Urban-Rural and Socioeconomic Differences in Patient Knowledge and Perceptions of Genomic Tumor Testing

Abstract

Purpose: Social determinants of health, such as rurality, income, and education, may widen health disparities by driving variation in patients' knowledge and perceptions of medical interventions. This effect may be greatest for medical technologies that are hard to understand and less accessible. This study explored whether knowledge and perceptions (expectations and attitudes) of patients with cancer toward large-panel genomic tumor testing (GTT), an emerging cancer technology, vary by patient rurality independent of other socioeconomic characteristics (education and income).

Methods: Patients with cancer enrolled in a large precision oncology initiative completed surveys measuring rurality, sociodemographic characteristics, and knowledge and perceptions of GTT. We used multivariable linear models to examine differences in GTT knowledge, expectations, and attitudes by patient rurality, education, and income level. Models controlled for age, sex and clinical cancer stage and type.

Results: Rural patients had significantly lower knowledge of GTT than urban patients using bivariate models (P = .025). However, this association disappeared when adjusting for education and income level: patients with lower educational attainment and lower income had lower knowledge and higher expectations (P ≤ .002), whereas patients with higher income had more positive attitudes (P = .005). Urban patients had higher expectations of GTT compared with patients living in large rural areas (P = .011). Rurality was not associated with attitudes.

Conclusion: Patients' education and income level are associated with knowledge, expectations, and attitudes toward GTT, whereas rurality is associated with patient expectations. These findings suggest that efforts to promote adoption of GTT should focus on improving knowledge and awareness among individuals with low education and income. These differences may lead to downstream disparities in GTT utilization, which should be explored in future research.

FIG 1.

Unadjusted bivariate relationships between rurality, education, and income and knowledge, expectations, and attitudes. The solid orange lines are regression lines fitted for the participants (education and income level are treated as discrete numeric variables). No regression lines were fit for rurality because it is a categorical variable with three groups. Each colored dot represents a participant's response; positions are slightly adjusted randomly to avoid overlapping. The white dots represent group means. *Bivariate statistically significant models. Att., attitudes; Avg., average; Bach, bachelor's; Coll, college; Expect., expectations; Grad, graduate; HS, high school; K, $1,000; L rural, large rural; NS, not significant; Pos., positive; Sm/iso rural, small and isolated rural.