SUPREME COURT OF QUEENSLAND

1. FRASER JA: The appellant was found guilty by a jury and convicted of the manslaughter of Rose Marie Williams, the 14 week old daughter of his girlfriend, Kimberley Williams, on 14 December 2007.  The appellant has appealed against his conviction on the following grounds:

“Ground 1:The verdict was unreasonable (unsafe and unsatisfactory) having regard to all of the evidence.

Ground 2:The Learned Judge erred in failing to provide proper directions of:

(a)The particulars of the circumstantial case against the Appellant.

(b)The use to be made of evidence of previous injuries to the deceased and the versions the Appellant gave of these.

(c)The lies.”

2. The appellant has also applied for leave to appeal against sentence. It is unnecessary to discuss that application because, for reasons to which I now turn, I have concluded that the appeal against conviction should succeed on the basis of grounds 2(b) and 2(c).  In order to explain that conclusion, and why I would not uphold ground 1, it is necessary to outline the Crown case and the evidence adduced at the trial.  The Court has the benefit of having considered the whole record with the assistance of extensive summaries of the evidence in the parties’ outlines, but it is necessary for present purposes only to summarise the more significant aspects of the evidence. 

The Crown case and the evidence

3. It was not in issue at the trial that the appellant had caused the death of the baby. In opening the case, the prosecutor acknowledged that the Crown could not prove precisely what happened to the baby, but suggested that the jury would be satisfied that the appellant had “either hit her in the head or hit her head against an object, and that’s there’s been an element of having her shaken as well.” It was not alleged that the shaking caused the baby’s death. Rather, the evidence relied upon for the allegation of shaking was led in support of the Crown case that the act which caused the death was conscious and willed and that the death did not result from an accident.

4. A nurse at the Proserpine Hospital, Karen Grace, gave evidence that at 8.50 am on Sunday, 9 December 2007 she received a telephone call from the appellant.  The appellant enquired whether he should be concerned about his little daughter who had a feed early in the morning at 5.30 to 6.00 am, and then went stiff and settled to sleep for a few hours.  The appellant said the baby was drowsy but easily rousable if he rubbed her feet or her hands.  The appellant told Ms Grace that the baby had a bump on her head two days ago but he did not think it was that.  The appellant said that one eye was dilated.  After speaking to a doctor, Ms Grace called the appellant back and told him to bring the baby into the hospital for assessment.

5. Dr Lewis gave evidence that she attended the emergency department at the hospital at about 10.20 am that morning. She saw the baby with the appellant, Kimberley, and the appellant’s mother, Cathie Shoesmith.  The baby was pale, unconscious, floppy, only responded to pain, had a very low pulse rate and very infrequent breathing.  Dr Lewis noticed that the baby had a linear (straight line) bruise on the left side of her forehead above the eyebrow, and two rows of three equidistant marks, slightly purplish in colour, on the tip of her nose.  She called in another doctor, Dr McCallum, who intubated the baby so that she could be ventilated.  Dr Lewis said that the appellant told her that the baby fell off a one and a half foot high bed two days earlier and landed on concrete. 

6. Skull x-rays ordered by Dr Lewis showed a fracture of the left parietal bone. The baby was subsequently transferred to Townsville General Hospital, where she died on 14 December 2007.  There was no challenge to the medical evidence to the effect that her death resulted from the force which caused the fracture to the parietal bone.

7. In cross-examination Dr Lewis agreed that the linear bruise was consistent with the baby’s head striking the edge of a coffee table. She confirmed her recollection that the appellant said that two days earlier the baby had fallen about a foot and a half from a bed and landed on concrete. She did not observe that the baby had a damaged fraenulum (a web of skin tissue inside the mouth joining the upper lip to the gum).  She agreed that a torn fraenulum was a possible side effect of the intubation process, although it would be very infrequent. 

8. Dr McCallum gave evidence that he had intubated several hundred newborn babies and he had no difficulties in intubating the baby. It was a straightforward intubation and he did not recall seeing any bleeding in the mouth that might suggest that he had caused trauma. He noticed the baby had indentations on the tip of her nose in a very regular pattern and the appellant suggested that they might match the pattern on the concrete floor of their house.  In cross-examination he agreed that he could not be 100 per cent certain that he did not damage the fraenulum in the intubation process. 

9. The appellant’s mother, Cathie Shoesmith, gave evidence that whilst the baby was being treated at the hospital the appellant told her that what had happened was that, while Kimberley was changing the baby’s nappy on their bed, she had leant on the end of the bed to throw something in the basket and the baby had fallen off the bed onto the cement floor in the bedroom. The appellant told her that the little marks on the baby’s nose were from a strip to which carpet was to be attached. Similarly, in a police interview on the same evening, the appellant said that on the morning of the previous day (Saturday 8 December) the baby had fallen from the bed onto the concrete floor of the bedroom when he was absent from the bedroom and Kimberley was dressing the baby.  The appellant told police that after that accident the baby was fine all day and perfect all night. 

10. In a subsequent police interview on Tuesday 11 December 2007, the appellant again told police about the baby falling off the bed in the bedroom when he was absent from the bedroom and Kimberley was changing the baby on Saturday, but he also referred to a subsequent accident on Sunday morning. The appellant said that, after feeding the baby, he was sitting on a chair with his legs up and his feet on the coffee table, with the baby lying on his legs facing away from him. He must have drifted off to sleep and he woke as the baby was falling between his legs. He tried to grab her as she fell, but that made it worse because he pushed her. The baby hit her head on the coffee table and she fell onto the floor. (I will call this “the defence scenario”).

11. The appellant was asked to demonstrate how the injury happened, using a doll. The video tape of that demonstration, as part of the police interview, was played to the jury. The demonstration suggested that the force of the baby’s fall was broken by the coffee table before she hit the concrete floor, rather than that the baby merely struck a glancing blow against the coffee table. (The respondent contrasted the demonstration with the evidence that the linear mark attributed to the baby’s head hitting the coffee table was far less serious than the fractured skull apparently attributed to her head hitting the concrete floor.)

