The annual report by the National Confidential Inquiry into Suicide and Homicide by People with Mental Illness also found:
A fall in suicide by mental health patients overall and a continued fall in suicide by in-patients. Suicide following absconding from the ward remains a serious problem. However, few serious incidents occurred following absconding from secure units.
The University of Manchester research, funded by the National Patient Safety Agency, examines suicide and homicide data for people with mental illness for England and Wales. The figures relate to 1997-2006 for suicide and 1997-2005 for homicide.
The report found that there had been an increase in the number of homicides committed by people with mental illness at the time of the offense from 54 in 1997 to over 70 in 2004 and 2005. There was also a rise in the number of homicides by people with schizophrenia, from 25 in 1997 to 46 in 2004 and an estimated 40 in 2005.
Professor Louis Appleby, Director of the National Confidential Inquiry into Suicide and Homicide by People with Mental Illness, said: "There has been an unexplained rise in the number of homicides by people with mental illness and we now have to try to understand why this has happened.
"It is important to emphasize that the increase has not occurred in mental health patients. It is also important to keep these findings in perspective. The risk of being a victim of homicide in England and Wales is around 1 in 1,000 and the risk of being killed by someone with schizophrenia is around 1 in 20,000."
The number of patient deaths by suicide has gone down to its lowest level since data collection began in 1997. In 2006, there were 185 fewer deaths than in 2005. The number of in-patient suicides has continued to fall from a high of 219 deaths in 1997 to 141 in 2006.
There have been no reports of in-patient deaths using fixed curtain rails since 2003, following an NHS directive calling for them to be removed from wards. The report also found that the number of suicides by patients in the community who have refused treatment or refused further contact with services has fallen.
However, in the ten years to 2006, there were 469 suicides by patients who had left a ward without permission. Most were on an open ward, only five had absconded from a secure psychiatric unit. There were no homicides by people who had absconded from a secure unit in the nine years to 2005.
Professor Appleby added: "Fatal incidents following absconding from secure units are rare. A more common event is the suicide of a detained patient following absconding from an open ward. Measures to prevent absconding from general wards include improvements to the ward environment and greater supervision and control of exits. Overall, in-patient suicides are falling twice as fast as suicides in the general population."
The report found that there had been an apparent increase in the number of sudden unexplained deaths of psychiatric in-patients, although the authors say it is unclear whether this reflects a true increase in numbers or is attributable to improved reporting.
In the eight years to 2006, there were a total of 338 unexplained deaths, equivalent to 42 per year. From 2002-6, 13 deaths occurred within 24 hours of restraint, though the authors point out that restraint may not have been the cause.