Two hospitals in the Philadelphia region perform complex heart surgery on newborn babies. But the institutions' results are vastly different, an Inquirer analysis of insurance claims data shows.
At St. Christopher's Hospital for Children, one in four babies less than a month old died after arduous, highly risky heart operations performed between 2009 and 2014, a death rate nearly triple that of Children's Hospital of Philadelphia.
The newspaper began a review of St. Christopher's after it declined last year to publicly reveal how many of its heart-surgery patients died - the only one of six hospitals not included in a first-ever state evaluation of such programs.
Late last month, the hospital stopped performing nonemergency heart surgery pending an internal review but did not say exactly what prompted the move.
In response to the Inquirer analysis, the hospital said Friday:
"While we are making progress growing our volumes and improving our performance, we are undertaking a comprehensive review of our cardiovascular surgery program."
Officials did not directly address The Inquirer's data analysis. Rather, they provided a four-year summary of heart operations on children of all ages, which includes procedures that are generally less risky than those done on newborns.
The hospital said its mortality rate for that broader group was 7.4 percent from mid-2011 to mid-2015 and had improved recently. The national mortality rate is 3.3 percent for all children undergoing heart surgery, according to the Society of Thoracic Surgeons.
The hospital did not provide data for newborns alone, but The Inquirer calculated their mortality rate to be 24 percent in surgeries performed from 2009 to 2014.
Paul F. Levy, a former Boston hospital chief executive who advises health systems on safety and quality, said a combination of high mortality and lack of details would give him cause for concern.
"What's striking to me, based on the numbers, and the fact they're not telling you the reasons, is the lack of transparency in and of itself would be concerning," said Levy, formerly of Beth Israel Deaconess Medical Center. "If I were a parent, I'd want to know why."
The Inquirer analysis comes amid a growing national call for hospitals to disclose how well they perform heart surgery on children - lifesaving measures needed to correct grave defects such as hearts with backward plumbing or missing chambers. It is the sort of information that has been publicly available for years on adult heart surgery.
But of the 125 or so U.S. hospitals that perform pediatric heart surgery, just 33 publicly share their results on a website maintained by the Society of Thoracic Surgeons. The Pediatric Congenital Heart Association, a national patient advocacy group, is holding a meeting with physicians and hospital administrators this month in Florida on strategies to improve transparency.
The state evaluation in which St. Christopher's declined to take part, conducted by the Pennsylvania Health Care Cost Containment Council, relied on clinical data not available to the public - information collected by nurses, physicians, and other caregivers.
To evaluate performance at St. Christopher's, The Inquirer relied on insurance claims data collected by the same agency.
Out of 121 newborn patients who underwent heart surgery at St. Christopher's from 2009 through 2014, 29 died, yielding the 24 percent mortality rate.
At CHOP, 67 of 784 newborns in the analysis - 8.5 percent - died during that period. CHOP's outcomes on the difficult operations generally meet or exceed national norms.
Just a few decades ago, most newborns with severe congenital heart defects died.
J. William Gaynor, one of the surgeons at CHOP who performs the complex procedures, cautioned that a count of operations based on claims data will not exactly match clinical numbers.
He praised the state report - in which St. Christopher's did not participate - which relied on clinical data. Five Pennsylvania hospitals and one Delaware hospital (no South Jersey facilities perform all the procedures) volunteered their data.
"I was very proud of the hospitals . . . that reported," Gaynor said. "I think that's a hugely important thing."
Jeffrey Jacobs, chief of cardiac surgery at All Children's Hospital in St. Petersburg, Fla., agreed that clinical data were the best measure, but he called claims data, also called administrative data, a "reasonable alternative."
"Large-scale differences in outcomes with administrative data are at least concerning and merit further investigation," said Jacobs, also a professor of cardiac surgery at Johns Hopkins Medicine.
Officials at the Pennsylvania Department of Human Services, which administers the Medicaid program for low-income children, promised to review the numbers.
"We'll take this very seriously," said David Kelley, chief medical officer for the department's office of medical assistance programs.