12. The appellant told police that after this accident the baby cried for about 10 to 15 minutes and then fell asleep.  He took the baby in to Kimberley in the bedroom to give the baby the rest of her bottle, but the baby would not take it and she fell asleep.  The appellant said that he and Kimberley also fell asleep and when they awoke the baby would not open her eyes and would not move; she was like jelly and, when the appellant opened her eyelids, he saw that one eye was more dilated than the other.  The appellant rang his mother and the hospital.  In this second police interview the appellant initially said that he had told Kimberley that he had dropped the baby, but he subsequently said that she did not know and that he had not told her that he had dropped the baby.  He acknowledged that he had earlier attributed the injuries to the baby having fallen off the end of the bed with Kimberley.  He said that he had not admitted to the baby hitting her head on the corner of the coffee table because he was scared and in shock.

13. On the Crown case, the defence scenario was a false account revealing a consciousness of guilt. The Crown also relied upon the failure of the appellant to give that account initially, and his attribution of the fatal injury to the baby falling from the bed whilst in the sole care of Kimberley.  A further point in the Crown case in this respect was that Tui Hergatt (who owned the property where the appellant and Kimberley lived) gave evidence that on 8 December the appellant took the baby to his house, showed him a bump on the baby’s head, and said that “he’d” changed the baby’s nappy and “as he went to dispose of the nappy when he came back he found [the baby] laying on the floor.” 

14. At the appellant’s police interview on 11 December, he volunteered that the weekend before the serious injury, the baby had “split her lip on the little web just inside her mouth.” The appellant said that this occurred as he leant across the microwave when the baby turned around and “smacked her mouth just above, under her nose and just above her lip on the microwave”. The appellant used half a box of wipes trying to stop the bleeding. The appellant made those statements before anyone else had noticed this injury to the fraenulum, so far as the evidence reveals.

15. Police searched the appellant’s house on 10 and 11 December 2007. Blood smears, which were admitted to be the baby’s blood, were found on the cement floor beside the coffee table and on a tissue in the bathroom. It was admitted that the height of the coffee table was approximately 40 to 45 centimetres.  A police officer gave evidence that the height of the bed in the bedroom was 36 centimetres. 

16. Professor Lamont, a radiologist with a speciality in paediatric radiology, reviewed CT scans of the baby taken between 9 and 11 December 2007 and an MRI scan taken on 11 December. He gave evidence of a Y shaped skull fracture in the parietal bone. In his opinion the impact causing the fracture would have occurred at the centre of the Y shape, most likely by the rounded surface of the skull impacting against a flat surface. The limbs of the Y extended upwards into other plates which was unusual and suggested that a very substantial force was applied. Professor Lamont gave evidence that the defence scenario did not fit the baby’s injuries. In his opinion, the major impact would have been on the frontal area of the skull where it struck the edge of the coffee table. He would have expected to have seen a very substantial soft tissue swelling over the front of the skull and a fracture at that site. In his opinion, the fall from the coffee table to the floor would not produce a severe, extensive fracture of the sort seen in the scans.

17. Dr Sive, a paediatrician, was involved in the care of the baby when she was brought to the Townsville General Hospital on 9 December 2007.  When he saw the baby he was concerned that the history he had been given that she had fallen one to one and a half feet two days earlier did not match her injuries.  To fracture a baby’s skull there would have to have been more than a fall from 40 or 50 centimetres.  In Dr Sive’s opinion, it would be highly unlikely for the defence scenario to cause sufficient force in the relatively soft pliable skull of the baby to lead to the degree of injury she suffered.  Dr Sive also referred to haemorrhaging he had seen in the back of the baby’s eye and considered it highly unlikely that the defence scenario accounted for those injuries.  He did not think that the eye injuries could have occurred from the fall itself.  In his opinion, the eye injuries would occur in three situations, namely, a “major crush injury; motor vehicle accidents often with a car rolling; and … acceleration, de-acceleration, forces applied to the baby … what we call shaking.”

18. Professor Williams, a pathologist, conducted an autopsy on 17 December 2007. He described the fracture to the parietal bone as extending towards the back of the head into the occipital bone in the shape of a dog’s hind leg and 8.7 centimetres long. He said that there was a significant subdural haemorrhage on the brain, brain swelling, and areas of contusion on the brain. He also described a 2.6 centimetre linear bruise on the left side of the forehead, with an adjacent small bruise, bruising to the left side of the scalp, and a torn fraenulum. The linear bruise was caused by forcible contact with a straight edge. The separate, quite small bruise may have been caused by some kind of other implement. He did not believe that the defence scenario explained the head injuries because there was insufficient height, and babies fall from a lot greater heights than that without bruising themselves, having a fracture or sustaining any injury. He expressed the opinion that the injuries were non‑accidental injuries, especially the torn fraenulum.

19. Dr Gole, a paediatric ophthalmologist, reviewed the retinal scans which had been taken of the baby’s left eye on 10 December 2007. He said that the scans revealed retinal haemorrhages and a fold in the retina. Dr Gole said that there were a large number of causes of retinal haemorrhages “but by far the commonest cause of extensive retinal haemorrhages which go from side to side inside the eye” is shaking. The more severe and widespread and extensive a retinal haemorrhage is, the more likely it is that it was due to a rapid to and fro shaking. In his opinion retinal folds were caused by a “rapid to and fro movement”. He considered that the commonest cause of retinal folds is a shaking injury to the head. They have also been seen in crush injuries to the head, including a large television falling onto a toddler’s head and a 63 kilogram sibling sitting down and crushing a four month old sibling’s head, fatal traffic injuries involving a car rolling over and killing a baby, and a baby falling 11 metres off a balcony and landing head first.  Retinal haemorrhages and a fold in the retina had not been described in short falls.

20. During the autopsy on 17 December 2007 Professor Williams noted that the fraenulum had been torn. Professor Williams gave evidence this injury was usually caused by a hand being slapped across the face, but it could be caused by other shearing forces causing the skin to slide across the jaw. Dr Sive gave evidence to similar effect.

21. The Crown also adduced evidence that the baby had suffered other injuries whilst in the appellant’s care and that he had given explanations of those injuries which were inconsistent with explanations attributed to him in other evidence. I will outline this evidence when I discuss appeal ground 2(b). I do not propose to take that evidence into account against the interests of the appellant when considering ground 1. It is necessary to note here merely that defence counsel adduced evidence in cross-examination of those Crown witnesses who had come into contact with the appellant that he seemed to have the primary care of the baby and that he behaved towards her as a loving and caring father. That evidence is relevant in the appellant’s favour under ground 1.

