England and Wales Court of Appeal (Criminal Division) Decisions

IN THE HIGH COURT OF JUSTICE
COURT OF APPEAL (CRIMINAL DIVISION)
ON APPEAL FROM the Crown Court sitting at Kingston on Thames
Her Honour Judge Matthews QC
T20087418

		Royal Courts of Justice Strand, London, WC2A 2LL
		15/12/2009

LORD JUSTICE HOOPER:

1. This appeal raises the issue of what directions should be given to the jury when a defendant charged with causing death by dangerous driving in one count submits that he did not cause the death of a person, but that a driver in another vehicle, whose death the defendant is also said to have caused in another count, did. Lawyers have traditionally used the rubric "novus actus interveniens" to describe the issue raised in this case. We shall use those words or translate them as new and intervening act or event.
2. The appeal also raises the issue of whether, in these circumstances, the verdicts on the two counts would have to be the same.
3. On 22 July 2009 in the Crown Court at Kingston-upon-Thames (Her Honour Judge Matthews QC) the appellant was convicted unanimously of causing the death by dangerous driving of Mr Preston Trewick, count 1. The jury were unable to reach a verdict on count 2, causing the death by dangerous driving of Catherine Cunningham. A retrial on count 2 has been ordered by the Crown Court.
4. At the conclusion of the hearing we announced that we were allowing the appeal, quashing the conviction and ordering a retrial. Mr Royle accepted that the retrial which we have ordered would be heard together with the retrial which the Crown Court has ordered.
5. Because we are ordering a new trial, it is not necessary to go into the facts in any detail. The unchallenged case was that the appellant's vehicle collided with a black taxi cab driven by Mr Preston Trewick at shortly after midnight on an unlit stretch of the west bound A3. There was an issue as to whether both rear lights of the taxi were illuminated (not an issue, however, about which the judge was to give specific directions). The weather was clear and the road damp. At the point of the collision the west bound A3 consists of three lanes with a slip road joining the A3, making four lanes. There is also a hard shoulder. The jury found, contrary to the appellant's case, that he was driving dangerously.
6. The collision propelled the taxi into the fast lane leaving it broadside on to the traffic going in the same direction. There was an issue as to how visible the taxi was to oncoming traffic. It was accepted that Mr Trewick was alive and moving at this time. Following the collision some cars passed the taxi, with or without difficulty, and some at least stopped and displayed hazard lights. The appellant's car was also there. There was debris in all three lanes. Not long after the collision between the appellant's car and the taxi and whilst the taxi remained broadside in the fast lane, the car which Catherine Cunningham was driving collided with the taxi killing both her and Mr Trewick. Samples of Catherine Cunningham's blood showed that it contained 55 milligrams of alcohol in 100 millilitres, less than the legal limit.
7. As to the cause of the death of Mr Trewick and Catherine Cunningham the defence case was, in the words of the judge when summing up the case to the jury:

But what the defence say – and you'll clearly have to think about it carefully – is that other drivers managed to steer round the taxi in their path and it was the fact that Miss Cunningham did not manage to steer round the taxi in her path that was the cause of the deaths. You'll have to give this careful consideration, and you'll be looking at the time, the circumstances, the location, what other witnesses have said about the unexpectedness of coming on the scene and finding that in front of them. The prosecution word used on Friday was the mayhem in front of them. But you're going to have to grapple with the fact that some drivers did escape colliding with the taxi, at least in the way that Miss Cunningham did, and consider what the reason and the explanation for that is. I remind you the defence have said that Miss Cunningham was shown to have had something to drink but was certainly not over the drink/drive limit, but that her driving, her reactions, may have been impaired. It may be that those who were passengers in vehicles were able to take on board more of what was in front of them than drivers who told you that they were distracted by the, what was going on to the left of the scene with the Audi and its lights flashing. So all those kind of considerations you're going to have to take on board when considering where the cause lies.

