The Grim Consequences of Closing Urban Hospitals

By Emily Badger for the Atlantic –

In 1995, Philadelphia had 19 hospitals with obstetric units, places where local women – about half of them on Medicaid – could go for prenatal care and, ultimately, to deliver their babies. Then these hospitals started closing, not one or two as occasionally happens in the life of many communities, but three, four, five of them…

By 2005, nine of these 19 hospitals in the city had closed, a group accounting for more than 30 percent of all of the deliveries that had taken place in Philadelphia in the mid-90s.

Now, some 15 years after the spate of closures started – brought on, by most accounts, because of financial pressure – we have some better sense of what this really meant for the city. A new study, published in the journal Health Services Research, concludes that Philadelphia’s newborn mortality rate temporarily increased by nearly 50 percent in the first few years after the closures began.

“Anecdotally, we had been hearing about the large academic centers seeing a pretty dramatic uptick in the number of deliveries that they were performing in a given year,” says Scott Lorch, a neonatologist at the Children’s Hospital of Philadelphia and the lead author of the study. The remaining hospitals, several of them associated with local universities, were clearly absorbing the influx of expectant mothers from around the city. But Lorch and his colleagues at the Center for Outcomes Research at the Children’s Hospital wanted to understand if the closures had caused more than overcrowding.

No one had studied this question before in a city with an unusually dramatic loss of healthcare infrastructure.

To answer it, the researchers looked at birth and death records in Philadelphia between January of 1995 and June of 2007, spanning more than 150,000 births. They then created two control groups, one from the five-county suburban area outside of Philadelphia (where medical practices and regulations are similar), and the other from an eight-county sample of urban populations elsewhere in Pennsylvania and California (where the demographics are similar). In total, they examined 3.1 million births, looking among them for babies who were already dead upon delivery as well as those who died during that first hospitalization.

In Philadelphia, compared to the control groups and accounting for other factors, the mortality rate increased by 49 percent in the first three years after the closures began, relative to the city’s mortality rate from 1995-1997. In other words, Philadelphia’s newborn mortality rate was higher than the surrounding suburbs even before these hospitals started closing, but that difference widened dramatically afterward. (For further context, America’s newborn mortality rate is also depressingly high relative to other countries, meaning the worst of American cities are in it particularly bad shape.)

“The size of this effect surprised us, and the fact that it happened fairly early in the process also surprised us,” Lorch says. “It’s been surprising to everybody how quickly the rates changed.”

As for why this happened, Lorch hypothesizes that some of these women may not have received the prenatal care that could have changed their outcomes. In Philadelphia, most of these hospitals don’t just deliver babies; their clinics provide women with prenatal checkups, too. (In contrast, a suburban woman is more likely to go to a doctor’s office for her prenatal care and a hospital for delivery.) The resultant overcrowding at OB units elsewhere in the city may also have impacted the women there, too.

There was some good news embedded in the findings. After the first three years, the mortality rate in Philadelphia leveled back off to what it had been before the closures. The city’s remaining medical infrastructure appeared to be absorbing the shock. That resiliency suggests that the city managed to cope with this loss of health care infrastructure, although that process requires painstaking planning.

“Yes, hospitals have made changes, adaptations, but it’s not easy to build a new OB unit,” Lorch says. He cites one hospital in town that needed several years to do so. It’s also worth noting that Philadelphia has lost four more obstetrics units since the study period concluded in 2007. Today there are only six hospitals delivering babies in all of Philadelphia.

The city’s situation has been so unusual that the researchers hesitate to draw implications from these findings for communities elsewhere. But it seems clear that hospital infrastructure matters for a city’s health. This study comes as close as possible – “as best as we can tell,” Lorch says – to illustrating a causal connection between hospital closures and the outcomes of babies who might have been cared for and born there.

“If there was something [else] that specifically happened in Philadelphia, and only in Philadelphia, at the exact same time that the closures were happening, we won’t be able to distinguish that from the closures,” Lorch says. “We haven’t come up with anything, but it’s always a possibility that we have to acknowledge is out there.”