Findings: Radiology Reporting Improvements Reduced Inappropriate Surgical Referrals

In a clinical audit tracking management practices for small, uncomplicated, deep intracerebral haemorrhage, radiology report recommendations for inappropriate surgical referrals fell from 55.6% to zero following a targeted educational and pathway intervention, according to findings.

Key Clinical Takeaways:
• Radiology report advice for inappropriate neurosurgical consultation dropped from 55.6% (10 of 18 cases) in the baseline audit to 0% (0 of 27 cases) following service improvement initiatives at University Hospitals of Leicester NHS Trust.
• Appropriate medical management recommendations in radiology reports increased from 16.7% (3 of 18 cases) to 51.9% (14 of 27 cases) across the two audit cycles.
• Actual surgical consultations requested for small, deep intracerebral haemorrhage patients declined from 55.6% to 14.8%, while conservative management remained the universal outcome for all referred cases.

University Hospitals of Leicester Audit Identifies Reporting Gaps

Spontaneous intracerebral haemorrhage accounts for 15% of all strokes and presents severe time-critical management challenges. Clinical guidelines recommend prompt blood pressure reduction and early stroke team involvement for small, deep haemorrhages, assigning no routine role for surgical intervention. However, informal observation at University Hospitals of Leicester NHS Trust revealed that initial radiology reports frequently advised neurosurgical consultation without accompanying guidance on guideline-directed medical therapy. The audit team noted that resident doctors in admissions units often relied on these report conclusions to guide patient escalation, leading to deferred stroke unit transfers and unneeded workloads for neurosurgical teams.

To measure and rectify this pattern, the clinical team conducted a two-cycle audit examining patients with small, uncomplicated, deep intracerebral haemorrhage. Audit Cycle 1 reviewed 18 patients between July and November 2024. The median patient age was 82 years, and 66.7% were male. Baseline data showed that 55.6% of reports contained inappropriate surgical referral advice, while only 16.7% provided appropriate medical management advice. Surgical consultation was sought for 10 of the 18 patients, though neurosurgical teams recommended conservative management for every referred case, with zero patients proceeding to surgery.

Educational Push and New Imaging Pathways Improve Reporting

Following the baseline findings, researchers implemented a targeted service improvement initiative in December 2024. The intervention combined an educational push—including teaching sessions for reporting clinicians—with a system-level change promoting a stroke-specific imaging request pathway. Audit Cycle 2 evaluated 27 patients between February and July 2025 following these changes. The median patient age in the second cycle was 71 years, with 48.1% male.

The distribution of management advice categories shifted significantly between cycles, yielding a p-value of less than 0.05 and a large effect size with a Cramér’s V of 0.66. Appropriate management advice in radiology reports rose to 51.9% of cases, while inappropriate surgical referral advice was completely eliminated at 0%. Reports containing no referral advice increased to 48%, with nearly 77% of those unassigned reports originating from the new stroke imaging request pathway. Actual surgical consultations fell from 55.6% to 14.8%, demonstrating a statistically significant reduction with a p-value of 0.007.

Operational Metrics and Antihypertensive Timing Across Audit Cycles

Alongside referral changes, the audit tracked the timeliness of acute medical therapies, specifically focusing on blood pressure reduction. Current guidelines recommend tight blood pressure control for acute intracerebral haemorrhage patients presenting with elevated systolic blood pressure. In Audit Cycle 1, 11 patients presented with an admission systolic blood pressure of 140 mmHg or above, and 55% of those individuals received antihypertensive therapy within one hour of hospital admission. In Audit Cycle 2, 11 patients presented with elevated blood pressure, and 55% received timely antihypertensive treatment within the one-hour window. This secondary operational metric showed a non-significant improvement between the two cycles, yielding a p-value of 0.714.

Researchers acknowledged methodological limitations, noting that case identification relied primarily on electronic patient records without an independent alternative search method, introducing potential case-selection bias. The small sample sizes across both cycles resulted in wide confidence intervals for principal proportions, requiring cautious interpretation of the quantitative shifts. Despite these limitations, the initiative demonstrates that targeted adjustments to radiology reporting workflows can effectively align clinical outputs with evidence-based stroke management protocols and reduce unnecessary surgical consultations.