The Role of County-Level Contextual Predictors in Multiple Myeloma Mortality

Background: Myeloma is the second most common hematologic malignancy in the US.. Prior studies elucidate that lower socio economic status (SES) and other social predictors such as education, poverty-line, income inequality and unemployment influence all-cause mortality and survival outcomes. We have explored the independent association of county-level contextual predictors on mortality among myeloma patients.

Methods: Using SEER data from 18 registries across the US, with ICD-O-3 and morphologic (9732/3) codes, mortality data was examined from 1973-2012. Analysis was done on individual (race, sex and age of diagnosis) and county-level (SES, high school education, poverty level and unemployment).  Differences in baseline characteristics at diagnosis across racial groups were analyzed using two-sided t-tests and ANOVA tests. Crude incidence and mortality rates and regression models were used for analysis. Statistics were computed using the National Cancer Institute SEER*Stat software, version 8.2.0. and SAS software, version 9.4 (SAS Institute Inc, Cary, NC).

Results: The study population consisted of 89, 867 cases (68701 white, 16364 black and 4,802 other patients). Median age at diagnosis was lower for blacks vs. whites (66 vs. 71 years, P<0.01) and lower for females vs. males (69 vs. 71 years, P<0.01). The age-adjusted incidence rates and mortality rates per 100,000 populations were: blacks- 11.9 (95% CI 11.6, 12.1, P-value<0.05) and 10.2 (95% CI 9.9, 10.4, p<0.05); whites- 5.1 (95% CI 5.0, 5.2) and 4.3 (95% CI 4.3, 4.4); others- 3.7 (95% CI 3.6, 3.9) and 3.0 (95% CI 2.9, 3.1), respectively. ANOVA tests show black and other race are less likely to have received high school education compared to whites (P<0.0001). Similarly, black and other race was significantly associated with living below poverty line and unemployment when compared to whites (P<0.0001). A multivariate model showed gender, race, county-level SES and year of diagnosis to be independently associated with survival outcomes.  Blacks showed reduced mortality (HR = 0.851, P<0.0001) as compared to whites. Conversely, middle and low SES were significantly associated with higher mortality compared to high SES (HR = 1.096 and 1.191; P <0.0001, respectively).

Conclusions: To our knowledge, no previous studies have analyzed the independent association of county-level contextual predictors on mortality. Black patients have higher crude mortality rates compared to whites and other myeloma patients. After controlling for county-level contextual variables; blacks had comparatively lower mortality rates signifying the role of confounders between race and myeloma mortality. Lower SES is an independent risk factor for increased mortality. Further studies examining both, the racial biological disease variants in conjunction with contextual factors, are needed to bridge the gap.