Objective: To compare 3 methods of clinical assessment of jaundice in newborns by home health nurses. Design: Prospective clinical trial. Setting: Homes of newborns living within 10 miles of a 340-bed community hospital where they were delivered. Participants: Home health nurses and newborn patients (≤2 weeks old). Interventions: The nurses examined the newborns and documented whether they detected jaundice. In newborns thought to have jaundice, the nurses estimated bilirubin levels, documented the extent of caudal progression of the jaundice, and determined the Ingram (Cascade Health Care Products, Salem, Ore) icterometer readings from the newborns' noses. Total serum bilirubin tests were obtained from all newborns studied. Outcome Measures: Nurse assessment of the presence of jaundice and its caudal progression, nurse estimates of bilirubin levels, icterometer readings, and bilirubin levels. Results: The nurses determined that 82 (50%) of the 164 newborns had jaundice. Their estimates of bilirubin levels were most highly correlated with serum bilirubin levels (Pearson correlation, 0.61). All 3 newborns with bilirubin levels greater than or equal to 291 μmol/L (≥17 mg/dL) were recognized by the nurses as having jaundice. These newborns had icterometer readings greater than or equal to 3.5 and had estimated bilirubin levels of greater than or equal to 274 μmol/L (≥ 16 mg/dL). Conclusions: The method of evaluation that each nurse was accustomed to using was the most accurate in determining the severity of newborn jaundice. These results suggest that postpartum home health nurses can effectively evaluate newborns for the presence and severity of jaundice.