Why are nephrologists initiating dialysis treatments for patients with end-stage renal disease (ESRD) at an earlier stage of disease than in the past? Aggressive early treatment with maintenancedialysis drives uphealth care costs anddoesnot usuallyprovidea clinical benefit.1 Fewdata are available to answer thequestionof howclinician andpatient preferences influence the decision to begin dialysis. In this issue of JAMA Internal Medicine, Wong et al2 report on a qualitative study of adults in theDepartment of VeteransAffairshealthcaresystemthatprovides insight into these decisions. The authors abstracted text from electronic medical records and identified themes related to the timing of dialysis initiation. They detected evidence of frictionbetweenpatientswho are reluctant to startdialysis andclinicianswhoappear tooverride these stated desires to avoid or delay therapy. After reviewing discussions about the initiation of dialysis, the authors suggest that nephrologists may need better training in shared decision making. Yet, training may not suffice; clinicians may also need different incentives. Discussions about treatment options for ESRDpresent an important opportunity for clinicians to elicit the values and preferencesof their patients,whileprovidingeducationabout their condition. The stakes are high. If discussions do not go well, the consequences for the patient can include loss of autonomy with unwanted dialysis treatment; poor preparation for dialysis, which may lead to harm, including catheter infections, congestiveheart failure, or electrolytedisarray;oruntimely death. However, patients and their clinicians bring distinct limitations toconversationsaboutESRDtreatments.Patientsmight have unrecognized knowledge gaps ormisconceptions about dialysis. Every nephrologist has heard patients assert that dialysis would ruin their lives. Although maintenance dialysis is associated with low quality of life and poor functional status, some patients adapt quite well; they endure the rigors of the therapy and still take pleasure from their families, activities, or work. Nephrology lacks tools to identify which patients will make this kind of successful transition to dialysis. The clinicianmay have limited insight into a patient’s experience with the disease, health literacy, and personal or cultural preferences. The clinician who encourages early initiationof dialysismaybedrivenby fear of failing thepatientwho hasmedical complications. Further confounding this clinical interaction is the problem that determining the right time to initiate dialysis involves judgment. Uremia is a syndrome of nonspecific symptoms, such as nausea, fatigue, andmalaise, thatmayemergeandprogress gradually,making it difficult for clinicians to decide exactly when renal replacement therapy is needed. Unfortunately, some nephrologists alsomay be influencedby the financial benefit of filling a dialysis chairwith another patient. Taken together, these observations suggest that cliniciansmay need incentives or feedback to encourage themtodo thedifficultworkof supportingpatientswithESRD as their disease progresses to a point when maintenance dialysis is definitively needed. For decades, nephrologists have faced these challenges of dialysis timing.However, somethinghaschanged.National registry data indicate that, for the past 20 years, patients with ESRD in the United States have been initiating dialysis earlier in the course of their disease, although this trendmay be leveling off.3,4 Early initiation of dialysis is concerning because of findings from the InitiatingDialysis Early and Late (IDEAL) trial,1 conducted in Australia and New Zealand, and published in 2010. The IDEAL trial randomizedpatients to a strategy of dialysis initiation at an estimated glomerular filtration rate (eGFR) of 10 to 15mL/min/1.73m2 vs a lower eGFR target of5 to7mL/min/1.73m2.Thestudy foundnodifference inmortality, but individuals in the lower eGFR arm received a median of 6 fewer months of dialysis, with substantial cost savings. Retrospective studies involving patients in the United States have similarly found no survival advantage to starting dialysis at a higher eGFR.5 Given the trend toward earlier initiation of dialysis, amajorunansweredquestion ishowpatientpreferences informthis decision.Wong et al2 took a novel approach to characterizing communication about dialysis. They randomly selected the electronic medical records of 1691 patients who initiated dialysis and abstracted paragraphs that described conversations about ESRD care. The exemplary quotes related to clinicianpractices suggest that thechallenges facingclinicianswho manage the myriad clinical problems (eg, hyperkalemia and volumeexpansion) causedbyESRDare important factors. The quotes related to motivating factors suggest that when patientspresentwithacuteclinicalproblems, cliniciansmayview that period of intense medical attention as a feasible or opportune time to commencedialysis, even in the faceof thepatient’s ambivalence. Most pertinent to concerns about paternalism, thequotesrelatedtopatient-cliniciandynamicssuggest that some clinicians considered their patients to be in denial when they expressed reluctance about dialysis, or to be noncompliant when they missed appointments intended to prepare for dialysis. Thestudy,however, cannotanswerallof the relevantquestions. For example, it cannot provide insight into situations in which patients with ESRD died without having received dialysis, selectedhospice care, or a kidney transplant. The study Author Audio Interview at jamainternalmedicine.com