An overview of "emerging or changing topics" for health and social policies

1️⃣ ANALYTICAL SUMMARY

Context and issues — social protection under uncertainty and constraint.

The report starts from a finding of instability: economic and demographic uncertainties (slowdown in productivity, decline in birth rates, effects of warming, technological leap of AI) that undermine long-term projections and the "sustainability" of the social model (pp. 16-20). Against this shifting backdrop, it gathers the growing expression of doubts from stakeholders — researchers, administrations, field professionals — regarding the ability of usual mechanisms to design, implement, and evaluate policies (pp. 30-40). The affected publics are broad: users in non-recourse, single-parent families, exposed workers, patients facing a neglected psychiatry, beneficiaries subjected to control algorithms.

Operational contributions — a reasoned inventory of tipping points.

The document does not provide tools but a structured inventory of subjects that "return with increased acuity" and those that are emerging. It provides stakeholders with precise vocabulary and benchmarks to situate their difficulties: effectiveness of social rights (pp. 41-44), form of benefits in the face of inflation of essential goods (pp. 44-48), crisis of work and of "meaning" (pp. 49-51), articulation of public health / occupational health (pp. 51-53), financialisation of operators (pp. 67-69), risks of bias and discrimination related to AI in social organisations (pp. 71-75). Its value is to objectify feelings and to offer a shared framework for discussion with decision-makers.

2️⃣ KEY POINTS OF THE DOCUMENT

1️⃣ Non-recourse is no longer just an information problem, but a mismatch of responses. The report shows that "zero non-take-up territories" and "moving towards" can become "buzzwords" if they do not address the fact that thresholds and amounts do not keep up with inflation, and that the objectives of recovering overpayments and combating fraud produce "contradictory injunctions" that themselves feed into subsequent non-take-up (pp. 41-42).

2️⃣ The "management by indicators" is being directly questioned. The systematic recourse to objectives and indicators of "performance", developed since the 1980s, is described as a possible "metric obsession": it polarises attention on the most "profitable" groups, measures only the effects already expected, and makes agents "report" on data whose relevance and validity they do not recognise (pp. 39-40).

3️⃣ Medical demographics are set to reverse — an anticipated "stop and go". After stagnation until 2027, medical staff numbers would rise sharply, with the standardised medical density becoming 31% higher than its current level by 2050. The issue is therefore no longer just the shortage but the management of a possible long-term overstaffing (p. 66).

4️⃣ Financialisation now affects both health and the medico-social. Beyond pharmaceuticals and biology, the phenomenon extends to dental centres, teleconsultation, nursing home networks, nurseries, and home services, with directives dictated by national profitability standards — public authorities having "not thought generally" about adapting their negotiation methods (pp. 67-69).

5️⃣ AI in social organisations reinforces the risks of algorithmic discrimination. The use of AI to enhance the targeting algorithms for controls (CAF, combating fraud) tends to consolidate in a "non-transparent" manner the criteria leading to the "most productive" controls, amplifying biases against beneficiaries in unstable situations. The report advocates for "explainability" and a mastery of the "new data prescribers" (pp. 71-75).

3️⃣ ACTION PATHS FOR LOCAL ACTORS

1️⃣ Document partial and temporary non-take-up, not just total. Use the analytical framework from pp. 41-43 to identify, in your active files, the rights breaches related to subsidiarity rules and "isolation" controls, and report these reasons as causes of non-take-up (pp. 41-42).

2️⃣ Protect the core business of social work from "administrative overflow". Rely on the findings from pp. 43-44 to objectify, to the hierarchy, the time taken by unblocking files at the expense of long-term support, and redefine the support scope for volunteers and mediators.

3️⃣ Recognise the health AND social role of the municipal level. Mobilise the arguments from pp. 35-36 on the entanglement of competencies to advocate for a sustainable funding (and not through successive project calls) for social connection actions and local "outreach".

4️⃣ Reassess locally the steering indicators. Use pages 39-40 to propose that indicators be treated as "warning signals" triggering investigation, rather than as objectives automatically cascaded down through all levels, and to voice the "real work" of agents.

5️⃣ Anticipate the entry of AI into your control and recruitment processes. Rely on pages 71-75 to demand explainability, traceability and stages of human control, and to include AI in social dialogue — a subject not currently identified as such in negotiations.

6️⃣ Map the financialisation in your territory. From pages 67-69, identify operators (nursing homes, nurseries, health centres, home support services) belonging to profit-driven groups, in order to adapt contracting, authorisation and control strategies accordingly — knowing that unmet needs may lead local actors to advocate for these groups.

Identified unmet needs: the report remains, by nature, at the diagnostic stage — it does not provide any tools, checklists or protocols. Operational adaptation remains entirely to be constructed by the actors.

