Short Falls: Part I

Draft, confidential – the first of two chapters on the short fall research, featuring John Plunkett

copyright Sue Luttner 2018

Back to book Table of Contents

Chapter 22. “You couldn’t get these injuries by shaking,  not without breaking the kid’s neck.”

1996            Watertown and Hudson, Wisconsin

While David Dodge was arguing Stephanie’s innocence in Michigan, and Helen Brinkman arguing her guilt, two child-care providers in nearby Wisconsin were fighting murder charges, accused of shaking babies to death in surprisingly parallel circumstances.

Eight-month-old Kayla had been dropped off at Heidi Spern’s house late that morning, with a jar of carrots because her dad thought she was probably hungry. When Heidi opened the carrots about twenty minutes later, she says, the baby “took a little bit, and then she gave a deep breath and stiffened up. I swear it was a seizure.” Heidi dialed 911.

By the time Kayla reached the hospital, brain swelling was out of control. She died early the following morning. Subdural hematoma and retinal hemorrhages confirmed the shaking diagnosis, and Heidi was accused of murder.

She and her carpenter husband got a second mortgage on their house to pay for an attorney, who found forensic pathologist John Plunkett in Hastings, Minnesota. Although he lived and worked only a few hundred miles away from the Sperns, Dr. Plunkett was probably the most qualified expert in the country at that time for defending a shaking case. He had encountered his first shaking diagnosis in 1986, when shaking theory was first entering the courtroom. By 1996 he’d read and re-read the literature and examined dozens of cases, and he had concluded, he told me when I contacted him in 2001, “You couldn’t cause these injuries by shaking, not without breaking the kid’s neck.” Plunkett believed then, and still believes, that Shaken Baby Syndrome is a complete misnomer. “These are impact injuries,” he says, “when they’re injuries at all. Disease processes can also cause these findings.”

In Heidi Spern’s case, Plunkett saw evidence of an old subdural hematoma, which he believed had triggered the child’s seizure and subsequent collapse. That’s what he told the jury, which believed him, and Heidi Spern was declared not guilty. The emotional and financial consequences to her family were staggering, but her marriage survived and she didn’t lose her own children or spend any time in prison.


What had inspired Dr. Plunkett to become an expert on infant head injury was a case he’d been called in on in 1986, which looked to him like an accidental death. Janet Ostlund said her toddler had fallen off the arm of a couch while reaching up to a shelf above her, but doctors at Minneapolis Children’s Hospital insisted that short household falls are never fatal for young children and, further, the brain findings proved that Maria had been shaken to death.

As a forensic pathologist, Dr. Plunkett regularly examined the results of both accidental and non-accidental trauma, and he’d seen adults die from the impact of a short fall. He studied the pathology reports, the imaging, and the preserved tissues, and he saw no reason to doubt the mother’s story. The little girl had died of brain swelling, a well-known consequence of the breathing problems that typically accompany head injury, and in the autopsy photos, under her scalp, above one ear, he saw what certainly looked like the kind of bleeding that implied impact. “At that time, we were told that a short fall can’t injure a child,” Plunkett says, “but I didn’t see why not.”

Dr. Plunkett is a robust man, with white hair, piercing blue eyes, and a down-home Midwest charm. He exudes competence, and he needs all the credence his air affords him, because he’s also been blessed, or cursed, with a mind that never stops asking the next question. That urge to understand the science behind his art was what had channeled him into forensic pathology, where he could know the biochemistry behind slide preparations and look up the failure rate on lab tests. When confronted with the Ostlund case, he went to the medical library, where he found the least scientific collection of research he had ever studied. A decade ahead of me, and with better comprehension, Dr. Plunkett started reading the literature on shaken baby syndrome.

To recap parts of Chapter 8, in 1971 British pediatric neurosurgeon A. Norman Guthkelch[1] proposed that the practice of shaking a crying baby, acceptable at the time in northern England where he lived and worked, was triggering subdural hematomas in the immature brains of babies. Dr. Guthkelch noted that the most common cause of subdural hematoma is trauma, and he encouraged doctors to “inquire, however guardedly or tactfully, whether perhaps the baby’s head could have been shaken.” He also recruited public health nurses into a campaign of educating parents against the practice.