8. It follows that the defence case was that Miss Cunningham should have seen what the other drivers saw and that she was therefore solely responsible for the accident and for the death of Preston Trewick and herself. It seems likely that at least some members of the jury decided that she was or may have been responsible for her own death otherwise the jury would have convicted on both counts.
9. Mr Connolly accepts that there was evidence that she and not the appellant was responsible for both deaths such that the issue had to be left to the jury and appropriate directions given. Mr Connolly, unsurprisingly, submitted that the judge should direct the jury that they could not as a matter of law come to different verdicts. The judge accepted that it was highly unlikely that the jury could come to different verdicts but decided, with the concurrence of Mr Royle, not to give the jury the direction which Mr Connolly was seeking. Mr Royle now agrees with Mr Connolly.
10. The judge directed the jury that they "must consider the case for and against the defendant on each of those counts separately".
11. Mr Connolly submitted that the judge gave the appropriate directions to the jury to apply when considering the defence case on causation. She did so, he submits, in the passage set out in paragraph 6 above and in her general direction on causation in which she said:

If you are sure that the driving was dangerous, then go on to consider whether that caused the deaths in Count 1 of Mr Trewick and in Count 2 of Miss Cunningham. As I've said, and it's on the handout, you do not have to be sure that the dangerous driving was the principal or substantial cause of the death, as long as you are sure that there was something more than a slight or trifling link.

12. Mr Connolly rightly did not dispute that this passage by itself was insufficient guidance for the jury when the issue of novus actus interveniens is raised. A person's driving could have caused the death of another applying this test albeit that there was a clear novus actus interveniens.
13. Mr Connolly accepted that a further direction was necessary to guide the jury about this issue. He submitted that the passage set out in paragraph 7 above was sufficient. We do not agree. It is not enough, in a case of this kind, to say what the defence case is without giving the necessary legal directions for resolving the issue.
14. The jury sent a note after about a day and a half from the start of their deliberations. The note read:

Please can we have paragraph four, the last paragraph of the legal directions, clarified?"

15. The fourth paragraph read:

You do not have to be sure that his dangerous driving was the principal or substantial cause of death, as long as you are sure that it was a cause and that there was something more than a slight or trifling link.

16. The matter was discussed with counsel. Mr Royle submitted to the judge unsuccessfully (as he had earlier before the start of the summing up) that she should direct the jury in accordance with Skelton [1995] Crim LR 635 (CA). The facts of that case have some similarity with the facts in the present case. The dangerous driving in that case consisted of driving an unroadworthy lorry on a motorway. The hand brake system was activated due to lack of pressure and the vehicle came to a halt in the near side lane of the motorway. A number of vehicles avoided the lorry but 12 minutes later another lorry crashed into the stranded lorry and the driver, S, was killed. The appellant was convicted of causing death by dangerous and driving. It was submitted on appeal that the danger in the driving was spent by the time that S was killed. Sedley J giving the judgment of the court said that although there may come a point in which the appellant's driving was too remote, that point had not been reached. Sedley J said, in a passage relied upon by Mr Royle, that the dangerous driving must have played a part, not simply in creating the occasion of the fatal accident but in bringing it about. Mr Royle told the judge that Skelton had been approved in Barnes [200] EWCA Crim 2726; [2009] RTR 21. We shall look at Barnes a little later.
17. Mr Connolly opposed the submission that the judge should direct the jury in accordance with Skelton on the grounds that such a direction would mislead the jury and lead to a wrongful acquittal. The judge agreed with Mr Connolly.
18. After reminding the jury of the contents of the note and the fourth paragraph, the judge said this:

The prosecution do not have to make you sure that it was his driving which literally killed them. By that I mean that it was his car that caused the injuries. They don't have to make you sure about that. But they do have to make you sure that it was the manner of his driving which set off a chain of events that was more than a slight or trifling link to their deaths in the second collision. It may help to ask yourselves this question: are you sure the Defendant caused one or both deaths – not the accident, but caused one or both deaths – by driving dangerously in leaving the taxi immobilised in the middle of the fast moving A3, which put that taxi driver at real risk of being hit by another vehicle, or not?

19. The judge subsequently reduced this direction into writing and the written direction was given to the jury. That read:

1. You do not have to be sure that his dangerous driving was the principal or a substantial cause of a death as long as you are sure that it was A cause and that there was something more than a slight or trifling link.

