« View all Articles


Marika Henneberg, Institute of Criminal Justice Studies, University of Portsmouth

By Marika Henneberg, Institute of Criminal Justice Studies, University of Portsmouth on 19/04/16


Forensic science, or “the application of science in the resolution of legal disputes”,  has become an integral part of the criminal justice systems in the United States of America (US) and in England and Wales (E&W). It is beneficial to the process of justice that scientific methods and techniques have the potential to draw links between victims, perpetrators and crime scenes.

Criminal and civil courts often rely upon scientific evidence provided by expert witnesses to settle different types of legal disputes. This inevitably increases the risk that ‘junk science’  may be admitted in courts in the US and in E&W.  

It is, undoubtedly, a hard task for judges and jurors to adjudge topics which may have taken years for a scientist to become proficient in.  Collecting, preparing, analysing and presenting scientific evidence all require high levels of training, expertise and skill, and it would be naive to suggest that the intricacies of any field of forensic science can be fully understood by judges, lawyers or jurors (or indeed scientists).  Nevertheless, judges need to decide whether or not to admit scientific evidence, lawyers need to be able to use expert evidence to support their case and effectively challenge the views of the opposing side’s experts during cross examination, and jurors have the important task of deciding, based on the evidence that they have heard, whether the defendant is guilty or not. 

In E&W, three categories of miscarriages of justice caused by expert evidence have been scrutinized by the courts and the Law Commission recently.

According to Phillips, these are cases where the expert (1) “deliberately misled the court,”  (2) testified beyond or outside their expertise,  and (3), where the expert has relied “on flawed or faulty forensic or diagnostic techniques.”  

The CACD has acknowledged that SBS/AHT cases are largely relying on expert evidence of a complex nature.  However, the adversarial system does not appear to be an objective ‘truth-finder’, and Roberts  supports the notion that the “adversarial procedure is sometimes suboptimal for truth-finding.”  This is also apparent when looking at what has been aptly described as “litigation-driven science”,  where arguments from both sides of the SBS/AHT debate have been exacerbated.  At the heart of this debate is whether a triad of symptoms, namely subdural hemorrhage, retinal hemorrhage and hypoxic-ischemic encephalopathy, is pathognomonic of SBS/AHT.  

The reliability of evidence relating to alleged SBS/AHT cases appears to be particularly problematic in the heavily polarized debate around SBS/AHT, especially in relation to the triad of symptoms thought to be pathognomonic of SBS/AHT.  In order to improve our understanding of the symptoms included in the triad, studies and experiments have been carried out in a variety of different disciplines, including anatomy, engineering, medicine and pathology.  

In 2003, Geddes et al suggested a new hypothesis around the triad of symptoms, which is often referred to as “Geddes Unified Hypothesis” or Geddes III.  The unifying hypothesis is a proposition that the triad of symptoms associated with SBS can be caused by severe hypoxia (lack of oxygen in the tissues).  This may lead to brain swelling.  This, combined with raised intracranial pressure, could produce both subdural and retinal hemorrhages and can be seen in natural deaths as well as in alleged SBS cases.  Accordingly it was asserted that any incidents of apnea (the cessation of breathing) set in motion a cascade of events leading to the triad of symptoms often associated with SBS.  This questions the validity of the triad as a diagnostic tool and suggests that apnea may have many different causes, including trauma or abuse. 

Squier and Mack researched structures of the cranium, including the dura, arachnoid barrier and the bridging veins, and how these developed in infants. Their findings challenged beliefs that bridging vein rupture was the cause of subdural hemorrhage in this particular age group, and suggested that rupturing of a rich plexus of intradural vessels may be the cause of subdural hemorrhage without trauma.  A similar study noted decreased levels of cerebrospinal fluid in subdural hemorrhage without trauma, and acknowledged that the dura was incredibly complex and that a better understanding of it would provide important insights into subdural hemorrhages.  

A study of ten teenage athletes who had suffered acute subdural hemorrhage drew multiple parallels between these teenage athletes and infants diagnosed with non-accidental trauma.  The study suggested that a “second impact syndrome”,  which occurs when a second injury is sustained before the first injury has healed, may be relevant to the SBS debate.  In addition, the higher representation of males in cases of SBS/AHT was considered in a study which provided several biological explanations for such an overrepresentation,  which further undermined the case for the triad as pathognomonic of SBS as these causes mimicked child abuse.   

Retinal hemorrhages in infants have also been studied. In 2009, Togioka et al analyzed 66 academic articles that covered the subject, and concluded that although the presence of retinal hemorrhages did not guarantee SBS, it was rare in cases of non-accidental head injury whereas it appeared to be common in abused children.  However, studies have since explored a number of alternative causes for retinal hemorrhages in infants.  For example, it was suspected that a five-week old child had sustained a non-accidental injury after experiencing a rapid onset of symptoms including drowsiness and hypotonia, unilateral retinal hemorrhages and an intracranial hemorrhage in the posterior fossa.  An arteriovenous malformation was discovered at a repeat neuroimaging scan weeks later, which would have contributed to the retinal hemorrhages and this removed the suspicion of a non-accidental injury.  The autopsy findings from four children aged three or younger found that retinal hemorrhages, which had been caused by fatal consequences of infection complications, mimicked some patterns in retinal hemorrhages associated with non-accidental trauma.  

In addition, a study which looked at two babies who had experienced fatal head injuries in utero following motor vehicle accidents, noted that both had extensive retinal hemorrhages and optic nerve sheet hemorrhages with peripapillary intrascleral hemorrhages.  The authors commented that, on the rare occasions that they were discussed in the relevant literature, peripapillary intrascleral hemorrhages were considered pathognomonic for abusive head trauma.  As the babies had been in utero a shaking motion was highly unlikely. Therefore, peripapillary intrascleral hemorrhages should not necessarily be considered conclusively the agnostic of abusive head trauma.    

