The recent trend to early initiation of dialysis (at eGFR >10 ml/min/1.73 m(2) ) appears to have been based on conventional wisdoms that are not supported by evidence. Observational studies using administrative databases report worse comorbidity-adjusted dialysis survival with early dialysis initiation. Although some have concluded that the IDEAL randomized controlled trial of dialysis start provided evidence that patients become symptomatic with late dialysis start, there is no definitive support for this view. The potential harms of early start of dialysis, including the loss of residual renal function (RRF), have been well documented. The rate of RRF loss (renal function trajectory) is an important consideration for the timing of the dialysis initiation decision. Patients with low glomerular filtration rate (GFR) may have sufficient RRF to be maintained off dialysis for years. Delay of dialysis start until a working arterio-venous access is in place seems prudent in light of the lack of harm and possible benefit of late dialysis initiation. Prescribing frequent hemodialysis is not recommended when dialysis is initiated early. The benefits of early initiation of chronic dialysis after episodes of congestive heart failure or acute kidney injury require further study. There are no data to show that early start benefits diabetics or other patient groups. Preemptive start of dialysis in noncompliant patients may be necessary to avoid complications. The decision to initiate dialysis requires informed patient consent and a joint decision by the patient and dialysis provider. Possible talking points for obtaining informed consent are provided.