4️⃣ ADDITIONAL REFERENCES

1️⃣ Court of Auditors, The implementation of the "Zero Non-Use Territories" experiment (flash audit, 30 October 2025). Directly extends part 3.1: interim assessment of the system, with the persistent black spot of data sharing between partners.🔗 https://www.ccomptes.fr/sites/default/files/2025-10/20251030-Mise-en-place-experimentation-Territoires-zero-non-recours.pdf

2️⃣ CNIL, The development of AI systems: the recommendations of the CNIL to comply with the GDPR (July 2025). Methodologically completes section 4.6 on biases, explainability and the regulation of algorithms in organisations. 🔗 https://www.cnil.fr/fr/developpement-des-systemes-dia-les-recommandations-de-la-cnil-pour-respecter-le-rgpd

3️⃣ Court of Auditors, Social Security 2025 — report on the application of the social security financing laws (May 2025). Updates sections 1 and 2: sustainability, drift of the Ondam, nearly quadrupling of general contribution reductions (77 billion euros). 🔗 https://www.ccomptes.fr/fr/publications/securite-sociale-2025

5️⃣ FREQUENTLY ASKED QUESTIONS (FAQ)

1️⃣ Why does this report not provide solutions?

Because it is a prospective framing mission: it identifies "emerging or changing" topics without specifying how the IGAS would address them, to serve as early warning (pp. 13-15).

2️⃣ What is "the effectiveness of social rights"?

A legal concept: a right is only effective if it is "justiciable" (possible recourse to a judge) and "enforceable / opposable". The report applies it to housing, child custody, access to care and support (pp. 43-44).

3️⃣ Why do highly targeted benefits ("vouchers") pose a problem?

They multiply threshold effects and non-take-up, and mark a return to targeted in-kind benefits, in tension with the universalist spirit of social insurance from 1945 (pp. 47-48).

4️⃣ How are occupational health and public health poorly articulated?

They fall under separate institutional frameworks. The report advocates for bringing them closer together, particularly for the employment retention of chronic patients and in the face of infectious and environmental risks (pp. 51-53).

5️⃣ What does the report say about single-parent families?

Despite the refocusing of benefits, their poverty remains a "black spot": poverty rate of 32.3% (under 65 years, 2021), and about 40% of poor children reside there (p. 58).

6️⃣ What is the duty of vigilance and the CSRD directive?

European obligations for "sustainability reporting" and prevention of human rights and environmental violations throughout the corporate value chain, applicable progressively from 2024 (pp. 61-63).

7️⃣ What uses of AI will be prohibited or monitored?

The AI Act prohibits the "social score", biometric categorisation, emotion recognition at work; it places "high-risk" AIs (medical devices, facial recognition…) under surveillance (p. 75).

6️⃣ REWRITING IN EASY TO READ LANGUAGE

What is this document about?

It is a report from IGAS. IGAS oversees health policies and social policies. The report discusses major recurring issues. It does not provide solutions. It helps to reflect.

The main ideas.

• Many people do not request the benefits they are entitled to. This is called non-take-up.
• Benefits are sometimes too small. Prices are rising. Benefits do not increase quickly enough.
• Social workers spend too much time on paperwork. They have less time to help people.
• Work is judged by numbers. These numbers do not show the real work.
• Soon, there will be many doctors. Today, there is a shortage. That will change.
• Large companies are buying up care homes, nurseries, and practices. They want to make money. This creates problems.
• Computers choose who to control. Sometimes, they are unfair. They often target poor people.

What to remember.

This report highlights the problems of tomorrow. It helps frontline workers to understand. It aids in discussions with leaders and partners.

7️⃣ CROSS-ANALYSIS — VALUES OF HEALTH PRACTICES

• Literacy : little present as an object; the report mainly points out the difficulty users face with the ‘verticalised’ dematerialisation of procedures (pp. 41-42).
• Empowerment : low; the report describes users undergoing controls and injunctions more than being involved in the design of systems.
• Participation : approached from the perspective of deficit — it criticises the lack of stakeholder involvement in the development of reforms (pp. 34-35).
• Community health : collective dimension present through public health, the social function of hospitals, and the ‘general interest health service’ at the local level (pp. 54-56).
• Ethics : central at the end of the report — algorithmic biases, digital ethics, costs of innovative therapies (QALYs, DALYs, ‘fair innings’) (pp. 64-65, 71-75).
• Human rights : driven by the issue of the effectiveness and enforceability of social rights and by the duty of vigilance of companies (pp. 43-44, 62-63).
• Intersectorality : strong theme — articulation public health / occupational health, health / social, municipalities / departments / ARS (pp. 35-36, 51-53).
• Partnership : described mainly in its limits (overlapping competencies, fatigue in response to calls for projects), with a call for territorial deliberation frameworks (pp. 35-36, 55-56).
• Fight against discrimination : explicitly addressed through algorithmic biases in recruitment and control, and the overexposure of chronic patients and single-parent families (pp. 58, 71-73).