Then in 1972, pediatric radiologist Dr. John Caffey[2] in the U.S. made a more definite statement. He shared x-rays illustrating the bone findings that he believed to result from whiplash shaking: chips and fragments at the ends of the arm and leg bones, presumably from the stress on the joints when children are grabbed by the limbs, and anomalous growths on the bone surfaces. He identified 27 cases that he thought represented brain damage from shaking without impact, including 15 by infant nurse Virginia Jaspers. Caffey’s 1972 paper identified three hallmarks of the condition: the long-bone findings, subdural hematoma, and retinal hemorrhages. He warned that habitual, casual shaking in the course of transport and play was robbing children of full brain development. This first paper by Caffey concluded with the quatrain:

Hark ye, good parents, to my words true and plain,

When you are shaking your baby, you could be bruising his brain.

So, save the limbs, the brain, even the life of your tot;

By shaking him never; never and not.

He followed up with a 1974 article[3] that treated shaking theory more confidently, opening with the reflection that “many supposedly battered infants were actually shaken infants.” That paper added encephalopathy to the list of defining features and repeated that the victims show no external signs of trauma. The second paper offered a new poem:

Guard well your baby’s precious head,

Shake, jerk and slap it never,

Lest you bruise his brain and twist his mind,

Or whiplash him dead, forever.

Over the decade that followed, child abuse experts adopted shaking theory, citing the original papers and apparently not noticing that the propositions had never been proven. A 1978 child-abuse text, for example, in a section titled “Subdural hematoma,” reported a causal connection between shaking and subdurals as an established and reliable fact:

Caffey and Guthkelch have described the whiplash injuries from shaking infants. Guthkelch pointed out that subdural haematoma is a rare condition in adults although it may occur following relatively minor trauma which can often be shown to be of the acceleration-deceleration type, e.g. a bob-sled or motor car accident. In the infant and young child, however, especially under 2 years… subdural haematoma is a common condition following head injury when it is usually bilateral, associated with eye haemorrhages and is often seen without any signs of trauma to the head. There may be evidence of grasping limbs or trunk and small fractures at the ends of long bones. Some cases are unsuspected until autopsy. Confessions of shaking the infant are usually obtained and such lesions are inflicted in almost 100% of cases.[4]

That last sentence had no footnote citation.

By the early 1980s, shaking diagnoses began entering both the courtroom[5] and the literature.[6], [7] No one seemed to be citing any objective research that confirmed the theory, and somehow, the common knowledge had veered from the original proposition to something much stronger: Guthkelch and Caffey had suggested that shaking a child could cause the named symptoms. They did not claim that shaking was the only possible cause of those symptoms, but that was the argument in the prosecution of Janet Ostlund in 1986. The local newspaper later summarized the state’s position regarding evidence of impact to Maria’s head:

The prosecution’s doctors said the bleeding had no significance with regard to the brain injury. They said it could have been caused by another blow about the same time that she was shaken.[8]

Dr. Plunkett realized that shaken baby theory had entered the courtroom before it was fully tested, or even critiqued. At the time of Janet Ostlund’s trial, he was willing to believe that shaking could cause the brain injury Guthkelch and Caffey were describing—but in his experience, so could impact.

So he turned from the shaking literature to the research on short falls. While the standard texts contained the same guidelines described in Chapter 18, the same guidelines that still prevail—minor trauma represents a minor event, like a short household fall; major trauma represents a major event, like intentional battering[9]—the support for that position in the literature was slim to absent.

Helfer 1977

The most cited paper seemed to be a 1977 report by pioneering child-abuse pediatrician Dr. Ray Helfer and colleagues on children who had fallen from beds, sofas, and tables.[10] The team was trying to fill in what they conceded was a gap in the scientific evidence for the widely accepted opinion that short falls do not cause serious injury in children:

Occasionally a child who is reported to have fallen out of bed presents with a skull fracture, cerebral edema, retinal hemorrhage, subdural hematoma, and/or epidural hemorrhage. These severe injuries are discrepant with the history; it is often this discrepancy that indicates child abuse. There are, however, few objective data on the incidence of skull fractures, head trauma, or other injuries in infants who fell out of bed and in whom there is no suspicion of abuse.

