When Unanimous Psychiatric Evidence Shows a Section 37/41 Regime Better Protects the Public, Sentencers Must Give Cogent, Evidence-Based Reasons to Prefer a Section 45A Hybrid Order

Case: R v Coulson Neutral citation: [2026] EWCA Crim 1097
Court: Court of Appeal (Criminal Division), England and Wales
Date: 21 August 2026
Coram: Popplewell LJ, Sir Gary Hickinbottom, The Recorder of Redbridge (HHJ Rosa Dean)

1. Introduction

The appellant (aged 20 at the time) carried out an unprovoked knife attack on two strangers at a petrol station, inflicting neck wounds that narrowly avoided catastrophic injury. He then left the area and avoided police for 11 days before surrendering. He later pleaded guilty to two counts of wounding with intent contrary to s.18 of the Offences Against the Person Act 1861 and one count of possessing a bladed article; further unrelated offences were before the court for sentence.

The case was dominated by psychiatric evidence that the appellant suffered lifelong schizophrenia with a poor prognosis, requiring sustained medication and robust relapse management. The sentencing judge imposed an extended determinate sentence (EDS) with a hybrid order under s.45A Mental Health Act 1983 (MHA). The appellant appealed, contending that the judge should instead have imposed a hospital order under s.37 MHA with a restriction order under s.41 MHA.

Central issues
  • How should a sentencing court choose between a s.37/41 hospital and restriction order and a s.45A hybrid order where s.37 criteria are met?
  • What weight should be given to unanimous psychiatric evidence about comparative public protection under competing release/recall regimes?
  • How should culpability be assessed where violent offending is closely linked to active psychosis and substance misuse may be “self-medication”?

2. Summary of the Judgment

The Court of Appeal allowed the appeal. It held that the judge committed an error of principle in rejecting unanimous expert evidence that a s.37/41 disposal would better protect the public than a s.45A hybrid order for this appellant, and in suggesting that accepting the experts’ view would render s.45A “worthless” legislation. There was no evidential basis for the judge’s conclusion that an EDS plus s.45A would be equally or more protective.

Re-doing the R v Vowles balancing exercise, the Court found:

  • the appellant needed lifelong treatment and remained a foreseeable risk beyond the EDS term;
  • the offending was driven by psychosis; culpability from substance use was low, given the context of unmanaged illness and self-medication;
  • public protection was distinctly better served by the indefinite duration and stronger conditional discharge/recall architecture under s.37/41.

The Court quashed all sentences (including those committed for sentence) and substituted a s.37 hospital order with s.41 restrictions, continuing the appellant’s detention in secure hospital.

3. Analysis

3.1 Statutory context: the real-world significance of “release regimes”

The judgment is an unusually detailed exposition of how statutory release/recall systems function in practice, and why those mechanics may be outcome-determinative in cases involving enduring psychosis and violence risk.

  • s.37 MHA (hospital order): detention in hospital for treatment, with discharge when clinically justified (subject to tribunal powers).
  • s.41 MHA (restriction order): adds an enhanced public protection layer; discharge (often conditional) is controlled by the Secretary of State (with tribunal involvement), and recall can occur swiftly if risk escalates.
  • s.45A MHA (hybrid order): a prison sentence plus a hospital direction and “limitation direction” (s.41-like restrictions) while detained in hospital; if fit for discharge from hospital before the custodial term ends, the offender is transferred to prison. Importantly, the s.41-type limitations lapse at the automatic release point (here, after 9 years) by virtue of s.50 MHA.
  • s.47/s.49 MHA: allow transfer of prisoners to hospital and attachment of restrictions by executive action, but this is contingent and does not replicate the structured, indefinite community risk-management typical of s.37/41 conditional discharge.

The Court identified (and applied) three practical deficits of prison licence compared to conditional discharge/community treatment pathways: time-limited supervision, less specialised mental health oversight, and narrower/less responsive recall triggers.

