Introduction

The Court of Appeal in Northern Ireland has quashed convictions for murder and attempted murder after concluding that the evidence of the prosecution psychiatrist could not safely be relied upon. The decision in King v ZX is a forceful restatement of what courts expect from expert witnesses. It deserves close reading by every clinician who prepares reports for litigation, whatever the jurisdiction or the instructing party.

The case and the statutory test

The appellant, identified in the judgment only by the cipher ZX, was convicted of killing her eight week old son and of the attempted murder of her two year old daughter. The facts were largely undisputed. The central question was her mental state, so psychiatric evidence was decisive.

Two psychiatrists examined her about ten months after the events. The defence expert diagnosed moderate to severe depression, observed features suggestive of coercive control, and considered that both diminished responsibility and infanticide were open to the jury. The prosecution expert accepted that the appellant was depressed but judged the condition to be milder. Infanticide requires that the balance of the mother’s mind was disturbed because she had not fully recovered from the effects of childbirth. Following R v Tunstill, childbirth need not be the sole cause, but it must be more than a minimal one, and once the issue is raised the burden rests on the prosecution to disprove it.

Deficiencies in the expert evidence

The difficulties emerged at trial. At the end of his evidence in chief, the prosecution psychiatrist told the jury that the most likely explanation for the killing was spousal revenge. This opinion appeared in neither his notes nor his reports. Cross examination then exposed further concerns. Crown counsel had supplied him with a research article on the subject. A telephone conversation of up to an hour between expert and counsel went unrecorded. A quotation attributed to the appellant, that her partner had destroyed her life so she destroyed his, appeared in full in his report. His handwritten notes, produced only during the trial, recorded merely the first half, and in the middle of the consultation rather than at its end.

There was more. The expert had not viewed the body worn video recorded when police attended an earlier domestic incident. He had misstated the legal definition of infanticide and recorded the history of antidepressant treatment inaccurately. His reliance on the tidy state of the appellant’s home as evidence against depression surfaced for the first time in the witness box. He had also received no written letter of instruction.

The standards applied

The court held that the evidence could not be, and be seen to be, the independent product of the expert. It concluded that he had aligned himself with the central pillars of the prosecution case, and that his evidence was not something a jury could rely on. The court adopted the familiar duties set out in R v Harris, which require objective and unbiased opinion, a clear statement of the facts and assumptions relied upon, consideration of material that detracts from the concluded view, and prompt notice of any change of opinion. It also drew the line between legitimate witness familiarisation and impermissible coaching, found an appearance of coaching in the circumstances, and held that prosecutorial conduct had fallen below recognised standards of good practice.

Outcome and retrial

All three grounds of appeal succeeded: the compromised expert evidence, the conduct of the prosecution, and the trial judge’s refusal to discharge the jury without hearing oral argument. Both convictions were quashed, including the attempted murder conviction, because of its close connection with the murder count. The court declined to substitute lesser verdicts, since the jury’s findings on mental state were not clear enough to support that course. In a postscript dated 20 March 2026, the court ordered a retrial, holding that the public interest in prosecuting serious charges prevailed. The prosecution undertook not to rely on the original expert, and the letter of instruction to a new expert was to be prepared on a blank page basis to avoid contamination.

Practical lessons for experts

Several lessons come from this case. Firstly, every opinion to be given in evidence should appear in a written report served in advance, and any change of view must be communicated without delay. 

Secondly, contemporaneous notes are evidence in their own right. They should be complete, accurate, and ready for disclosure, with quotations recorded verbatim and placed where they occurred. 

Thirdly, experts should review all relevant material, including recordings and medical records, or state clearly why they have not. Also, the court said that the practice of instructing criminal experts without a written letter must change immediately, so experts should insist on one. 

Finally, contact with instructing lawyers should be recorded, ideally with a solicitor present, and literature supplied by counsel to support a new theory should be treated with particular caution.

Conclusion

The court observed that the defects might never have come to light without the persistence of defence counsel. That is a sobering reminder that independence is ultimately the expert’s own responsibility. An opinion that bends toward the party who instructs the expert damages the case, the profession, and, in the end, the administration of justice.