8️⃣ EVALUATION OF THE RELIABILITY OF THE RESOURCE

Scientific relevance — high. Institutional reference source (IGAS), recognised author, explicit method (≈ 190 interviews and round tables, list of individuals met pp. 81-89). Rich and solid footnote apparatus (Insee, Drees, Dares, OFCE, High councils, identified researchers). Limit acknowledged by the author herself: the report reproduces debates without resolving or deepening them, and does not provide any new primary data.

Operational relevance — low to medium. It is its nature: monitoring and framing document, without any directly usable tools (no checklist, protocol, ready-made key indicator). Its value for the field is argumentative and reflective, not instrumental. Data dated January 2024: on the financial, demographic and regulatory aspects (AI, CSRD, social security), an update via the references in section 4 is necessary.

Internal inconsistencies noted (reported, not corrected):

• Incorrect reference p. 28 to a "point 3.7" that does not exist; the section on inequalities between children is actually 3.6.3.
• Slight discrepancy in figures between the summary ("over 30% in 2050", p. 5) and the body ("over 31%", p. 66).
• Passage clearly corrupted / unfinished p. 38: "on the verge of drying up or at 'migué'" (likely a typographical error).
• A few duplicates and typographical errors ("indeed indeed" p. 6; "the the devices" p. 16; "conditiond" p. 48; "economisque, sosial" p. 82) — without impact on the substance.

9️⃣ STRATEGIC HASHTAGS

#SocialPolicies #NonUse #SocialProtection #PublicHealth #AIandSocial #Financialisation #Prospective #healthpractices

What has changed since the report 

📈 What has increased / worsened or been confirmed

The decline in birth rates — the report's "structural" hypothesis is validated. INSEE confirms the continuation of the decline: after 1.61 in 2024, the current fertility indicator is set at 1.56 children per woman in 2025, with 645,000 births and a natural balance becoming negative for the first time since 1945. The report was hesitant (cyclical or structural?): the turning point is now clear (pp. 19-20). INSEEThe Quadrature du Net

Financial sustainability — confirmed slippage. The social security deficit has doubled in two years, reaching €21.6 billion in 2025, the highest since 2012 excluding covid, and the Court of Auditors judges that the Ondam no longer ensures effective regulation since the health crisis. This is exactly the concern of parts 1 and 2. INSPINSP

The risk of algorithmic discrimination — moved from diagnosis to litigation. What the report indicated as a risk (pp. 71-75) has become a legal battle: a case before the Council of State brought by 25 organisations denounces the CAF algorithm, reinforced by an opinion from the Defender of Rights which sees a presumption of “indirect discrimination”, and an internal study from the CNAF in 2025 acknowledged the discriminatory effects of the algorithm. Also to note: since 1 January 2025, automated control extends to people on RSA via France Travail. CAF – La Quadrature du Net +2

Psychiatry “in decline” — still untreated. Mental health was the National Great Cause 2025, but stakeholders point out that no measures regarding mental health are present in the PLFSS 2026. The observation on p. 55 still holds. Ministry of Solidarity

📉 What has decreased / found a beginning of response

The “family and health banners” — the recommendation of the report has been implemented. The report advocated for capping and redeploying these reliefs (pp. 23-25). It is done: since 1 January 2026, the former triptych “Fillon reduction” + health band + family band is replaced by a Unique Gradual General Reduction, with a monitoring committee placed with the Prime Minister responsible for evaluation until 2029. This point is therefore no longer a “blind spot” but an open project. UnivadisKlesia

Financialisation — moved from the blind spot to regulation. The Senate produced its reference report (“A takeover of health?”, Sept. 2024), then the LFSS began to act: the PLFSS 2026 extends the price regulation measure for procedures introduced in 2025 to other financialised sectors such as ophthalmology and dental care, and stakeholders are now calling for an observatory on the financialisation of the health system. The subject remains lively but is no longer ignored (pp. 67-69). Info.gouv.frMinistry of Solidarity

Non-take-up — moved to evaluation. The “Zero Non-Take-Up Territories” experiment became effective in mid-2025, with an evaluation report expected in June 2026. The diagnosis of the report (pp. 41-44) remains accurate, but we have entered the assessment phase.

🌱 Emerging / new points since January 2024

The suspension of the pension reform. A subject that the report had deliberately set aside, but which reshapes the context: the PLFSS 2026 suspends until January 2028 the schedule for raising the legal age for the generations 1964 to 1968, financed notably by a lower revaluation of pensions in 2027. Risqual

The operationalisation of AI regulation. The AI Act, still under discussion in the report (p. 75), is adopted and being deployed; in parallel, the CNIL has published its AI/GDPR recommendations (July 2025) and is assigned in 2025 a role as AI regulator, notably as the supervisory authority for high-risk systems in employment and biometrics. We move from “caveat” to regulation. Accordia

De-smicardisation as a framework of thought. The reform of exemptions has reactivated a debate that the report only touched upon indirectly: the reform, designed by economists Bozio and Wasmer as a lever for de-smicardisation of wages, has been taken up by the Government with an economic objective — the “accordion effect” of the minimum wage (p. 21) finds there a partial and contested response. Apicil