3.2 Precedents cited and how they shaped the outcome

R v Calocane [2024] EWCA Crim 490

The Court treated Calocane as the modern anchor for comparing s.37/41 and s.45A. It adopted its identification of the “two key differences” between regimes: (i) hospital discharge leads to community release under s.37/41 but return to prison under s.45A (before minimum term expiry), and (ii) who controls release (Secretary of State vs Parole Board). The Court also invoked Calocane at [85]–[86] for the proposition that the strength of the causal link between illness and offending is a key factor in disposal choice and that, for some dangerous psychotic offenders, hospital/restriction can better protect the public.

R v Vowles [2015] EWCA Crim 45

Vowles supplied the governing framework: where s.37 criteria are met, the judge must decide the appropriate disposal by considering (1) treatment need, (2) attribution of offending to disorder, (3) punishment, and (4) public protection and the release regime, giving “sound reasons” for departing from the usual penal sentence. The Court in Coulson did not fault the judge for citing Vowles; rather, it held the judge misapplied the fourth factor by rejecting unanimous expert evidence without a cogent evidential explanation.

R v Edwards [2018] EWCA Crim 595

The Court relied on Edwards for two linked propositions:

  • the penal element matters because sentencing must address the full set of purposes in s.142 Criminal Justice Act 2003, not rehabilitation alone;
  • assessing culpability for serious offending in the presence of mental disorder is difficult but mandatory, and the “graver the offence and greater the risk”, the more weight must be placed on public protection and release regime considerations.

In Coulson, those propositions supported re-balancing towards s.37/41, because the public protection benefits were concrete and the case was one where illness was the principal driver of grave violence.

R v Nelson [2020] EWCA Crim 1615 and R v Rendell [2019] EWCA Crim 621; [2020] MHLR 60

These authorities were used to ground the Court’s practical assessment of risk. They support the reality that returning psychotic offenders to prison can precipitate medication non-compliance, relapse, exposure to drugs, and repeated transfers (“yo-yo” patterns) that are harmful both to the offender’s stability and to public safety. Nelson was also used to underline that hospital-based conditional discharge typically provides more specialist, intensive monitoring than prison licence supervision and that recall from conditional discharge can be rapid and clinically led.

R v Walker [2023] EWCA Crim 548

Walker was cited alongside Vowles, Nelson, and Calocane as part of an emerging line of decisions recognising that, for dangerous offenders with enduring psychotic disorders requiring long-term medication and substance abstinence, the hospital/restriction architecture can offer distinctly superior public protection.

3.3 The Court’s core legal reasoning

(a) Error of principle: rejecting unanimous expert evidence without an evidential footing

The Court accepted the orthodox position that disposal is for the judge, not the experts. However, it held that where experts unanimously explain—by reference to how the statutory schemes operate in practice—why one regime better protects the public in the specific case, the judge must either accept that analysis or provide cogent, evidence-based reasons for rejecting it.

Here, the judge’s statement that he was “unpersuaded” that s.37/41 offers more protection was insufficient because:

  • the view was unanimous across three experienced forensic psychiatrists;
  • it was consistent with accepted reasoning in the appellate authorities;
  • the judge identified no evidential basis to displace it, and instead relied on an abstract statutory inference (“Parliament wouldn’t enact a worthless provision”).

(b) Misstep in statutory logic: s.45A is not “nullified” by preferring s.37/41 on these facts

The Court rejected the notion that accepting the experts’ risk-management evidence would entail that s.45A can never be appropriate. It explained that public protection is only one factor, albeit important; culpability and punishment can legitimately tilt the balance towards a hybrid order in other cases. The judge’s “worthless legislation” reasoning therefore revealed a misdirection: it treated the existence of s.45A as implying equivalence of protective effect, when s.45A was designed as a flexible option, not a guarantee of parity.

