By THOMAS HARGROVE and LEE BOWMAN

Every day in America at least 10 babies die suddenly and mysteriously.

Yet some of the more than 4,000 victims of sudden infant death each year could be saved if there was a simple national standard for infant death investigations, a seven-month review by Scripps Howard News Service has found.

In fact, we are getting further away from solving the mystery of Sudden Infant Death Syndrome because of sloppy procedures, manipulation of statistics, misguided efforts to protect the feelings of grieving parents, and deliberate attempts to make SIDS go away, at least on paper.

The Scripps review of 40,000 infant deaths going back to 1992 revealed that the quality of infant death investigations, the level of training for coroners, and the amount of oversight and review vary enormously across the country. In many cases, professional bias -- both for and against a diagnosis of SIDS -- trumps medical evidence.

As a result, the odds that an infant's death will be correctly diagnosed are often determined by geography rather than science. In other words, the same death might be called SIDS in one county and called something else just down the road.

Theresa Covington: As director of the National Center for Child Death Review Policy and Practice at the University of Michigan for the past decade, Theresa Covington is a strong advocate for more thorough and professional investigations of infant deaths (SHNS photo) "There's no rhyme or reason to what medical examiners are diagnosing as SIDS, suffocation, strangulation or undetermined," said Theresa Covington, director of the National Center for Child Death Review Policy and Practice at the University of Michigan. "The variability is across the country and within the states."

The confusion comes with a very high price: the deaths of more babies who might have been saved through medical research.

The deeply muddled approach even has many experts questioning if a much-celebrated decline of SIDS deaths since the early 1990s was actually as significant as first believed.

"If we had a standard approach to investigating and classifying these deaths, our approach to prevention and research could be a lot clearer," said Dr. James Kemp, a leading SIDS researcher at St. Louis University. "The whole reason for keeping count is to figure out how to avoid the next infant death."

Yet questionable statistics from this haphazard system continue to guide public policy and outreach campaigns, as well as government research efforts that have devoted more than $110 million to SIDS research in the past five years from the National Institutes of Health alone, plus millions more from foundations seeking to understand why babies continue to die.

"You have to worry about the quality of this data (from death certificates), but there are researchers still using them," said Covington. "I simply don't put any credibility on any research that uses those numbers anymore."

According to standards set by the World Health Organization and the Centers for Disease Control and Prevention, SIDS should be diagnosed when an infant less than 1 year of age dies suddenly and unexpectedly and no clear cause of death is found after a thorough investigation that includes an autopsy, examination of the death scene and review of the child's clinical history.

Scripps conducted an extensive study into how infant deaths are investigated in the United States using records provided by the Centers for Disease Control and Prevention. The records detail the sudden deaths of 40,239 infants, half of whom died in the 1990s and the rest in a five-year period from 2000 to 2004.

Map of states reporting SIDS as casue of infant death: Analysis of some 20,000 official records of sudden infant death in the United States from 2000 to 2004 indicates huge irregularities in how states report infant mortality.The records of the most recent infant deaths, from 2000 to 2004, can be examined in the first searchable database of its kind at www.scrippsnews.com/sids/database.

The review found enormous variation in how the deaths of infants are investigated and classified. The SIDS rate, according to the data, is 12 times higher in Mississippi than in New York. Most experts agree that the big differences are caused by how the deaths are classified, not by how the babies died.

Variations are sometimes even greater from county to county within a state because coroners take widely different approaches to how they determine the cause of infant deaths in their areas.

Federal records show a dramatic decline in reported cases of SIDS, dropping from 4,895 cases in 1992 to only 2,247 in 2004, the most recent year for which complete data is available.

The records examined by Scripps showed that cases of SIDS virtually disappeared in some states and cities over the last several years, but closer examination of the data makes it evident that thousands of those lives have not been "saved," but rather lost under another name.

Coroners and medical examiners said SIDS was responsible for nearly 80 percent of all sudden infant deaths 15 years ago and only 55 percent in 2004. What increased during this time were diagnoses that CDC statisticians labeled as "threats to breathing" and "other ill-defined causes of mortality."

Some researchers think that this "code shifting" of infant death causes has substantially overstated the success of public health efforts against SIDS. The diagnosis of SIDS has been replaced on death certificates by new and vague terms like "undetermined cause," "sudden and unexplained death" and "other ill-defined and unknown causes of mortality."

The result is that, while deaths attributed to SIDS are down, the overall number of sudden infant deaths has remained steady, and even ticked up in some years, since 2000.

Next chapter: Death by another name >>