Fistula Care Continuum and Individualized Vascular Access Models for Older Adults Undergoing Hemodialysis

Annually over 50,000 patients who are 65 years and older initiated hemodialysis in the United Studies. Choosing a vascular access for older dialysis recipients is challenging. Although current clinical guidelines recommend arteriovenous fistula as the optimal type of vascular access, older hemodialysis patients have significantly worse fistula outcomes. For those whose created fistulas are not usable in a timely matter, advantages of fistula attempts are greatly compromised. Few studies have tracked fistula development and maintenance as an integrated care process in older hemodialysis patients. Evidence on the association of age and AVF outcomes is scarce in the literature. There is also a call to identify older patients who are more likely to benefit from a fistula placement. The first contribution of this dissertation is the description of arteriovenous fistula development as a process of care from the initial step of fistula placement to achieving the goal of continuous fistula patency. We examine the proportion, timing, and geographic variability of fistula construction and outcomes among older hemodialysis recipients and find only a small proportion of patients have completed the sequential stages of fistula care in the United States and disparities in practice patterns needed to be addressed to improve fistula outcomes. The second contribution of this dissertation is the precisely estimation of the effect of age on arteriovenous fistula construction and outcomes. We find increasing age is significantly associated with lower probability of fistula placement and maturation but not fistula primary and secondary patency loss. We conclude the likelihood of fistula maturation should be the most important consideration for vascular access planning and fistula might not the best vascular access option for patients approaching eighty years old. The third contribution of this dissertation is the externally validation of the Lok’s risk equation for fistula primary failure which has achieved good prediction accuracy previously. We show the Lok’s model is invalid to predict fistula primary failure in the U.S. older hemodialysis patients. Finally, we use random survival forests to identify important predictors for fistula maturation and find patient’s gender might be considered as the most important predictor for fistula maturation