Social Determinants of Health and Treatment Receipt Among Privately and Publicly Insured Adults Diagnosed With Cancer in the United States

OBJECTIVES: The ICD-10-CM diagnosis codes ranging from Z55 to Z65 (Z codes) capture individual-level social determinants of health (SDOH); however, they may be under-coded in the health record. This analysis describes the utilization of Z codes among adults diagnosed with cancer and explores the relationship with cancer treatment.

METHODS: We conducted a cross-sectional study using claims data from the 2016-2020 IQVIA PharMetrics^® Plus for Academics and the 2016-2019 SEER-Medicare 5% Cancer File to identify individuals diagnosed with breast, colorectal, lung, or prostate cancer. We calculated the number of individuals with any Z code claims and examined the distribution of specific Z codes. For those diagnosed with metastatic cancer, we used logistic regression models to examine the relationship between Z code utilization and cancer treatment, adjusting for individual-level variables and state-level SDOH measures from PolicyMap.

RESULTS: We identified 35,099 individuals with cancer in the PharMetrics Plus data and 19,625 in the SEER-Medicare data. Individuals with any Z code claims had more comorbidities and a higher prevalence of metastatic cancer compared to those without Z code claims. The proportion of individuals with any Z code claims ranged from 0.6% to 0.9% in the PharMetrics Plus data and from 1.4% to 2.3% in the SEER-Medicare data. The proportion of metastatic patients receiving cancer treatment ranged from 51% to 64% in the PharMetrics Plus data and from 16% to 25% in the SEER-Medicare data. The odds ratios of treatment for individuals with Z code claims were not statistically significant. Residence in states ranked high on average out-of-pocket medical costs and percent of fast-food restaurants was statistically significantly associated with lower odds of treatment using the PharMetrics Plus data.

CONCLUSIONS: The under-coding of Z codes, along with the role of state-level SDOH in the treatment models, underscores the need for additional research on individual-level SDOH and treatment receipt.