The Crown’s framing
The Crown’s opening and closing speeches framed the 2015–2016 Countess of Chester neonatal unit mortality as exceptional: a rise from low single-figure annual deaths in previous years to substantially more in 2015 and 2016. The jury was invited to read the rise as evidence that something beyond ordinary clinical variation was at work.
Why absolute numbers are not informative
The Countess of Chester neonatal unit is a small unit. In any small unit, year-on-year variation in absolute deaths is substantial just from chance. The single most important methodological move in interpreting a cluster is to ask whether the rate (deaths per admission, or deaths per gestation-adjusted admission-day) is out of range for comparable units.
That move was not systematically made at trial. The jury was shown absolute numbers against a local historical baseline that did not adjust for acuity, admission volume, or gestation-mix.
The UK National Neonatal Research Database
The UK maintains a national neonatal outcomes database (the NNRD), founded by Prof. Neena Modi at Imperial College London (see her biography). It is the canonical source for UK NICU outcome baselines. A properly baselined reading of the Countess of Chester cluster would place the unit’s 2015–2016 mortality against the NNRD distribution for comparable units, in comparable years, for comparable acuity mix.
The independent perinatal-epidemiology reading available on the site is that, once the baseline is adjusted for the unit’s specific admission profile during the cluster period, the mortality is within the expected distribution for a struggling Level 2 unit admitting infants beyond its design acuity.
The acuity-mismatch adjustment
The Countess of Chester in 2015–2016 was routinely admitting infants whose acuity exceeded its Level 2 designation. Extremely preterm babies (23–25 weeks) have multi-fold higher mortality risk than the typical Level 2 admission. Multiple pregnancies (twins, triplets) carry elevated risk. If a Level 2 unit is admitting a Level 3 patient mix, its mortality will look high compared to other Level 2 units — without anything unusual happening on the ward.
When the unit’s admission mix is adjusted to the acuity it was actually caring for, the mortality rate falls within the expected range for that acuity.
The outbreak-and-infrastructure adjustment
The Guardian’s September 2024 investigation documented:
- A multi-drug-resistant bacterial outbreak on the unit during the cluster period.
- Recurrent sewage and plumbing failures (see the Mansutti defence-witness testimony).
- Chronic shortages of consultant and middle-grade paediatric staff.
- A unit caring for extremely preterm infants below its Level 2 designation.
Each of these factors independently raises mortality. Together they are a well-recognised epidemiological configuration for a NICU mortality cluster. The clustered deaths are what one would expect under these conditions.
What this means for the convictions
The Crown’s pattern argument rested in part on the implicit claim that the mortality rise was anomalous and therefore required a criminal explanation. If the rise is within expected variation for the unit’s actual conditions, the claim fails. That does not itself acquit the individual counts — each is separately contested on its own medical merits — but it removes the statistical anchor that gave the pattern argument its emotional weight.