PhD student in public health (2nd year) · Jurist · Entrepreneur · Special needs educator
From supporting people to public health research, my path has led me to explore the different scales of the health system: from individual realities to organisations, all the way to public policy and governance.
These experiences, at the crossroads of practice, management, law and research, have shaped a conviction: transforming health systems means bringing human experiences, scientific knowledge, collective dynamics and political choices into dialogue.
My ambition is to open doors between these worlds that sit side by side without always understanding one another, to break down silos between perspectives and bring research closer to action, so as to foster self-determination, inform collective decisions and contribute to a fairer, healthier society.
PhD research currently being conceptualised, co-supervised by Nadia Sourial (PhD) and Géraldine Layani (MD) GSK Research Chair in optimal management of chronic diseases CRCHUM FRQ doctoral scholar (B2)
Drag to exploreClick a door to enterSwipe the doorsTap to enter
Drag to look left or right, or move the mouse towards the edges. Click a door or section to enter.
Swipe through the doors or choose a number. Tap a door or a section to enter.
Thesis in progress · Topic under construction, subject to major change
Provisional title
How is the intersectoral response organised today around older adults at home who combine social needs and morbidity: according to which profiles, with which resources and through which networks?
Design: multiphase mixed methods (Creswell & Plano Clark, 2017), pragmatic stance; O1 scoping review (JBI, PRISMA-ScR) and descriptive analysis of 211 data; O2 quantitative observational ecological study; O3 explanatory sequential mixed design (QUAN → qual), egocentric network analysis.
The research question
How is the intersectoral response organised today around older adults at home who combine social needs and morbidity: according to which profiles, with which resources and through which networks?
In Quebec, home care and support (SAD) covers the care and services provided in people’s homes by the public network, mainly by the team of the CLSC (local community service centre). In 2024-2025 it served 395,783 people, for 37.2 million hours; 17,883 people were waiting for a first service. Among people aged 65 and over, 80% live with at least one chronic disease and half with several (multimorbidity) (Ministère de la Santé et des Services sociaux [MSSS], 2026a, pp. 4 and 6).
Share of needs met by the public network, 2023
Home care: 25.4 of the 234.7 million hours required (CSBE, 2024, p. 11).
People waiting for a first home care service
An increase of about 30% in three years (CSBE, 2024, pp. 2 and 96).
Long-term care users and resources
Residential care receives 62% of funds for 16% of users (CSBE, 2024, p. 7).
Health of people aged 65 and over in Quebec
And 91% of people over 55 identify with “at home for life” (CSBE, 2024, p. 7; MSSS, 2026a, p. 4).
After the four volumes of Bien vieillir chez soi (2023-2024), the Health and Welfare Commissioner (CSBE) concludes that the home care ecosystem is complex, poorly integrated and underperforming. In 2023, the public network provided only 25.4 of the 234.7 million hours required at home: a response rate of 10.7%, compared with 98.9% in CHSLDs. Home care receives 38% of long-term care resources for 84% of users, and people waiting for a first service rose from 13,250 (2019) to 17,226 (2022). The rest relies on family caregivers, community organisations and the private sector, that is, on sectors outside the public network.
In the study area (CCSMTL, 8 CLSC territories), 48,575 people are aged 65 and over in 2026 (ISQ). In 2021, 28.6% of them lived below the low-income threshold, up to 41.8% in Pointe-Saint-Charles and 38.7% in Centre-Sud; 66.9% of households rent (Statistics Canada). At 211, the needs of people aged 70 and over concern first personal and community support (21.9%) and housing (13.4%), two domains where their needs remain unmet twice as often as for all ages. Among people aged 60 and over, the most frequent reason is “service not available” (324 needs), followed by eligibility criteria (210) and lack of transport (66). Only 44 requests from people aged 60 and over come from a CLSC.
