Christopher Berry’s troubles started long before he killed his infant son in 2013. After surviving a suicide bombing and returning from Afghanistan with post-traumatic stress in 2011, Berry racked up arrests for allegedly shoving his teenage girlfriend, deliberately running over pigeons, and stealing from his employer.

Yet, when state social workers got a report that the Lowell couple was neglecting two-month-old William James Berry in the spring of 2013, records show they assigned the family to the “lower risk” category of state protection for children they believe are not in immediate danger. These families are targeted for increased social services rather than a full abuse investigation.

A month later, Christopher Berry lost his temper over the baby’s crying, shaking him for 30 seconds until his body went limp.

“I was holding him, and I was like, ‘Oh, my God, oh my God, what did I just do?’ ” Berry told police in recorded confession.

William is one of at least 110 children 17 and younger whose deaths were linked to abuse and neglect between 2009 and 2013 in Massachusetts, a third of whom had at some point been under the watch of the state Department of Children and Families. Records obtained by the New England Center for Investigative Reporting show that the vast majority of the dead were under the age of three, beaten, drowned, smothered or otherwise abused or neglected by caretakers. And their numbers have steadily increased, records show, from 14 reported abuse and neglect deaths in 2009 to 38 in 2013 – and state officials say numbers will likely remain elevated when the 2014 death toll is made public.

Most of these children’s stories have gone untold, either because their plight wasn’t known to the state until they died, or because the state’s missteps and failures to protect dozens of them was long concealed by confidentiality laws and secrecy. An examination of these sad cases shows that mistakes occur at all levels of the child welfare process – from at-risk youths the system failed to catch, to infants with open social services cases who fell through the cracks, to infants like William who were funneled into a program meant for lower-risk youths that couldn’t save them.

DCF already has faced harsh criticism for failing to protect children under its watch, including Jeremiah Oliver, the Fitchburg toddler who disappeared and was later found dead by a highway in 2014 and 7-year-old Jack Loiselle of Hardwick who fell into a coma in July after his father allegedly starved and beat him. Just Sept. 18, DCF faced more criticism with the revelation that Baby Doe, the child found dead in a garbage bag on Deer Island this summer, was Bella Bond, a 2-year-old who had twice been under supervision of state social workers.

Earlier this month, Gov. Charlie Baker held a press conference to say that DCF “has many systemic problems and we are going to fix them ... No one is standing here and saying everything is fine.”

But many child specialists worry the state swings from one tragedy to the next without learning from past mistakes or implementing lasting reform. The state’s own child fatality data is faulty, review teams set up to analyze fatalities don’t often meet, and DCF social workers say they often are kept from learning anything about what went wrong when a child dies, the Center’s review found.

“It’s a very dysfunctional system. Not only is DCF failing, but the other eye of the state, the child fatality review teams, are largely nonfunctional,” said Dr. Robert Sege, vice president at the Boston-based nonprofit Health Resources in Action who sits on a county-level review team in Suffolk County that has not met for over a year. “How do you make improvements if you don’t open your eyes and look at what is going on?”

The New England Center and the Boston Globe obtained information about child abuse and neglect deaths caused by parents and caretakers through a public records request that took seven months to complete and cost nearly $4,500 to obtain.  Center staff also spent months reviewing court and police records, interviewing families and child experts for this story, and found that:

• Thirty-eight children who died between 2009 and 2013 had received services from state social workers, and 26 of those were under state supervision at the time of their deaths. Other deceased children undoubtedly had contact with DCF, either to receive voluntary services or because their family was the subject of a complaint that social workers dismissed. But DCF declined to release information about complaints that had been rejected.

• A six-year-old DCF intake system for maltreatment complaints – opposed by the union that represents social workers – divides children into high-risk and lower-risk categories, with less risky cases assigned to workers with less required training. Between 2009 and 2013, 10 children on the lower risk track died, including 7 in 2013, records show, raising questions about whether the system has enough safeguards to protect children.

• The DCF screening system does not require social workers to do criminal background checks of a child’s caretakers when analyzing neglect and abuse complaints – an oversight that some child advocates say leaves a huge gap in assessing risk.