2. Prosecution do not have to make you sure that it was his driving which literally killed them, ie that it was his car that caused the injuries which killed them;

3. The prosecution must make you sure that it was the manner of his driving which set off a chain of events that was more than a slight or trifling link to their deaths in the second collision.

4. May help to ask yourselves this question

Are you sure the defendant caused one or both deaths, not the accident, but caused one or both deaths by his dangerous driving leaving the taxi immobilised in the middle of the fast moving A3 whereby he put the driver of the taxi at real risk of being hit by another vehicle, or not?

20. Not long after the jury convicted the appellant of count 1. About two hours later the jury informed the judge that they were unable to reach a majority verdict on count 2 and they were discharged.
21. Before us Mr Royle repeated his submission that the judge should have directed the jury in accordance with Skelton. Mr Royle submitted that the phrase "set off a chain of events" obscured the distinction between creating the occasion of the accident and actually bringing it about. He complains of the lack of guidance as to what "chain of events" meant in the light of the fact that other drivers had avoided the taxi before the fatal collision. Nor did Mr Royle like the phrase "put the driver of the taxi at real risk" in the last sentence of the direction, in which the judge is suggesting a question which the jury might like to ask.
22. Mr Connolly accepts that the verdict should have been the same on both counts in the light of the direction. Mr Royle submits that the verdicts are logically inconsistent and reveal that the jury did not understand the directions.
23. Mr Connolly opposes the proposed Skelton direction, saying that it was "such a subtle concept that it would not have assisted the jury". He went on to argue in his skeleton argument that the jury should perhaps have been directed "Was the only cause of each death the action of Catherine Cunningham, or some failure on the part of Preston Trewick (eg to turn his lights on)?"
24. At the conclusion of the hearing counsel were asked to consider what was said about novus actus interveniens in Smith and Hogan, Criminal Law, 12ed., page 78 (something which they had not done) and submit, if possible, an agreed direction. Subject to the underlined words which are not agreed, the proposed direction reads:

...

The prosecution do not have to make you sure that it was his driving that literally killed them – i.e. that it was an impact with his car that inflicted the injuries that killed them.

The prosecution must make you sure that the dangerous driving caused the death, not just the preceding collision.

You do not have to be sure that his dangerous driving was the only cause of death or even that it was a substantial cause of the deaths. There may be a number of causes, including for example the negligent actions of the deceased, or other drivers at the scene.   It is no defence to say that if others had been more careful or more alert it could have been avoided but you must be sure that:

1.   It was a cause, and

2.  That no other cause had effectively replaced it, and

3.  That there was something more than a trifling link between his dangerous driving and the deaths

It is the Crown's case that the defendant's dangerous driving set off a chain of events which was more than a slight or trifling link to their deaths in the second collision.

The Defence say that regardless of the manner in which he drove, the prosecution cannot prove to you that the deaths were caused by the chain of events set off by his driving because there may have been intervening acts of other persons that effectively replaced it ("broke the chain of causation"). There was or may have been an act or omission on the part of Catherine Cunningham or Preston Trewick that was so significant that the contribution of the effect of the Defendant's driving was so trifling that it should be discounted.

The prosecution must make you sure that the defendant's conduct remained a cause and that, despite the actions of Ms Cunningham or Mr Trewick it was still a cause of the death that was more than trifling.

There are two separate allegations simply because there are two deaths. Because they arose out of the same circumstances and the one impact it follows that if the defendant is guilty of one count he would be guilty the other and accordingly if not guilty of one count then not guilty of the other.

25. We turn to Barnes.  The appellant, convicted of causing death by dangerous driving, was driving a truck which was carrying what the jury concluded was an unsafe load, namely a sofa. With him was a passenger helping him to move the sofa from one property to another. In the words of Hallett LJ:

3. The appellant's journey took them along the A1139, a dual carriageway with a speed limit of 70 miles per hour. The road conditions were good, and the appellant's speed was estimated variously at between 50 and 60 miles per hour. A short distance along the road the sofa suddenly flew out of the back of the truck, probably taken into the air by the wind. It landed on the inside lane of the dual carriageway.