22. The appellant called evidence in his defence from two forensic pathologists, Professor Hilton and Professor Duflou. Professor Hilton considered that the defence scenario was consistent with the injuries that the baby sustained. In relation to the retinal haemorrhage, Professor Hilton agreed that a substantial body of paediatricians believe that retinal haemorrhage was strongly presumptive of physical abuse, but that it had been recognised that increased intra-cranial pressure is a cause of retinal haemorrhages in a number of cases; the swelling of the baby’s brain relayed into cranial pressure in this case.

23. Defence counsel referred Professor Hilton to Dr Gole’s evidence attributing the retinal haemorrhages to acceleration and deceleration of the baby’s head and noted that Dr Gole made reference to a fold in the retina as well as the extensive retinal haemorrhages, but Professor Hilton was only asked to comment upon the likely cause of the retinal haemorrhage (although it may be that the retinal fold was treated as an example of a retinal haemorrhage). Professor Hilton referred to a hypothesis widely held within the ophthalmic community that the haemorrhage is “entirely or partly caused by an oscillation of the eyeball in association with oscillation of the head” and commented that there is “a fairly reputable and fairly respectable body of evidence which refutes this hypothesis.” In a subsequent answer Professor Hilton noted that when the retinal haemorrhages were observed there was severe brain injury and evidence of brain swelling and that “the interpretation of the significance of these retinal haemorrhages has to be made - taking cognisance of being aware of the fact that there is brain swelling and its possible association with retinal haemorrhage.” The proposition that retinal haemorrhage could only have been caused through an acceleration and deceleration of the baby’s head was “something that has to be considered, but it’s by no means a conclusion that one can lead to.” In Professor Hilton’s opinion, the fact that the forces generated when an infant’s head impacts on the ground tend to exceed or greatly exceed the forces that can be induced by even a strong person shaking a baby is one of the weaknesses which the earlier proponents of “shaken baby syndrome” failed to face up to. The absence of other injuries associated with that “syndrome” persuaded Professor Hilton that the baby was not the subject of “baby shake”. He referred to the absence of injuries to the muscles or any of the other soft structures of the neck, fingertip or thumb bruises on the skin of the chest, or fractures of the ribs near the backbone. He added that whether that ruled out shaken baby syndrome or not was a moot point.

24. When it was put to Professor Hilton in cross-examination that a retinal fold had not been reported in any literature as being caused by anything other than a fatal crush injury, a motor vehicle accident, or shaking of children, Professor Hilton did not disagree. He added that the most widely reported one had been in association with “shaken baby, whatever that may be.”

25. Professor Duflou expressed the opinion that it was “possible” that the defence scenario fitted in with the baby’s injuries and it was “reasonably possible” that the fall was “capable of explaining the injury …”. He also gave evidence that there were many theories about the cause of retinal haemorrhage; extensive shaking of an infant can cause it, but the presence of any one of the “triad” of retinal haemorrhage, bleeding over the surface of the brain, and a sudden collapse in brain function did not mean that the infant had been shaken. In his opinion retinal haemorrhage was not a sign necessarily of an abused child. Five years, and probably even two years ago, Professor Duflou would have agreed with Dr Gole in ruling out any other causes of the retinal haemorrhage, but he now considered that extensive retinal haemorrhage occurred in a wide variety of different cases and that retinal folds have also been seen in various different types of cases “on occasion”. In this case, a possible cause was a very severe head injury.  Also, the medical records referred to superficial cortical vein thrombosis and Dr Plunkett (a forensic pathologist) had presented a paper pointing out that thrombosed blood vessels over the surface of the brain were strongly associated with retinal haemorrhage.  Although Professor Duflou did not know enough about cortical vein thrombosis to have a firm opinion, he wondered if this baby’s superficial cortical vein thrombosis was linked with the extensive retinal haemorrhage.  Professor Duflou thought that the impact of the baby’s head and the associated head injury was an “entirely reasonable” cause of the retinal haemorrhages and retinal fold observed in this case.

26. In cross-examination Professor Duflou agreed that a retinal fold had not been reported in any literature as being caused by any mechanism other than by a motor vehicle accident, a fatal crush injury, or violent shaking of children. In relation to a fatal crush injury, Professor Duflou observed that a paper by Lantz indicated that a television set toppling onto your head can cause it, but until that paper was written, that was said to have been impossible. 

27. Professor Duflou said he did not know whether the appellant’s statement to police that the baby had knocked her top lip on a microwave would have been a very forcible impact and it was difficult to know how much force would have been required to cause the torn fraenulum. A slap in the face can cause a torn fraenulum, but other mechanisms such as the pushing in of a bottle can do it as well. In his opinion the fact that there was a torn fraenulum was not necessarily indicative that the baby had been physically abused by way of a slap or something similar.

28. Professor Duflou gave evidence that Professor Williams’ post mortem examination would not have passed the quality assurance program in his department. In particular, Professor Williams should have examined all the layers of the eye and optic nerve sheath. Professor Duflou had not seen optic nerve sheath haemorrhaging in the absence of significant head trauma, and such haemorrhaging could cause retinal haemorrhaging. Professor Duflou would have also examined the spinal cord because with “so-called shaken baby syndrome” the major problem was a whiplash type injury in the lower brain stem or upper cervical spine at the very back of the head.

29. In cross-examination Professor Duflou agreed that the best current estimate of mortality rate for short falls affecting infants and young children was below .48 deaths per one million per year.  It was rare for a short fall to cause these sorts of severe head injuries resulting in death.  Professor Duflou thought that the baby must have hit a glancing blow against the edge of the coffee table, because a direct strike against the edge of the table with a propulsion (presumably, the appellant attempting to grab the baby) would produce a more severe injury to the forehead than the baby sustained.

Appeal Ground 1: the verdict was unreasonable

30. Ground 1 raises the question whether, in terms of s 668E(1) of the Criminal Code 1899 (Qld), the verdict of the jury should be set aside on the ground that it is unreasonable or cannot be supported having regard to the evidence.  The test is whether upon the whole of the evidence it was open to the jury to be satisfied beyond reasonable doubt that the accused was guilty.^[1]  The Court must conduct an independent review of the evidence, but it must also bear in mind that the jury had the benefit of seeing and hearing the witnesses give their evidence and it must accord respect to the jury’s resolution of the contested factual questions reflected in the guilty verdict.^[2]  However, in light of my conclusion that ground 2(b) is made out, I think it inappropriate in this case to attribute any weight to the jury’s verdict.