• The state keeps shoddy data on child deaths and its child fatality review system is crippled by lack of funds and resources. The New England Center found 10 children who were not included in state data even though their deaths were ruled to be homicide and, in most cases, parents or other caretakers were implicated.

Commissioner Linda Spears, named to run DCF in January, would not discuss individual cases that predated her tenure, but said that, much as a hospital emergency room has to determine the patients in most urgent need, DCF has to better identify and protect the most vulnerable children.

DCF faces a daunting task: responding to more than 92,000 child abuse complaints last year alone and figuring out which situations are so dire that children need to be removed from home, even though that could mean sending to foster homes that have their own problems. Last year, social workers substantiated 62,452 maltreatment complaints, a 34 percent increase over 2013, records show.

What needs refinement, Spears said, is “how do we make decisions based on risk factors that we know in the case ... I’m taking a very broad systemic view.”

In contrast to the headlines about Oliver and Loiselle, most abuse and neglect victims die with little public notice. That includes Dejalyse Alcantara of Boston, who was put under state watch at birth in 2011 because of her mother’s drug abuse. She died six months later in an overheated car, her mother asleep or unconscious in the front seat. Two-year-old Yarelis Rosario-Pereyra of Boston died allegedly of abuse and neglect in 2013 even though social workers had confirmed that she suffered bouts of maltreatment throughout her life. No one yet has been charged in her case.

Peter MacKinnon, DCF chapter president of the SEIU union local 509, said social workers are devastated when a child dies on their watch, but seldom learn from their managers about what went wrong or how they could improve their work.

“If you are truly looking to get a sense from DCF about what you did well, what you might have missed, you need to see what that analysis is,” he said. “If you don’t know what you are doing, how can you fix it? It goes into this black hole.”

Child fatalities – from natural and abuse-related causes – are supposed to be reviewed by a panel of experts, but that system has ground to a halt. The state review team, chaired by the Office of the Chief Medical Examiner, has filed only four reports since its launch 15 years ago, even though state law requires it to file findings and recommendations annually.

State officials say the review teams lack funding to do their work, but regular appeals for more money from the state Office of the Child Advocate, have gone nowhere.

Informed of 10 homicide cases of children not included in the 2009-2013 data, DCF officials acknowledged that they sometimes miss maltreatment deaths entirely – because they don’t always hear about them. In some cases, they said, medical examiners did not always alert DCF when a child’s death was linked to abuse and neglect as required by law. As a result, the agency has undercounted child abuse deaths and may be leaving abusive parents with other children.

Spears called the increase in child maltreatment deaths “tragic but not surprising,” blaming the jump on affects of the state’s opioid crisis as well as an increase in reporting of infants who die suddenly due to unsafe sleep practices, like sleeping with an adult, which state officials consider a form of neglect. Sixteen child deaths in 2013 were sleep-related, state records show, five of those had histories of maltreatment.

And Spears said she expects 2014 death levels, not yet finalized, to remain elevated. “I don’t think anything in the caseload and the community would give me any indication that the number will go down,” she said.

***

William James Berry’s shaking death, some say, points to weaknesses in a system launched in 2009 to help social workers separate cases where children are in imminent danger from those where the family simply needs help.

The policy, part of a national movement, was quickly embraced in Massachusetts: in 2013, 38 percent of child abuse reports were assigned to the lower risk group, DCF records show.

The higher risk cases, including allegations of sexual or serious physical abuse or neglect, are referred to social workers whose “primary purpose” is to investigate and “determine the safety of the reported child,” state documents show. Social workers are supposed to “engage and support families” when the child is in the lower risk group.

The state social worker’s union opposed the state’s two-tier system from the onset, MacKinnon said, because of concerns that families considered lower risk may get short shrift. The social workers who do the full child abuse investigations are provided more training on how to interview children and ferret out signs of child abuse, he said, leaving people with less specialized training to handle lower risk cases.

