Outcomes and Care Associated with Residential Area Life Expectancy in Patients with ESRD

The general population of the United States faces significant geographic variation in health status due to complex factors that include regional differences in health indicators such as morbidity, mortality, and physical and mental disability. There is also regional variation in social determinants of health that include environmental and behavioral risk factors, access to healthcare and a healthy diet, education level, and socioeconomic status.

For example, the estimated life expectancy from birth in rural West Virginia and Mississippi is approximately 65 years, compared with Gunnison, Colorado, and Fairfax, Virginia, where life expectancy is approximately 82 years. Results of empirical studies have suggested that the primary sources of this variation are factors such as individual behavior and environmental  conditions.

In individuals with end-stage renal disease (ESRD), studies have assessed the association of area-level (ecologic) factors including median household income level, social deprivation, and income inequality on ESRD outcomes. Results of those studies found modest associations between socioeconomic indicators and ESRD mortality. However, there are few data available on the association of these factors and processes of care and outcomes across treatment modalities, including renal transplantation.

Jesse D. Schold, PhD, and colleagues recently conducted a retrospective cohort study designed to assess the association of residential area life expectancy with outcomes and processes of care for patients with ESRD in the United States. The researchers sought to test the hypothesis that patients who resided in areas with lower life expectancy had poorer outcomes and processes of care compared with patients living in areas with higher life expectancy independent of standard demographic and clinical conditions. Results of the study were reported in the American Journal of Kidney Diseases [2018;72(1):19-29].

The study population included all patients 18 to 70 years of age with incident ESRD (defined as initiation of dialysis therapy or receipt of a pre-emptive kidney transplant) from 2006 to 2013 in the US Renal Data System (n=606,046). The primary exposure variable was sex-specific life expectancy in patients’ residential areas based on the county (derived from the patient’s zip code) in the year of ESRD onset. Mean age of the cohort was 54.5 years, 58% were men, 33% were black, 25% were Hispanic, and 44% had body mass index ≥30 kg/m2. The number of incident ESRD patients increased each year; there were 72,751 patients in 2006 and 81,782 patients in 2013.

For women, median life expectancy in residential counties was 80.5 years; for men, median life expectancy was 75.8 years. Patients were categorized into quintiles of life expectancy. The 20th, 40th, 60th, and 80th percentiles for women were 78.7, 80.0, 81.0, and 82.1 years, respectively; for men, the quintiles were 73.6, 74.2, 76.4, and 77.7 years, respectively. There was a modest increase in life expectancy during the study period: life expectancy in 2006 for women was 79.8 and 74.7 for men; in 2013, life expectancies were 80.7 for women and 76.1 for men.

Across quintiles, the distribution of age and sex was relatively similar. African Americans were significantly more likely to reside in quintile 1 (Q1, lowest life expectancy) than in Q5 (highest life expectancy) regions. Asians and Hispanics were more likely to reside in Q5 regions than in lower life expectancy quintiles. Most comorbid conditions were relatively evenly distributed by quintile, with the exception of hypertension, which was more frequently the primary disease cause of ESRD in Q1 regions. Tobacco use, ie, current smoker, was more common in Q1 regions.

Data on specialist nephrology care was available for 86% of patients. Of those patients with data on specialist care, 69% of patients in Q5 regions were receiving care from a nephrologist, compared with 63% in Q1 regions. Following risk adjustment, the association remained (adjusted odds ratio [aOR], 0.83; 95% confidence interval [CI], 0.81-0.85) for patients in Q1 relative to Q5 regions.

Overall, 15% of patients reported an arteriovenous fistula as the first access type for outpatient dialysis; in Q5 regions, 16.4% reported fistulas as initial access, compared with 13.9% of patients in Q1 regions, a statistically significant difference (aOR, 0.90; 95% CI, 0.88-0.93) in patients in Q1 regions relative to patients in Q5 regions.

In the total cohort, 81.7% of patients reported being informed of options relating to kidney transplantation. The proportion was lower among patients in Q1 regions compared with those in Q5 regions (80.4% vs 83.7%); the difference was statistically significant in multivariable analyses (aOR, 0.93; 95% CI, 0.91-0.95).

There was a significant association between the probability of mortality following onset of ESRD and residential life expectancy. Eight-year mortality ranged from 67% in patients in Q1 regions to 57% among patients in Q5 regions. The association remained consistent in the multivariable Cox proportional hazard model with a dose-response association for mortality on dialysis (adjusted hazard ratio[aHR], 1.20; 95% CI, 1.18-1.22) for patients in Q1 regions relative to those in Q5 regions.

Following adjustment for demographic factors, disease cause, and multiple comorbid conditions, compared with patients in Q5 regions, among those in the lowest quintile aHRs were 0.68 (95% CI, 0.67-0.70) for placement on the transplant waiting list; 0.53 (95% CI,0.51-0.56) for time to living donor transplantation; 1.35 (95% CI, 1.27-1.43) for time to post-transplantation graft loss; and 1.29 (95% C I, 1.19-1.39) for time to post-transplantation mortality. Following adjustment for zip-code level median income, population size, and urban-rural locality, results were consistent.

Study limitations cited by the authors included basing the findings on retrospective data and the inability to evaluate direct causal inferences, basing the effects of residential life expectancy on county-level measures, and the need to further delineate factors affecting residential life expectancy.

“In summary, results of the study demonstrate  a significant association of regional life expectancy on outcomes and processes of care for a national cohort of patients with ESRD. These data suggest that noncodified factors are important considerations for risk assessment and treatment strategies. Further delineation of the factors explaining life expectancy variations by county and development of interventions tailored to patients who reside in lower life expectancy communities are needed,” the researchers said.

Takeaway Points

  1. Researchers conducted a national retrospective cohort study to examine the association between residential area life expectancy and outcomes and processes of care for patients with end-stage renal disease (ESRD) in the United States.
  2. Patients were categorized into quintiles based on life expectancy in residential counties: patients in quintile 1 (Q1) had lower life expectancies compared with patients in Q5.
  3. There was an independent association between residential area life expectancy and mortality and processes of care measures for patients with ESRD.