4. The appellant and Mr Wilson said they had been conscious of the fact that the sofa might come loose and had been watching it carefully. They had ensured that there was no distracting noise in the cab. When they realised the sofa was missing they stopped some 105 metres further on in the inside lane. ... The appellant seems to have parked as best he could with his nearside wheels off the carriageway. He put on his hazard lights, and his passenger, Mr Wilson, got out to try to retrieve the sofa from the carriageway.

5. A number of vehicles came upon the sofa in the carriageway and tried to avoid it. They included a motorcyclist, a Mr Darren Wildman. He successfully avoided the sofa, but having done so, he must have been distracted or turned to warn others behind him. In so doing, he either failed to see the appellant's stationary vehicle at all or failed to see it in time. A matter of seconds after the Mitsubishi had stopped, Mr Wildman drove into the back of it. Tragically he died at the scene.

26. A little later Hallett LJ said:

11. The first ground is that HHJ Enright erred in refusing the submission of no case to answer, and in particular that he was wrong to find that there was evidence upon which a jury could conclude that the appellant's dangerous driving caused Mr Wildman's death. Mr Bridge relied principally on the decision of this court in Skelton [1995] Crim LR 635, of which we have a transcript of the judgment delivered on 25 November 1994 (94/2914/W5). Mr Bridge relied upon the following principle (see page 10 of the transcript per Sedley J):

"…the dangerous driving must have played a part, not simply in creating the occasion of the fatal accident but in bringing it about."

12. Albeit the judge directed the jury in accordance with this principle, Mr Bridge argued the judge himself never properly addressed this question. Had he done so, he would have felt obliged to withdraw the case from the jury. We disagree. In our judgment, this was a case very properly left to the jury. We accept Mr Bridge may have had at his disposal many very powerful arguments, but it was for the jury to decide whether or not they found the prosecution witnesses and, in particular, Miss Ferguson reliable and accurate. If they did, they were entitled to conclude that the appellant was driving a truck with a heavy load at or over 50 miles per hour along a busy road where the speed limit was 70 miles per hour. On Miss Ferguson's version of events, the sofa was obviously working itself free as he drove along. It must have been banging noisily in the back of the Mitsubishi. If so the appellant could and should have driven in a different fashion, or he should have stopped earlier than he did. Instead, he continued on his journey until his load fell off and into the path of oncoming vehicles. He stopped with his vehicle still partly on the carriageway. His actions made both the Mitsubishi and the sofa immediate hazards to oncoming traffic.

13. The jury was entitled to find that the appellant put other road users at risk by driving dangerously. He drove with a load which was insecure. Had he not done so the sofa would not have fallen off, and Mr Wildman would not have been forced to drive round it. He would not have been distracted by it or turned to warn others coming behind him. The appellant's car would not have been stopped in the carriageway and Mr Wildman would not have driven into the back of it. Whatever criticisms Mr Bridge could properly make of Mr Wildman's driving, in our judgment all those circumstances are such that it was open to the jury to find that his dangerous driving played more than a minimal role in bringing about the accident and the death.

14. We turn therefore to the further criticisms made of the judge by Mr Bridge. The second ground of appeal is that the judge, it is said, failed adequately to sum up the law in respect of causation. The judge summed up the law in this way:

"Now the words 'thereby caused the death'. You have to be sure the dangerous driving was a cause of death, not the only cause of death or the main cause of death, but a cause of death which was more than just trivial. This means you must be sure that not only the defendant's dangerous driving created the circumstances of the fatal collision but it was an actual cause in bringing about the death of Mr Wildman. And the defence say here, you might be satisfied the defendant had created the circumstances of the collision but  and they say, and they recognise it is an unattractive argument  and they say it is nonetheless right  the only cause of death was Mr Wildman failing to keep a proper lookout. And if that is so, or may be so, I direct you to acquit."

