There is a major difference in the death rates at hospitals around England, a report has found.
The researchers looked at data for hospitals across England
Patients at the hospital with the worst rate were twice as likely to die as those at the top ranked hospital.
The work by Dr Foster Research, an independent health information company, estimated 7,400 could have been saved if mortality rates were standard.
But the Department of Health warned it was "impossible" to reflect a hospital's safety in one statistic.
Dr Foster, which carried out the study for the Daily Telegraph, looked at Hospital Standardised Mortality Ratios for 152 hospital trusts across England in 2005-2006.
This is a measure which takes into account the differences in the patients being treated at each hospital, and which measures unexpected deaths.
The ratio shows whether the number of patients who died was higher or lower than expected.
The research took into account risk factors such as age, sex, social demographics, the level of deprivation in the area and whether a patient had any other illnesses.
Information for Wales and Scotland has only just become available and is not robust enough to reach conclusions, the researchers said.
Of the trusts analysed, 56 were found to have high mortality rates, 45 had low mortality and 51 were average.
Nearly all of them - 145 - had rates which had gone down over the past five years.
The Royal Free Hospital NHS Trust in north London had the lowest mortality rate, which stood at 26% below expected levels.
The hospital said it was the second year it had come top in the mortality ratings.
But the George Eliot Hospital NHS Trust in Nuneaton had the highest mortality rate, 43% above that which was expected.
The hospital acknowledged its poor rating, but said there had been problems in the way patient's illnesses had been recorded, which had affected their rating.
Dr Peter Handslip, the trust's medical director, said: "Systems are now in place to ensure that every aspect of a patient's care is recorded on the appropriate forms, which in turn has helped to significantly reduce the trust's mortality rate in just a few months."
He added that the area had a high rate of heart and lung disease, plus an ageing population, which had to be taken into account when looking at the data.
Roger Taylor, the research director at Dr Foster Research, said: "Now that people are being given greater choice over where they are treated, hospitals will need to reassure patients that they offer the highest standards of care."
Katherine Murphy of the Patients Association said:" Patients looking at these figures will be alarmed to learn that they are more likely to die in some hospitals than in others.
"It is essential that poor performing trusts should learn from the experience of better ones.
"The NHS is very bad at doing that."
Nigel Edwards, director of policy at the NHS Confederation which represents over 90% of NHS organisations, said: "Variations can be caused by a number of different reasons from differences in local patients┐ needs to the very nature of gathering statistical information.
"Where variations occur due to differences in quality of care, this report will help NHS trusts to identify where improvements need to be made."
And a spokesman for the Department of Health said: "We need to be careful in using just one measure to assess safety in our hospitals.
"Some have higher mortality rates because they take on more complex work, but obviously patient safety will always be the key priority.
"At present it can be difficult for patients to get a true picture of a hospital's overall performance from the available data because it covers so many different areas.
"We will, however, be bringing all the available data together in the new website which will help patients make informed choices about their health and healthcare."