Hospital death rate study reveals wide variations and stresses importance of Registered Nurses

Jan 14 2007

Hospital death rates can be reduced by employing more Registered Nurses and the routine use of care maps or protocols, according to a study in the latest UK-based Journal of Advanced Nursing.

A research team from the University of Toronto and the Institute for Clinical Evaluative Sciences in Ontario, Canada, studied 46,993 patients admitted to hospital with heart attacks, stroke, pneumonia and blood poisoning.

They discovered that deaths within 30 days of admission varied considerably between the 75 hospitals in the study – ranging from ten per cent to 28 per cent and averaging just under 17 per cent.

When they added in the survey results from nearly 3,886 nurses at the hospitals - together with official discharge and death rates, population statistics and insurance plan data – they discovered that a number of factors accounted for 45 per cent of the variation in death rates.

"Our research underlines the need for hospitals to look as carefully at staffing structures and care processes as they already do at accurate diagnosis and appropriate and effective interventions" says lead author Dr Ann Tourangeau.

19 variables were examined to gauge their effect on 30-day death rates. Key findings included:

- A ten per cent increase in the proportion of Registered Nurses employed was associated with six fewer deaths per 1000 discharged patients.

- The death rate also went down by nine per 1000 discharged patients when the number of Baccalaureate-prepared (university graduate rather than diploma qualified) nurses went up by ten per cent.

- A ten per cent increase in adequate staffing and resources (as reported by nurses) was associated with 17 fewer deaths per 1,000 discharged patients.

"An important finding of our study was the effect that the routine use of care maps or protocols had on lowering 30-day death rates" adds Dr Tourangeau.

"A ten per cent increase in the use of care maps in hospitals, as reported by nurses, was associated with ten fewer deaths for every 1000 patients.

"Our findings contribute to the mounting evidence that structures and processes in hospital nursing care have an impact on patient mortality and survival. They clearly have implications for hospital management, clinical practice and future research" she adds.

"We specifically recommend greater use of care maps or protocols to guide patient care during their time in hospital. These could be shared, as a matter of good practice, on the Internet and adapted around the world to provide patients with culturally sensitive services.

"Although we were able to identify what caused 45 per cent of the variance in 30-day death rates, more than half of the variance remains unexplained.

"Death rates are a complex issue and nursing care is just one factor that influences survival rates. Future research should look at elements such as access to in-patient and out-patient care, the healthcare environment and the impact of hospital management and leadership on outcomes."

Data sources for the study included the Ontario Canada Discharge Database for 2002/3, the Ontario Hospital Reporting System 2002/3 and the 2003 Ontario Nurse Survey.

All Ontario teaching and community hospitals in operation during the 2002/3 study period were included. Small hospitals discharging fewer than 100 acute medical patients a year and specialty hospitals not providing care for the four key conditions studied were excluded.

5980 nurses working in medical and combined medical-surgical clinical areas across the study hospitals were surveyed and 65 per cent responded - an average of 52 from each of the 75 study hospitals.

Other study findings included:

• The percentage of Registered Nurses ranged from 36 per cent to 100 per cent, with an average of 66 per cent.
• Full-time nursing staff ranged from 35 and 92 per cent, with an average of 61 per cent.
• Nurses had spent between three and a half and fourteen and a half years on their current clinical unit, with an average of just over eight years.
• On average, 13 per cent were Baccalaureate-prepared (graduate) nurses. This ranged from zero to 62 per cent across the hospitals.
• The use of care maps and protocols varied from 29 per cent to 85 per cent, averaging 63 per cent.
• An average of 40 per cent of nurses felt their staffing and resources were adequate, with figures ranging from 19 to 52 per cent.

"This study provides a valuable insight into the nursing factors associated with the 30-day death rate for four common but serious illnesses and adds to the growing body of evidence linking hospital death rates with nurse staffing levels" says Professor Alison Tierney, Editor-in-Chief of the Journal of Advanced Nursing.

"Its observations about the nursing skill mix and use of care protocols is particularly relevant and we hope that hospitals worldwide will study these findings as they address issues that have relevance to both the international nursing community and hospital care in general."