Currently, caseworkers who handle lower risk cases are less likely to interview the child away from parents -- often a key to getting at the truth, explained Taunton DCF social worker Laurie Cyphers. They are less capable, she said, of pushing parents to cooperate if they refuse state help. She worries that social workers with less experience and less training won’t be able to accurately assess safety risks.

“They don’t have the training and they don’t have the experience to fall back on,” said Cyphers, a 14-year DCF veteran who mainly oversees lower-risk cases.

There is no national data tracking deaths of children who had been placed on the lower risk track. But there have been enough incidents, here and elsewhere, to lead some child welfare advocates to question the idea of a two-tier system. In Minnesota, for example, the murder of a 4-year-old boy who had been placed on the lower risk track prompted statewide scrutiny and recommendations to narrow, and perhaps do away with, the program.

In Massachusetts, 3-year-old Alyvia Navarro was put on the less severe track months before the autistic preschooler drowned in a pond behind a Wareham trailer home in a death DCF ascribed to neglect, state records show. There’s also 10-year-old Isaiah Buckner from Athol who died from abuse and neglect-related injuries in July 2013, according to DCF, a case that remains unsolved. At least four other children on the lower risk track died of what DCF determined are neglect-related unsafe sleep issues, records show.

Sharon Crawford, Buckner’s maternal grandmother, said she was not aware that her daughter was being visited by social workers, much less in what is considered a lower risk track. She said the state should have taken special care since her grandson was deaf and legally blind. She’s angry that social workers never reached out to her, since she was very involved in Isaiah’s life.

“Something is not right here,” said Crawford, 53, who lives in Whitinsville. “Why would he be (placed on the lower risk track) if he couldn’t hear and couldn’t talk?”

Before taking charge of DCF, Spears last year oversaw a critical report on the agency by the Child Welfare League of America. The report found that DCF’s budget cuts, lack of staff support and growing caseloads compromise the effectiveness of the two-tier system. The report also said DCF needs to put a higher priority on a “child’s right to basic safety.”

Now, Spears says the two-tier program needs to be tightened up. She noted that, when the system works properly, children can be shifted to the higher risk group as social workers learn more about the families.

“We may walk in and find that something else is going on, at which point the case can then go back over to the investigation response track,” said Spears. “The paths are not so distinct.”

A new DCF review on Loiselle, the Hardwick boy now in a coma, showed that social workers had dismissed multiple allegations of abuse and neglect as far back as 2008. But when they finally opened a case on the boy in February 2015 in response to two new complaints, the social workers placed him in the lower risk category, records show.

Elizabeth Bartholet, a Harvard law professor and national critic of the two-track program, said the Hardwick report “screams out” that social workers involved with the family were more concerned with keeping the family together than ensuring the boy’s safety.

Especially for children in the lower risk category, Bartholet said, “Best interest of the child is clearly not the standard.”

William Berry’s case, which did not get the same kind of public scrutiny as Loiselle’s, also raises questions about how closely social workers studied the baby’s home life before concluding he was at low risk of harm.

When a maltreatment call comes in, individual caseworkers must decide which track to place a family on based largely on agency files and phone conversations, according to DCF documents. They can also request a criminal background check – a “Criminal Offense Record Information” or CORI – though it’s not required.

DCF won’t say whether staffers checked Christopher Berry’s CORI when they received a neglect complaint in 2013. If they did, the record would have shown Berry was facing a series of pending criminal cases, including an allegation that he repeatedly shoved Tabatha Cupan, his 18-year-old girlfriend who was pregnant with William, during a dispute in their Lowell apartment.

Mary McGeown, president of the Boston-based Massachusetts Society for the Prevention of Cruelty to Children, said simply knowing that Berry was a veteran should have prompted a closer look at the family because so many people return home suffering from mental issues, leading to increases in domestic and child abuse.

In the end, DCF assigned the family to the lower risk group a month before Berry killed his son, state records show.

Boston pediatrician and child abuse expert Eli Newberger said he was “appalled” to learn that social workers are not required to request a person’s criminal history as part of a screening – saying that the state is ignoring key evidence that puts a child at risk.