15. Mr Bridge had a number of complaints to make about this passage. He argued that the judge failed in a number of respects. He suggested the judge failed to explain in the kind of detail the jury required in a case "as complex as this" what the difference was between driving which created the circumstances of the fatal accident as opposed to driving which was an actual cause in bringing about the death of Mr Wildman. He gave examples in his own address to the jury, and in his submission it was incumbent upon the judge to do much the same in the summing up. He also complained about the fact that the judge told the jury that the defence recognised their argument might appear unattractive given they appeared to be blaming Mr Wildman, the deceased, for his own death.

16. Finally on this passage, Mr Bridge criticised the judge for focusing on the defence argument that the cause of death here was Mr Wildman's failure to keep a proper lookout. This he argued would have led to the members of the jury being distracted from their main task which was to focus on the issue of whether or not the appellant's driving was the cause of Mr Wildman's death.

17. We accept that, in principle, the distinction between dangerous driving which creates the circumstances of a fatal collision and dangerous driving which is the actual cause of a death may not be an easy concept to grasp. There may well be circumstances in which it would be preferable if a judge went into a little more detail than the judge did here. However, we note that on the facts of this case the stark issue was whether or not the cause of Mr Wildman's death was his own driving. That issue was left fairly and squarely to the jury. The judge directed them specifically that they had to be sure that if they found the appellant drove dangerously with an insecure load they should consider the second element. Was the dangerous driving with a load that might work itself free and flip out onto the road a cause of Mr Wildman's death which was more than just trivial? We reject, therefore, Mr Bridge's criticism of the judge's directions on the question of causation. (Underlining added)

27. We agree with the underlined words in paragraph 17. We do not think that the passage in Skelton relied upon by Mr Royle would give a jury much help in resolving a novus actus interveniens issue.
28. Barnes was concerned with the conduct of the victim. Did the failure on the part of the motor cyclist to see the truck parked with its rear nearside wheels off the carriageway break the chain of causation? The Court of Appeal, in paragraph 13, referred to the fact that the motor cyclist, but for the presence of the sofa on the road, would not have been distracted by it or turned to warn others coming behind him. The Court also said that it was open to the jury to find that his dangerous driving played more than a minimal role in bringing about the accident and the death. The jury, so the Court of Appeal held, should answer that by asking themselves: "Whether the dangerous driving with a load that might work itself free and flip out onto the road was a cause of Mr Wildman's death which was more than just trivial".
29. Although there was in the present case an issue about the taxi driver's rear lights, the important issue concerned the conduct of a third party, Catherine Cunningham whose collision with the taxi was the immediate cause of the taxi driver's death. But given that the appellant in count 2 was alleged to have caused her death too, the issue, in this respect, was similar to the issue in Barnes.
30. What, in our view, the jury needed in this case was a simple test to decide whether the driving of Catherine Cunningham was, as the defendant contended, such a new and intervening act that it could not be said that the appellant caused either the death of the taxi driver or her death.
31. In Pagett (1983) 76 Cr.App.R.279 the Court approved the following passage from Hart and Honoré, Causation in the Law (2^nd edn., 1985), Chapter 12, Criminal Law: Causing Harm at p. 326:

The free deliberate and informed intervention of a second person, who intends to exploit the situation created by the first, but is not acting in concert with him, is normally held to relieve the first actor of criminal responsibility.

32. Examples of new and intervening acts may be found in Kennedy [2007] UKHL 38; [2008] 1 AC 269; [2008] 1 Cr App R 19 and in Rafferty [2007] EWCA Crim 1846 (neither of which were cited in argument before us). Kennedy concerned the appellant's conviction for manslaughter in the following circumstances:

3. ... The appellant lived in a hostel in which Marco Bosque and Andrew Cody, who shared a room, also lived. On 10 September 1996 the appellant visited the room which Bosque and Cody shared. Bosque was drinking with Cody. According to Cody, Bosque told the appellant that he wanted "a bit to make him sleep" and the appellant told Bosque to take care that he did not go to sleep permanently. The appellant prepared a dose of heroin for the deceased and gave him a syringe ready for injection. The deceased then injected himself and returned the empty syringe to the appellant, who left the room. Bosque then appeared to stop breathing. An ambulance was called and he was taken to hospital, where he was pronounced dead. The cause of death was inhalation of gastric contents while acutely intoxicated by opiates and alcohol.