31. The appellant acknowledged that, as I would conclude, the effect of the expert evidence given by Professor Williams, Dr Gole and Dr Sive was that the defence scenario was not a reasonable hypothesis to account for the serious fractures to the skull, the retinal haemorrhages, or the retinal fold, and that the torn fraenulum was a key indicator that the child had been physically abused.  The appellant argued, however, that concessions made by those Crown witnesses and conflicts between their evidence and the evidence of Professor Hilton and Professor Duflou meant that there was a reasonable doubt about the guilt of the appellant.

32. I will first discuss the concessions made by the Crown witnesses to which the Court was referred.

33. Dr Sive agreed that: the exact science as to the cause of retinal haemorrhages had not been determined; that there was now more reason to look for causes of retinal haemorrhages other than “baby shake” than was the case even as recently as 10 years ago; causes such as impact head injury can result in retinal haemorrhages; and in the absence of knowledge of any particular cause, the “primary diagnosis” would be one of child abuse or inflicted head injury.  However, whilst Dr Sive agreed that brain swelling was one of the causes of retinal haemorrhage, Dr Sive gave evidence in cross-examination that not all forms of retinal haemorrhage were the same.  He said that in this case the retinal haemorrhages extended to folds, the causes of those folds were major crush injuries, motor vehicle accidents, and inflicted trauma with shaking, and he had not seen reports of those sorts of haemorrhages from raised intra-cranial pressure secondary to brain swelling from a subdural haemorrhage. 

34. The appellant referred to Dr Sive’s subsequent answer that “I can’t be 100 per cent certain, but it seems that the kind of retinal haemorrhage, as that is described here, do not result from brain swelling”, but it is apparent that Dr Sive was considering the question of scientific certainty. A jury is not bound to find a reasonable doubt merely because of the absence of scientific certainty.^[3]  That is true also in relation to Dr Sive’s agreement with the propositions put in cross-examination that: short falls in infants, although uncommon, “have been known to cause severe and sometimes lethal head injury”; the height of the fall is “not necessarily a prohibitor for a child to sustain severe and lethal head injury”; a fall from a short height “can result” in a severe fracture; and he could not rule out the scenario described by the appellant as one of the causes for the skull fracture.  A further answer by Dr Sive, that in the “vast majority or [sic] instances” a height of some 40 or 50 centimetres was not sufficient to generate the forces necessary to fracture the skull, might not itself have been sufficient to exclude a reasonable doubt, but it is necessary to consider that answer in the context of Dr Sive’s other evidence and the evidence of the other witnesses.

35. The appellant referred to Professor Lamont’s agreement with the proposition in cross-examination that there were “any number of variables” which would dictate the rate at which the baby would make contact with the floor, but he added that it was “just hard to envisage a variable which would actually accelerate the child significantly between the coffee table and the floor.”  Professor Lamont was cross‑examined about the possibility that the baby struck the coffee table in a glancing blow.  He disagreed with that scenario referring to the absence of grazing or other marks suggesting that kind of impact.  When cross-examined about the significance of the absence of such marks, Professor Lamont agreed that would not always be the case but his evidence was that in medicine “nothing is always always”, and he would still expect to see some grazing in that scenario.  He agreed that the whole of medicine confirmed defence counsel’s proposition that he could express an opinion as to whether something was likely or unlikely, but could not be more confident.  Again it seems clear that in this evidence Professor Lamont was concerned with scientific certainty, the absence of which does not necessarily produce a reasonable doubt.  He went on to explain that he had seen many cases of accidental and non-accidental injuries to the skull in babies in his 25 years as a paediatric radiologist and that, whilst short linear fractures of the skull were common and frequently seen, “these major fractures … the amount of force that causes those, you’re talking about road traffic accidents, you’re talking about babies falling from a height of two storeys … not these little … sofa heights.”

36. The appellant referred to the cross-examination of Professor Williams in which he agreed that the fracture of the skull of itself would be consistent with the child falling and hitting a hard surface but that his difficulty in accepting the defence scenario was the height from which the child fell. Professor Williams felt that the height was “a bit too low, really.” When it was put to him that severe and lethal head injury following short falls in infants is not common but does occur, Professor Williams said that it was “ascribed” but frequently there is an “unreliable history in that, you know, you’re looking at these studies and people who seem quite smart about what’s happened may give you a false indication of what the height actually was.” When it was put to him that falls from even relatively small heights involving infants can result in severe head fractures and death, Professor Williams responded that it was “thought so” but he tended to believe that it had to be more than a metre in height to have caused a fracture. He disagreed that the height the baby fell from in this case could account for the head fracture and the underlying subdural haemorrhage and brain swelling. In that context, Professor Williams’ subsequent answer accepting that it was “at least a possibility” did not require the jury to harbour a reasonable doubt upon the basis of his evidence.

37. Dr Gole accepted that there was controversy about the cause of retinal haemorrhage, but he gave evidence that the current understanding was supported by the “overwhelming preponderance of evidence both based on confessions of perpetrators and then correlating with what you see with mechanical dolls et cetera”, and that the current understanding was that retinal haemorrhages were a “shearing injury caused by rapid to and fro mechanical movement”. Dr Gole said that it was necessary to examine the particular retinal haemorrhage, and the more widespread it was the more likely it was that there had been a shaking injury. He said that subdural haemorrhages, fractures and retinal haemorrhages remain “highly suggestive of an abusive injury to the child” and that the most prominent of the people who had postulated other causes had been “thoroughly discredited.” In his opinion, swelling of the brain rarely causes retinal haemorrhages in children; where it did (in a condition called “Terson Syndrome”) there were only two reported cases in children, each of which resulted from aneurisms. Furthermore, Dr Gole did not retreat from his evidence that retinal folds had not been reported in the literature as arising from short falls. In the result, Dr Gole’s concessions in cross-examination did not materially weaken his evidence to the effect that the particular extent and nature of the retinal haemorrhage and retinal fold in the baby’s left eye strongly suggested that the injury was caused by rapid to and fro movement rather than by the defence scenario.