The doctors looked at 304 total falls by 246 children, collecting their cases in two ways:

  • Eight participating pediatricians asked parents who visited their offices to answer a questionnaire, from memory, about incidents in which their children under the age of five years had fallen from beds or sofas.
    This strategy yielded reports of 161 children who suffered a total of 219 falls, none of which produced any serious, life-threatening injuries. Most of the falls, 176 of them, resulted in no visible injuries at all; 37 falls produced “bumps, lumps, bruises, scratches, and the like”; and six falls resulted in fractures, two of them skull fractures but without any apparent brain damage. Most of the falls that produced no visible injuries, 169 of them, were from less than 90 cm (36 in); five were from about 120 cm (48 in), and two were from about 150 cm (60 in). All of the serious injuries were from falls of less than 90 cm. When the parents reported a “potentially serious injury”—presumably the fractures—the doctors checked their own records to confirm the report.
  • Researchers reviewed incident reports from a large children’s hospital over a six-year period, identifying 85 incidents of children’s falling to the floor from beds, cribs, or examination tables, a distance of approximately 90 cm (36 in).
    Out of 85 reported incidents, 57 resulted in no apparent injury and the remainder produced 37 small cuts, scratches, bloody noses, bumps, and bruises—some incidents resulted in more than one injury—and one skull fracture with no apparent soft-tissue injury.

The authors noted that taking x-rays in these cases would not be cost effective, citing $33 for each image and pointing out that even in the rare cases of skull fracture in these short falls, treatment was not necessary. No one seems to have looked for retinal hemorrhages, and the doctors apparently relied on the lack of seizures, slurred speech, or other neurologic symptoms to eliminate the possibility of subdural hematoma, cerebral edema, and epidural hemorrhage. The introduction implied that all of these findings, plus skull fracture, raised the suspicion of abuse, but the conclusion addressed only skull fracture, presumably because the only suspicious findings present were skull fractures—but those were also the only findings anyone had explicitly looked for, and they did occur. Doctors already understood at that point that skull fracture is not a good measure of brain damage, as an impact can produce both skull fracture and trauma to the brain, or skull fracture without trauma to the brain. There can even be trauma to the brain without skull fracture.[11]

The numbers in this study are very small, of course, but Helfer et al. did not make absolute statements, merely offering their advice and conclusion:

The physician should be extremely suspicious of child abuse if he/she examines a child with serious head injury, with or without skull fracture, when the cause of the injury is reported to be a fall from a bed, sofa, or crib…

…From this study we must conclude that severe head injury and CNS damage or injury of any type are extremely rare when children, aged 5 years or less, fall out of bed.

I think “extremely rare” is a fair statement from these numbers, but that’s a long way from “it can’t happen,” which was the medical testimony that convinced a jury to convict Janet Ostlund of murder, despite the testimony of Dr. Plunkett, one forensic pathologist with no special training in child abuse against the sincere opinions of the treating physicians at the local children’s hospital.

Plunkett remained puzzled by the diagnosis: Janet’s daughter Maria had no fresh or healing fractures, no unexplained bruising, and no grip marks, only the brain findings that had somehow become proof of shaking. Janet claimed Maria had fallen from a standing position on the arm of a couch—a distance probably twice as far as the “fall from a couch,” presumably the seating portion of the couch, that the child abuse texts routinely rejected as a cause of serious injury.

Bilmire 1985

Pediatrician Dr. Elaine Billmire, for example, had recently come out with a study that would join Helfer’s in the footnotes for decades to come, “Serious Head Injury in Infants: Accident or Abuse?”[12] Dr. Billmire reviewed the charts of infants admitted to a Midwest children’s hospital with a diagnosis of head injury and/or abnormal findings on CT scans. She found 54 cases, 30 of which had been diagnosed as abuse based on these criteria:

… a history or confession of abuse (ten cases), the presence of multiple injuries (12 cases), or an inadequate explanation for the infant’s injuries (six cases)… major discrepancies between the history and the physical findings

Of the 28 cases reported to authorities, all were “subsequently substantiated by protective service and law enforcement investigations.”