33. Lord Bingham giving the considered opinion of the appeal committee that the appellant was not guilty of manslaughter said:

14. The criminal law generally assumes the existence of free will. The law recognises certain exceptions, in the case of the young, those who for any reason are not fully responsible for their actions, and the vulnerable, and it acknowledges situations of duress and necessity, as also of deception and mistake. But, generally speaking, informed adults of sound mind are treated as autonomous beings able to make their own decisions how they will act, and none of the exceptions is relied on as possibly applicable in this case. Thus D is not to be treated as causing V to act in a certain way if V makes a voluntary and informed decision to act in that way rather than another. There are many classic statements to this effect. In his article "Finis for Novus Actus?" (1989) 48(3) CLJ 391, 392, Professor Glanville Williams wrote:

"I may suggest reasons to you for doing something; I may urge you to do it, tell you it will pay you to do it, tell you it is your duty to do it. My efforts may perhaps make it very much more likely that you will do it. But they do not cause you to do it, in the sense in which one causes a kettle of water to boil by putting it on the stove. Your volitional act is regarded (within the doctrine of responsibility) as setting a new 'chain of causation' going, irrespective of what has happened before."

In chapter XII of Causation in the Law, 2nd ed (1985), p 326, Hart and Honoré wrote:

"The free, deliberate, and informed intervention of a second person, who intends to exploit the situation created by the first, but is not acting in concert with him, is normally held to relieve the first actor of criminal responsibility."

This statement was cited by the House with approval in R v Latif [1996] 1 WLR 104, 115. The principle is fundamental and not controversial.

15. Questions of causation frequently arise in many areas of the law, but causation is not a single, unvarying concept to be mechanically applied without regard to the context in which the question arises. That was the point which Lord Hoffmann, with the express concurrence of three other members of the House, was at pains to make in Environment Agency (formerly National Rivers Authority) v Empress Car Co (Abertillery) Ltd [1999] 2 AC 22. The House was not in that decision purporting to lay down general rules governing causation in criminal law. It was construing, with reference to the facts of the case before it, a statutory provision imposing strict criminal liability on those who cause pollution of controlled waters. Lord Hoffmann made clear that (p 29E-F) common sense answers to questions of causation will differ according to the purpose for which the question is asked; that (p 31E) one cannot give a common sense answer to a question of causation for the purpose of attributing responsibility under some rule without knowing the purpose and scope of the rule; that (p 32B) strict liability was imposed in the interests of protecting controlled waters; and that (p 36A) in the situation under consideration the act of the defendant could properly be held to have caused the pollution even though an ordinary act of a third party was the immediate cause of the diesel oil flowing into the river. It is worth underlining that the relevant question was the cause of the pollution, not the cause of the third party's act.

 16. The committee would not wish to throw any doubt on the correctness of Empress Car. But the reasoning in that case cannot be applied to the wholly different context of causing a noxious thing to be administered to or taken by another person contrary to section 23 of the 1861 Act. In R v Finlay [2003] EWCA Crim 3868 (8 December 2003) V was injected with heroin and died. D was tried on two counts of manslaughter, one on the basis that he had himself injected V, the second on the basis that he had prepared a syringe and handed it to V who had injected herself. The jury could not agree on the first count but convicted on the second. When rejecting an application to remove the second count from the indictment, the trial judge ruled, relying on Empress Car, that D had produced a situation in which V could inject herself, in which her self-injection was entirely foreseeable and in which self-injection could not be regarded as something extraordinary. He directed the jury along those lines. The Court of Appeal upheld the judge's analysis and dismissed the appeal. It was wrong to do so. Its decision conflicted with the rules on personal autonomy and informed voluntary choice to which reference has been made above. In the decision under appeal the Court of Appeal did not follow R v Finlay in seeking to apply Empress Car, and it was right not to do so.

