When two medical witnesses disagree at trial and nothing has been done beforehand to narrow the difference, the party who cross-examines the other side’s doctor takes on a risk that is easy to underestimate. The answer can land worse than the report did. Once it is in the pool of evidence, the trial judge is entitled to act on it.

That is the working lesson of Barbina v McKenzie [2026] QCA 134, delivered on 17 July 2026. The Court of Appeal, constituted by Bond JA, Brown JA and Freeburn J, refused leave to appeal from a modest damages award in a motor accident claim. The result is unremarkable, but the reasoning deals with two questions that recur in almost every contested injury trial in Queensland: how far the rule in Browne v Dunn (1893) 6 R 67 (HL) actually reaches, and why the author of a clinical record did not have to be called on these facts.

What the Court decided

The applicant was injured in a collision at Bahrs Scrub on 3 June 2020. Liability was never the issue. RACQ Insurance Limited was the second respondent, the action having been brought against driver and insurer as joint defendants under Motor Accident Insurance Act 1994 (Qld) s 52, and it accepted liability for the driver’s negligence. The fight was about whether the continuing back symptoms were a permanent injury caused by the collision or the product of a pre-existing arthritic condition.

The trial judge preferred the respondents’ medical evidence and awarded a modest sum. Because the judgment fell below the Magistrates Courts jurisdictional limit, an appeal required leave under District Court of Queensland Act 1967 (Qld) s 118(3). The Court of Appeal refused it, holding there was no reasonable argument that the primary judge had erred.

The disagreement the case turned on

Each side had an orthopaedic report. The applicant’s doctor put the back injury at 8 per cent whole person impairment, on the basis that a pre-existing L5/S1 disc bulge had been made significantly worse by the collision and was impinging the exiting right L5 nerve root. The respondents’ doctor assessed nil impairment. His view, as the primary judge recorded it at [15], was that the lumbar aggravation was temporary; his musculoligamentous strain finding at [14] concerned the cervical spine, and what the lumbar scans showed was age-related degeneration.

Both accepted the disc problem was not created by the collision. The real question was whether the aggravation was temporary or permanent. Critically, the Court recorded at [12] that by the time they gave evidence the experts had had the opportunity to consider each other’s opinion, but no joint report was ever produced recording where they agreed and where they did not. That omission set up everything that followed.

The answer that arrived under cross-examination

The applicant was self-represented by the time of trial, assisted by her son as a McKenzie friend, who was given leave to ask questions. He cross-examined the respondents’ doctor about whether that opinion was consistent with the radiology.

The answer was that it was not anatomically possible for the disc bulge to be impinging the nerve root the applicant’s doctor had identified, because that nerve root exits the spinal canal above the level of the disc. Dr Morgan drew the distinction expressly in the passage reproduced at [18], separating compression of the canal from impingement of the exiting root. On this evidence the narrowing visible on the scan was caused by longstanding facet joint arthritis, not by the bulge. The witness added that the applicant’s doctor had misread the scan and was in error.

The Court described that as a powerful explanation, and at [24] described the risk that had been run: the obvious hazard of investigating the point while the witness was in the box was that the witness might give an answer damaging to the applicant’s case, and the trial judge is ordinarily entitled to consider a favourable or an unfavourable answer alike.

What the rule in Browne v Dunn actually requires

The first ground was that the trial judge should not have accepted the anatomical evidence, because it had never been put to the applicant’s own doctor. That is an appeal to the rule in Browne v Dunn. The Court rejected it on three bases, and the third is the one worth carrying forward.

First, there was no breach. The applicant already had notice of the nature of the case being run against her doctor’s opinion, because the disagreement was live on the face of the competing reports. At [31] the Court put the limit plainly: the rule does not require that every point on which a witness’s evidence might be used against them be put in cross-examination, and it is not a rule designed to encourage or condone excessive cross-examination.

