THE RATIO. One judgment, decoded twice: once for anyone, once for the profession.

Headnote

Four members of one household were convicted of allowing a young woman living with them to suffer catastrophic, permanent brain injury, after a jury concluded they knew she was already at serious risk of harm within that home. An appeal court quashed their convictions, reasoning that the specific act which caused her brain injury was carried out by such a different method from an earlier act of abuse she had already suffered that no jury could find it fell within what the defendants had reason to foresee. The Supreme Court has restored the convictions, holding that the law asks whether the surrounding circumstances of a later act of harm were foreseeable, not whether its precise method matched an earlier one.

Factual Background

Ambreen Sheikh came to England in 2014 to join her husband’s family in Huddersfield, following an arranged marriage. She lived in their household with her husband, his parents and his sister. In August 2015, an ambulance was called to the house after she was found unconscious and unable to breathe properly. She was taken to hospital in a critical condition, with severe swelling of the brain, and has never recovered consciousness.

While treating her, hospital staff discovered a severe, untreated wound on her lower back, one that must have been extremely painful and must have been apparent to everyone in the household, yet for which no medical help had ever been sought. Other injuries were also found. Items recovered from the house, including clothing and bedding bearing traces consistent with her injuries, had been hidden rather than disposed of normally, and the clothes she was found wearing when the ambulance arrived showed no such traces at all.

The prosecution’s case was that the wound on her back had been inflicted deliberately, using a caustic substance, some days before her collapse, and that her brain injury was most likely caused separately, by administration of a small amount of a diabetes medication belonging to one member of the household, which can cause a dangerous drop in blood sugar in someone who does not have diabetes. The prosecution could not prove which specific member of the household had committed either act, so relied on a provision of English law built for precisely that situation.

The Question for Determination

Section 5 of the Domestic Violence, Crime and Victims Act 2004 was created to solve a specific, longstanding problem in English criminal law. Where a vulnerable person is seriously harmed by someone in their own household, and more than one person had the opportunity and possible reason to have done it, the prosecution has traditionally had to prove which individual person was responsible, or that they acted together. At common law, if it could not, everyone had to be acquitted, even where it was certain that someone in the household was guilty of something very serious.

Section 5 allows a defendant to be convicted without the prosecution having to prove which of two alternatives applies: that the defendant caused the harm directly, or that the defendant knew, or ought to have known, of a significant risk of serious harm to the victim from someone else in the household, failed to take reasonable steps to protect against it, and the harmful act occurred in circumstances of the kind the defendant foresaw or ought to have foreseen. This appeal turned entirely on that last requirement. Does an act of harm fall within circumstances a defendant should have foreseen only if it resembles, in method, an earlier act of harm the defendant already knew about, or can it fall within foreseeable circumstances even where the method used is completely different, provided the wider pattern and setting are the same?

The Court’s Reasoning

The Court agreed with the Court of Appeal on one central point. If simply knowing that a household carries some general risk of violence were enough, on its own, to satisfy the foreseeability requirement, the requirement would do no real work at all, since a jury would already have to find that general risk proved under the earlier stages of the provision. Parliament plainly intended this final requirement to offer a defendant some additional protection beyond that, not merely to repeat what had already been established.

Where the Court departed from the Court of Appeal was in how that correct principle had actually been applied to the facts. The Court of Appeal had held that because the injury which caused the coma was inflicted by such a different method from the earlier back injury, chemical poisoning as against a caustic substance applied directly to the skin, no reasonable jury could find the poisoning occurred in circumstances of the kind that had been foreseeable. The Supreme Court held this went too far. The relevant question under the statute is whether the circumstances in which the harmful act occurred were of a foreseeable kind, not whether the specific method used matched a method previously seen. The manner in which harm is inflicted can be part of the relevant circumstances, but it is not, on its own, decisive of the question.

Applied here, the jury had been entitled to weigh a fuller picture: evidence of ongoing hostility toward the victim within the household, an earlier act of serious, deliberately inflicted harm that everyone in the household must have known about, a further injury that may have involved the same substance, and evidence of prolonged, serious neglect. Taken together, this was capable of showing a continuing, foreseeable pattern of risk to the victim from within her own household, regardless of whether the specific mechanism of the final act of harm matched anything that had gone before. Whether it in fact did so was a question for the jury, not one the appeal court was entitled to resolve for itself by treating the difference in method as automatically decisive.

The Court also upheld the trial judge’s approach to directing the jury. Rather than attempting to paraphrase or add legal gloss to the words of the statute, she had directed the jury in the statute’s own terms, and asked them to decide, on the whole of the evidence, whether those words fitted the facts as they found them. The Court held this was correct: where a statute uses ordinary language, applying that language to the facts is a task for the jury, not a question of law for the judge to resolve in advance by redefining the words. The Court of Appeal’s criticisms of the summing up, and of several other aspects of the trial judge’s handling of the case, were found to be unfounded.