Experiments to improve our understanding of the symptoms of SBS/AHT have been carried out. These have had varying results, but no experiment has unequivocally confirmed that the triad of symptoms is pathognomonic to SBS/AHT. For example, the shaking of a computational model of an infant eye suggested that shaking alone was enough to cause retinal hemorrhaging.  The shaking of an automotive dummy showed that biomechanical investigations of SBS provided more accurate results “if the skull with paediatric features and the brain-skull interface are correctly represented.”  The vigorous shaking of anesthetized lambs found only minimal axonal injury, subdural hemorrhages and retinal hemorrhages, although there was widespread neuronal perikaryal amyloid precursor protein expression.  Finally, a computational model of an infant brain was subjected to periodic motion and the findings “provided an enhanced understanding of the effects of parameter uncertainty on the dynamics of SBS.”  

As early as 2006, it was widely recognized, that there were many “mimics” of SBS/AHT.   In a 2011 article, Squier discussed alternative and common causes for the triad of symptoms, including accidental falls, cortical vein and sinus thrombosis, inflicted injury, vitamin D deficiency, second impact syndrome, aneurysm rupture, rare genetic conditions, and resuscitation related injuries.  

In contrast, also in 2011, Narang argued that evidence-based medical literature supported the scientific soundness of AHT and, therefore, satisfied all the criteria under Daubert.  Narang further disregarded Geddes’ Unified Hypothesis and Squier and Mack’s dural immature vascular plexus theory, pointing out that studies existed that showed that retinal hemorrhages and subdural hemorrhages were statistically significantly correlated with AHT.    

Findley et al responded to the criticism and identified methodological problems with Narang’s article, including statistical misinterpretations and observer bias, and went on to clearly state the main scientific points behind the “new paradigm.”  Furthermore, as Findley et al point out, “Changing the name of the syndrome from SBS to AHT does not, however, resolve the disagreement.”  

What emerges from the above discussions is that this is a field in crisis. Medical, legal, academic and other scholars have voiced concern about the use of the SBS/AHT triad of symptoms as a tool for diagnosis and prosecution, as significant medical and scientific evidence discredits its very existence. Each of the symptoms contained in the triad have several natural and accidental causes in addition to abuse, so it is therefore difficult to see how the triad could be considered reliable in a medical or scientific sense. Maintaining that the triad is reliable in a legal sense is simply absurd. 

Furthermore, it needs to be questioned how objective the science around the triad and identification of SBS/AHT really is. Although the identification of symptoms might be objective (i.e. symptoms are present and correctly identified), concluding that these symptoms are indeed evidence consistent with non-accidental injuries appears to be a completely subjective exercise.   

In the US case of Del Prete v. Thompson in 2014, the court was explicitly suspicious of the SBS/AHT diagnosis.  In a footnote to the opinion, the court wrote that the testimony and evidence presented “arguably suggests that a claim of shaken baby syndrome is more an article of faith than a proposition of science.”  A year later, in early 2015, an open letter signed by an international group of 34 doctors, medical professionals, and international experts outlined the problems with how cases of SBS/AHT were being prosecuted by the courts in the US and E&W.  In the letter, the authors noted that “SBS has never been proved as anything more than an hypothesis”,  and urged both criminal and civil courts to listen to both sides of the disagreements around the diagnosis.

Despite the controversies outlined above, in E&W the Crown Prosecution Service (CPS) still rely on guidelines from March 2011 on how to prosecute cases of “non accidental head injury” (NAHI).  Changing the term from SBS to non accidental head injury does not change the fact that it is still the triad of symptoms that will be relied upon for a prosecution. The CPS states that “cases will usually be diagnosed by a Triad of internal head injuries”  and “will usually require the Triad of injuries plus supporting evidence.”   It is also states that Geddes’ Unified Hypothesis has not been endorsed by the CACD.  Although a positive step has been taken by acknowledging that the triad alone should usually not be the sole basis for a prosecution, there are still dangers associated with the supporting evidence, as this also needs to be reliable, relevant and objective. As discussed by Henneberg and Loveday in this special issue, the mere finding of the triad of symptoms may influence the police investigation, leading to tunnel vision and case construction.  Case construction occurs when the police and prosecution only focus on evidence which will support their hypothesis that a particular suspect is guilty, and any evidence to the contrary is excluded or its importance is minimized.  This may result in the prosecution having insufficiently reliable evidence against a suspect admitted in court in order to bolster its case.   

Studies referred to in this article show that medical and scientific evidence believed to be pathognomonic of SBS/AHT need to be considered as highly unreliable.  Edmond and San Roque state that adversarial legal systems assume that science and medical evidence is “epistemologically robust” although when scrutinized, a lot of such evidence “is either unreliable or of unknown reliability.”  Litigation-driven science may be part of the problem, at least in the US, as it keeps the SBS/AHT debate extremely polarised. 

The Law Commission identified a number of problems relating to the reliability of scientific evidence in criminal courts.  A contradiction can be seen as although courts have been shown to have a lax attitude towards evidentiary reliability, for example by allowing weak science such as ear prints into court, they appear to be holding on to outdated science for far too long. The question of admissibility then becomes a double-edged sword, and this is a problem in both the US and in E&W. To avoid this, courts need to recognize and consider significant developments relating to SBS/AHT, and use this to improve the quality of medical and scientific evidence before it is admitted into court.