(c) Reassessment of culpability: psychosis-driven violence, flight, and substance misuse

The Court’s re-evaluation is significant in three respects:

  • Psychosis as the principal driver: the further reports supported that the appellant’s actions were driven by active psychotic symptoms (persecutory delusions/hallucinations), and that prescribed medication at the time was not effectively controlling symptoms.
  • Post-offence evasion: the Court held that flight/avoidance did not necessarily show meaningful retained culpability at the time of the stabbing; it could be consistent with disordered thinking and fear-based responses in the context of major mental illness. It also noted the point had not been raised at sentencing for submissions.
  • Substance misuse as “self-medication”: while long-term drug use may have contributed to the disorder’s development and symptom severity, the Court treated culpability for such use as low in context—given the absence of diagnosis at the time, unmanaged symptoms, and the known phenomenon of schizophrenic self-medication.

(d) Public protection: duration and robustness of control beyond the EDS term

A decisive feature was the foreseeable persistence of serious risk beyond the 14-year EDS (and, in practical terms, beyond the expiry of licence). A s.37/41 order is potentially indefinite, and conditional discharge conditions can mandate medication compliance and forensic mental health engagement with swift recall. By contrast, once the EDS ends, there are no enforceable penal-supervision conditions, and under s.45A the restriction-like limitations lapse at the automatic release stage. For a lifelong relapsing psychotic illness, that structural difference mattered.

3.4 Impact and significance

The decision consolidates and operationalises an appellate trend (seen in Nelson, Walker, and Calocane) towards rigorous scrutiny of release regime realities in serious violence cases involving enduring psychosis.

Practical impacts
  • Sentencing judges must engage with the mechanics of release/recall and give reasoned, evidence-based findings when disagreeing with unanimous expert evidence on comparative risk management.
  • Hybrid orders (s.45A) are not to be treated as inherently equivalent to s.37/41 on public protection; the court must evaluate the specific offender’s illness trajectory, treatment compliance risk, and the time horizon of risk beyond determinate supervision.
  • Culpability analysis should not mechanically infer retained responsibility from post-offence conduct (e.g., evasion), and should recognise the “self-medication” phenomenon when weighing substance misuse in psychosis-linked offending.
  • Case preparation: parties should expect appellate scrutiny where sentencing remarks suggest a disposal choice was driven by abstract statutory assumptions rather than evidence about how the regimes function for the offender.

4. Complex Concepts Simplified

  • Hospital order (s.37 MHA): the court orders detention in hospital for treatment instead of prison.
  • Restriction order (s.41 MHA): added controls for public protection; discharge is tightly controlled (often by the Secretary of State), commonly via conditional discharge with enforceable conditions; recall can be rapid if risk increases.
  • Hybrid order (s.45A MHA): a prison sentence plus an initial hospital direction; if the patient becomes well enough before the custodial term ends, they go back to prison, and the restriction-like regime can lapse at the statutory point (s.50 MHA).
  • Extended determinate sentence (EDS): a custodial term plus an extended licence period for dangerousness; release during part of the custodial phase is Parole Board-dependent, and licence conditions end when the sentence ends.
  • Conditional discharge vs prison licence: conditional discharge can include mental-health-specific conditions (including medication compliance) and allows swift recall for deterioration; prison licence typically cannot compel medication and is time-limited.

5. Conclusion

R v Coulson is a pointed reminder that the disposal choice between s.37/41 and s.45A is not a theoretical debate about whether Parliament “must have meant” both are equally protective. It is an evidence-led assessment of how the statutory regimes will operate for the particular offender, especially where the offender presents a long-term psychosis-driven violence risk dependent on medication adherence and abstinence from drugs.

  • Unanimous expert evidence on comparative public protection cannot be dismissed with a bare expression of being “unpersuaded”; cogent, evidential reasons are required.
  • Release regime realities—duration of control, specialist forensic supervision, and recall responsiveness—may properly be determinative where serious risk is lifelong and extends beyond determinate sentence horizons.
  • Flight after offending is not automatically a reliable marker of culpability at the time of the offence in cases of major mental disorder.
  • Substance misuse in unmanaged psychosis may carry low culpability where it reflects self-medication and the illness is the primary driver of the violence.