Hospitals in Newfoundland and Labrador and Nova Scotia have some of the highest death rates in the country, finds a new report that, for the first time, compares mortality rates at hospitals across the country.

Saskatchewan, on the other hand, has some of the lowest hospital death rates, according to the report released Thursday by the Canadian Institute for Health Information.

The information comes from a calculation tool being used for the first time in Canada, called the Hospital Standardized Mortality Ratio (HSMR). The tool compares the death rates of patients treated at 92 larger acute care hospitals and 42 health regions in Canada, outside Quebec.

(Quebec results are not available due to historical differences in hospital data collection. And seven of the larger hospitals chose not to have their results published this year, because of concerns over their own data collection methods.)

The number of deaths in a hospital or region are compared with the average Canadian experience, after adjusting for several factors, such as the age of the patient, their health status upon admission and their gender.

An HSMR of 100 is the overall average rate. So a hospital or region with an overall HSMR greater than 100 has a mortality rate that is higher than the average; if it's less than 100, that suggests that its mortality rate is lower than the average.

The Cape Breton Healthcare Complex, in Nova Scotia, for example, had the worst score of any hospital that chose to have its results included on the list, with an overall HSMR of 137 -- though its ratio has dropped from 143 in 2004-05, to 129 in 2006-07.

The Regina General Hospital in Saskatchewan had one of the best overall HSMR in Canada, at 71.

Other hospitals with poor scores were:

• Grand River Hospital, KW Health Centre, in Kitchener, Ont. HSMR: 130
• Scarborough Hospital General site, in Toronto, Ont. HSMR: 129
• St. Catharines General Niagara Health System, in St. Catharines, Ont. HSMR: 129

Other hospitals with good scores were:

• St. Mary's General Hospital, in Kitchener, Ont. HSMR: 81
• Thunder Bay Regional Health Sciences Centre, in Thunder Bay, Ont. HSMR: 81
• Moncton Hospital, in Moncton, N.B. HSMR: 83

On a region-by-region basis -- which included all hospitals in a jurisdiction, not just the larger health centres -- the five hospitals of the Cape Breton District Health Authority reported the poorest score with the highest overall HSMRs.

The nine hospitals in the Newfoundland and Labrador Central Regional Integrated Health Authority also reported a high average mortality rate.

The seven hospitals in Saskatchewan's Sunrise Regional Health Authority had some of the best scores. Second best was the province's Regina Qu'Appelle Regional Health Authority.

Aim is to see death rates fall

The hope is that the data from this report can be used to cut the number of preventable "adverse events" -- those hospital errors that can sometimes lead to unexpected death.

A recent study in 2004 found that 7.5 per cent of adult medical or surgical patients had adverse events in hospital -- about one-third of which were deemed preventable.

Most patients recover from these errors, but each year, between 9,250 and 23,750 Canadian adults experience a "preventable" adverse event in hospital and later die, according to the study, which was funded in part by CIHI.

Phil Hassen, CEO of the Canadian Patient Safety Institute says the HSMR will be an important new measure of patient outcomes.  

"What we want to do is have hospitals pay attention to this and begin to do work to reduce the number of... unnecessary deaths," he told CTV. "HSMR is an indicator that will focus people's attention This will then allow us to see a reduction in lives lost that should not have happened."

Already, the overall average HSMR in Canada has fallen by six per cent since the data began to be collected three years ago. This translates to just over 2,500 fewer in-hospital deaths.

But trends vary by patient group. For example, death rates for patients with heart attacks fell faster than those for patients with pneumonia. The five illnesses that had the highest numbers of deaths were:

• heart attack,
• heart failure,
• pneumonia,
• chronic obstructive pulmonary disease
• septicemia (whole body infection)

Those patients who had higher odds of dying in hospital were:

• Men
• Those with chronic health problems (such as kidney disease or AIDS) in addition to their main diagnosis
• Those who came in as urgent/emergency admissions

These factors were all taken into account in calculating the results.

The HSMR is calculated based on the illness groups that account for 80 per cent of all deaths in acute care hospitals. The death rates are then adjusted for other factors that would affect mortality, such as advanced age, having multiple illnesses at once, and whether or not patients were transferred between hospitals.

The hospitals and health regions included in Thursday's report are already aware of how their facility or region compares to others across the country. Some have already begun to make changes.

For example, the Saskatoon Health Region has begun introducing rapid response teams to prevent deaths in patients who find themselves in trauma outside of intensive care settings. They are also adopting practices to prevent ventilator-associated pneumonia, to reduce medication problems and infections and to standardize wound and skin care.

Yet some hospitals have seen their numbers rise in the last three years. Windsor Regional Hospital has seen its HSMR go from 118 in 2004-05, to 109 the next year, to 133 in 2006-07.

The hospital chose not to have its results published in the report but did decide to hold its own news conference to explain what it says were mitigating factors that explain their numbers. The hospital says, for example, that its high death rate is become it often accepts transfers of the sickest patients across the region to its health centre.

Hospital CEO Martin Girash says the hospital is not content with its mortality rates and will strike a committee to review the numbers and will work to improve its score.

History of the HSMR

The HSMR has been used by hospitals in several countries throughout the world as part of efforts to track mortality rates over time and target areas for improvement.

Every country uses similar data and calculation methods, but each country's results are based on their own national mortality experience, not compared to one another.

The HSMR tool was developed by Sir Brian Jarman at the Imperial College in the United Kingdom. He says that since Britain started tracking its hospitals' death rates, numbers have fallen across the board.

"I think it's been very successful, because about 80 per cent of hospitals in England now use the data on a monthly basis," he told CTV News.

"You can pick out problems to do with mortality and monitor the changes you implement and then look at how you're getting on with that."

Hospitals can use the data to identify which departments of their hospitals have unusually high death rates compared to other hospitals' similar departments, he says. Hospitals can then begin to make changes by reviewing patient charts and organizing teams to review their hospital's clinical practices.