If the department confirms there is a problem, he said, it would ask: "With that high of a mortality rate, what are our next steps, and whether or not we want those types of procedures to be performed there and paid?"
Questions have arisen before at the North Philadelphia hospital, where more than 80 percent of patients are eligible for Medicaid. In a whistle-blower lawsuit filed against St. Christopher's in September 2012, a former employee alleged that from April 2007 through April 2009, 10 of the hospital's heart-surgery patients died or suffered neurological complications due to deficient care. That case, filed by a specialist who operated the heart-lung bypass machine during operations, is pending, and the hospital has denied wrongdoing.
The 140-year-old hospital, which plans to expand from 189 beds to 226 this spring, is owned by Tenet Healthcare Corp. of Dallas. In mid-August, a West Palm Beach, Fla., hospital owned by the same for-profit corporation shut down its pediatric heart-surgery program after questions arose about its death rate.
A high mortality rate could be due to problems during surgery or to other shortcomings, said Karl F. Welke, a pediatric cardiac surgeon at Children's Hospital of Illinois. A perfectly executed surgery can be for naught without proper care afterward.
That is especially true in the days after a procedure called the Norwood, in which the patient's body must learn to cope with a radically rebuilt cardiac plumbing system, Welke said.
"The care that is required is very intensive and very complex that has nothing to do with whether A was hooked to B correctly," Welke said.
Poor follow-up care was alleged to have played a role in at least two deaths at St. Christopher's, according to lawsuits pending in the Philadelphia Court of Common Pleas. The hospital has stood behind its care in both cases.
La'Kel Alford-Robbins underwent surgery in January 2012 to correct a variety of congenital abnormalities, including a double-outlet right ventricle and a large ventricular septal defect - a hole between the heart's left and right pumping chambers.
The baby had a second operation March 23, 2012, and initially was listed as stable. But later that day, La'Kel's blood pressure dropped sharply over a period of three hours during which nurses failed to notify physicians, his parents alleged in a 2014 lawsuit. They say their baby went into cardiac arrest as a result.
He was revived and placed on extracorporeal membrane oxygenation (ECMO), a form of life support. But then he went into cardiac arrest again because the device was operated by a nurse who was not certified to use it, the complaint alleged. He could not be revived.
In another lawsuit, the mother of baby Adrian Wilson alleged that her son was given too high a dose of calcium, leading to cardiac arrest June 6, 2011.
Clinicians resuscitated Adrian and placed him on ECMO. But his condition grew worse, and he died at the hospital June 15. In reports filed with the court, expert witnesses retained by the hospital have said the calcium was not responsible for the boy's death; experts for the family say it was a contributing factor.
Bernard W. Smalley, the lawyer who filed both the whistle-blower case and the individual lawsuits, declined to comment on them, and said the families would not comment either. Attempts to reach them were unsuccessful.
Counting up mortality rates from insurance claims data is not straightforward, as some surgeries are designated by multiple codes, one for each of several complex steps.
The Inquirer tally used software from the U.S. Agency for Healthcare Research and Quality, which relies on a method developed at Boston Children's Hospital.
Though the analysis found St. Christopher's to have a markedly higher mortality rate for surgery on babies in the first weeks of life, outcomes for older children were within national norms.
For operations on infants - age 1 month to 1 year - the mortality rate for both St. Christopher's and CHOP was just under 3 percent during the six years in question. For children ages 1 to 17, rates for both hospitals were under 1 percent.
The Inquirer used insurance data because clinical data are not publicly available unless hospitals disclose them. St. Christopher's said Friday it would put its overall data set on its website but declined to break down results by age group.
The best option would be for hospitals to make their complete track records public, said Jacobs, the Florida surgeon.
"Why would a program not share their clinical outcomes?" Jacobs said. "Patients and families have a right to know the outcomes of the treatment they are going to receive."
Bradley S. Marino, a cardiologist at Ann and Robert H. Lurie Children's Hospital of Chicago, said that is the only way parents can make an informed choice.
"I can do that now for my restaurant," Marino said, referring to online sites such as Yelp.
"Why can't I do that for my hospital that's going to open my child's chest and do heart surgery?"