Socioeconomic Status Bears Little Impact on Survival Rates

Primary cancers of the lung, esophagus, and pancreas—all of which have a poor prognosis—account for >35% of all cancer deaths in the United States annually. Although the survival rates of these cancers have not improved over time, the intensity of treatment has increased, according to data presented at the 2011 ASCO conference in Chicago, Illinois.

In the study, the authors used SEER-Medicare data from 1992 to 2002 with follow-up to 2005 to examine cancer outcomes and aggressiveness of treatment based on patients’ socioeconomic status (SES). The analysis was limited to fee-for service Medicare beneficiaries aged >65 years, as more than two-thirds of patients with lung, esophageal, and pancreatic cancers are Medicare beneficiaries.

The lead author of the study, Sandra L. Wong, MD, assistant professor of surgery for the Division of Surgical Oncology at the University of Michigan Health System in Ann Arbor, explained the genesis of the study by noting, “These poor-prognosis cancers account for a disproportionate share of cancer mortality and cancer costs. We hypothesized that socioeconomic status adversely affects treatment aggressiveness and outcomes in poor-prognosis cancers.”

Using methodology developed by Diaz-Roux and colleagues, Wong’s team of investigators constructed summary measures of SES based on US ZIP code data, using 6 variables from 2000 Census data. The primary variables noted were wealth/income, education, and employment. Each Medicare beneficiary in the study cohort was assigned an SES summary score based on his/her ZIP code of residence in the Medicare files.

Sandra L. Wong, MD

The study included 68,167 patients with lung cancer, 4350 patients with esophageal cancer, and 12,779 patients with pancreatic cancer. Median household incomes by SES group were $28,247 (SES 1), $37,316 (SES 2), $44,198 (SES 3), $53,895 (SES 4), and $73,365 (SES 5).

The study found that the lowest SES patients were more likely to be black, more likely to require urgent/emergent admissions, more likely to be treated at low-volume hospitals, and “much more” likely to be treated at nonteaching hospitals. These findings were similar across all 3 of the cancer cohorts.

“We connected our analyses on a patient level and on a hospital level,” said Wong. “For the hospital-level analysis, patients were attributed to a hospital based on where the plurality of patient admissions took place or where hospital-based services were performed. In the uncommon [event] where neither of these two situations were seen, we assigned hospitals based on where the beneficiary’s primary care physicians tended to admit their patients.”

Wong and her team evaluated cancer staging and evaluation and found that most of the imaging conducted involved CT scans. There was some increased specialized testing (ie, PET scans) in the SES 5 group for lung (n = 1396) and esophageal cancers (n = 371), but not for pancreatic cancer.

The investigators also examined patterns of cancer-directed treatments. In the highest SES group (SES 5), there was higher utilization of cancer-directed treatments across all 3 of the cancer types (lung, esophageal, pancreatic) and across all modalities of care (ie, cancer-directed surgery, chemotherapy, and radiation).

The results showed that a greater number of patients in the lowest SES groups (specifically SES 1) received no cancer-directed treatment compared with those in the highest SES group (SES 5). For example, 60% of the patients with pancreatic cancer in the SES 1 group had no cancer-directed treatment compared with 50% of the patients with pancreatic cancer in the SES 5 group.

The hazard ratio (HR) for mortality between the SES 5 and SES 1 groups were marginally significant in the lung cancer group when adjusted for patient characteristics (95% confidence interval; HR = 1.02 [0.96-1.10]), but were not significant for esophageal and pancreatic cancers.

Wong noted that there was a variation in cancer treatment across SES groups. The lowest SES patients were more likely to receive no cancer-directed treatments, whereas the highest SES patients were more likely to undergo cancer-directed surgery, chemotherapy, and/or radiation.

Most patients in the highest SES groups received multi-modality treatment. There were no differences in 2-year survival rates for esophageal and pancreatic cancers, despite more aggressive treatments between the lowest and highest SES patients. There were modest differences in 2-year survival for lung cancer between the lowest and highest SES patients. However, within the lowest and highest SES hospitals, survival rates were similar across patients’ SES groupings.

Wong acknowledged that the study had several limitations. First, because it was based on SEER-Medicare data, the findings may not be generalizable to younger patients. Secondly, the SES classification for the study was based on area-level measures, so a lack of patient-level data may lead to SES-status misclassification for some patients. Finally, patient preferences were not able to be captured with administrative claims data so the team was unable to determine its effect on treatment preferences. Nevertheless, the authors felt that this wouldn’t be enough to account for all the differences that were seen in this study.

In conclusion, SES did not greatly impact survival rates for poor-prognosis cancers; it did, however, affect treatment intensity. Wong noted that differences in care were seen at the hospital level, which suggests that “reducing variation in cancer treatment strategies between socioeconomic groups may improve healthcare efficiency.”

Wong SL, Gu N, Banerjee M, Birkmeyer JD, Birkmeyer NJ. The impact of socioeconomic status on cancer care and survival. J Clin Oncol. 2011. (suppl: abstr 6004).