Perlman also suggested that a child smothered to death by his or her parents in bed has not necessarily died of neglect if the parents’ behavior met a “socially acceptable threshold” — in other words, if other parents place their children in an adult bed at night.

McIntosh, the Jacksonville doctor, took particular issue with Perlman’s claim that investigators must find “intent” to leave a child at risk when a child drowns or is suffocated.

“This proposal,” he wrote, “assumes that a caretaker could credibly claim not to know that it was dangerous to leave a child unsupervised around a body of water.”

Death investigators, McIntosh added, should not consider whether it is “socially acceptable” to place infants in bed with their parents, adding that just because many parents engage in a risky practice “does not make it safe or right.”

“There was a time when parents did not have to buckle their children into car seats, during which time thousands died annually in car accidents,” McIntosh wrote. “Those deaths are now prevented. We are tasked with identifying avoidable death hazards that need to be corrected, not simply accepting the way they are.”

McIntosh added: “Under-reporting and under-verification [of deaths] compromise the validity of the statistics related to child abuse and neglect, make it more difficult to ascertain true progress in combating these problems and, most importantly, defeat efforts to identify causes of preventable death that could be addressed through education, product redesign and legislative regulatory action.”

In a draft of the statewide committee’s annual report, death team members noted that DCF had eliminated more than half of the drowning deaths from 2010 — verifying neglect in 42 of the 91 drownings that were reported to the state. As a result, they stated, the other 48 were never studied and their absence from the sample made it appear as if neglectful drownings had declined dramatically.

In the report language, they used the word “only” in referring to the 42 verified drownings.

That one world — “only” — raised hackles at DCF. Christie Ferris, a member of the committee from DCF, fired off an email on Dec. 30 that said: “The over-use of the word ONLY...implies DCF is under-verifying the reports and/or is not being transparent.”

Another line in the draft report also angered DCF. The agency threatened to vote against release of the report if the full committee refused to delete the line — one that said the dramatic decline in abuse and neglect deaths was owing to DCF “modify[ing] their criteria for verification of certain neglect deaths.”

Ferris, in her email, insisted “no criteria was changed or modified and therefore this is a false statement.” She said the sentence “must be deleted to gain a YES vote from DCF.”

In the end, the sentence was removed as DCF demanded.

Manatee County Sheriff’s Major Connie Shingledecker, who chaired the committee last year, said she supported the concept of achieving consistency throughout the state in the investigation of child deaths: Some regions, she said, considered most drownings a result of neglect, while others required other circumstances, such as the caregiver’s use of drugs or alcohol, as a factor in their poor supervision.

Perlman’s memo, she added, seemed to go too far in excusing poor — and fatal — decisions by parents, unless the parents intended to harm their children.

Both Shingledecker and other team members strongly suspected the reduction in deaths resulted from DCF’s decision to change its criteria. But the denials from Ferris and others prompted her to agree to remove the objectionable language.

“I couldn’t prove it,” Shingledecker told The Herald. “We were told they had not made any changes.”

Going forward, Shingledecker hopes the group will be able to study every case in which a child dies, because such inquiries are the best way to prevent tragedies, she said.

“You never know about the lives you’ve saved,” Shingledecker said. “You only know about the ones you don’t.”