Kidney disease trials for the 21st century: innovations in design and conduct

Compared to other specialties, nephrology has reported relatively few clinical trials, and most of these are too small to detect moderate treatment effects. Consequently, interventions that are commonly used by nephrologists have not been adequately tested and some may be ineffective or harmful. More randomized trials are urgently needed to address important clinical questions in patients with kidney disease. The use of robust surrogate markers may accelerate early-phase drug development. However, scientific innovations in trial conduct developed by other specialties should also be adopted to improve trial quality and enable more, larger trials in kidney disease to be completed in the current era of burdensome regulation and escalating research costs. Examples of such innovations include utilizing routinely collected health-care data and disease-specific registries to identify and invite potential trial participants, and for long-term follow-up; use of prescreening to facilitate rapid recruitment of participants; use of pre-randomization run-in periods to improve participant adherence and assess responses to study interventions prior to randomization; and appropriate use of statistics to monitor studies and analyse their results. Nephrology is well positioned to harness such innovations due to its advanced use of electronic health-care records and the development of disease-specific registries. Adopting a population approach and efficient trial conduct along with challenging unscientific regulation may increase the number of definitive clinical trials in nephrology and improve the care of current and future patients. The field of nephrology has conducted fewer trials than other medical specialties. Here, the authors discuss how innovations in trial design and conduct could help achieve the goal of conducting a greater number of larger renal trials. Key points Nephrology has the potential to benefit from large streamlined trials similar to those that have led to advances in cardiology and diabetology. As effective interventions are developed and population-level risks fall, larger trial sample sizes are often needed to ensure sufficient statistical power to demonstrate the effects of new therapies. When moderate effect sizes are anticipated, real-world evidence from association studies cannot provide a reliable estimate of the effect of an intervention; only ‘randomization’ guarantees the elimination of moderate biases. Potential surrogate outcomes in renal