38. In relation to the torn fraenulum, in the context of the far more serious injuries confronting the doctors at the hospital it does not seem especially surprising that it might have been present but not noticed until the autopsy. There was no expert evidence that, in the usual course of the medical investigations into the more serious injury, it should have been noticed earlier, or that the failure to notice it suggested that it was not present. Dr McCallum’s evidence was to the effect that he had not damaged it in the intubation process. His concession that he could not be 100 per cent certain about that did not require the jury to harbour a doubt about it. The jury was also not bound to accept the appellant’s version that this injury occurred by accidental contact with the microwave. The jury might have concluded that when the appellant changed his version in the police interview of 11 December he must have appreciated that the blood soaked tissue which he mentioned in that interview would be found where he had left it. In that context, the significance of the change in the appellant’s version supplied a reasonable ground for discounting the appellant’s statement that the blood resulted from an accident in which the baby came into contact with the microwave. The possibility that the injury was caused accidentally was a matter the jury was required to consider, but the jury might reasonably have thought that it was merely speculative. It was open to the jury to take into account the expert evidence adduced in the Crown case about the nature of this injury in considering whether the Crown had proved beyond reasonable doubt that the fatal head injury was not sustained in an accident.

39. Notwithstanding the concessions relied upon by the appellant, the evidence which I have outlined, if accepted by the jury, was capable of establishing beyond reasonable doubt that the defence scenario could not account for the nature and extent of the baby’s injuries. 

40. The evidence of Professor Duflou and Professor Hilton provides some support for the appellant’s argument, but it does not follow that a verdict of guilty was not reasonably open when the whole of the evidence is considered. Neither Professor Hilton nor Professor Duflou suggested that the fall described in the defence scenario was likely to cause the baby’s head injury. Professor Duflou’s evidence on that topic was given in the heavily qualified terms quoted in [25] of these reasons.  It is also significant that neither Professor Duflou nor Professor Hilton contradicted the evidence of Dr Gole that retinal folds had not been reported in short falls or otherwise than as a result of fatal crush injuries, motor vehicle accidents, or violent shaking of children, and in one case where a child fell 11 metres.  Professor Duflou conceded in cross-examination that there was no other relevant report of retinal folds, and Professor Hilton did not refer to any other such report.  The jury could also take into account that Professor Hilton was not an ophthalmologist and that, when he was asked whether he agreed that brain swelling does not cause massive haemorrhages, he responded in the highly qualified terms that he could not be in “unqualified agreement”, “it has been certainly postulated”, but that he “would not pretend to know everything about other retinal haemorrhages, nor the forces that produce them.”

41. The appellant submitted that there were conflicts in the expert evidence which necessarily created a reasonable doubt. This feature of the evidence called for care in framing directions to the jury about the expert evidence,^[4] but this is not a case in which the conflicting evidence itself necessarily required the jury to hold a reasonable doubt.  It was not a case which depended wholly upon conflicting expert opinions on scientific questions which were incapable of critical evaluation by the jury.^[5]  The jury could analyse the bases of the expert opinions, which were explained in detail in each case.  In addition, there was other evidence.  There was evidence that the appellant delayed in disclosing the defence scenario in circumstances in which, the jury might have concluded, the appellant would have appreciated that might be important information for the treating doctors and medical staff.  That topic was canvassed with the appellant in the second police interview.  Notably, there was also evidence that the appellant admitted that he had initially given a false explanation that the injuries occurred whilst the baby was in the care of her mother.

42. The appellant also referred to the evidence discussed in English decisions reported in 2006 which were treated as test cases.^[6]  Those decisions were not referred to for any legal proposition, but rather as evidence concerning the reliability of medical opinion about retinal haemorrhaging, encephalopathy, and subdural haemorrhaging, as clinical indicators of shaken baby syndrome.  To the extent that the expert opinions discussed in the cases remain current, it was open to the parties to adduce similar evidence at the trial.  Reference to the reports of the cases is not a substitute for evidence which was not adduced at the trial. 

43. For the reasons I have given there was a reasonable basis in this case for the jury to accept the critical aspects of the evidence of the Crown experts, notwithstanding the evidence of the experts called by the appellant. Professor Hilton and Professor Duflou are very well qualified and their evidence called for anxious consideration, but it remained open to the jury to prefer the expert evidence adduced by the Crown about the great force required to produce the baby’s injuries, and the significance of the torn fraenulum, the retinal haemorrhage, and the retinal fold. Since there was acceptable evidence that the baby suffered the fatal injury whilst in the appellant’s sole care, it was reasonably open to the jury to find that the Crown had proved beyond reasonable doubt that the appellant caused the baby’s injuries by a conscious and willed act and that the baby’s death was not the result of an accident. 

Grounds 2(b) and 2(c): failing to provide proper directions of the use to be made of evidence of previous injuries to the deceased, the versions the appellant gave of these, and the lies

44. A general medical practitioner, Dr Carr, gave evidence that the appellant and Kimberley brought the baby to his surgery on 28 November 2007.  The appellant said that the baby did not seem to want to lift her arm above her head and he was bringing her in to check out why.  On examination he noticed a lump on the top surface of her left clavicle which was not tender to touch.  He asked whether the baby had sustained any injuries and the appellant replied that the baby had sustained no injuries.  He diagnosed a fracture of the left clavicle which was well into the process of healing.  Dr Carr advised the appellant and Kimberley that the injury seemed most consistent with a fracture that was in the process of healing but that there was a possibility that it could be a growth or tumour so it needed follow up.  He printed out a referral for the baby to the x-ray department.  In cross-examination Dr Carr estimated that the injury must have occurred on or before 7 November, that is, three or more weeks before he saw the baby.

45. The Crown adduced evidence from Kimberley’s mother, Mrs Williams, that on 1 December 2007 the appellant told her that, whilst he was pushing the pram and Kimberley was walking ahead of him, the baby fell out of the pram onto her arm when the pram wheel got caught in a rut in the ground.  The appellant gave the same explanation for the broken collarbone to police when interviewed on 11 December 2007, but the Crown relied upon two further statements in that interview.  First, the appellant told police that the doctor had said that “it was fine.”  However, the appellant added that he had been given a referral for an x-ray.  Secondly, the Crown relied upon the fact that the appellant had not taken the baby for an x-ray despite the medical advice to do so.  The appellant told police that he and Kimberley did not follow that advice, initially because they could not get into town (neither the appellant or Kimberley drove), then they forgot about it, and subsequently because the appellant’s mother later said that if they took the baby for an x-ray there might be an investigation by Child Services into the safety of the baby.  The appellant’s mother gave evidence but she did not say that she had given any such advice.

