Volume 53 | Number 5 | October 2018

Abstract List

Jessica K. Roydhouse B.A., M.P.H., Roee Gutman Ph.D., Nancy L. Keating M.D., M.P.H., Vincent Mor Ph.D., Ira B. Wilson M.D., M.Sc.


To assess the association of proxy‐specific covariates with proxy‐reported patient cancer care experience, quality rating, and quality of life.

Data Sources/Study Setting

Secondary analysis of data from the Cancer Care Outcomes Research and Surveillance (Can) study.

Study Design

Cross‐sectional observational study. The respondents were proxies for patients with incident colorectal or lung cancer.

Data Collection/Extraction Methods

Analyses used linear regression models and adjusted for patient sociodemographic and clinical characteristics. Outcomes included patients’ experiences with medical care, nursing care, and care coordination, overall quality ratings, and physical and mental health, all scored on 0–100 scales (0 = worst, 100 = best). Independent variables included the proxy's relationship with the patient and engagement in patient care.

Principal Findings

Of 1,011 proxies, most were the patient's spouse (50 percent) or child (36 percent). Although most proxies (66 percent) always attended medical visits, 3 percent reported never attending. After adjustment, on average children reported worse experiences and poorer quality care than spouses (4–9 points lower across outcomes). Proxies who never attended medical visits reported significantly worse medical care (−11 points, 95 percent  = −18 to −3) and care coordination (−13 points, 95 percent  = −20 to −6).


Collecting data on proxy engagement in care is warranted if proxy responses are used.