Commissioner Spears said that DCF hasn’t traditionally believed every neglect case requires that level of scrutiny, but agrees that the agency needs to re-examine the role CORI checks play in evaluating abuse and neglect complaints.

“We need to look at when CORI should be done and we should make those things routine,” Spears said.

Of course, even a fullscale investigation by DCF is no guarantee children will be safe. Dejalyse Alcantara, for instance, died in March 2012 even though she had been under state supervision since birth because of her mother’s substance abuse, DCF documents show.

D.J. Alcantara of Boston, Dejalyse’s father, had separated from the girl’s mother before the child’s birth to deal with his own drug problems. But now he can’t stop thinking of what he could have done to save his baby, who died of heat exposure in the back seat of a car while her mother was seemingly passed out in front. He said he told Boston police a month before Dejalyse’s death about his concerns about his daughter’s safety, because of the mother’s drug use. But the police report shows the concern was not relayed to DCF.

Marivette Morales, the mother, declined requests to comment for this story. But Alcantara wonders how social workers could have failed to see that his daughter was in danger. He said the baby didn’t even have a crib and slept on the couch for months.

Some argue that, until DCF makes a clear commitment to put child safety above all else, including keeping families together, child deaths like Dejalyse’s will continue to be a troubling problem.

“Strengthening families and keeping children safe are both vital, but child safety must always take precedence,” said Gail Garinger, former head of the state Office of the Child Advocate. “In some cases it may not be possible for vulnerable infants, especially those born prematurely or with drugs in their systems, to be safely maintained in their homes.”

Perhaps as troubling as children who die under the watch of social workers tasked to protect them are the stories of the children who died of abuse and neglect between 2009 and 2013 that were never brought into the state system at all – there were 72 of them. 

Social workers either dismissed reports of alleged maltreatment or never heard from concerned teachers, police, hospital workers or other mandated reporters at all, records and interviews show. 

Some cases reviewed by NECIR include clear signs that the state missed opportunities to save them. For example, state officials knew that Alexis Medina Sr. had repeatedly assaulted his baby daughter – he served 18 months behind bars for that crime. But, released on probation, he suffocated his three-month-old son from a new relationship in 2013. State officials at the time said Medina’s case had been closed and social workers were not aware he was again living with children.

It’s also unclear how many times social workers dismissed abuse and neglect claims about children who later died of abuse and neglect. The state, citing legal issues, refused to provide this information for the children who died between 2009 and 2013. NECIR is appealing this decision to the Secretary of State. But it’s likely that many had been known to state agencies: A 2013 state report, for example, found that in 65 percent of all child maltreatment deaths between 2001 and 2010, families were known to DCF.

For Laura Cyphers, the Taunton social worker who handles low risk cases, each new tragedy is a painful reminder of problems that front line workers know only too well. She wishes the agency would be more transparent and introspective, but, in her experience, that has not been the case.

When a child dies, she says, co-workers are interrogated by higher ups about what happened, but never see a final report or learn about findings – unless it involves losing their jobs.

“They go in saying, ‘I did nothing wrong,’ and they come out devastated,” she said. “If we can learn from something, it is important.”

NECIR interns Shan Wang, Bianca Padró Ocasio, Weiwen Zhao, Tessa Roy, Brittany Comak and Jordan Abosch, contributed research to this project. Boston Globe staff reporter Todd Wallack also contributed to this report.  The New England Center for Investigative Reporting is an independent nonprofit investigative reporting newsroom based at Boston University and WGBH News and supported in part by New England news outlets.

Jenifer McKim can be reached at jmckim@bu.edu.

INFORMATION DENIED, CHILDREN ENDANGERED

By Jenifer McKim

It is a federal mandate for states to publicly disclose information about child fatalities caused by abuse or neglect so that such tragedies can be prevented in the future.

But obtaining that grim data can be difficult because of the time it takes and money it costs.

A new report tabulating and describing publicly, for the first time, the deaths of  children linked to abuse and neglect between 2009 and 2013 in Massachusetts grew out of a public records request filed under  the federal Child Abuse and Prevention Treatment Act, which requires states to provide certain information about the children’s histories when asked.

