52 children died, and many could've been saved.

In some instances, the city washed its hands of a case not because an endangered child was safe - but simply because the parent refused to cooperate with social workers.

In others, workers skipped visits or ignored warnings that a youngster's plight was growing more dangerous.

In still others, the city misinterpreted or didn't understand important medical information.

And in the end, children died needlessly despite the mission of the city Department of Human Services to protect them.

Between 2001 and last year, 52 children whose families were known to DHS died - and "a number of those fatalities could have been prevented," according to a panel that reviewed the deaths as part of an examination of the agency.

The group said preventable deaths were the worst result of widespread inadequacies: confusion about when a child was in danger, too few meetings with children, supervisors who failed to supervise, and arbitrary decision-making.

The panel uncovered a statistic it found "shocking," that more than half of all children who died from abuse had parents who themselves had grown up in homes overseen by DHS.

Yet, the panelists said, the agency seemed blind that child abuse could repeat - and lead to the death of a child. This ignorance, the report said, opened a door to "predictable, intergenerational, devastating, dysfunctional behavior."

Mayor Street formed the panel after a series of Inquirer articles last fall raised questions about the deaths of children within the city system.

Among other deaths, the newspaper looked into the cases of 14-year-old Danieal Kelly, found dead of neglect and weighing just 46 pounds; 1-year-old Alayiah Turman, beaten after she interrupted a video game; and 11-week-old Marrieon Currie, doused in hot water and beaten.

While it leveled some tough criticism, Street's panel didn't name names or point fingers in its examination of deaths - omissions that were both by choice and because of big gaps in agency records.

In all, the panel looked at 52 children who died between the summers of 2001 and 2006 - all of them in families which at some point had been under DHS scrutiny. (Excluded were deaths unrelated to the family, such as teens shot by outsiders.)

Of the 52 deaths, DHS had no files for five victims. Two files had been expunged; three had been lost.

The remaining files were often a jumbled and mislabled mess in which some documents were duplicated and others missing entirely. Case information was "disorganized, internally inconsistent and incomplete," the panel found.

Still, the experts chose not to interview the assigned social workers or the private case workers who handled most of the week-to-week contact with the troubled families. As the panel wrote, "Our work involved paper reviews only."

The panelists also reviewed the more detailed so-called "death review" for each case. The agency has refused to make those reviews public.

The report does not identify any of the child victims, detail which deaths were preventable, or name DHS social workers or outside contractors involved in the death cases.

Panelist Carol Tracy, executive director of the Women's Law Project, said the panel, in order to see the records, signed a pledge agreeing it would not disclose confidential information in the files.

She said state law limits access to DHS documents. The panel urged the state legislature to change the law to make the agency's operations far more open.

And, Tracy said, the panel set out to be "future-oriented."

The tally of 52 child deaths was not an unusual toll for a city of Philadelphia's size, panelists said.

Even so, the report's statistical overview was chilling.

It found that of the 52 dead children, 20 had been killed deliberately - beaten, strangled, starved or drowned.

City pathologists could not determine the cause of death for eight other children, but seven of these had previously been beaten up or injured in intentional assaults, the report said.

Twelve children died from unsafe sleeping, such as "co-sleeping" deaths in which an infant is smothered by a parent. Such deaths often stem from drinking or drug abuse by a parent, panelists said.

"All of these co-sleeping deaths are preventable," said panelist Cindy W. Christian, a doctor at Children's Hospital of Philadelphia.

The remaining children died of natural causes or from sudden, unexplained infant deaths.

Among the cases reviewed, about two-thirds of the victims were under a year old.

As is true nationwide, the panel noted, the city's victims were disproportionately young.

"The children who are most vulnerable due to their age are most likely to be alleged victims of fatal child maltreatment," the report said.

Christian said she reviewed DHS files of deaths in which infants had previous fractures.

"There were cases where children had fractured bones, broken bones," Christian said.

Nonetheless, she said, those injuries didn't give DHS "a sense of the danger to the child."