46. The appellant’s sister, Kylie Shoesmith, gave evidence that the appellant explained to her that the baby’s arm was sore and that he believed that was because Kimberley’s mother, Mrs Williams, would pick up the baby from the top of the arms close to the shoulders.  However Kylie Shoesmith’s answers in cross-examination suggested that this sore arm might have been suffered before the time when, on Dr Carr’s evidence, the baby’s collarbone was probably broken.

47. There was also evidence that the baby sustained a bruise to her face at a large party held by the appellant’s mother on 30 November. The appellant’s mother and Kimberley’s mother gave evidence that the appellant attributed the bruise to an accidental blow to the baby by a man called Shaun.  Shaun Piggins gave evidence that he did not cause the injury.  Tui Hergatt gave evidence that the appellant told him that the appellant’s mother was holding the baby when one of the guests at a party accidentally bumped into the baby.  The appellant’s mother gave evidence that the appellant said that he was holding the baby at the time.  Kimberley’s mother gave evidence that the appellant told her that his mother was holding the baby at the time, and the bruise was not visible by 4 December 2007.  The appellant told police on 11 December 2007 that on Saturday 8 December there was “a little bit on her left cheek from the bruise that was already there.”

48. The relevance of the evidence of previous injuries to the baby was the subject of debate at a pre-trial hearing on 1 March 2011 when defence counsel applied to exclude evidence which the Crown foreshadowed would be given by Professor Lamont that x-rays taken on 9 December 2007 showed fractures to the baby’s legs.  Such evidence had been adduced in a previous trial at which the jury failed to reach a verdict.  The primary judge ruled against the admissibility of the evidence of the fractures to the legs and that evidence was not adduced at the trial.  At the preliminary hearing there was also discussion about the evidence of the bruise and the fractured collarbone.  Defence counsel did not submit that this evidence was not admissible.  Although the discussion about this lacked clarity, it seems that the primary judge accepted that this evidence would be put before the jury on the basis that the appellant told lies about it, that the jury might conclude that the appellant was responsible for those injuries, and that the evidence was therefore admissible for various possible reasons, including to exclude accident.

49. In subsequent discussions during the preliminary hearing, reference was made to the possibility that a jury might conclude that the appellant had given a false account in relation to the earlier injuries. The trial judge observed that the jury would have to be instructed that they could not reason from a conclusion that the appellant had been responsible for the earlier injuries that he must therefore have been responsible for the fatal injury, but rather, it was evidence the jury could take into account in considering the question whether the fatal injuries occurred accidentally.  Such a warning about impermissible propensity reasoning had been given at the first trial where the Crown adduced evidence about the fractures to the baby’s legs.  However, no such warning was given to the jury at this trial.

50. The appellant submitted that the absence of any such warning was particularly significant because of the use which the prosecutor made of the evidence of previous injuries in final address.

51. The respondent submitted, that the jury would not adopt impermissible propensity reasoning because the prosecutor did not contend that the earlier injuries were deliberately inflicted. Those were minor injuries consistent with accidents, and the prosecutor did not seek to contradict the evidence that they were caused by accidents.  The respondent submitted that the prosecutor’s argument was merely that the jury should reject the appellant’s version in his 11 December police interview about how the baby sustained her fatal injury.  As the respondent acknowledged, if the evidence related only to the credibility of the appellant it was questionable whether the evidence was admissible in the Crown case,^[7] but the respondent submitted that the evidence was also admissible to rebut the appellant’s case that he behaved as a loving and caring father towards the baby.  That is a legitimate argument, but that ground for the admission of the evidence is consistent with the Crown relying upon it to demonstrate that the appellant had wilfully caused the earlier injuries, with an associated risk that the jury might reason that the appellant was guilty for that reason or because he was the kind of person who would commit the offence.

52. Reference to the prosecutor’s final address makes good the appellant’s submission that it was necessary for the trial judge to warn the jury against impermissible forms of propensity reasoning.  Although the prosecutor did not say that the jury should find that the appellant was responsible for the earlier injuries, or that this suggested that he was guilty of the offence, the jury might have understood the submission to encourage such an approach.  The prosecutor sought support for the Crown case in the earlier injuries and in alleged lies in the appellant’s statements about them and other matters.  I will set out the relevant section of the prosecutor’s address, with the most significant passages highlighted:

“Now, Sir Walter Scott, he was a Scottish poet, and he once penned, ‘What a tangled web we weave when we first we practise to deceive.’ We practise to deceive. Well, you can be sure of one thing, your web of lies will catch you out, and you may think that the accused’s lies have done just that to him.

Let’s look at the litany of lies he has told. Rose Williams was born on the 10th of September 2007. What does the accused tell his mother, Cathie Shoesmith, knowing it not to be true?  ‘She’s my baby.’ It’s a lie.

Rose Williams is given a clean bill of health on the 10th of - 9th of November 2007 by Dr Dodman. She’s examined, there are no marks on her, yet a little over four weeks later she’s dead.

On the 28th of November 2007, Rose was presented to Dr Carr – Dr Carr’s surgery. Remember this is the collarbone injury.  Dr Carr asked if Rose had suffered any injuries. The accused, who doctor said - Dr Carr said did the majority of the talking, what did he say? ‘No, no injuries.’ Yet three days later, on the 1st of December, he tells Sue Williams that Rose fell from the pram when the pram got struck in a rut, and Kimberley was in front of him. He told Kimberley - told the police that Kimberley was behind him. Caught in the web of inconsistency?  Matter for you.

Associate Professor Lamont dates the injury at three weeks, around the same time the accused’s moved into Tui’s and the pram incident. He used that bit of evidence to say that Lamont is - is - is a credible witness.

Now, remember what the accused said to police on the 11th of December, that Dr Carr said Rose was fine. He said, ‘I didn’t think it was a big deal.’ Not a big deal? Dr Carr also said there was a possibility of a tumour, yet it’s not a big deal?

