increased positive predictive value to 0.60–0.69 in high-risk strata, suggesting that proactive outreach based on predicted risk could more efficiently concentrate psychosocial resources among survivors most likely to require sustained support. These findings may also inform women’s health programs seeking to improve supportive care access for women facing psychosocial or socioeconomic vulnerability after breast cancer. Supplementary Information: The online version contains supplementary material available at https://doi.org/10.1186/s12905-026-04726-2.
SN - 1472-6874
ER -