The survival rate for bowel cancer surgery varies widely between hospitals, a study suggests.
It is the first time the number of people dying within 30 days of an operation has been reported across England for any cancer.
Some 6.7% of the 160,920 who had surgery from 1998 to 2006 died but some hospitals had much higher mortality rates, the report in Gut journal says.
Cancer Research UK said the situation was "worrying".
Bowel cancer is the third most common cancer in the UK.
The outcomes of all 160,920 people who had surgery to remove a tumour between 1998 and 2006 were analysed.
The overall mortality rate in that period dropped from 6.8% to 5.8%.
Looking at data for the last three years of the study, five hospital trusts had significantly higher death rates and three had significantly lower ones.
The proportion of patients dying within a month of surgery ranged from 15.6% at Burton Hospitals NHS Foundation Trust to 1.7% at Central Manchester University Hospitals NHS Foundation Trust.
Preliminary data suggests the five hospitals which appeared to perform poorly had all improved by 2008, by when they all had achieved mortality rates close to the average.
Consultant surgeon at Burton Hospitals Trust, James Eccersley, said: "The Trust is aware that constant vigilance and improvement has reduced our mortality rate after surgery. We always strive to take on board new innovations to make surgery even safer."
The explanation for the variation remains unclear.
Professor Paul Finan, from Leeds General Infirmary, said possible reasons included: "the degree of surgical specialisation, how guidelines were followed, the quality of high-dependency units (after-care surgery), anaesthetic services and whether the surgery was laparoscopic, which is less invasive."
In heart surgery, outcomes improved significantly after death rates were published. By finding out what more successful hospitals do differently, it is hoped bowel cancer care can be improved across the country.
Professor Philip Quirke, consultant pathologist at the University of Leeds, said: "In the past there's not been enough emphasis on routine clinical practice. We can learn a lot, especially from the best. Why are they so good?"
The government's national cancer director, Sir Mike Richards, wants statistics to become available for other cancers.
He said: "It's taken time to get this far, but we've demonstrated this as a proof of principle and are working out how to do the same with other cancers."
It is difficult to make international comparisons as other countries do not have data on a full national scale.
Studies in Scandinavia, Canada and the US have shown mortality rates between 2.7% and 5.7%. However, researchers cautioned against making direct comparisons.
Mark Flannagan, chief executive of Beating Bowel Cancer, said: "This study shows worrying variations in bowel cancer survival that need to be addressed.
"Bowel cancer is the UK's second-biggest cancer killer, yet there are still significant inequalities in survival which need to be better understood before they can be tackled."