A report on the scheme, which concluded it was "very successful", has been published at last after a four-year delay by the Government.
The project involved five support workers, who were sent into seven schools in North Yorkshire between 1996 and 1999 to work with pupils at risk of exclusion.
Researchers from York university picked 10 of the pupils to examine in-depth how the project helped them to improve their behaviour.
Although the names in the report have been changed, one of the pupils studied was Robert Fuller, a student at Rossett high school in Harrogate, who was assigned a support worker in 1997. The pupil, then aged 13, regularly disrupted lessons and had been cautioned by police for stealing alcohol from a public house.
Robert's mother said that the support worker was a "lifesaver". However, in spite of the extra help, the teenager's behaviour worsened.
On January 28, 1998, Robert and a group of friends assaulted a 10-year-old autistic boy from a nearby primary school and filled his clothes with gravel.
Researcher Graham Vuillamy, one of the report's authors, was interviewing Robert's mother on the afternoon when the teenager attacked the autistic child.
Robert was suspended from school after the attack and temporarily excluded again later in the year for threatening and harassing a fellow pupil.
But, after his original support worker was replaced by a temporary one, his behaviour was said to have improved so much that he no longer needed extra help. However, on January 17, 1999, Robert and a friend, 14-year-old Daniel Gill, lured another boy from Rossett high, 13-year-old Ashley Murray, to the remote beauty spot Birk Crag.
There they stabbed him 18 times and left him for dead in a bin bag on the bottom of a cliff.
Both Robert and Daniel were later convicted of attempted murder and sentenced to six years in youth custody.
The court heard that the teenagers had been obsessed by the film Scream and had been exposed to drugs, witchcraft and knives when they regularly visited the home of a jailed drug supplier, Paul Aurens.
The York University report reveals that Robert's support worker had earlier identified Mr Aurens' "potentially damaging influence".
The evaluation, which was submitted to the Home Office in 2000, said the case showed how valuable the support worker had been to the teenager's mother.
"Whether continued support worker involvement might have contributed to avert the situation (Robert) is in now is an unanswerable question," it said.
"In this and other similar projects staff turnover and the end of project funding can withdraw support from vulnerable youngsters and their families at times when they may be most in need of it."
The researchers said that the project had been a success overall because the support workers had saved teachers' time and helped reduce permanent exclusions in the schools by a quarter.
However, they found that the support workers had not had any impact on reducing the pupils' offending behaviour.
Occasional disagreements were also reported between the support workers and teachers because of a "clash between teaching and social work cultures".
The report is believed to have influenced the Government's decision to provide at-risk pupils with extra support through such programmes as extended schools.
A Department for Education and Skills spokeswoman said the report's delay was due to administrative reasons rather than sensitivities surrounding Robert Fuller's case. The report was transferred from the Home Office to the DfES because of changes in the departments' responsibilities.
A Multi-Agency Approach to Reducing Disaffection and Exclusions from School is at www.dfes.gov.uk