The point of the study was to identify the “spectrum of injuries” caused by accidents versus child abuse, but it was clear to the skeptical eye that the authors had determined the outcome by assuming from the beginning that any explanation they didn’t believe was a lie intended to hide abuse. Their discussion section offered:

Our findings concur with those of Helfer… that accidental trauma rarely, if ever, causes intracranial injury in infants


[O]nly motor vehicle accidents accounted for injuries comparable to those seen in battered children. We found no instances of disease processes mimicking the intracranial abnormalities seen in battered children.

Then in their conclusion, Billmire and Myers added a new twist to the literature, rejecting cautions from earlier authors who warned, for example, that an abuse diagnosis requires “absolute certainty, to avoid conviction of innocent parents and unwarranted removal of children from their homes.”[13]

Instead, Billmire and Myers wrote:

In fact, however, the “burden of proof” rests not upon the physician but upon local protective services and law enforcement agencies legally mandated to investigate such cases.[14]

The authors had noted earlier that social services and law enforcement had “substantiated” all reported cases, which implies first that the doctors never reported any borderline cases and second that investigators had found independent evidence confirming the diagnosis, aside from medical opinion.

But Dr. Plunkett had read the police report on Janet Ostlund’s case. The police were conducting a murder investigation, with the understanding that the perpetrator was with Maria Ostlund when she lost consciousness. Once police and social services confirmed that Janet was home alone with her daughter, she was the only suspect. The only truly damning evidence against her at trial was the medical opinion.


Just a few months after Janet Ostlund was convicted and sent to prison, Dr. Anne-Christine Duhaime and colleagues published their controversial data from laboratory tests of dummies, described in Chapter 10, that questioned whether manual shaking of an infant could even cause the symptoms commonly diagnosed as shaken baby syndrome. The first scientific attempt to test shaking theory had rejected it—in fact, the researchers concluded that the findings more likely represented impact, Plunkett’s own diagnosis in the Ostlund case.

Astonished by the gap between the scientific record and what had happened at trial, Dr. Plunkett sent the article to Janet’s attorney, Steve Meshbesher. He later received the first in what has become a sobering collection of heart-wrenching letters from prisoners. Janet Ostlund had been convicted of murder in the wake of her daughter’s death, the state was trying to terminate her parental rights to her son, and her appeal had just been denied, but she wrote:

Living with 130 females has taught me tolerance and humor which I hope to put to good use in my life. God is giving me opportunities to use my time in constructive ways so I’m thankful for His direction…

The holiday season is approaching and my wish for you Dr. Plunkett is that you can experience the kindness and warmth from people as I have [from you] the past year. Your helping me has restored my self-esteem and has given me faith to trust people.

The least he could do, Plunkett thought, was to stay on top of the literature, which grew only stranger. For the next ten years, Dr. Plunkett followed each paper as it came out, traveling forward on a path that I would trace back, through the footnotes, a decade later.

Joffe 1988

In 1988, emergency-medicine pediatricians Dr. Mark D. Joffe and Dr. Stephen Ludwig published the results of their collected data on 363 stairway falls by children up to 18.7 years of age who visited the Children’s Hospital of Philadelphia (CHOP) emergency department between April and October of 1985, a time period that saw no serious injuries from pediatric stairway falls.[15] Like Helfer et al., Joffe and Ludwig found a few skull fractures, especially in the youngest children, “but these patients did well without neurosurgical intervention.” Most of the fractures, still a small number, were in the extremities. A majority of the patients sustained minor soft tissue injuries, most commonly bruising and swelling on the head and neck. Children younger than four years old were significantly more likely than older children to suffer head trauma. The severity of the injuries was unrelated to the total number of steps in the fall, leading the authors to conclude:

Our data suggest that most stairway falls consist of an initial mild to moderately severe impact followed by a series of low-energy noninjurious falls. The predominance of single injuries and the absence of more severe or more numerous injuries in patients falling a greater number of steps is predicted by this model. Physicians can use the data from this study and the model for stairway falls to judge the likelihood that particular injuries resulted from falls down steps.

