Objective: A recent rapid decrease in the duration of neonatal hospita
l stay in Ontario, Canada, enabled us to study the association between
healthy infants' age at neonatal hospital discharge and subsequent re
admission rates. Design: (1) Population-based and (2) single hospital-
based retrospective studies. Setting and study population: (1) A total
of 920,554 healthy infants with a birth weight of 2500 gm or more, bo
rn in Ontario from 1987 to 1994. (2) Infants with a birth weight of 25
00 gm or more, born during the same period, and readmitted before 15 d
ays from home to the Hospital for Sick Children far jaundice or dehydr
ation. Measurement: Duration of neonatal stay in the hospital and read
mission rates were measured, and diagnostic codes were analyzed. Sever
ity of illness was evaluated in infants readmitted to our hospital. Re
sults: In Ontario the mean length of stay decreased from 4.5 days to 2
.7 days (p = 0.000), and the readmission rate during the first 2 weeks
of life increased from 12.9 to 20.7 per 1000 (p = 0.000). Increased r
ates of readmission were most marked for jaundice and dehydration. The
infants readmitted to our hospital had evidence of increased severity
of illness as indicated by higher serum bilirubin and sodium concentr
ations. Two deaths occurred in infants with hypernatremic dehydration,
one in 1992-1993 and another in 1993-1994. Conclusions. In Ontario, s
horter neonatal hospital stay was associated with increased readmissio
n rates for conditions that may not give rise to symptoms or signs on
days 1 to 3 of life. In our hospital the severity of jaundice and dehy
dration in readmitted infants increased. The severity-of-illness data
raise the question of whether shorter neonatal hospital stay of appare
ntly healthy infants is always safe. Decisions to discharge infants sh
ould be based on rigorous evaluation of individual infants.