An older person may live at the same time with one or more morbidities (health problems, acute or chronic) and several social needs: unsuitable housing, too little income, isolation, no transport, an exhausted caregiver. These are called co-occurring social needs. Together they are complexity factors: each makes the others harder to solve, and the person’s care pathway becomes less fluid (Safford et al., 2007; Grembowski et al., 2014; Schaink et al., 2012; Johnson & Bacsu, 2018). The Mieux chez soi policy recognises this: when needs span several areas of life, several practitioners “must work together”, and chronic diseases make home care needs more complex and more intensive (MSSS, 2026a, p. 4).
These situations are hard to detect, because no single service sees all of a person’s needs. They are also hard to coordinate: the Health and Welfare Commissioner (CSBE) describes the home care ecosystem as complex, poorly integrated and underperforming. A joint, tailored and coordinated response is not always possible. It is plausible that, in the same territory, organisations do not know each other or do not work together, and that a resource is missing while no one knows that another one, from another sector, could take its place. No study shows this directly: it is an inference, suggested by research on service networks, on the capacity of community organisations and on how hard it is to find available help (Lemieux-Charles et al., 2005; Varda et al., 2020; Hogg-Graham et al., 2021; Macleod et al., 2017). At 211, the most frequent reason why needs of people aged 60 and over remain unmet is “service not available”.
The Mieux chez soi policy wants the CLSC to be the main entry point (MSSS, 2026a, p. XII). In practice, an older person enters the network through many doors: the CLSC home care team; the day hospital; a hospital stay; the emergency department after a fall; home care from the CISSS or CIUSSS (the regional health and social services institutions); the family medicine group (GMF); the pharmacy; the municipality; social housing; paratransit; a community organisation; a family caregiver; 211, the information and referral service. Each door sees part of the situation; none guarantees that all needs will be taken into account.
The government policy on first-line health and social services (MSSS, March 2026) attaches each person to a CLSC territory, which becomes the reference unit for coordinating first-line care (p. 15). It gives the CLSC four roles: navigator, coordinator, intersectoral leader and co-facilitator (p. 42). The accompanying ministerial orientations find that the CLSC, designed as a territorial hub, has become devitalised and cut off from its original mission (p. 1): the aim is to restore that role.
Several arrangements already bring stakeholders together. PRISMA, a Quebec model of integrated services for frail older people, combined a single entry point, case management, a shared assessment tool and partner coordination; its evaluation showed less functional decline and fewer unmet needs (Hébert et al., 2003, 2010; Couturier et al., 2021). The MSSS Integrated Local Approach for Older Adults (AIPPA, 2024) relies on collaborative practices and on identifying people with needs (p. 8), on ITMAV outreach workers and the community organisers of health institutions (p. 28), and on communities of practice (p. 59; Ranmuthugala et al., 2011). The community organiser, attached to the community organisation service of the CISSS or CIUSSS, is the link between the institution and local organisations (Bourque & Lachapelle, 2010).
Intersectoral action is not new in Quebec: the 2003 policy Chez soi : le premier choix already called for it, and the Université de Montréal RUISSS (2022) proposed a six-sector diagram (public network, first line, parapublic, community, city, housing) and four levels of collaboration (information exchange, cooperation, coordination, integration; after Bilodeau, Parent and Potvin, 2019). The MSSS AIPPA approach and its intersectoral liaison committee at the CCSMTL extend this work. What is missing is a way to know, territory by territory, which sectors to mobilise for which combinations of needs.
This diagram, together with the literature, was used to delimit the sectors studied in the thesis (MSSS, 2026b, p. 42; Bilodeau et al., 2019).
These arrangements are valuable, but the intersectoral response remains hard to see. It often rests on local partnerships: a community of practice, a coordination table, one person acting as facilitator. If that person leaves or funding stops, the response weakens: it is not sustainable and often incomplete. Above all, none of these arrangements says which combinations of needs recur most, which resources exist for each in a territory, or how the network organises itself today to respond. That is exactly what the thesis addresses.
It shows that the social needs of older adults living at home are frequent and cumulative (McManus et al., 2025; Tingvold et al., 2026), that they cluster (Kreuter et al., 2021) and that profiles of met and unmet needs exist, but mostly based on the CANE and on caregivers’ views (Janssen et al., 2020; Kraake et al., 2024). It documents referral, navigation and social prescribing (Hébert et al., 2010; Percival et al., 2022) and mapped service supply (Kim, 2026; Álvarez-Pérez et al., 2025).
