Measuring the Spillover Effect of Cancer on Family Members' Health: A US-Based Empirical Analysis
Author(s)
Suning Zhao, MPH1, Ruixi Yu, MA1, Xin Hu, PhD2, Changchuan Jiang, MD, MPH3, Xiaoyu Pan, MD4, Boshen Jiao, MPH, PhD1;
1University of Southern California, Alfred E. Mann School of Pharmacy and Pharmaceutical Sciences, Los Angeles, CA, USA, 2Emory University School of Medicine, Department of Radiation Oncology, Atlanta, GA, USA, 3UT Southwestern Medical Center, Division of Hematology and Oncology, Department of Internal Medicine, Dallas, TX, USA, 4UT Southwestern Medical Center, Department of Physical Medicine and Rehabilitation, Dallas, TX, USA
1University of Southern California, Alfred E. Mann School of Pharmacy and Pharmaceutical Sciences, Los Angeles, CA, USA, 2Emory University School of Medicine, Department of Radiation Oncology, Atlanta, GA, USA, 3UT Southwestern Medical Center, Division of Hematology and Oncology, Department of Internal Medicine, Dallas, TX, USA, 4UT Southwestern Medical Center, Department of Physical Medicine and Rehabilitation, Dallas, TX, USA
Presentation Documents
OBJECTIVES: Cancer affects not only patients' health but also the well-being of their family members, imposing substantial physical, emotional, and financial burden. This effect intensifies as patient health declines. Understanding and incorporating spillover effects into cost-effectiveness analyses (CEAs) is crucial for accurately assessing the value of cancer treatments. However, the lack of systematic data on cancer-related spillover effects remains a significant barrier. This study aims to estimate health-related quality of life (HRQoL) loss among family members as a function of patients' HRQoL decline.
METHODS: We leveraged data from the Medical Expenditure Panel Survey from 2008 to 2019 to create a panel dataset of cancer patients and their family members. Families, typically including parents and children, without cancer patients served as a control group. HRQoL outcomes were measured using EQ-5D utility values derived via the time trade-off method. A zero-one inflated beta regression model was employed to assess the association between family members’ HRQoL, cancer status, patient EQ-5D scores, and age, adjusting for other demographic characteristics. The fitted regression model was then used to predict family HRQoL loss across varying levels of patient HRQoL loss and age.
RESULTS: 10,282 family members of cancer patients exhibited a mean EQ-5D utility value 0.016 [p-value < 0.001] lower than that of the control group. Spillover effects increased as patients’ HRQoL loss arouse. For example, when a patient’s HRQoL loss was 0.1, the corresponding family member’s loss was estimated at 0.044[95% CI: 0.0433-0.0441]. When the patients’ HRQoL loss reached 0.2, the family member’s loss increased to 0.060[95% CI: 0.0594-0.0603]. These effects increased with patients' age, except for a substantial impact observed under age 30.
CONCLUSIONS: This study provides key estimates for integrating family spillover effects into CEAs for cancer interventions, enabling a more comprehensive evaluation of their societal value.
METHODS: We leveraged data from the Medical Expenditure Panel Survey from 2008 to 2019 to create a panel dataset of cancer patients and their family members. Families, typically including parents and children, without cancer patients served as a control group. HRQoL outcomes were measured using EQ-5D utility values derived via the time trade-off method. A zero-one inflated beta regression model was employed to assess the association between family members’ HRQoL, cancer status, patient EQ-5D scores, and age, adjusting for other demographic characteristics. The fitted regression model was then used to predict family HRQoL loss across varying levels of patient HRQoL loss and age.
RESULTS: 10,282 family members of cancer patients exhibited a mean EQ-5D utility value 0.016 [p-value < 0.001] lower than that of the control group. Spillover effects increased as patients’ HRQoL loss arouse. For example, when a patient’s HRQoL loss was 0.1, the corresponding family member’s loss was estimated at 0.044[95% CI: 0.0433-0.0441]. When the patients’ HRQoL loss reached 0.2, the family member’s loss increased to 0.060[95% CI: 0.0594-0.0603]. These effects increased with patients' age, except for a substantial impact observed under age 30.
CONCLUSIONS: This study provides key estimates for integrating family spillover effects into CEAs for cancer interventions, enabling a more comprehensive evaluation of their societal value.
Conference/Value in Health Info
2025-05, ISPOR 2025, Montréal, Quebec, CA
Value in Health, Volume 28, Issue S1
Code
EE515
Topic
Economic Evaluation
Disease
SDC: Oncology