The Prevent anti-terrorism scheme missed chances to protect the public from the attacker who assassinated the MP Sir David Amess, and from the youth who murdered three young girls at a Southport dance class, an official report has found.

The report is by David Anderson, the interim independent reviewer of Prevent, the official scheme to spot potential terrorists and turn them away from violence.

Lord Anderson finds that Prevent suffered from “a long string of failings” as it tried to deradicalise the man who went on to stab Amess, closing his case after one meeting to assess his danger when he was supposed to have had seven.

The new revelations led Amess’s family to demand a full inquiry, which the government has refused.

The veteran Tory MP was stabbed to death in October 2021 as he held a constituency surgery meeting with the attacker, Ali Harbi Ali, who was motivated by Islamist terrorism.

Ali was referred to Prevent in 2014, aged 18, and his case was dealt with by the Channel programme, which handles those of greatest concern.

Anderson finds six main areas of failings and notes that Ali was given a mentor but adds: “The programme of mentoring that was planned for him was allowed to peter out when it had hardly begun.”

The mentor was supposed to meet Ali seven times but only one took place, over a coffee at a McDonalds in Croydon, south London.

The report finds “a long string of failings”, some consistent with practice at the time.

But it adds, of the errors, that “most … were the product of poor judgment, poor communication and lack of follow-through.”

Ali had pretended to be complying and Anderson finds he still could have been missed as a danger, the crime being committed five years after Prevent declared him safe.

The government has promised a further review into whether chances were missed from 2016.

The Amess family want a full inquiry into the failings, with the power to call and cross-examine witnesses.

A family spokesperson said: “This is a grieving family … They have every right … to have it properly explained to them why the man who killed him was allowed to slip through the state’s safeguarding net, who is accountable for that, and what is being done to ensure no other family suffers in the same way. That is why they have called for a public inquiry.”

As well as the Amess case, the report looks at how Prevent handled the case of the Southport murderer Axel Rudakubana, and finds a new potential missed opportunity.

As first revealed in the Guardian, Prevent declined to take on the case three times, after teachers raised concerns about him from 2019 to 2021, three years before he committed the atrocity.

His first referral was from his teachers after he admitted bringing a knife into school on 10 occasions to “stab someone”, researched massacres of children at US schools and made “graphic” comments about violence. He later said “the [2017] terrorist attack on the MEN [Manchester Arena] was a good thing”.

While there was concern about his interest in violence, Prevent concluded there was no sign it was driven by a terrorist ideology.

The report says one expert concluded Rudakubana’s case should be examined by Channel, but that did not happen.

Anderson says: “Had this happened, he might have been offered a mentor who could have made a difference.”

A public inquiry into the Southport killings is already under way.

Rudakubana, who was 17 at the time of the attacks, is serving a life sentence with a minimum term of 52 years for the three murders, and the attempted murders of eight other children and of two adults who tried to protect them.

Anderson concludes: “Several years before the attacks, both the perpetrators had been referred by their schools to Prevent … Prevent’s Channel programme for early interventions had the capacity to address concerns of the kind that were raised in these referrals. But in neither case did it do so.”

The report recommends a potential dramatic expansion of Prevent, to include “individuals who have no fixed ideology but a fascination with extreme violence or mass casualty attacks”.

Anderson also says Prevent “could function better if formally connected to a broader safeguarding and violence protection system”.

The report recommends: “A Cabinet Office task force should be established to lead exploratory work into the possibility of formally connecting Prevent to a broader safeguarding and violence prevention system.”

​Anderson says: “I commend the teachers who were concerned that each of these youths was on a dangerous path, years before their crimes were committed. They did the right thing by referring them to Prevent. But in neither case did Prevent do what was needed to engage with them and protect the public.

“Success is never guaranteed, but the process can work with the help of skilled and committed practitioners, and often does so.

“When Prevent is offered a chance and fails to take it, it is vital that all possible lessons are learned.​”

Anderson says Prevent reforms undertaken and planned should make such failings harder to recur and adds:

“Prevent needs to up its game in the online world, where most radicalisation now takes place. It needs to get better at information sharing, and be more open with the public to gain the trust on which it depends.”