For comprehensive protection of children — From prevention to reconstruction: building a continuous protection chain

📜🔗LINK to the source

1. ANALYTICAL SUMMARY

A robust right, a broken chain

Ten years of reforms have provided France with a dense legal corpus — law of 14 March 2016, law of 7 February 2022, extension then sliding of the limitation period (2018, 2021), facilitated incrimination of incest, prohibition of educational violence (2019), offence of school bullying (2022) — and effective mechanisms: 119, UAPED, OFMIN, EVAR(S), honourability checks (pp. 7-8). The central observation is nonetheless one of a discontinuity : a child is protected differently depending on their territory, the institution they approach and the location where the violence occurs (p. 11). The data confirms this: 73% of sexual violence is intrafamilial and 81% of perpetrators are acquaintances (p. 28), nearly 88,000 reports of sexual violence against minors are awaiting processing (p. 37), and the UAPED network has reached 151 units open or in creation for 164 targets (p. 37). Digital technology is treated as a cross-cutting dimension — both a site of violence commission and a lever for protection (pp. 68-77).

Twenty measures, five axes, a budgetary framework

The document outlines a doctrine of "integrated protection chain" in five axes — act early, lose no signal, protect without delay, support sustainably, make childhood a shared national priority (p. 15) — and proposes that the cost of inaction, estimated by CIIVISE at €9.7 billion per year, of which €6.7 billion is attributable to the lack of therapeutic care, serves as a reference for all budgetary arbitration (pp. 12, 49). For field actors, the operational contribution consists of four blocks: a unique inter-ministerial service for checking criminal records organised by the public to be protected and not by sector (pp. 19-20); a mandatory national foundation for training in detection, deployed in digital format from the start of the 2026 school year (p. 31); a right to an adapted hearing in UAPED or Mélanie room (p. 39); and a "child victim package" opening ten sessions of psychotrauma care funded by Health Insurance, renewable once (p. 51). Annex 1 (pp. 83-102) provides for each measure the principle, content, normative vehicle, deployment timeline, and performance indicators.

2. KEY POINTS OF THE DOCUMENT

1. The honourability check changes its logic: from the sector to the child. The document proposes a unique inter-ministerial service for background checks, backed by a secure national information system, accessible to administrations, local authorities, associations, and organisations hosting minors according to harmonised rules (pp. 19-20). The numerical assessment is already significant: 968,626 certificates issued as of 21 May 2026 since the generalisation on 1 October 2025 for early childhood and child protection, and 5,126 individuals excluded, including 175 registered with the FIJAISV (p. 83). Two blind spots are explicitly recognised: the cultural sector and babysitting when the employer is the parent (p. 20).

2. A massive and documented training deficit on identification. Two-thirds of educational professionals working in children's homes or living spaces have never been made aware of or trained in the identification of sexual violence against minors, and 79% of professionals report having felt powerless in the face of a situation of violence (p. 88). Hence measure 6: a mandatory national training framework, developed by six ministries with the support of specialised associations, to be rolled out digitally from the start of the 2026 school year and then integrated into initial and ongoing training — with adaptation to local languages and cultural realities of the territories (p. 31).

3. Access to care for psychotrauma becomes a right, decoupled from the criminal procedure. Measure 13 makes the prescription of a psychotraumatic assessment systematic at the end of the collection of speech, particularly in UAPED, opening a pathway of three assessment sessions followed by a "victim child package" of ten sessions funded by Health Insurance, renewable once, with teleconsultation and mobile teams for remote areas (pp. 51-52, 95). Professionals estimate the average duration of necessary support to be two years (p. 51) — the gap between the proposed package and this need is a point to be discussed in the team.

