The present study, entitled Trial on Education And Clinical outcomes for Home PD patients (TEACH), investigated the effect of frequent retraining at home on the outcomes of peritoneal dialysis (PD). TEACH is a multicentre, open-label, randomised, controlled trial with parallel arms. Patients starting PD were randomized into either the conventional retraining group (CG) or the frequent retraining group (FG). Patients in the FG were given more frequent home visits for retraining. The primary endpoint was exit site infection (ESI). Secondary endpoints were peritonitis, any PD-related infections, hospitalization, technique failure, and patient survival. A generalised estimating equations (GEE) approach was employed for the adjusted effect of training level on the outcomes. Cox regression was employed for peritonitis and other secondary outcomes. The subjects were randomised to either the FG (n = 51) or the CG (n = 53). Although the time of initial training did not differ between the 2 groups, the total time of training was longer and the frequency of training visits was higher in the FG. In the GEE model, the p-values for interactions between groups and time were significant for both ESI and any PD-related infections, suggesting that the event rates of the two groups significantly changed over time. The event rates for the FG decreased over time, and the event rates for the CG increased after month 12. In the older subgroup (age ≥ 60), frequent retraining had a significant effect in the risk reduction of the first episode of peritonitis (adjusted HR 0.01 [0.001-0.35], p = 0.01). Frequent retraining at home reduced the risk of PD-related infections.