GeraClinic / Hospital Mortality Index / Methodology
How the Gera Hospital Mortality Index is calculated
The index is fully reproducible from the real NHS SHMI value and a single, disclosed transformation. This page gives the exact formula, worked examples, how the NHS models the underlying figures, the provenance and the limitations.
Read this first: what the SHMI is not
The SHMI is NOT a measure of quality of care. NHS England is explicit that a higher-than-expected SHMI should be treated as a "smoke alarm" — a prompt to investigate, not evidence of poor care — and that an "as expected" or "lower than expected" value should not be read as proof of good care. Many factors outside a hospital’s control (local population health, the prevalence of palliative-care coding, deaths in hospices) affect the figure.
What the NHS SHMI measures
The Summary Hospital-level Mortality Indicator (SHMI) is published by NHS England (NHS Digital). For each acute NHS trust it is the ratio of observed deaths — patients who died in hospital or within 30 days of discharge — to the expected deaths for that trust. Expected deaths are modelled by the NHS from the characteristics of the patients each trust treats: their age, sex, method of admission, the diagnosis group they were admitted with, and other factors. A trust treating older or sicker patients is expected to have more deaths, so the ratio adjusts for case mix. A SHMI of 1.00 means observed deaths equalled expected deaths.
The formula
where SHMI = observed deaths ÷ expected deaths
The only transformation Gera applies is the × 100 — putting the ratio on a scale with a national baseline of 100 so a lay reader can immediately see how far a trust sits above or below the England expectation. It is monotonic in the published SHMI: no hidden model, no re-weighting, no re-banding. The value-add is legibility, naming and the per-trust presentation, all disclosed here.
Worked examples
From the November 2024 to October 2025 release:
- National: the SHMI is constructed so the England average is 1.00 → index 100 (deaths as expected).
- University Hospitals of Morecambe Bay NHS Foundation Trust: SHMI 1.396 × 100 = 139.6 — 1,895 observed vs 1,360 expected deaths (banded higher than expected).
- Imperial College Healthcare NHS Trust: SHMI 0.718 × 100 = 71.8 — 2,165 observed vs 3,015 expected deaths (banded lower than expected).
How the NHS bands each trust
The NHS does not band trusts on the raw ratio alone. It places control limits around each trust’s expected count (a funnel plot that accounts for how many patients the trust treats) and assigns one of three bands: 1 — higher than expected (above the upper limit), 2 — as expected (within the limits), and 3 — lower than expected (below the lower limit). We reproduce this banding exactly. In the November 2024 to October 2025 release, 11 trusts were band 1, 99 band 2 and 8 band 3, out of 118 trusts.
What this does NOT show
- It is not a measure of the quality or safety of care. A higher figure is a smoke alarm for investigation; a lower figure is not proof of good care.
- It does not adjust for every difference between hospitals. How a trust codes palliative care, and how many deaths occur in hospices rather than hospital, materially affect the figure — a known limitation the NHS documents.
- It covers acute English NHS trusts only — not specialist, community, mental-health or independent providers — and is a trust-level, not hospital-site or ward-level, indicator.
- It is not clinical advice about where any individual should be treated.
Provenance & licence
Contains public sector information published by NHS England and licensed under the Open Government Licence v3.0. The index was last recomputed on 3 July 2026 against the release published 12 March 2026.
Contains public sector information published by NHS England and licensed under the Open Government Licence v3.0. Source: Summary Hospital-level Mortality Indicator (SHMI) — Deaths associated with hospitalisation, England, November 2024 - October 2025 (November 2024 to October 2025, published 12 March 2026).
NHS methodology in full: About the Summary Hospital-level Mortality Indicator (SHMI). Gera source reference: Gera Hospital Mortality Index — derived from the NHS SHMI open data.