17. In his article already cited Professor Glanville Williams pointed out (at p 398) that the doctrine of secondary liability was developed precisely because an informed voluntary choice was ordinarily regarded as a novus actus interveniens breaking the chain of causation:

"Principals cause, accomplices encourage (or otherwise influence) or help. If the instigator were regarded as causing the result he would be a principal, and the conceptual division between principals (or, as I prefer to call them, perpetrators) and accessories would vanish. Indeed, it was because the instigator was not regarded as causing the crime that the notion of accessories had to be developed. This is the irrefragable argument for recognising the novus actus principle as one of the bases of our criminal law. The final act is done by the perpetrator, and his guilt pushes the accessories, conceptually speaking, into the background. Accessorial liability is, in the traditional theory, 'derivative' from that of the perpetrator."

This is a matter of some significance since, contrary to the view of the Court of Appeal when dismissing the appellant's first appeal, the deceased committed no offence when injecting himself with the fatal dose of heroin. It was so held by the Court of Appeal in R v Dias [2002] 2 Cr App R 96, paras 21-24, and in R v Rogers [2003] EWCA Crim 945, [2003] 1 WLR 1374 and is now accepted. If the conduct of the deceased was not criminal he was not a principal offender, and it of course follows that the appellant cannot be liable as a secondary party. It also follows that there is no meaningful legal sense in which the appellant can be said to have been a principal jointly with the deceased, or to have been acting in concert. The finding that the deceased freely and voluntarily administered the injection to himself, knowing what it was, is fatal to any contention that the appellant caused the heroin to be administered to the deceased or taken by him.

 18. The sole argument open to the crown was, therefore, that the appellant administered the injection to the deceased. It was argued that the term "administer" should not be narrowly interpreted. Reliance was placed on the steps taken by the appellant to facilitate the injection and on the trial judge's direction to the jury that they had to be satisfied that the appellant handed the syringe to the deceased "for immediate injection". But section 23 draws a very clear contrast between a noxious thing administered to another person and a noxious thing taken by another person. It cannot ordinarily be both. In this case the heroin is described as "freely and voluntarily self-administered" by the deceased. This, on the facts, is an inevitable finding. The appellant supplied the heroin and prepared the syringe. But the deceased had a choice whether to inject himself or not. He chose to do so, knowing what he was doing. It was his act."

34. Applying Kennedy, if the immediate cause of the death is the "free, deliberate, and informed" intervention of another person, then the chain of causation will be broken. There are, it seems to us, problems in applying that test in this case. Miss Cunningham's driving which led to the collision with the taxi driver cannot readily be described as falling within that category even if her driving was careless or dangerous (and we are not saying that it was).
35. If the only test were to be the "free, deliberate, and informed" test, then a driver who by his dangerous driving causes a minor accident pushing another car on to the hard shoulder would be automatically be liable for causing death by dangerous driving if the driver of a vehicle travelling at speed on the hard shoulder accidentally collided with the car on the hard shoulder whatever the circumstances, provided the jury took the view that the "more than a trifling link" test was satisfied. In our view juries needed a tailored direction to help them decide, in the appropriate case, whether there has been a new and intervening act or event.
36. In devising such a test we remind ourselves that the offences of causing death by dangerous and careless driving will punish the conduct of a person who has not intended or necessarily foreseen the consequences of his driving. Such a person is in a very different position to a person who has intended to kill or cause serious bodily harm or who has the mens rea for manslaughter. As Lord Hoffman said in the Empress Car case (page 29 F and 31E):

... common sense answers to questions of causation will differ according to the purpose for which the question is asked. Questions of causation often arise for the purpose of attributing responsibility to someone, for example, so as to blame him for something which has happened.

... one cannot give a common sense answer to a question of causation for the purpose of attributing responsibility under some rule without knowing the purpose and scope of the rule.