4. Minor offenders emerge from the unthought. 30% of perpetrators of sexual violence against minors are themselves minors, and 72% of these minor offenders have been victims of sexual abuse (p. 53, source FFCRIAVS 2025 report cited p. 96). Measure 14 proposes a systematic module on emotional, relational, and sexual life education in PJJ follow-ups, adapted educational assistance measures, and UAPED as an entry point to psychiatric evaluation — without prejudice to criminal sanction, the document specifying that prior trauma does not diminish either the severity of the facts or responsibility (p. 53).

5. Digital technology shifts from a peripheral risk status to that of a design standard. The digital focus (pp. 68-77) documents the shift: about one in two children under 10 will own a smartphone by 2026, 65% of primary school students have an account on a social network despite the minimum age of 13, 72% of teenagers report having used an AI-based conversational companion by the age of 15, 89% of grooming victims are said to be girls, and a quarter of the victims who contact 3018 are under 11 years old (p. 69). The proposed response is the doctrine of "default digital protection": effective age verification, security settings activated by default, disabling algorithmic recommendations for minors, removal of dark patterns (p. 74), complemented by a public index of child protection (p. 75).

3. ACTION TRACKS FOR LOCAL ACTORS

1. Audit your credibility chain now, including its blind spots. Take the list of targeted sectors on p. 20 (education, early childhood, sports, animation, culture, medico-social, child protection, leisure) and check structure by structure who is actually controlled — occasional volunteers, external contributors, service providers included. The document itself identifies culture and babysitting as uncovered (p. 20): these areas should be addressed through internal procedures without waiting for the law.

2. Build an internal training plan before the national framework is imposed. The foundation of measure 6 will be available in digital format from the start of the 2026 school year (p. 31, sheet p. 88, announced deadline 6 months). Until then, use the figures from p. 88 (two-thirds untrained, 79% feeling of helplessness) to objectify the need to your management or funder, and target the four announced contents: identification, welcoming speech, reporting obligations, psychotraumatic mechanisms.

3. Map your emergency entry points and display them at child height. The sheet measure 5 (p. 87) lists the numbers to know: 119 (child in danger), 112 (European emergency), 114 (emergency by SMS), 3018 (harassment), 116 000 (disappearance), 3114 (suicide prevention). While waiting for the national application announced in a year, produce a display support readable by a child and deploy it in places of passage — the document itself suggests mobilising tourism, hospitality, transport, local businesses (p. 25).

4. Anticipate the information protocol for families in case of a protective measure. Measure 8 (pp. 33, 90) provides for immediate, strictly factual information, coordinated with the judicial authority, without naming the agent, when a protective measure is taken for sexual offences. Collective receptions for minors welcome 3 million children per year and 210 Parisian facilitators were suspended between 2025 and 2026 (p. 90): drafting a standard letter template and a validation circuit now avoids improvisation in a crisis situation.

5. Review your reporting procedures in light of the four identified vulnerabilities. The pp. 28-29 list: heterogeneity of reporting methods, disparity in the functioning of CRIP between departments, inequality in the conditions for gathering testimony, non-consolidation of data. Cross this diagnosis with one's own practices, identify the departmental CRIP contact, and verify the existence of an identified referent and a simplified alert circuit — as the document recommends for all sports, cultural, extracurricular, out-of-school and medico-social structures (p. 35).

6. Open the issue of minors as perpetrators and problematic sexual behaviours. The p. 53 explicitly identifies as a priority the training of professionals in identifying problematic sexual behaviours in young children, particularly in situations of proximity. This need is currently largely unmet in existing training plans. The FFCRIAVS 2025 report (reference no. 2 below) provides the missing operational material.