The convictions on the counts concerning causing or allowing serious physical harm were reinstated.

Critical Assessment

The Court is clear about what “circumstances of the kind” does not mean, general household hostility alone is not enough, but considerably less clear about precisely where the line sits once more specific evidence, like the earlier injury here, enters the picture. Having criticised the Court of Appeal for treating a difference in method as automatically decisive, the Court declines to say how much weight a difference in method should carry against the other factors it lists, hostility, pattern, neglect, beyond stating that this balancing exercise belongs to the jury. That is doctrinally sound respect for the jury’s role, but it leaves future trial judges facing a no case to answer submission with meaningfully less guidance than the judgment’s confident tone might suggest, particularly in cases without the unusually strong antecedent evidence present here.

There is a related concern worth naming plainly. The factors the Court accepts as relevant circumstances, general hostility, a pattern of prior mistreatment, evidence of neglect and isolation, are also, in substance, close to the kind of facts already needed to prove the earlier stages of the offence, awareness of risk and failure to protect. The Court insists this does not collapse the final requirement into the earlier ones, since the earlier stages ask about awareness of risk in general while this one asks about the circumstances of the specific act that caused the harm. That distinction holds in principle. In a case with less extensive antecedent injury evidence than this one, it may prove harder in practice to keep the two questions meaningfully separate, which is precisely the danger the Court itself warned against when it rejected the prosecution’s broadest submission.

Set against that, the Court’s insistence on directing the jury in the statute’s own words, rather than attempting a more elaborate judicial definition, deserves real credit. Given how genuinely difficult “circumstances of the kind” is to pin down with precision, the temptation to offer juries a tidier paraphrase is understandable, but any such paraphrase risks quietly narrowing or widening what Parliament actually wrote. Trusting a jury’s ordinary understanding of ordinary words, rather than manufacturing a false sense of legal precision the statute does not itself provide, is honest and disciplined judicial craft, even where it means the boundary of the offence will keep being worked out case by case rather than settled once and for all.

Implications

If you’re not a lawyer: if you are concerned about someone living in a household where they may be at risk from more than one person, this case explains why English law created an unusual offence specifically for that situation. A family member can be convicted of allowing serious harm even where nobody can prove exactly which person in the house caused it, as long as it can be shown they knew of a real risk, failed to act on it, and the harm that followed fits the kind of danger they already had reason to fear. This case confirms that a family member cannot escape that responsibility just because the specific way harm was ultimately caused turns out to differ from what came before, if the wider pattern of danger in the home was already something they should have recognised.

For the profession: this is now the leading authority on section 5(1)(d)(iii) DVCVA 2004, and confirms that the provision retains independent force, foreseeability of circumstances is not automatically satisfied by proof of general awareness of risk under section 5(1)(d)(i), while also confirming that a difference in the method of infliction between an antecedent act and the charged act is a relevant but not determinative factor. Practitioners resisting a no case to answer submission on behalf of the prosecution should marshal the fullest available picture of pattern, hostility and neglect, rather than relying on similarity of method alone, since the Court has confirmed that method similarity is neither necessary nor sufficient on its own. Defence counsel should press for precision in how the prosecution frames its case on circumstances, since the Court’s own reasoning leaves real room to argue, on a thinner evidential record than existed here, that the circumstances relied upon are really no more than the general household risk already required at the earlier stages of the provision, which the Court itself said cannot be enough. The Court’s endorsement of directing juries in the statute’s own language, following Brutus v Cozens, is also worth noting for any case turning on an ordinary English phrase left undefined by Parliament.

Case Details

  • Citation: [2026] UKSC 28, on appeal from [2025] EWCA Crim 38
  • Court: UK Supreme Court. Judgment given by Lord Lloyd-Jones, with Lord Reed, Lord Burrows, Lady Rose and Lady Simler agreeing
  • Heard: 25 June 2026 · Decided: 27 July 2026
  • Governing provision: Section 5, Domestic Violence, Crime and Victims Act 2004, as amended by the Domestic Violence, Crime and Victims (Amendment) Act 2012
  • Key authorities discussed: R v Uzma Khan [2009] EWCA Crim 2 · Brutus v Cozens [1973] AC 854 · R v Lane and Lane (1986) 82 Cr App R 5
  • Key paragraphs: [42]-[46] (why section 5(1)(d)(iii) must retain independent force) · [48]-[49] (the Court of Appeal’s error: treating a difference in method as decisive) · [58]-[59] (directing the jury in the statute’s own words)