He says he was worried about Child Services. This is a man, by his own admission, had Rose and Kimberley relying on him, and he’s more worried about Child Services than a tumour - possible tumour.

You may recall on the same occasion on the 1st of December Sue Williams inquired about the bruise on Rose’s face, and was this accused’s answer, ‘Well, while Rose was being nursed by his mother, Cathie Shoesmith, she’d accidentally been hit by Shaun Piggin.’

The accused told Tui Hergatt that Rose was being nursed by his mother, Cathie Shoesmith. Cathie Shoesmith also asks what happens and happened, and he says, ‘Shaun accidentally hit Rose.’ Remember her evidence. She wanted to confront Shaun, and the accused said, ‘No, no.’ You heard Shaun Piggin last week, and he said, ‘No, didn’t touch the child. The only thing was put my finger out and she grabbed it.’ You may think another lie told by the accused. So the accused shifts responsibility to Shaun Piggin.

Recall what his sister, Kylie Shoesmith, said when he asked her how Rose hurt her arm, said, ‘Sue Williams, the way she picks her up.’ You may think, that’s another diversion. It’s a matter for you.

Now, still on the bruise, you may recall when my learned friend asked Mrs Williams about that bruise, she said, ‘That bruise had started to fade when Rose, Kim and herself returned from the Clarion Hotel in Mackay, a three day stay.’ Said, ‘The bruise had started to face.’ Interesting then that the accused told the police that the bruise in the photograph taken on the 8th of December 2007, that was a bruise from the party. Now, you may think it doesn’t stack up. Yet another one of the - yet one for the web of deceit.

The 8th of December 2007; Tui says the accused brought Rose over about midday, worried about a bump on the head. The accused said that Rose had fallen from the bed after he had changed her nappy. To police on the 11th of December, he said it was Kimberley was in the room, and the incident occurred early in the morning. He also tells Sergeant Taylor on the 9th of December, that Kimberley was the person present when Rose fell from the bed.

Now, on the 9th of December, Rose is presented to Proserpine Hospital. Cathie Shoesmith asks him, ‘What happened to Rose?’  In the absence of Kimberley, what does the accused say? He told Cathie Shoesmith that Rose fell from the bed, and yes, Kimberley was there. You may recall that he also said to Sergeant Taylor about her profanities.

You may again think the shifting of responsibility, the weaving of the web of deceit, the injuries to the nose, carpet strip, concrete or hospital blanket. Yet, another weaving of the web. You may think that every time there is an injury to Rose this accused has an explanation.

…

All of Rose’s injuries lead to one place, that’s to this accused. You will, I submit, that - find that this accused cannot be believed. You will be satisfied that because of his lies and the medical opinion, that you cannot believe his final version as to what happened to Rose Williams. That’s the lies.

You may think they’re linked with the injuries as well, but let’s look at the injuries and you’ve heard one or more of the doctors say that you need to look at all of the injuries.” (errors as in original)

53. In relation to the collarbone injury, the prosecutor’s argument was that the appellant was “[c]aught in the web of inconsistency” because he told Dr Carr that the baby had not suffered any injuries, his version to police was that the baby was injured falling out of a pram whilst Kimberley was behind him, and he told Mrs Williams that Kimberley was in front of the pram when the baby fell out.  The jury may well have understood the alleged lies to comprehend the appellant’s version to police that he did not tell Dr Carr about the injury because his mother mentioned that Child Services might become involved, particularly because the appellant’s mother did not corroborate that version.  Although the prosecutor did not make this explicit, there was an available implication that the appellant’s statements about this injury were illustrations of the prosecutor’s opening submission on this topic that a “web of lies will catch you out”.

54. In relation to the facial bruising, the prosecutor’s argument was that the jury should find that the appellant “shifts responsibility” for the accident away from himself to another having regard to the appellant’s alleged inconsistent versions about who was nursing the baby at the time she was injured, the alleged inconsistency between the appellant’s versions and Shaun Piggin’s evidence, and the evidence that he discouraged his mother from confronting Shaun Piggins. Similarly, the prosecutor asked the jury to find that the appellant was “shifting … responsibility” for the baby’s bump on the head on 8 December 2007 because the appellant told Tui Hergatt that the baby fell after he changed her nappy but he told police and his mother that Kimberley was with the baby when she fell.

55. The prosecutor concluded this section of his address with two arguments. First, he argued that “[a]ll of [the baby’s] injuries lead to one place, that’s to this accused.” Secondly, he argued that, because the appellant had lied about the earlier injuries and other matters, the jury would not believe the appellant’s final version to police about the fatal injury. The second argument concerned the effect of the appellant’s alleged lies about the earlier injuries (and other topics mentioned by the prosecutor) upon the credibility of his final version to police, but the first argument was not confined to that topic. Consistently with the earlier references to the appellant shifting responsibility for the earlier injuries, the prosecutor’s first argument encompassed an invitation to reason towards guilt from a conclusion that the appellant was responsible for the earlier injuries.

56. The trial judge gave the following directions about how the jury should use any conclusion that the appellant had lied:

“Now, let me say something about the subject of lies. You’ve heard a good deal said about what are said to be lies on the part of the accused, and it is clear from what is said in the interviews, that some of the things he had said were not the truth when speaking to the police and at least one of the people at the hospital.

It is for you to decide what significance those lies have in relation to the issues in this case. You may decide that if you find the accused has lied, that that only affects his credibility, and indeed that is the position, it seems to me, in this case. Do not follow a process of reasoning to the effect that because a person is shown to have told a lie about something, that that is evidence of guilt. The mere fact that an accused tells a lie is not in itself evidence of guilt. A person may lie for many reasons, for example, to bolster a true defence, to protect someone else, to conceal disgraceful conduct on his part, short of the commission of the offence or out of panic or confusion.

If you think that there is or may have been an innocent explanation, that is innocent in terms of the offences for which he’s charged, you should take no notice of the lies and in this case, as I have said, the lies, if you find them to be as such, do not provide any evidence of guilt against the accused.”

