Introducing HSMR+: pioneering a new era in mortality data analysis

HSMR+, an enhanced and robust methodology for analysing hospital performance, is setting a new standard for the healthcare industry.

Since 2001, Telstra Health UK – and Dr Foster before it – has led the way in mortality data analysis with the Hospital Standardised Mortality Ratio, a pioneering model based on Hospital Episode Statistics data. For over two decades, the HSMR has provided invaluable insights into hospital performance by employing a risk-adjusted model that considers patient case mix and sub-population differences. This methodology has been instrumental in monitoring not only mortality rates but also other critical outcomes such as long length of stay and readmissions across more than 250 diagnosis groups and 200 procedures.

 Why HSMR+?

To keep pace with evolving medical practices and the changing healthcare landscape, we have undertaken a comprehensive review of the HSMR methodology. Customer feedback and advancements in data practices have led us to develop the new HSMR+, an updated model designed to offer more robust, fair, and unbiased benchmarking.

Key enhancements in HSMR+

Cohort updates

Diagnosis group revisions: The HSMR+ model now includes updated diagnosis groups to reflect current mortality data more accurately, reducing the list from 56 to 41 groups and adding in the viral infections diagnosis group which includes covid.

Exclusion of stillbirths: Stillbirths have been removed from the model, aligning with the latest methodologies and ensuring that the expected death calculations are more precise.

Variable adjustments

Covid-19 inclusion: A new covid-19 sub-group has been added within the viral infections diagnosis group, ensuring more specific risk adjustment for the pandemic’s impact.

Deprivation metric update: The Carstairs deprivation index has been replaced with the more comprehensive Index of Multiple Deprivation, offering a deeper understanding of socio-economic factors.

Comorbidity index enhancement: The Charlson Comorbidity Index has been upgraded to the Elixhauser-Bottle Comorbidity Index, a superior predictor of mortality, improving model accuracy and considering a broader range of comorbidities.

Global Frailty Index addition: Introducing the Dr Foster Global Frailty Index, which accounts for frailty in the model by looking across seven groups of frailty syndromes. This is a significant predictor of mortality and adds depth to patient risk profiles.

Removal of palliative care: Following extensive review and customer feedback, palliative care has been excluded from the model, addressing inconsistencies and potential biases.

Methodological innovations

Elimination of backwards-stepwise variable elimination: This change enhances the robustness and explainability of our models, reducing potential inconsistencies in variable selection.

Methodology and performance

The new HSMR+ model maintains the core principle of expressing hospital mortality as a relative risk, where 100 represents the national average. Observed deaths are compared to expected deaths calculated using a logistic regression model that now incorporates the latest variables and statistical techniques. Meanwhile, 95 per cent confidence intervals are used to classify HSMRs as “high”, “low” or “within the expected range” and alert trusts of potential areas for concern.

Impact and future directions

The transition to HSMR+ represents a notable advancement in mortality data analysis. By incorporating new variables like frailty and the Elixhauser-Bottle Index, and excluding palliative care, the model has been refined to better align with the NHS’s requirements. This updated methodology has led to changes in performance banding for nearly a third of trusts in the 2022-23 financial year, creating a more balanced distribution of performance assessments.

Telstra Health UK is committed to advancing healthcare analytics, and the launch of the HSMR+ underscores our dedication to innovation and excellence. By reflecting the latest coding practices, incorporating comprehensive variables, and employing cutting-edge statistical methods, HSMR+ sets a new precedent for hospital performance measurement.

Independent review

Telstra Health UK commissioned Simon Jones, professor of population health at New York University Grossman School of Medicine, to carry out an independent review of the inclusion of palliative care as a risk factor in the Hospital HSMR.

Professor Jones said: “Telstra Health UK’s introduction of HSMR+ marks a significant advancement in mortality data analysis and sets a new benchmark in international best practice. This enhanced methodology continues their tradition of excellence, building on over two decades of success with the original [HSMR].

“The meticulous updates in HSMR+ reflect current medical practices and incorporate critical variables like the Elixhauser-Bottle Comorbidity Index and the Global Frailty Index. These enhancements ensure more accurate, fair, and unbiased benchmarking of hospital performance. The inclusion of covid-19 data and the updated socio-economic deprivation metrics demonstrate a comprehensive approach to modern healthcare challenges.

“I’m impressed by Telstra Health UK’s commitment to innovation and the robustness of their new model. HSMR+ is poised to set a new industry standard, providing invaluable insights for improving patient care and hospital efficiency. Telstra Health UK leads the way in healthcare analytics, and HSMR+ is a testament to their dedication to advancing the field.”