37. What we need is a form of words which sets out a test (comparatively easy to apply) which places an outside limit on the culpability of a driver in circumstances where there is more than a trifling link between the dangerous (or careless) driving and a death. It seeks to exclude consequences which are simply "too remote" from the driver's culpable conduct.
38. Help in devising such a test comes from the passage in Smith and Hogan to which we have made reference. In paragraphs 4.5.6 and 4.5.6.1 (at pages 78-79) the author states that a person will not have caused something to have happened if a natural event which is not reasonably foreseeable was the immediate cause of the event. The test receives support from Clause 17(2)(c) of the Draft Criminal Code for England and Wales Law Comm. No 177 (1989). It also has similarities to the test to be applied when the defendant's unlawful and violent acts are said to have caused another person to suffer injuries in the course of escaping from those acts, see R v Williams (1992) 95 Cr App R 1.
39. On the other hand, in the Empress Car case (referred to in paragraph 33 above) which involved an offence of strict responsibility, Lord Hoffman said (at page 34E) that "foreseeability is not the criterion for deciding whether a person caused something or not. People often cause things which they could not have foreseen". We note that in this passage Lord Hoffmann is referring to the defendant's foresight whereas the test which we are considering is concerned with what could reasonably have been foreseen.
40. Lord Hoffman preferred a test which made an alleged polluter liable for "acts and events which, although not necessarily foreseeable in the particular case, are in the generality a normal and familiar fact of life" but not for "acts or events which are abnormal and extraordinary" (page 34F). Lord Hoffmann said in his conclusions (page 36B):

If the defendant did something which produced a situation in which the polluting matter could escape but a necessary condition of the actual escape which happened was also the act of a third party or a natural event, the justices should consider whether that act or event should be regarded as a normal fact of life or something extraordinary. If it was in the general run of things a matter of ordinary occurrence, it will not negative the causal effect of the defendant's acts, even if it was not foreseeable that it would happen to that particular defendant or take that particular form. If it can be regarded as something extraordinary, it will be open to the justices to hold that the defendant did not cause the pollution.

41. It is important to note that Lord Hoffmann was of the view that the foreseeability test was too favourable to the alleged polluter, taking into account that the offence was one of strict responsibility. Causing death by dangerous (or careless) driving are not offences of strict responsibility.
42. In practice we do not think that there is much difference in cases like the present between a reasonable foreseeability test and the test propounded by Lord Hoffmann but, on balance, we prefer the objective test of reasonable foreseeability in a case like the present where the defendant's case is that there was a new and supervening act or event. We bear in mind that Lord Bingham, in Kennedy, said in the passage which we have already set out in full (paragraph 33 above) that:

The House was not in that decision purporting to lay down general rules governing causation in criminal law

43. We are of the view that the words "reasonably foreseeable" whilst apt to describe for a lawyer the appropriate test, may need to be reworded to ease the task of a jury. We suggest that a jury could be told, in circumstances like the present where the immediate cause of death is a second collision, that if they were sure that the defendant drove dangerously and were sure that his dangerous driving was more than a slight or trifling link to the death(s) then:

the defendant will have caused the death(s) only if you are sure that it could sensibly have been anticipated that a fatal collision might occur in the circumstances in which the second collision did occur.

The judge should identify the relevant circumstances and remind the jury of the prosecution and defence cases. If it is thought necessary it could be made clear to the jury that they are not concerned with what the defendant foresaw.

44. We prefer such a test to the test suggested by counsel in this case (paragraph 24 above).
45. We do not think that the test which the learned judge gave is close enough to our formulation to make the conviction safe. She came nearer to it with the last paragraph of the oral and written direction given whilst the jury were in deliberation:

4. May help to ask your selves this question

Are you sure the defendant caused one or both deaths, not the accident, but caused one or both deaths by his dangerous driving leaving the taxi immobilised in the middle of the fast moving A3 whereby he put the driver of the taxi at real risk of being hit by another vehicle, or not?

But the judge left that to the jury as a question which they may like to ask and we think it insufficient to ask whether the defendant merely put the driver at real risk of being hit by another vehicle.

46. There remains the additional problem that the jury were not sure that the appellant caused the death of Miss Cunningham. If they properly applied the suggested test, it is, to say the least, very difficult to see how the jury could be sure that the appellant caused the death of Mr Trewick but not the death of Miss Cunningham. The fact that they could not be sure that the appellant was guilty of both counts suggests that they may have had difficulty with the causation direction.
47. We do agree with counsel in their suggested direction that the jury should have been told that if the defendant is guilty of causing the death of Mr Trewick he would also be guilty of causing the death of Miss Cunningham.
48. For these reasons we allow the appeal, quash the conviction and order a retrial.