It does not yet say which combinations of needs recur among people receiving home care, nor which remain unmet; it rarely links these combinations, within a territory, to accessible resources; and its network analyses record whether a tie exists, rarely its level of collaboration, almost never starting from the home care teams of several CLSCs. The O1 scoping review starts from 60,957 records from four databases (MEDLINE, Embase, APA PsycInfo, PubMed): 8,938 published before 2006 removed, 52,021 imported into Covidence, 25,412 duplicates removed, 26,609 studies to screen by two reviewers.
Some studies have already mapped older adults’ needs and resources: in Quebec, the study led by Anne-Marie Séguin (INRS) for the Société d’habitation du Québec (2018) described the met and unmet needs of people aged 75 and over by type of area, then local resources and people’s strategies. Juliette Duc and colleagues mapped, with QGIS and R, the supply of and demand for services for people aged 65 and over across Quebec: fewer services per older person in rural areas and strong local variability (Duc et al., 2026). Others have described need profiles among people with dementia living at home (Janssen et al., 2020) or analysed service networks with social network analysis, SNA (Urbanoski et al., 2018). No study identified yet links co-occurring social needs and one or more morbidities as complexity factors at the scale of a CLSC territory.
The literature is not empty on each component of the question, whether the measurement of multiple social needs, their co-occurrence, patients’ desire for assistance or the services that respond to them. However, no synthesis identified brings together, for adults in primary care, the unmet nature of needs, their co-occurrence, their prioritisation by patients, their interrelations, the response network and the trajectory to resolution. The gap therefore lies in the link between how needs are configured and how the cross-sector response is configured, and in the mechanisms that explain why some needs persist despite being identified.
For methodological accuracy, the wording used is ‘no synthesis identified in the PubMed and Cairn searches conducted’ rather than ‘no synthesis exists’.
The home care social worker assesses a person’s situation, refers and supports them: this is an individual intervention, and it is essential. An intersectoral response to the co-occurring social needs of a person with one or more morbidities is something else: it is the organisation, in a territory, of several sectors (health, social services, community, municipality, housing, transport) acting together, durably, for the same combination of needs. Today the issue is not practitioners’ skills but the lack of a shared organisation: each practitioner rebuilds, case by case, a network that could be known in advance. Social work assesses, refers and coordinates case by case; an intersectoral response organises several sectors in advance around a profile (Couturier et al., 2022; Bilodeau et al., 2019). The challenge: moving from “who knows whom” to an organised, lasting response.
| Social work today | Intended intersectoral response |
|---|---|
| Assesses one person’s situation | Organises several sectors in advance around a needs profile |
| Refers to a resource | Health, social services, community, municipality, housing, schools, daycares |
| Coordinates a file, case by case | Also relieves caregivers: respite, day centres, meals on wheels, transport |
| Depends on “who knows whom” | Organised, known and lasting in each territory |
Based on Couturier et al. (2022) and Bilodeau et al. (2019).
The 2026 policy makes the CLSC territory the reference unit. Its accompanying Orientations state that this territory is no longer a mere administrative division but a living space, and that the 166 current territories are not ends in themselves (MSSS, 2026c, summary and p. 4; MSSS, 2026b, p. 42). Yet CLSC territories were first defined to assign a population to each institution (MSSS, recommendations on managing territorial data); in Montreal, public health has divided the island into 101 neighbourhoods closer to communities, and administrative units do not always match lived environments (Di Méo, 1998; Riva et al., 2008; Macé, 2010; Breton et al., 2009). As a secondary question, the thesis will check whether the resources people actually use follow the CLSC boundaries or cross them.
Hence the starting idea: identify the most frequent profiles of co-occurring social needs among older adults receiving home care; for each profile, locate the resources available in each CLSC territory; then understand how the response is organised today, in order to improve it. The thesis concerns older adults with one or more morbidities who live at home and receive home care in the eight CLSC territories of the CIUSSS du Centre-Sud-de-l’Île-de-Montréal (48,575 people aged 65 and over in 2026).