4. ADDITIONAL REFERENCES

🔍➕ For more information, see the articles referenced by "Practices in Health" on the theme of Child Protection ➡️🔗 https://www.pratiquesensante.com/protection-enfance

1. ONPE — Child Protection and Abuse: State of Play 2025 (December 2025). First edition of a state of play in 14 sheets crossing data from ASE, justice, security services, and research. Methodologically complementary: where the El Haïry report mobilises shock figures with little sourcing, the ONPE documents the sources, definitions, and limits of comparability. Sheet 12 specifically addresses intrafamilial sexual violence experienced during childhood. 🔗 https://onpe.france-enfance-protegee.fr/document/protection-de-lenfance-et-maltraitances-etat-des-lieux-2025/ (verified page, downloadable PDF online)

2. FFCRIAVS — Pathway of minors who commit sexual violence. Orientation report and proposals 2025 (September 2025). Derived from a public hearing conducted according to HAS methodology (public session of 19-20 June 2025, Ministry of Health). 45 recommendations structured into four priorities: better understanding, evaluation, support, prevention. This is the direct source of the figures presented on p. 53 and p. 96 of the analysed report, and the only document providing the operational details of measure 14. 🔗 https://www.ffcriavs.org/nos-actions/laudition-publique-2025/ (verified page, report and booklet downloadable; also accessible via vie-publique.fr, report no. 300180)

3. Éduscol — Implement the EVAR/EVARS programme (updated page June 2026). Official entry point of the programme implemented at the start of the 2025 school year: programmes by level, educational booklets with ready-to-use sessions (nursery, CP, CE1, CM2, 6th, second, first), two sheets “guidelines for implementation”, FAQ from October 2025 designed to address family concerns, seven information flyers for parents, directory of academic steering teams. This is the concrete toolkit linked to measure 2 (pp. 21, 84).🔗 https://eduscol.education.gouv.fr/5916/mettre-en-oeuvre-le-programme-evarevars (verified page; the old address eduscol.education.fr/2083 redirects to this canonical URL)

5. FAQ

1. Does this report have binding value for my organisation?

No. It is a document of recommendations submitted to the Prime Minister. Each measure sheet indicates the required normative vehicle: circular, decree, ordinary law, organic law or constitutional revision (for example measure 1, p. 20; measure 11, p. 44; measure 17, p. 61). No new obligation directly arises from it. However, several measures rely on texts already in progress — notably the draft law relating to the protection of children, of which article 5 already broadens the scope of suitability checks (p. 20).

2. Does the report address Child Social Assistance?

No, and it is explicit: "this strategy does not address the specific issues related to Child Social Assistance, considering that these are already the subject of specific treatment by the Government" (p. 9). Only the interfaces are addressed: reporting, health, justice, suitability, digital, support, coordination of pathways. If your activity falls under Child Social Assistance, this document does not answer your substantive questions.

3. What concrete deadlines are announced for the measures that concern me?

The sheets in annex 1 indicate them. 3 months: family information protocol (p. 90), Victims' Council (p. 102). 6 months: national training framework (p. 88), prevention structures for acting out (p. 85), adapted hearing / UAPED (p. 91), child victim package (p. 95), status of ad hoc administrator (p. 97), specialised justice (p. 92), non-prescriptibility (p. 93). 1 year: unique suitability service (p. 83), group of inspectors (p. 86), national application (p. 87), reporting in ENT (p. 89), national file (p. 94), child-friendly courts (p. 98). 2 years: EVAR(S) training for teachers (p. 84), Child Code (p. 100).

4. What is the German Dunkelfeld programme that measure 3 is based on?

Created in 2005 and supported by the Charité Institute of Sexology under the name "Kein Täter werden" ("not to become an offender"), it offers free and confidential therapeutic support to individuals sexually attracted to children who have not committed an offence (p. 104). The report suggests using it as a model for an experiment in three departments linked to the CRIAVS, with reinforcement of the national STOP number (p. 22). An important clarification on p. 85: the system stops if a victim is identifiable or an imminent danger is detected.

5. What is the actual coverage of the UAPED today and what does measure 9 change?

The 2023-2027 plan aims for 164 units, one per court; 151 were open or in the process of being created by August 2025 (pp. 37, 91). In 2025, 11,963 hearings of child victims were conducted across 90 UAPED (p. 28). Measure 9 proposes to enshrine in law a right to be heard in appropriate conditions, with the generalisation of filmed hearings, systematic recourse to UAPED and Mélanie rooms, the creation of national governance, and the deployment of mobile teams (p. 39). Training in scientifically validated methods, particularly the NICHD protocol, is explicitly recommended (p. 40).

