Co-morbidity and predictors of health status in older rural breast cancer survivors

Abstract

Purpose: More than 66% of the 200,000 newly diagnosed annual breast cancers in the US occurs in women over 55 years. Treatment advances result in excellent survival, yet older breast cancer survivors with co-morbidity may live longer, but not better after cancer. Decline in physical function, increased social isolation, and diminished economic resources increase vulnerability among older women. Rural women represent an underserved population. The purpose is to examine associations between comorbidity and predictors of health status among older rural breast cancer survivors.

Methods: Baseline data of 331 BCS age 55-90 years enrolled in the Rural Breast Cancer Survivors Study. Four surveys were used for data collection. Self-reported prescription medications were used as proxy for co-morbidity. Bivariate tests of association and multivariable recursive partitioning techniques were used for analysis.

Results: Mean number of prescription medication categories reported was 3.68 (SD = 2.3; range = 0-12). Common prescription categories were: anti-hormonal, anti-hypertensive, and cholesterol- reducing agents. 69% was overweight or obese. BMI >31 was significantly associated with both poorer physical and mental health. Multivariate analyses indicated physical health status was predicted by BMI, comorbid conditions, social support, and adverse changes in economic lifestyle. The same variables, with the exception of BMI, were predictors of mental health status.

Conclusions: Assessing co-morbid conditions, mental health status, social support, and economic burden after breast cancer treatment may better inform cancer survivorship care and comprehensive geriatric assessment.

Figure 1

Conditional inference tree model for SF-36 Physical Component Summary (PCS) score (n=331). Notes: Correlation between actual and predicted= 0.47, P<.0001; Predicted SF-36 PCS per terminal node: Node 4= 52.54; Node 5= 45.72; Node 7= 36.70; Node 8= 46.47; Node 9= 38.46.

Figure 2

Conditional inference tree model for SF-36 Mental Component Summary (MCS) score (n=331). Notes: Correlation between actual and predicted= 0.48, P<.0001; Predicted SF-36 MCS per terminal node: Node 2= 40.18; Node 5= 49.06; Node 6= 54.21; Node 8= 50.43; Node 9= 41.53.