A research team at Stanford University School of Medicine and the Veterans Affairs Palo Alto Health Care System states that more than 20 percent of low weight new- born deaths between 1991 and 2000 in California, could have been prevented.
The babies would have survived if they had been born in different hospitals. Risky deliveries should be referred to highly experienced regional medical centers. This would afford the best chance of survival for newborns, who weigh less than 3.3 pounds.
Although the findings show that such a proposal is feasible, it flies in the face of a growing trend of decentralization in which community hospitals with lower-level neonatal intensive care units are shouldering an increasing number of complex patients.
"We can do better," said Ciaran Phibbs, PhD, a health economist at the Health Economics Resource Center at the Veterans Affairs Palo Alto Health Care System and an associate professor of pediatrics at Stanford's School of Medicine. "Although being born in a place with a mid-level neonatal intensive care unit is better than a place with no neonatal ICU at all, we found that these extremely tiny newborns were significantly more likely to survive in hospitals that cared for at least 100 such infants each year."
Phibbs is the lead author of the study, which will be published in the May 24 issue of the New England Journal of Medicine. The risks are real. Although very-low-birth-weight infants account for between 1 and 2 percent of births each year in California, they make up 51 percent of infant deaths.
Neonatal ICUs are classified according to the kind of care they can provide. The neonatal ICU at Lucile Packard Children's Hospital at Stanford is a level-3D center - the highest level reserved for hospitals that perform on-site cardiac surgery and a type of heart/lung bypass for newborns.
Packard Children's cares from more than 100 very-low-birth-weight infants each year. Other high-level, high-volume neonatal ICUs in the Bay Area are at the University of California-San Francisco; California Pacific Medical Center in San Francisco; Alta Bates Summit Medical Center in Berkeley and Oakland, and Good Samaritan Hospital in San Jose.
Only about one in five very-low-birth-weight infants in the study were born at such centers, despite the fact that more than 90 percent of the births occurred within urban areas with at least 100 very-low-weight births. The move away from regional centers is growing as increasing numbers of neonatal ICUs spring up at community hospitals around the state. Many of these new neonatal ICUs offer lower levels of care that, although sufficient for many infants, may not be optimal for very small newborns.
"Most of the babies admitted to the neonatal ICU have very little risk of death," said Phibbs. "Clearly, a 35-week-gestation newborn who doesn't need mechanical ventilation doesn't need to be in a big tertiary center. The extremely premature kids, however, do need that extra level of care."
Phibbs and his co-workers linked birth certificates and hospital discharge records of more than 48,000 very-low-birth-weight infants born in California between 1991 and 2000 with fetal and infant death certificates. They found that infants born at lower-level, lower-volume neonatal ICUs were between 1.2 and 3.1 times more likely to die than those born at higher-level, higher-volume centers.
It's not just that practice makes perfect. Regardless of a staff's skill, smaller neonatal ICUs that treat fewer infants don't have the same economies of scale as larger hospitals. And little things matter. A lack of dedicated, 24-hour anesthesia coverage for a woman needing an emergency caesarean section can make the difference between a live, but struggling, premature infant and a baby who dies before delivery. Attending physicians available around the clock can also respond at a moment's notice to any signs of trouble, for example.
Although low-level community neonatal ICUs are required to partner with higher-level centers, the study suggests that the relationships offer little benefit to tiny babies unless the laboring mother is transferred before the delivery to a facility with sophisticated obstetrical and neonatal support. "It's not good to move these kids around," said Phibbs. "The best way to transport a very premature infant is inside the mother.
"It's not that we think the community services are bad or poorly staffed," he added. "But a regional structure of services may be best able to help these kids."