Using routinely collected data, a recently developed tool can accurately predict the risk of death for patients within a year of admission to hospital, according to a study in CMAJ (Canadian Medical Association Journal).
An accurate assessment of risk of death, particularly if that risk is high, could motivate and inform discussions between patients and physicians regarding goals of care, states Dr. Carl van Walraven, a researcher at the Ottawa Hospital and the University of Ottawa (Ottawa, Ontario), who developed the tool.
Researchers in Canada and the United States looked at data for patients who had been admitted to hospitals in Ontario (2,862,996 people) over an eight-year period, as well as data on patients admitted to hospitals in Alberta (210,595) and to Brigham and Women's Hospital (66,683) in Boston, Massachusetts over shorter time periods. The study aimed to determine if the Hospital-patient One-year Mortality Risk (HOMR) model can accurately predict the one-year risk of death in a diverse group of patients. HOMR was developed to help predict risk of death from any cause in the year after hospitalization. It considers age, sex and living status (e.g., whether someone is living independently at home, with home care or in a nursing home), number of illnesses and severity of illness, and the number and type of admissions to hospital. The tool was previously validated using data from Ontario patients only over a one-year period.
The study included a large group of adults from Ontario, Alberta and Boston who had been admitted to hospital for non-psychiatric conditions. Patients in Alberta were younger and those in Boston had a greater number of chronic diseases. The overall risk of death within one year of hospital admission was 8.7%, but varied substantially according to patient characteristics. The predicted risk of one-year mortality based on the HOMR score accurately reflected the observed risk of death.
The HOMR score can be used with confidence to predict the risk of death within one year after admission using health administrative data, the authors conclude. Further research is required to determine whether the HOMR model can be used on the frontlines to predict survival during routine care.
Media Relations Officer
University of Ottawa
Office: 613-562-5800 ext. 2529