MeeTooSaives. Integrating machine learning (ML) and natural language processing (NLP) into the MeeToo app to create a state of the art suicide prevention solution for children and young people (CYP) aged 11+.

MeeTooSaives。

• 批准号: 10021282
• 金额: $ 50.18万
• 依托单位: MEETOO EDUCATION LIMITED
• 依托单位国家: 英国
• 项目类别: Collaborative R&D
• 财政年份: 2022
• 资助国家: 英国
• 起止时间: 2022 至 无数据
• 项目状态: 已结题
• 来源: https://gtr.ukri.org/projects?ref=10021282
• 关键词: MeeTooSaives.; Integrating; machine; learning; ML

Integrating machine learning (ML) and natural language processing (NLP) into the MeeToo app to create a state of the art suicide prevention solution for children and young people (CYP) aged 11+.The MeeToo app delivers free, safe, anonymous, pre-moderated, peer support to 56k CYP. MeeToo has been independently evaluated and demonstrated to improve mental health in CYP (AFC, 2021).In 2020, MeeToo supported 794 CYP expressing suicidal ideation. When suicidal ideation becomes suicidal intent, CYP don't seek support because they feel they have 'gone past help' (Biddle\_2020). Because 40% of CYP who die by suicide have no contact with the mental-health system (Rodway\_2020), suicide prevention tools which only focus on CYP in crisis exclude the majority who go on to take their lives (Kessler\_2019). Successful suicide prevention solutions must deliver increased support before distress becomes acute.Our innovation, MeeTooSaives, developed in collaboration with Bristol University Medical School and University of Sussex Department of Informatics, integrates ML and NLP into the MeeToo app to interrogate passive as well as high risk, data to identify risk patterns, analyse semantic content and predict risk.When risk patterns are detected, CYP are automatically stepped-up into a suicide support stream for clinical supervision and trained responder peer support. NLP integrates smart links to topic matched resources. Crucially, CYP doesn't have to ask for additional support. When distress deescalates, support steps down and CYP re-join their peers. AI continues to monitor risk.Because most suicides occur between midnight and 4 a.m. (Perlis\_2016) mirror provision in Australia facilitates 24/7 global delivery.MeeTooSaives is important because suicide is the leading cause of death in UK CYP (ONS\_2018) and for every suicide, there are 20 non-fatal suicide attempts (WHO\_2018). By the age of 17, 7.4% of UK CYP have attempted suicide (Millenium\_Cohort\_Study\_2021). In 2017, 72% of UK CYP suicides were male; 14% were black/minority ethnic; 5% were LGBT (NCISH\_2017).The lifetime cost of one youth suicide is £9.75 M (Platt\_2006) yet the estimated 10 year ROI for every £1 invested in suicide prevention is £39.11(McDaid\_2017). Late intervention in youth mental health costs public services an estimated £17bn a year (Oppenhim, 2015). The UK spends £49.7 M PA treating paracetamol poisoning alone (Bateman\_2014). This high risk, technically challenging innovation could provide the 2M UK CYP with mental health issues with 24/7, 365-day, access to crisis and continuous support for <£40M P.A.

将机器学习(ML)和自然语言处理(NLP)集成到Meetoo应用程序中，为11岁以上的儿童和年轻人(CYP)创建最先进的自杀预防解决方案。Meetoo应用程序为56K CYP提供免费、安全、匿名、预审核和同行支持。Meetoo已在CYP(AFC，2021)中被独立评估并证明具有改善心理健康的作用。2020年，Meetoo支持794个表达自杀意念的CYP。当自杀意念变成自杀意图时，CYP不会寻求支持，因为他们觉得自己已经“失去了帮助”(比德尔2020)。由于自杀死亡的CYP中有40%与心理健康系统没有接触(Rodway 2020)，仅关注危机中CYP的自杀预防工具排除了大多数继续自杀的人(Kessler 2019年)。成功的自杀预防解决方案必须在痛苦变得尖锐之前提供更多的支持。我们的创新MeeTooSaives与布里斯托尔大学医学院和苏塞克斯大学信息学系合作开发，将ML和NLP集成到Meetoo应用程序中，以询问被动和高风险的数据，以识别风险模式，分析语义内容和预测风险。当检测到风险模式时，CYP会自动增强为自杀支持流，以进行临床监督和受过培训的响应者同行支持。NLP集成了指向主题匹配资源的智能链接。至关重要的是，CYP不必要求额外的支持。当痛苦降级时，支持下台，CYP重新加入他们的同龄人。人工智能继续监测风险。因为大多数自杀发生在午夜到凌晨4点之间。(Perlis\_2016)澳大利亚的镜面服务促进了全天候全球分娩。MeeTooSaives很重要，因为自杀是英国CYP的主要死因(ONS\_2018)，每一次自杀，有20次非致命的自杀企图(WHO\_2018)。到17岁时，7.4%的英国CYP有过自杀未遂(千禧年队列研究2021)。2017年，英国青年自杀者中72%是男性，14%是黑人/少数民族，5%是LGBT(NCISH 2017)。一个青少年自杀的终生成本为975万GB(Platt 2006)，但每投入1 GB预防自杀估计10年ROI为39.11(McDaid 2017)。对青少年心理健康的晚期干预每年花费公共服务约170亿英镑(Oppenhim，2015)。英国仅治疗扑热息痛中毒就花费了4970万英镑(贝特曼2014)。这一高风险、具有技术挑战性的创新可能会为200万英国CYP提供全天候、365天全天候精神健康问题、危机应对和持续支持&lt；GB 4000万P.A.