But the “Materials and Methods” section of Joffe’s paper includes this revelation:

Cases of child abuse definitively identified by social workers or physicians were withdrawn from the study sample.

The way I read this last sentence is that if a family brought a child to the hospital during the study period with a report that the child fell down the stairs but with injuries that the doctors considered incompatible with that explanation, that case was excluded from the study. The researchers’ goal was laudable: To consider only cases that were known to be not abusive. But the result of their strategy was that their own expectations determined the outcome, by eliminating any falls with serious injuries from the study. The paper did not specify how many such cases were excluded, although the discussion section listed some of the findings that pointed to abuse, such as multiple, bilateral skull fractures and a spiral fracture. Spiral fractures were still considered highly indicative of abuse in the early 1990s, but that guideline was largely abandoned by about 2000.[16]

Hobbs 1984

When filtering his cases for the stairway fall paper, Dr. Joffe had relied in part on the work of Dr. C. J. Hobbs at St. James’s University Hospital in Leeds, who in the early 1980s had examined the skull fractures of 89 children, 60 of them injured accidentally and 29 injured by abuse.[17] Dr. Hobbs carefully described the locations, sizes, and features of the skull fractures found in the two groups, and offered this advice:

In children aged under 2 years with skull fracture after alleged minor accidents, abuse is suggested by one or more of the following features:

  • Multiple or complex fracture
  • Depressed fracture
  • Maximum fracture width greater than 3.0 mm
  • Growing fracture
  • Non-parietal fracture
  • Associated intracranial injury

The more features present, the more confidently can abuse be diagnosed.

The summary of Hobbs’ paper made the claim, “The results suggest that in skull fracture in young children where a minor fall is alleged, it is possible to recognize abuse by consideration of the fracture alone.”

But under the heading “Children studied,” Dr. Hobbs explained how the Leeds child abuse team had sorted the cases into abuse and accidental injury:

The diagnosis of abuse relied on the history and clinical findings. The history in 19 cases was of a minor fall from a few feet at most. In one case it was claimed that a door had been opened and the child’s head had been struck inadvertently. In the remainder the history was confused, unclear, or unrelated to the injuries…

The diagnosis of abuse was assisted by the presence of multiple injuries characteristic of abuse in 27 of the children, including 14 children with other fractures and 22 with multiple bruises.

That is, the authors had used the presence of injuries they believed to be “characteristic of abuse” to sort their cases in the first place. A child with serious injuries from a reported short fall would be identified as the victim of abuse. In Joffe’s study, these cases were merely excluded; in the Hobbs data set, these cases were added to the evidence base as known abuse.

Hall 1991

In 1991, pushing back against the common opinion that children are never injured in short falls, critical-care surgeon Dr. John R. Hall and his colleagues reviewed four years of records from the medical examiner’s office of Cook County, which encompasses Chicago, Illinois, and several of its suburbs. Between January 1983 and December of 1986, Cook County pathologists had autopsied 2,066 children 15 years and younger. Forty-four of the deaths were due to falls, 18 due to falls of three feet or less, which the researchers described as “minor.”[18] In yet another illustration of the lucid interval in pediatric head injury (see Chapter 18), the authors noted that nine of the children “were initially normal after their falls but did not seek medical care until there was neurological deterioration (range, 1 hour to 3 days).”

Conceding the possibility of abuse, the researchers wrote:

While some of the “minor” falls may have been secondary to abuse despite negative investigations (all of these had intense police investigation to rule out abuse), it is important that two of these falls did occur under medical observation.* “Minor falls” can be lethal, especially in a toddler, and must be evaluated.[19]

This assertion drew objections from both Dr. Helfer and Dr. Joffe, each citing his own work. Both questioned whether the supposed minor falls really happened, arguing that the researchers had not done enough to exclude abuse.