57. Those directions related to things the appellant told to the police and at least one of the people at the hospital. That overlooked the significance of the other alleged lies constituting the “web of deceit”. When the trial judge subsequently summarised the evidence, his Honour did not refer back to the direction about lies. But because the prosecutor implied that all of the alleged lies about the earlier injuries pointed to the appellant’s guilt, it was necessary for the trial judge to give directions about each of them in accordance with Edwards v The Queen:^[8]

“Thus, in any case where a lie is relied upon to prove guilt, the lie should be precisely identified, as should the circumstances and events that are said to indicate that it constitutes an admission against interest. And the jury should be instructed that they may take the lie into account only if they are satisfied, having regard to those circumstances and events, that it reveals a knowledge of the offence or some aspect of it and that it was told because the accused knew that the truth of the matter about which he lied would implicate him in the offence, or, as was said in Reg. v. Lucas (Ruth), because of ‘a realization of guilt and a fear of the truth’.

Moreover, the jury should be instructed that there may be reasons for the telling of a lie apart from the realization of guilt. A lie may be told out of panic, to escape an unjust accusation, to protect some other person or to avoid a consequence extraneous to the offence. The jury should be told that, if they accept that a reason of that kind is the explanation for the lie, they cannot regard it as an admission. It should be recognized that there is a risk that, if the jury are invited to consider a lie told by an accused, they will reason that he lied simply because he is guilty unless they are appropriately instructed with respect to these matters. And in many cases where there appears to be a departure from the truth it may not be possible to say that a deliberate lie has been told. The accused may be confused. He may not recollect something which, upon his memory being jolted in cross-examination, he subsequently does recollect.” (footnotes omitted)

58. The trial judge did not identify any of the alleged lies or give the directions which the High Court has held are necessary in such a case. His Honour also did not explain how the jury could legitimately use the evidence of the earlier injuries or give any direction in relation to the prosecutor’s argument that “[a]ll of [the baby’s] injuries lead to one place, that’s to this accused.” The absence of those directions may have resulted in a miscarriage of justice.

59. Furthermore, the prosecutor apparently intended to advance a case that the appellant’s statements about the earlier injuries amounted to deliberate lies which revealed some unspecified consciousness of guilt connected with those injuries.^[9]  Precisely what inferences arose and how they implicated the appellant in the offence were not clearly articulated, but some such case was implicit in the prosecutor’s address.  In light of the generality of the prosecutor’s argument, the jury might have understood the Crown case to be that the evidence of the baby’s earlier injuries and the appellant’s alleged lies about them revealed that the appellant had a propensity to injure the baby intentionally, or at least with culpable carelessness, and, for that reason, tended to prove that the baby’s death was not accidental.

60. If the evidence was capable of having that effect, it might have been admissible under s 132B of the Evidence Act 1977 (Qld) even if it did not pass the test in Pfennig^[10] that there is no rational view of the evidence consistent with the innocence of the accused.^[11]  However s 130 of the Evidence Act preserved the trial judge’s discretion to exclude the evidence if its probative value did not exceed its potential prejudice to the appellant.^[12]  At the trial the appellant did not object to the admission of the evidence and the notice of appeal did not contend that the evidence was not admissible for any purpose.  Accordingly, it is not appropriate to rule upon the questions whether the evidence was admissible and, if so, whether it should have been excluded in the exercise of the trial judge’s discretion.  The evidence having been admitted for the purposes encompassed within the prosecutor’s address, however, it is necessary to consider whether the trial judge should have given a direction about propensity reasoning.

61. In Roach v The Queen,^[13] French CJ, Hayne, Crennan and Kiefel JJ observed:

“The importance of directions in cases where evidence may show propensity should not be underestimated. It is necessary in such a case that a trial judge give a clear and comprehensible warning about the misuse of the evidence for that purpose and explain the purpose for which it is tendered. A trial judge should identify the inferences which may be open from it or the questions which may have occurred to the jury without the evidence. Those inferences and those questions should be identified by the prosecution at an early point in the trial.”

62. An allegation that an accused has wilfully killed a baby in his care is bound to excite emotions. In this case there were also allegations that the baby had previously suffered injuries whilst in the appellant’s care and that he had lied about the causes of those injuries. The very nature of the Crown case called for particular care in framing directions to ensure a fair trial. Clear and specific directions about the significance of alleged lies and the use of propensity reasoning were demanded by the prosecutor’s argument that the appellant had been caught out in his attempts to shift responsibility for the earlier injuries, and that the jury should find that all of the injuries led to the appellant.

63. In my respectful opinion, the trial judge’s directions were insufficient to avoid the risk of a miscarriage of justice occurring in this case as a result of the jury adopting an impermissible form of propensity reasoning. In the result, the appellant has not had a fair trial.

64. In reaching that conclusion, I have taken into account my conclusion that defence counsel’s omission to ask for more comprehensive directions about the alleged lies and for directions about propensity reasoning was not reasonably explicable by any rational forensic objective. In light of the extensive body of evidence devoted to the earlier injuries and the appellant’s statements about them, the trial judge’s detailed reference to that evidence, and the emphasis in the prosecutor’s submissions that the evidence pointed to the appellant’s guilt, a request for redirections would not have worked to the appellant’s disadvantage by unduly emphasising that aspect of the evidence.

65. The respondent’s counsel accepted that, if the Court considered that the trial judge’s directions were inadequate to guard against impermissible propensity reasoning, a new trial should be ordered.  That was a responsible concession.  The Crown case was not overwhelming.  The baby’s mother was not called to give any evidence which implicated the appellant and the Crown case depended in part upon expert evidence which conflicted with expert evidence adduced by the appellant.  The appeal could not be dismissed under s 668E(1A) of the Criminal Code on the ground that no substantial miscarriage of justice has actually occurred.

66. Accordingly, I would allow the appeal on this ground.

67. I would add that, at an early point in any new trial, the prosecutor should clearly identify each inference which is alleged to be open in relation to the evidence of the baby’s earlier injuries and the appellant’s statements about them.  That exercise should cast light upon the admissibility of the evidence, the purpose for which the evidence is admissible, and appropriate directions concerning any alleged lies and any impermissible form of propensity reasoning by the jury.

Proposed orders

68. I would allow the appeal, set aside the conviction and sentence, and order a new trial.

69. MULLINS J:  I agree with Fraser JA.

70. DOUGLAS J:  I agree with the reasons of Fraser JA and with the orders proposed by his Honour.