Barnsley Hospital NHS Foundation Trust has halved its mortality review workload, improved diagnostic accuracy and strengthened learning from deaths after adopting a modern benchmarking approach.
The trust reports that monthly time spent reviewing mortality cases has reduced from 13.5 to 5.5 hours, while the number of patient-level reviews has fallen from around 40 to around 20. The change has released clinical, administrative and analyst time for higher-value activity, including deeper Structured Judgement Reviews (SJRs), more targeted quality-of-care investigations and enhanced data-quality work.
The advanced benchmarking solution is provided by CHKS, a British provider of healthcare analytics, benchmarking and quality-improvement services for the NHS and wider health sector – both in the UK and internationally.
Previously, a shift in mortality indicators often triggered extensive retrospective review.
Associate Medical Director for Mortality, Dr Susie Orme, recalls one example:
“We once spent 10 full days reviewing 120 sets of notes, only to find no care issues. We would never need to do that now. Our reviews are focused, proportionate and based on much better data.”
One of the trust’s most significant breakthroughs came when modern benchmarking highlighted a high number of false Finished Consultant Episodes (FCEs). Staff movement within the electronic record system was unintentionally creating very short episodes of care – some lasting just 8–12 hours – which distorted expected mortality calculations.
Diagnoses were being captured too early in the pathway, often selecting symptom codes such as abdominal pain, UTI or URTI instead of the patient’s final underlying condition.
Tracey Radnall, Associate Director for Patient Safety & Quality Improvement, said benchmarking played a key role in exposing the issue:
“The tool showed we had far too many short episodes compared with similar trusts. Once we corrected that, diagnostic capture improved immediately. We stopped seeing misleading alerts and started seeing a much more accurate picture of mortality.”
The trust retrained ward clerks, adjusted workflows and added real-time checks to ensure genuine clinical episodes were captured in full. This restored clinical realism to the data and stabilised expected mortality modelling.
Dr Orme explains why this matters in practice. A patient may present with abdominal pain and weight loss and only receive a cancer diagnosis after CT imaging on day two.“If the episode is split too early, the system grabs the symptom, not the cancer. That changes expected mortality and triggers completely misleading alerts. Longer episodes fix the problem.”
Barnsley has also strengthened comorbidity recording, especially important in a population with high levels of frailty, multiple long-term conditions and significant multimorbidity across ex-mining communities. The trust now automatically pulls all historical comorbidities into each admission.
“I saw a patient this morning with 25 diagnoses,” Susie says. “Several were terminal in their own right. No patient could list them all from memory – now the system does it for them. It makes a huge difference.”
Mortality assurance has become more targeted and sustainable. Tracey describes the new approach:
“We follow a clear sequence: data quality first, then coding, then case mix – and only then clinical review. A real flag is when an alert in CHKS correlates with what we’re seeing through medical examiner reviews, SJRs or patient safety intelligence. That’s when escalation is justified.”
CHKS Director Bevin Manoy said Barnsley’s results reflect the value of combining analytical capability with strong internal processes.
“Barnsley shows what’s possible when teams focus on data quality, good clinical coding practice and proportionate investigation. The result isn’t just fewer reviews – it’s better learning and a clearer picture of quality.”
Barnsley’s leaders say the approach aligns well with national expectations set out in Learning from Deaths, the Patient Safety Incident Response Framework (PSIRF), the CQC Single Assessment Framework and Data Saves Lives. Rather than reacting to statistical fluctuation, the trust now focuses on where learning will add real value.
The result is a calmer, more credible and more sustainable model of mortality surveillance: fewer unnecessary reviews, stronger diagnostic accuracy, better comorbidity capture, reduced noise in mortality indicators and more time spent on learning that matters.