Dr. Helfer pulled no punches, stating clearly that he rejected these reports of serious injury from a short fall, based on his own 85 cases of in-hospital falls:[20]

I have considerable concern that too much reliance has been placed on the accuracy of the “intense police investigation.” The results are so strikingly different. Since the falls in the 1977 study were actually confirmed as in-hospital incidents, I must seriously question the occurrence of severe head trauma resulting in central nervous system damage and death as a result from [sic] falls of 3 feet or less.

Dr. Joffe stopped short of saying serious injury is impossible from a short fall, but he said he felt “compelled to challenge” some of Hall’s conclusions. He argued that his paper and Helfer’s had established the facts, and he pointed out that guilty parents are known to delay seeking medical care:[21]

Severe injuries are extremely uncommon after minor falls. This is supported by several studies with prospective designs. Unfortunately, the interpretation of poor-quality retrospective data led the investigators to erroneous and potentially dangerous conclusions…

The authors observed the high frequency with which parents delayed seeking care for their fatally injured children… The often-reported association between inflicted injuries and delays in seeking care is just staring us in the face.

In reply, Dr. Hall wrote:[22]

We do not argue with the statement that severe injuries are extremely uncommon after minor falls, yet while the risk is rare, they do happen. Eighteen deaths occurred in a 4-year time period in Cook County. Using the first writers’ estimate of 292,000 with a nominator of 4.5 per year, one needs to examine 64,889 children to find one death. It is thus not surprising that a study of 246 or 363 children finds no deaths.

Chadwick 1991

In a different approach to finding the truth in hospital records, researchers at the Children’s Hospital of San Diego, now Rady Children’s Hospital, looked at all visits to their trauma center over 42 consecutive months between 1984 and 1988, with the purpose[23]

[T]o determine the relationship of historical fall height to mortality, to assess the reliability of historical fall height, and to determine what types of fall histories might be used in establishing a database for children’s injuribility.

The researchers collected 317 reported falls by children ranging in age from infants to “13+” years—I’m not sure exactly what the plus sign means in the text—and of distances ranging in height from one foot to 45 feet. As reported in Table 1, the hospital recorded seven deaths among the 100 children who reportedly fell one to four feet, but no deaths among the 65 children reported to have fallen distances of five to nine feet, and only one death among 118 falls of ten to 45 feet. The children in all seven of the short-fall fatalities had cerebral edema, and five of the seven also had, in various unnamed combinations, retinal hemorrhages, subdural bleeding, and “associated injuries” like bruises and old fractures.

Fall Height Number of Deaths Number of Falls Fatality Rate
1–4 7 100 7.0
5–9 0 65 0.0
10–45 1 118 .8
Total 8 283 2.8

Table 1. Fatality rate by fall height, from Chadwick 1991

The doctors concluded that these numbers were illogical, as the fatality rate was highest among the children who had fallen the shortest distances. Clearly a short fall should not be more dangerous than a long fall. Therefore, they reasoned:

Falls of less than 4 feet are often reported in association with children’s head injuries that prove to be fatal, but such histories are inaccurate in all or most such cases. Long falls outside of buildings are more likely to provide accurate data points for studies of children’s injuribility, and research on children’s injuribility should utilize these longer falls rather than short indoor falls witnessed by just one person.

Again, the researchers pointed to the common belief that abusive parents do not seek medical care immediately:

[D]elay in care is a common feature in inflicted injuries to children, and was probably a factor in all seven children who died with histories of short falls.

To begin with that point: Falls of four feet or less happen with some regularity, off beds and changing tables, from swings and porches, out of the arms of parents. If a child seems fine after such a fall, few parents would seek medical care—unless the child later developed troubling symptoms. It seems to me that at least some of the delay in seeking treatment could have been due to belief that a short fall wouldn’t cause serious injury, the same assumption cited in Hall’s paper as the reason for delayed treatment.