The common thread fits in three words: profiles → resources → network. O1 says which combinations of needs recur; O2 says, territory by territory, what exists to respond and what is missing; O3 says who already works together and which links are missing. Each step gives the next what it needs.
My thinking comes from my path across several sectors and from observing home care policies, on which I worked at ANAP. As a specialised educator, I saw how hard it is to organise a joint response to the most complex situations, when several services intervene without coordinating. This thesis seeks to make visible, and therefore organisable, what today rests mostly on everyone’s goodwill.
The resources below inform conceptualisation. The title, research question, thesis argument and objectives in their dedicated sections constitute the current framing of the project.
I am at the conceptualisation stage: a corpus of prioritised publications, the gaps it reveals and the training readings that support it.
The corpus was assembled from Ovid (Embase, MEDLINE), PubMed and Cairn, and each publication was assigned a priority according to its relevance to the thesis question.
For each of the nine dimensions examined, the table shows the state of the literature, the significance of the gap and a few key references.
Present but not yet settled
Studies indexed in PubMed increasingly distinguish social risk, social need and desire for assistance. French-language work in Cairn adds a vocabulary of precariousness and social vulnerability that is often multidimensional but conceptually broader than unmet need.
Gruß 2021; Tucher 2026; Tudrej 2019; Karran 2023
Abundant but heterogeneous
Instruments differ markedly in their domains, thresholds and mode of administration. Multidimensional deprivation scores found in Cairn, such as EPICES and PRECAR, are not equivalent to unmet social needs questionnaires.
Zein 2026; Karran 2023; Tudrej 2019; Rogers 2025
Well documented for at least one need
Many studies describe prevalence by domain and cumulative burden, but estimates are difficult to compare when the definition of need and the instrument differ.
Jih 2023; Bleacher 2020; Fiori 2021
Emerging but now demonstrated
The coexistence of several social needs or risks is common, and clusters, classes and cumulative burden measures have been proposed. A recent study also shows that multiple risks sharply increase the likelihood of wanting assistance.
Vest 2025; Kreuter 2021; Zeng 2024; Tucher 2026
A few direct studies, but a field still underdeveloped
The literature mainly measures desire for assistance, negotiated goals, acceptability or relative priority, and recent work reinforces the distinction between detected risk and desired help. The e-CARE studies show support adjusted to the patient’s pace and priorities, without producing a standardised ranking of needs.
O’Loughlin 2022; Gruß 2021; Tucher 2026; Vasan 2026; Ouanhnon 2022
Mostly described as co-occurrence or overall complexity
Quantitative analyses mainly document co-occurrence and profiles. Qualitative French-language work in Cairn describes how social, administrative, economic and health difficulties are intertwined, but rarely their direction, sequence or causal mechanism.
Vest 2025; Kreuter 2021; Marron-Delabre 2015; Ouanhnon 2022
Growing but dominated by intermediate indicators
The field increasingly distinguishes screening, referral, connection, receipt of a resource and resolution. French-language e-CARE studies also track concrete outcomes such as access to entitlements, a regular doctor, housing and autonomy in navigating care, but without a standardised measure of whether each need is resolved.
Steeves-Reece 2022; Hausmann 2024; Ouanhnon 2022; Agarwal 2025
Substantial literature, but often separate from patient data
PubMed documents navigation, community health workers, medical-legal partnerships and closed-loop referral, while Cairn documents social units in general practice and French primary care and territorial coordination structures (MSP, CPTS, DAC). Yet the configuration of the network is rarely linked to need profiles and their resolution.
Hogg-Graham 2021; Tevaearai 2022; Yatim & Sebai 2021; Lessi 2023; Ouanhnon 2022
Not identified in the PubMed and Cairn audits
No synthesis located brings together, for adults attending primary care, the unmet nature of needs, multi-domain measurement, patterns of co-occurrence, the priority patients give to their needs, interrelations between needs, the configuration of the mobilisable network and the trajectory to resolution.
Karran 2023; Vest 2025; Chang 2026; Tudrej 2019; Ouanhnon 2022
Research question
How is the intersectoral response organised today around older adults at home who combine social needs and morbidity: according to which profiles, with which resources and through which networks?