6. What does "default digital protection" concretely cover?

Seven design requirements listed p. 74: effective age verification, security settings enabled by default, limitation of features promoting addictive use, disabling algorithmic recommendation systems for minors, simplified reporting procedures, removal of manipulative interfaces ("dark patterns"), priority withdrawal of content involving a child. The report presents it as complementary to the proposed law banning social networks for those under fifteen, adopted in the first reading at the National Assembly on 26 January 2026 (p. 73): where the ban sets an age threshold, the default protection sets a design standard.

7. Does the document say anything about the least visible children in institutions?

Yes, p. 34: very young children, children with disabilities, children accompanied by child welfare services, unaccompanied minors, victims of domestic violence and sexual exploitation. The proposed response is better coordination between maternal and child protection services, school medicine, child psychiatry, medico-social establishments, youth judicial protection, and departmental councils. The report also recommends increased vigilance during medical consultations for children who have attempted suicide, regarding the existence of violence, and a reform of the reporting system by doctors (p. 35). Note: this passage remains brief in light of the issue, and disability is mentioned only once.

6. REWRITING IN EASY TO READ LANGUAGE

What is this document about?

This document is about the protection of children in France.

It was written by Sarah El Haïry.

She works for the State. She looks after children.

She gave this document to the Prime Minister in July 2026.

The problem

In France, there are many laws to protect children.

There are also many services that help children.

But these laws and services do not work together enough.

A child is not protected in the same way everywhere in France.

It depends on where they live.

It also depends on the service they call.

It is unfair.

What the figures say

Many children suffer from violence.

A child dies every 5 days due to abuse from their parents.

Most sexual violence occurs within the family.

Many children do not speak. Or they speak very late.

The proposed solution

Sarah El Haïry proposes 20 measures.

These measures form a chain.

Each link in the chain protects the child at a moment.

Here are the 5 main ideas:

• Prevent violence before it happens.
• Not to miss any alerts.
• Protect the child immediately.
• Help the child long after.
• Make childhood a priority for the whole country.

The most important measures

1. Check the backgrounds of adults who work with children.

Today, each sector does its own checks.

Tomorrow, a single service will do all the checks.

It will be simpler and safer.

2. Train all adults who work with children.

Many professionals do not know how to identify violence.

They do not know what to do when a child speaks.

Mandatory training will be created for everyone.

3. Create an application to request help.

Today, there are many different numbers.

It is complicated for a child.

A single application will provide the right service in one click.

4. Treat the child immediately.

Today, the child often waits until the end of the trial to be treated.

Tomorrow, they will see a professional as soon as they have spoken.

Social security will pay for 10 sessions. You can request 10 more.

5. Listen to the child in an appropriate place.

There are special places: UAPEDs.

In these places, the child tells what happened only once.

These places do not yet exist everywhere. More need to be created.

6. Protect children on the Internet.

Many children are in danger on the Internet.

Adults seek to harm them.

Websites must protect children from the start.

Attention

These 20 measures are not yet laws.

They are proposals.

Some will be implemented quickly. Others will take 2 years.

Some require changing the law.

7. CROSS-ANALYSIS — VALUES OF HEALTH PRACTICES

Literacy : low. The document does not propose any tools suitable for varying levels of understanding; it only plans to adapt online training modules to local languages and cultural realities of the territories (p. 31) and to modernise terminology deemed stigmatising — "placement", "educational assistance in an open environment" — in the future Child Code (p. 62). No accessible version of the report itself is provided.