Meanwhile, if a child falls from a second-story window or off a roof, the parents are likely to seek medical advice. Most children who fell 10 feet or more anywhere in the San Diego area during the study period were probably taken to a hospital for a once-over, whether they had symptoms or not. The short falls that reached the trauma center, on the other hand, were self-selected to be those with the most serious consequences. I consider it disingenuous of Chadwick et al. to compare these numbers side by side, as they do in their discussion:

The data in the present study show an astonishing concentration of risk of death in the group with the shortest falls… If the histories of short falls are accepted as correct, the conclusion would be reached that the risk of death is eight times greater in children who fall from 1 to 4 feet than for those who fall from 10 to 45 feet.

Then there are the scant facts about the seven fatal fall reports believed to be cover stories for abuse. In one of them, the babysitter said she found an 11-month-old at the bottom of the stairs. The researchers counted this case as a short fall “because falls down stairs are believed to be a series of short free falls,” with a footnote to Joffe. Two of the cases were reported to be falls by adults who were carrying the children, one of them who said he fell against the crib and one who said she was going upstairs with the baby. Under these circumstances, I don’t see why the doctors are so sure that the accident could not have produced some of the “associated” injuries, especially “bruises on trunk or extremities” or “two head impact sites.”

To my reading, the authors’ speculations about why parents so frequently lie about short falls say more about their confidence in their position than it does about the accuracy of their conclusions:

In children whose injuries are inflicted, parents typically invent accident histories which they hope will be accepted by health care providers. Since most falls of over 10 feet usually require that the fall occur outdoors (from a window, balcony, or other such location), caretakers may not wish to risk the possibility that a history could be proven false by a neighbor or passerby. It is also very possible that many lay persons believe that short falls may be fatal for children and are surprised to encounter skepticism. The best explanation of the findings is that for the seven children who died following short falls the history was falsified.

To the authors’ credit, their footnotes included an intriguing commentary from outside the child abuse community, a letter to Pediatrics from playground safety consultant Theodora Briggs Sweeney, urging doctors to support the installation of resilient surfacing around children’s play structures.[24] Citing 23 playground deaths recorded in 15 months of records from the National Electronic Injury Surveillance System (NEISS), Sweeney wrote:

Tests conducted on the energy-absorption properties of common playground surfaces, such as asphalt and concrete, indicate that even at heights as low as 1 ft a fall directly onto the head can prove fatal (based on the fatal impact threshold of 160 g). A fall onto packed earth from as little as 3 ft is in this same impact range.

I found Sweeney’s report riveting, but Chadwick et al. were making a different point, issuing a caution “that the NEISS database led Sweeney to conclude that children may die in falls as short as 1 foot, and that [the database] may be seriously contaminated with inflicted injuries that are not screened out in the data collection process.”



Every new paper Dr. Plunkett read left him more exasperated than the last. By the mid-1990s, when he was contacted about the Heidi Spern case in Wisconsin, he had assembled his collection of relevant journal articles, and he’d set up a system for tracking the boxes of medical records that kept arriving from defense attorneys.

None of those boxes contained the records for Natalie Beard, however, the infant who collapsed in the care of Wisconsin child care provider Audrey Edmunds. Like Heidi Spern, Audrey Edmunds was a stay-at-home mom who took care of a few neighborhood children as well as her own. Like Heidi Spern, she was pregnant with her third daughter when the accusations were raised. And like Heidi Spern, she was with the child for less than an hour before calling 911 for help with a choking baby.

Unlike Heidi Spern, though, Audrey Edmunds was convicted of murder and sentenced to 18 years in prison. Although Audrey Edmunds’ attorney had not heard about Dr. John Plunkett in neighboring Minnesota, Dr. Plunkett eventually heard about Audrey Edmunds, and that has made a tremendous difference to this story.



