The purpose of this study was to examine ethnic differences in the use of peritoneal dialysis (PD) – an alternative treatment for end stage renal disease (ESRD) to hemodialysis – in a cohort of new ESRD patients. The analysis controlled for demographic and socioeconomic characteristics, comorbidity, functional status, and nutritional status.

Subjects for this study were African American or white patients who began treatment for ESRD at dialysis centers in three states and reported to ESRD Network 6 from 1989 through 1991. Data were derived from staff questionnaires and medical record reviews. The racial distribution of the sample was 58.9% African American and 41.1% white. PD was given as initial treatment for 21.8% of the sample. African Americans were 57% less likely to be initially treated with PD than whites (OR=0.43; 95% confidence interval=0.39-0.47).

Gender (women received PD less often than men), age (use of PD declined with increasing age), diagnosis (those with hypertensive ESRD were less likely to use PD), albumin level (those with high levels were less likely to use PD), functional level (those with low levels were less likely to use PD), comorbidity (more conditions were associated with less PD use, except peripheral neuropathy), and SES (low SES was associated with less PD use) were all associated with PD use. With the exception of patients who used illicit drugs (for whom there was no association between ethnicity and PD use), African Americans were uniformly less likely to use PD for initial treatment across levels of these covariates. After controlling for the confounding characteristics, the odds ratio of initial treatment with PD for African Americans compared with whites was 0.45 (95% confidence interval=0.38-0.52). The results were similar controlling for dialysis facility, which differed in the on-site availability of PD.

The authors state that “the explanation for these persistent differences is not evident from our data” but conclude that the findings “underscore the likelihood that complex relationships exist between non-clinical factors and the use of health services and may persist after the removal of economic barriers to care.” They suggest that cultural factors in health beliefs, self-perception, and patient preferences might be responsible. They also indicate “that a systematic bias on the part of providers may play a role in the observed association,” as well as factors related to communication between patients and providers.