Hospital Standardised Mortality Ratio

Hospital Standardised Mortality Ratio
Name	Hospital Standardised Mortality Ratio
Caption	Conceptual schematic of standardised mortality metrics
Type	Healthcare quality indicator
Introduced	20th century
Developer	Health services researchers
Related	Standardized Mortality Ratio

Hospital Standardised Mortality Ratio The Hospital Standardised Mortality Ratio is a comparative indicator used to evaluate inpatient mortality by comparing observed deaths to expected deaths derived from statistical models. Developed in health services research and biostatistics, it informs policy, regulation, and hospital management across systems such as the National Health Service, Medicare, and private hospital networks. Prominent organizations and figures in its evolution include World Health Organization, Organisation for Economic Co-operation and Development, Institute for Healthcare Improvement, and scholars associated with Johns Hopkins Hospital, Mayo Clinic, and Harvard Medical School.

Definition and Purpose

The metric quantifies mortality performance by expressing the ratio of observed to expected deaths, enabling comparisons between institutions like Guy's and St Thomas' NHS Foundation Trust, Cleveland Clinic, Karolinska University Hospital, Mount Sinai Hospital (New York), and Royal Free Hospital. It is intended for use by regulators such as Care Quality Commission, payers like Centers for Medicare & Medicaid Services, and research units at University College London, Imperial College London, Stanford University, and University of Toronto. Stakeholders include policymakers from Department of Health and Social Care (UK), administrators at NHS England, and patient-safety advocates associated with The King's Fund and Robert Wood Johnson Foundation.

Calculation Methodology

Calculation typically combines administrative datasets from sources like Hospital Episode Statistics and claims data from Medicare (United States). Models use logistic regression or other predictive algorithms developed by teams at University College London Hospitals NHS Foundation Trust, Imperial College Healthcare NHS Trust, Massachusetts General Hospital, University of California, San Francisco, and vendors advising NHS Digital. Statistical foundations trace to methods used by Office for National Statistics and epidemiologists at Johns Hopkins Bloomberg School of Public Health, using covariates sourced from electronic health records at institutions such as Mayo Clinic and registries like National Surgical Quality Improvement Program. Software implementations have been performed with packages from SAS Institute, R (programming language), and tools used by World Health Organization collaborators.

Risk Adjustment and Case-Mix Factors

Risk adjustment models incorporate variables such as age, sex, comorbidities coded via systems like International Classification of Diseases, admission type present in datasets maintained by NHS Digital and Centers for Disease Control and Prevention, and diagnostic groupings used by All-Payer Claims Database projects. Comorbidity indices informed by work from Elixhauser comorbidity measure and Charlson comorbidity index are commonly applied by researchers at Johns Hopkins Hospital and Brigham and Women's Hospital. Case-mix complexity is assessed using classification systems influenced by Diagnosis-Related Group frameworks developed in the United States and adaptations by Australian Institute of Health and Welfare, Canadian Institute for Health Information, and Eurostat collaborators.

Uses and Applications

Hospitals and regulators employ the ratio for benchmarking across networks such as NHS Trusts, Veterans Health Administration, Bupa, HCA Healthcare, and academic centers like Yale New Haven Hospital and University Hospital Zurich. It supports quality-improvement initiatives championed by Institute for Healthcare Improvement and informs commissioning decisions by bodies such as Clinical Commissioning Groups (England). Researchers at Harvard Pilgrim Health Care Institute and RAND Corporation use the metric for studies on performance, while patient advocacy groups including Patients Association (UK) and Healthwatch (England) refer to it in transparency reporting.

Limitations and Criticisms

Critiques have been raised by academics at London School of Hygiene & Tropical Medicine, University of Oxford, and think tanks like Nuffield Trust regarding residual confounding, coding variation, and gaming risks. High-profile debates involving Keogh Review-style inquiries and scrutiny by Care Quality Commission highlighted cases where unadjusted interpretations led to misleading conclusions about institutions including Mid Staffordshire NHS Foundation Trust. Methodological objections also come from statisticians associated with Royal Statistical Society and health services researchers at University of Melbourne and Monash University.

International and National Implementations

Nationwide implementations exist in systems managed by NHS England, Australian Commission on Safety and Quality in Health Care, Health Quality & Safety Commission New Zealand, Centers for Medicare & Medicaid Services, and agencies in Canada such as Canadian Institute for Health Information. Cross-country analyses have been performed by Organisation for Economic Co-operation and Development and academic consortia including researchers from Karolinska Institutet, University of Sydney, McGill University, and Université de Montréal.

Statistical Interpretation and Reporting

Interpretation uses confidence intervals and funnel plots popularized by analysts at Institute for Fiscal Studies and statisticians in publications by BMJ and The Lancet. Reporting conventions are influenced by guidance from National Institute for Health and Care Excellence and data standards set by Health Level Seven International. Misinterpretation risks prompt recommendations from Royal College of Physicians and Academy of Medical Royal Colleges that stress complementary indicators such as readmission rates and patient-reported outcome measures used by NHS England and Patient-Reported Outcomes Measurement Information System.

Category:Health care quality indicators