[1] Guthkelch A. Norman. “Infantile Subdural Haematoma and Its Relationship to Whiplash Injuries,” British Medical Journal 1971, 2: 430–431

[2] Caffey John. “On the Theory and Practice of Shaking Infants: Its Potential Residual Effects of Permanent Brain Damage and Mental Retardation,” American Journal of Diseases of Children, 19782 124(2): 161–169

[3] Caffey, John. “The Whiplash Shaken Infant Syndrome: Manual Shaking by the Extremities With Whiplash-Induced Intracranial and Intraocular Bleedings, Linked With Residual Permanent Brain Damage and Mental Retardation,” Pediatrics 1974, 54(4): 396–403

[4] Smith, Selwyn M., ed. The Maltreatment of Children, University Park Press, Baltimore, 1978, p 29

[5] Benstead JG. “Shaking as a Culpable Cause of Subdural Hemorrhage in Infants,” Medicine Science & the Law, 1983, 23(4): 242–244

[6] Ludwig S, Warman M. “Shaken Baby Syndrome: A Review of 20 Cases,” Annals of Emergency Medicine, 1984 13(2): 104–107

[7] Frank Y, Zimmerman R, Leeds NMD. “Neurological Manifestations in Abused Children Who Have Been Shaken,” Developmental Medicine & Child Neurology 1985 27: 312–316

[8] Cassano D, “Alteration of Child’s Death Certificate Is Defended,” Star Tribune, Jan. 22, 1997

[9] “Fatal Child Abuse: The Pathologist’s Perspective,” Kirschner RH and Wilson HL in Reece RM, Child Abuse: Medical Diagnosis and Management, Lea & Febiger, 1994, pp 325–357.

[10] Helfer Ray E, Slovis Thomas L, Black Mary. “Injuries Resulting When Small Children Fall Out of Bed,” Pediatrics 1977 Vol. 60 No. 4, 533–535

[11] Harwood-Nash DC, Hendrick EB, Hudson AR. “The Significance of Skull Fractures in Children,” Pediatric Radiology 1971 101:151–155

[12] Billmire M Elaine, Myers Patricia A. “Serious Head Injury in Infants: Accident or Abuse?” Pediatrics 1985 75;2:340­–342

[13] Hobbs CJ. “Skull Fracture and the Diagnosis of Abuse,” Archives of Disease in Childhood 1984 59: 246–252.

[14] Billmire M Elaine, Myers Patricia A. “Serious Head Injury in Infants: Accident or Abuse?” Pediatrics 1985 75;2:340­–342

[15] Joffe Mark, Ludwig Stephen. “Stairway Injuries in Children,” Pediatrics 1988 Vol. 82 No. 3 Pt. 2 September 457–461.

[16] Scherl Susan A, Miller Lisa, Lively Nicole, Russinoff Scott, Sullivan Christopher M, Tornetta Paul. “Accidental and Nonaccidental Femur Fractures in Children,” Clinical Orthopaedics and Related Research 2000, 376, 96:105

[17] Hobbs CJ. “Skull Fracture and the Diagnosis of Abuse,” Archives of Disease in Childhood 1984 59: 246–252.

[18] Hall John R, Reyes Hernan M, Horvat Maria, Meller Janet L, Stein Robert. “The Mortality of Childhood Falls,” The Journal of Trauma 1989 Vol. 29, No. 9 1273–1275.

* Years later, in a letter exchange in Pediatrics, Hall would specify that one child fell off a chair in a medical waiting room, and another, hospitalized for appendicitis, fell while running down the corridor. Hall John R. “Short Falls Can Be Lethal,” [eLetter] Pediatrics 13 August 2009

[19] Hall John R, Reyes Hernan M, Horvat Maria, Meller Janet L, Stein Robert. “The Mortality of Childhood Falls,” The Journal of Trauma 1989 Vol. 29, No. 9 1273–1275.

[20] Helfer Ray E. Letter to Editor, The Journal of Trauma 1990, Vol. 30,k No. 11 1422.

[21] Joffe Mark, Diamond Paul. Letter to Editor, The Journal of Trauma 1990, Vol. 30,k No. 11 1421–1422.

[22] Hall John R, Reyes Hernan M, Horvat Maria, Meller Janet L, Stein Robert. Letter to Editor, The Journal of Trauma 1990, Vol. 30,k No. 11 1423.

[23] Chadwick DL, Chin S, Salerno C, Landsvert J, Kitchen L. “Deaths from falls in children: how far is fatal?” The Journal of Trauma Oct 1991; 31(10):1353–5

[24] Sweeney TB. “X-Rated Playgrounds?” Pediatrics 1979 65:6 961