Second, the rule requires only that a witness be given an opportunity to comment or explain. The evidence had not closed, so if the applicant was genuinely surprised she could have contacted her doctor or sought to have him recalled.

Third, and decisively, even if the rule had been breached the consequence is a discretion in the trial judge: to reject the evidence, to permit rebuttal, or to permit a recall. The primary judge was asked for none of those things. There was no evident consternation or surprise, and the evidence was simply absorbed as a difference between the experts for the judge to resolve. On those facts the Court concluded at [34] that it could hardly be said the discretion had miscarried. That is a conclusion about how a discretion was exercised in these circumstances, not a rule that a failure to object always forecloses appellate relief.

Clinical records, and why the author did not need to be called here

The remaining grounds concerned a single line in a general practice record from 2006, noting “long standing low back pain”. The respondents’ doctor relied on it. The applicant argued it was an opinion, and that the doctor who wrote it should have been called to explain its basis.

The records had been tendered through the practice manager and admitted as a book of account under Evidence Act 1977 (Qld) ss 83–85. Those provisions are broader than the name suggests. The definition in s 83 extends to any document used in the ordinary course of an undertaking to record anything dealt with by or held on behalf of it, which is wide enough to catch patient records. Section 84 makes an entry evidence of the matters recorded, and s 85 sets the condition of admission: proof that the book was one of the ordinary books of the undertaking and that the entry was made in the usual and ordinary course.

The Court held at [46] that the note had not been shown to be an opinion or a diagnosis at all, and that in its context it was likely to be nothing more than a record of what the practitioner was told. That is a finding about this entry on this evidence, reached in part because the applicant had not established otherwise. No challenge had been made to the tender at trial, and there was no basis to disturb the admission. It also mattered that the applicant had ultimately accepted in evidence that she probably did have longstanding back pain she had forgotten about, so no unfairness could arise from a note confirming what she had already conceded.

What it changes for running an injury trial

Nothing here is new law. Every authority applied, from Allied Pastoral Holdings Pty Ltd v Federal Commissioner of Taxation [1983] 1 NSWLR 1 to MWJ v The Queen [2005] HCA 74 and Wang v Hur [2024] QCA 126, was settled before it. What the case supplies is a clean appellate statement of three things practitioners get wrong.

A joint report is not a formality: where the medical difference is technical, isolating it in writing before trial converts an ambush into a known issue. A cross-examiner owns the answer, and a question asked in hope is still a question asked. And the remedy for genuine surprise is procedural and immediate. Ask for the evidence to be rejected, or for rebuttal, or for a recall, while the evidence is still running. Left unraised, the complaint on appeal becomes that the trial judge’s discretion miscarried, which is a materially harder argument than an objection made at the time.

Published 4 August 2026.

Where this leaves the law

The decision breaks no new ground and does not purport to. The principles it applies, from Allied Pastoral Holdings and MWJ v The Queen, are stated in general terms. Whether they bear differently on a self-represented litigant is a question this Court did not reach: an application to add a ground alleging the primary judge failed in his duty to a self-represented litigant was made during the appeal hearing and refused at [8]. What the Court did do was acknowledge at [24] that the applicant may not have acutely appreciated the forensic choices being made, and hold her to their consequences. What the case supplies is an unusually clear illustration of how those principles operate once a technical medical difference has never been isolated before trial. Whether that pushes the profession towards treating a joint expert report as standard in such matters rather than optional is a question of practice rather than of law, and this judgment does not answer it. The full judgment is on AustLII, and the provisions discussed can be read on the current reprint of the Evidence Act.

Need advice?

Time limits and key decisions in injury matters can be strict and easy to miss. If this is your situation, contact Fraser Lawyers to understand your rights and where you stand, with no obligation.

Get in touch →

This article is general information only and is not legal advice. Queensland law and the figures referred to can change, and every situation turns on its own facts. Contact Fraser Lawyers for advice specific to your circumstances.

If you would like to discuss your matter, you can book a consultation or call (07) 5554 6116.