New DISH Surgical Classification System Standardises Wound Outcome Reporting

Surgical site reporting has long fixated on infection, leaving three other common complications largely unrecorded. A new international framework, published in the British Journal of Surgery, aims to change that. The DISH surgical classification system grades dehiscence, inflammation or infection, seroma, and haematoma in a single composite score, so a wound’s severity and its management are captured together.

The system responds to a specific gap. The CDC definitions for surgical site infection, the most widely used standard, focus on infection alone and rely on subjective clinical judgement that varies between institutions. They describe what a wound looks like without recording how it is then treated, even though these outcomes are costly to hospitals and patients.

Why infection-only reporting falls short

Dehiscence, seroma, and haematoma each carry real consequences. A dehisced abdominal or sternotomy wound can be life-threatening and often needs prolonged intervention. Postoperative seromas are common after mastectomy and may delay adjuvant chemotherapy or radiotherapy when they demand repeated drainage. Haematomas can become infected and progress to abscess or sepsis, sometimes requiring emergency surgery.

Ad

The authors, led by Giles Bond-Smith of Oxford University Hospitals NHS Foundation Trust, argue that standardising both presentation and management would let hospitals compare practice and benchmark outcomes, feeding directly into quality improvement.

How the DISH surgical classification system works

Each of the four domains is scored on two independent scales: five grades of clinical intervention (0 to 4) and four grades of clinical presentation (a to d). The result is a snapshot such as D1a I2b S0 H0, which shows the management received and the appearance of the incision for every domain at once.

The work was sponsored by Johnson & Johnson, with study design and facilitation by Costello Medical.

Built through global consensus

The classification emerged in stages. A targeted literature review surfaced inconsistent definitions, then a 32-member expert panel from 22 countries rated concept statements across three anonymous voting rounds. An in-person symposium debated the draft, and its feedback drove a fourth and final voting round that split clinical intervention from clinical presentation so the two could be scored separately.

Strong agreement in validation

Validation took place at a second symposium in Madrid, Spain, on 19-20 February 2026, with 64 surgeons and infection control specialists attending and 60 completing both rating sessions. Participants scored 19 patient case vignettes generated through a rule-based artificial intelligence pipeline.

Agreement was high on every measure. Mean patient-level inter-rater agreement, measured with a double-entry intraclass correlation coefficient (DE-ICC), reached 0.93 (95% CI 0.91-0.96; median 1.00). Intra-rater reproducibility came in at 0.95 (95% CI 0.93-0.96; median 1.00). Kendall’s W averaged 0.96 (range 0.93-0.97) and Gwet’s AC2 averaged 0.96 (range 0.94-0.99), with all confidence interval lower bounds above 0.90.

Mean accuracy across domains was 93% (range 87%-97%), and balanced mean accuracy was 87% (range 77%-95%). A sensitivity analysis that excluded zero-valued observations lowered accuracy in several domains, most notably dehiscence (intervention, 78%) and infection and seroma (presentation, 77% and 75%), prompting minor wording changes.

What it means for surveillance and cost

For health economics and outcomes research teams, the appeal is granularity. Because the score records both observed severity and the intervention delivered, it can be tied to resource use and downstream cost more directly than an infection-only flag. The authors prescribe no fixed collection time, advising the day of discharge and routine follow-up visits, and recommend at least one year of follow-up after implant surgery to catch delayed complications.

Existing tools such as the CDC classification, the Southampton Wound Assessment Scale, and the ASEPSIS wound score either omit non-infection outcomes or are hard for non-surgeons to apply. The DISH system is meant to give the whole surgical team a common language for surgical complications, which the authors see as a route to better prevention and more reliable benchmarking.

Limitations and next steps

The authors position the classification as a starting framework rather than a finished tool. The case vignettes represent ideal cases, while real-world notes are likely to be messier; patient involvement was not included; fringe cases were excluded; and some specialties were underrepresented. Clinical validation and prospective studies are still needed, with pilot exercises already planned or underway across specialties and regions.

Source: Bond-Smith G, Boermeester MA, Leaper DJ, et al. Dehiscence, Infection, Seroma, Haematoma (DISH): Development and Validation of a New Classification of Surgical Site Outcomes. BJS, 2026.