Empowerment : present but late. Measure 20 creates a Council of victims of violence during childhood, composed of adults with lived experience, consulted on bills and national strategies (pp. 64, 102). Article 388-1 of the Civil Code and the specific remedies for children are recalled (p. 57). However, the children themselves do not sit on any body created by the document.

Participation : announced mechanisms, not formalised. The report claims constant exchanges with associations, collectives, professionals, institutions, families, and children (p. 2), and plans a consultation of children, families, researchers, associations, and operators to define the modalities of default digital protection (p. 74). No co-construction methodology is described, and no consultation report is attached.

Community health : real collective dimension but top-down. The doctrine is based on a "minimal common culture" and a "collective reflex" of protection (p. 2), with mobilisation of economic actors — tourism, hospitality, transport, local businesses — and communication "at child level" in transit areas (p. 25). The logic remains that of national dissemination to the territories, not of community empowerment.

Ethics : attentive to legal tensions, little to social biases. The document precisely addresses conflicts of principles: presumption of innocence and confidentiality of the investigation versus information for families (p. 33), freedom of expression versus the offence of glorification (p. 63), non-retroactivity versus imprescriptibility (p. 44), data protection versus new files (p. 83). It explicitly dismisses public registers of sexual offenders of the SORNA type as unconstitutional in France (p. 105). However, cultural and social biases in detection are not analysed.

Human rights : structuring. The whole is based on the 1989 International Convention on the Rights of the Child and proposes to constitutionalise the best interests of the child (pp. 61, 99). The 130 recommendations addressed to France by the UN Committee on the Rights of the Child in 2023 are cited (p. 99). The requirement for territorial equity — mainland and overseas (p. 39) — is constant.

Intersectorality : it is the best-treated axis. The training foundation associates six ministries (p. 31), the group of inspectors brings together IEN, IA-IPR and IJS jointly trained by IGESR and IGJ (p. 23), the interoperability of Justice-Interior-departmental councils is set as a condition (p. 72), and the coordination of PMI / school medicine / child psychiatry / ESMS / PJJ / departments is explicitly recommended (p. 34).

Partnership : formalised mainly in inter-institutional. Coordination protocols UAPED-ARS-Health insurance-health professionals (p. 52), 'Violence against minors' hubs modelled on the intrafamily violence hubs of decree no. 2023-1077 (p. 58), cooperation OFMIN-NCMEC-Europol-Interpol and WeProtect alliance (pp. 38, 65). The partnership with the local associative fabric is mentioned as support, rarely as co-decision.

Fight against discrimination : partial and unequal. The report addresses gender inequalities, the risks of masculinism (p. 24), the strongly gendered nature of digital violence (89% of grooming victims would be girls, p. 69), and the objective of the EVARS programme including the fight against discrimination and equality of rights (p. 84). Disability only appears under the title of less visible groups (p. 34) and an increase in aid in the German system (p. 104). Unaccompanied minors are mentioned once (p. 34). No mention of LGBTQI+ children, nor of discrimination related to origin or migration status in access to protection. Non-judgment is, however, assumed for individuals presenting a sexual attraction to children before any act, with confidential and free support (p. 22).

8. EVALUATION OF THE RELIABILITY OF THE RESOURCE

Scientific relevance: average, with sourcing weaknesses that must be reported in dissemination.

Key points: the legal framework is precise and verifiable (dates of laws, decree numbers, code articles, order of 5 May 2017 for the FOJ framework p. 46, decree no. 2023-1077 of 23 November 2023 p. 58). Annex 3 (pp. 109-110) provides a substantial and recent bibliography: parliamentary reports, CIIVISE, ONPE, HCFEA, CESE, HAS, FFCRIAVS, associative reports. Annex 2 (pp. 104-107) offers a documented and dated international comparison of ten countries. The document explicitly refers to HAS for the development of recommendations that are lacking — clinical identification of abuse in infants (p. 25), framework for managing child psychotrauma with CN2R (p. 57), medical care for children victims of sexual violence (p. 57) — which reflects methodological clarity.