The request, filed by the New England Center for Investigative Reporting to the Department of Children and Families, took months of negotiations and nearly $4,500 to produce what ultimately totaled 110 cases.

That’s a price too high and a process too complex to allow the kind of public scrutiny of data that might save children’s lives, say many child advocates.

“When it is so difficult, many groups will simply give up, especially when they don’t have the resources,’’ said Christina Riehl, a senior staff attorney with the nonprofit Children’s Advocacy Institute at the University of San Diego School of Law. 

Indeed, nationwide, many states fail to provide adequate public information about child abuse and neglect deaths that is required under CAPTA, according to the advocacy institute’s recent report titled “Shame on Us,” which faults the federal government’s lack of oversight of the estimated 1,500 maltreatment deaths a year.

Advocates and officials nationwide are pushing for more government transparency related to maltreatment deaths. Change can’t happen and fixes can’t be made if the public is denied information, said Michael Petit, an advisor for the Washington, D.C.-based child advocacy group Every Child Matters and a member of the new federal Commission to Eliminate Child Abuse and Neglect Fatalities.

Petit said the public is denied information by government officials “who use confidentiality laws as a shield to protect agencies.”

Massachusetts receives about $500,000 annually through CAPTA for improving child protective systems. Yet the commonwealth is one of only two states in the US that did not provide timely child fatality data for the federal government’s report, “Child Maltreatment 2013.” State officials attributed the delays largely to waiting for death information from the Office of the Chief Medical Examiner, which has notoriously long delays in its system.

The New England Center’s effort to obtain fatality data illustrates the protracted challenges in extracting such information.

The nonprofit news center first requested the data in December, 2013, filing a public records request for three years worth of child fatality information –  including the age and gender of child who died,  previous  abuse and neglect reports involving the child, and any state services provided to that child.

A month later, the Center revised its request, asking for just one year of data in hope of speeding up the process. But DCF said it would need to charge $2,023 to search for records and redact confidential information.

The Center appealed the fee in April 2014 to the Massachusetts Secretary of State, arguing that the high price tag posed a “significant barrier” to obtaining information and was tantamount to a denial – an appeal soon rejected.

After several more months of negotiations, NECIR joined forces with the Boston Globe and expanded the request to five years’ worth of CAPTA data.  By the summer of 2014, the state issued a new price of $4,468 – estimating the data collection and redactions would take 123 hours to complete. The payment was made in August.

But by January, 2015, when Gov. Deval Patrick was handing the reins of the state to newly elected Gov. Charles Baker, DCF had still not released any information.

When the Center contacted the new Baker administration in January to call attention to the long-delayed public records request, the state apologized. But, internal DCF emails obtained by NECIR, show officials discussed the timing of a release through a political lens.

In one email, DCF chief of staff Ryan FitzGerald said to several DCF communications staff that top officials likely wouldn’t be too concerned about the release of data from the prior administration because, “anything that comes out of it won’t reflect on them,’’ he wrote. “And on the flip side, I’m concerned it only looks significantly worse for us and the previous administration, the longer this drags on.’’

In the end, the state released the first three years of data on Jan. 21 and the next two years of data in phases through the end of March. However, DCF declined to provide any record of allegations of abuse and neglect that were dismissed by the state before a child died, citing legal restraints. 

In April, NECIR appealed this decision to the Secretary of State, claiming such information is key to learning about what went wrong in cases where the state knew about a troubled family but dismissed concerns. Shawn William, supervisor of records, ruled on Sept. 14 that the state either has to provide the information or provide “with specificity” why records are being withheld. DCF quickly responded – again denying the request. NECIR plans to appeal again. 

It does appear state officials plan to be more open with at least limited information in the future. In July, Baker announced "administration-wide measures to improve transparency and public access to government records and information, including a reduced and streamlined fee structure and more efficient communications and responses to requesters." 

Earlier this month, NECIR requested 2014 and 2015 fatalities data. State officials said results should be provided in eight weeks and – considering the time it took to provide earlier data – the information will be free.