Publications

Artificial intelligence (AI) promises significant advancements in nephrology. While current use of AI in clinical nephrology practice is limited, there is scope for application in transplant allocation, remote dialysis monitoring, and CKD management.Yet, despite its potential, the use of AI in nephrology raises important ethical issues. Challenges include concerns about trustworthiness and reliability of AI systems, accountability for inaccurate results, injustice from the perpetuation or exacerbation of individual and systemic biases and inequities, and risks to patient privacy arising from the need for AI to access to sensitive health data.In addition, there are unresolved questions about the application of current disclosure and informed consent practices for patients when AI is used in their care and about the environmental and sustainability effect of this energy-hungry technology in an already resource intensive specialty. This article explores these ethical challenges and gives recommendations for addressing them, both at an individual clinician and health care system level.

IntroductionPregnancy may unmask or worsen kidney disease, in some cases leading to postnatal kidney failure. Outcomes in these women remain poorly defined.MethodsUsing linked perinatal (1991–2013), hospital and Australia and New Zealand Dialysis and Transplant Registry (ANZDATA) (1970–2016) records from 4 Australian jurisdictions, we explored characteristics and outcomes among women who commenced kidney replacement therapy (KRT) within 12 months postchildbirth.ResultsAmong 761 women who commenced KRT postchildbirth (N = 1196 babies), 44 started KRT within 12 months postchildbirth (n = 48 babies). Of these, 38 women began hemodialysis (HD), 4 began peritoneal dialysis (PD), and 2 received preemptive transplants. Median maternal age at childbirth was 29 years. The median time from childbirth to KRT initiation was 142 (interquartile range [IQR] 17–216) days, with one-third starting within 42 days postpartum. Among the HD group, 9 of 38 women (23.7%) transitioned to home dialysis within the first year. One-third of women died during follow-up (median: 13.7 years), with a median age at death of 31 (IQR: 28.5–36.1) years. Women who commenced HD died younger and sooner (median: 3.4 years postchildbirth, IQR: 2.8–7.1 years) than those who started on PD. Nearly 90% of babies were liveborn. Median gestational age was 34 (IQR: 31–37) weeks and birthweight was 2120 g (IQR: 1690–3090 g). Composite adverse pregnancy outcomes (≥ 1 of preterm birth, low birthweight, or hypertensive disorders) occurred in 75% of pregnancies.ConclusionWomen commencing KRT within 12 months postchildbirth represents a high-risk group with complex medical needs. Maternal death during early childhood years is an underrecognized phenomenon and warrants further research.