How is the intersectoral response organised today around older adults at home who combine social needs and morbidity: according to which profiles, with which resources and through which networks?
For older adults receiving home care, the intersectoral response cannot be organised need by need: it must be organised by profile of co-occurring needs and by CLSC territory, based on the resources actually available and the networks actually mobilised around home care teams.
To produce, by profile and by CLSC territory, guidance for organising the intersectoral response with the resources actually available.
1 · Gap
The co-occurring social needs of older adults living at home are documented, but in a fragmented way: definitions, measures and profiles vary. No study identified derives profiles from home care assessments (OEMC, interRAI-HC) or from 211 requests (Wilks et al., 2026; Janssen et al., 2020; Kraake et al., 2024; Kreuter et al., 2021).
2 · Objective
Identify the different types of profiles of co-occurring social needs among older adults with morbidity or comorbidity factors who are followed in home care services (CLSC).
3 · Final output
Final output: a typology of profiles of co-occurring social needs, met or unmet, translated into vignettes. What it gives us: knowing which combinations of needs recur most, so the response can be organised around them.
1 · Gap
Analyses of supply start from a directory fixed in advance, without spatially cross-referencing it with expressed and unmet needs. Directories are often inaccurate; it is not known where supply is absent, inaccessible or saturated (Kim, 2026; Janda et al., 2021; Garcia et al., 2025).
2 · Objective
Map, in space, the resources in each CLSC territory that could provide an intersectoral response to these needs.
Six sectors: public network, first line, parapublic, community, municipality, housing.
3 · Final output
Final output: an atlas by CLSC territory, a needs × resources matrix and a gap index. What it gives us: knowing, for each profile, where the response exists, where it is missing, and why (absent, inaccessible or saturated).
1 · Gap
Responses to older adults’ social needs are mostly studied as referral processes. Interorganisational networks are most often measured without a level of collaboration and rarely start from several CLSCs (Hébert et al., 2010; Bevc et al., 2015; Richter et al., 2026).
2 · Objective
Identify the organisational trajectories of the intersectoral resources mobilised to respond to the different profiles, in several contrasting CLSC territories (SAD).
With the home care teams of the CIUSSS du Centre-Sud-de-l’Île-de-Montréal.
3 · Final output
Final output: a network map for each CLSC and each profile, with the missing links. What it gives us: knowing who already works together, who should, and which links to act on.
Population: people aged 65 and over. For 211, the main analysis covers people aged 70 and over, because 211 only publishes the 60-69 and 70-and-over age groups; people aged 60 and over serve as a sensitivity analysis.
At least two social needs from different domains, present at the same time in the same person.
(Kreuter et al., 2021)
Any health problem, acute or chronic. A person may have only one or several.
The health and social care and services provided at home by the team of the CLSC territory.
(MSSS, 2026)
At 211, a need is unmet when no adequate resource could be offered, with a reason. It is measured; it is not made an inclusion criterion.
(Alderwick and Gottlieb, 2019)
“Unmet” changes meaning depending on the source. At 211, a need is “unmet” when no suitable resource could be offered, with a reason; for Statistics Canada or INSPQ, it is a perceived need without service received; in the CANE, a problem without adequate help according to an assessor. The thesis uses the 211 definition and compares met and unmet needs in all three objectives (Alderwick & Gottlieb, 2019).
The method at a glance
Scoping review and 211 requests
ProductTypology of profiles and vignettes
R in Positron, maps in QGIS
ProductAtlas by CLSC and gap index
Ego-centred network of home care teams, then interviews
ProductNetwork map and missing links
Population: people aged 65 and over with one or more morbidities, living at home and followed by home care support (SAD). At 211: people aged 70 and over (211 publishes 60–69 and 70 and over); people aged 60 and over serve as a sensitivity analysis.
Bernier (2026), O1 scoping review; PRISMA-ScR (Tricco et al., 2018).
Overall design: pragmatic stance; multiphase mixed methods (Creswell and Plano Clark, 2017), O1 → O2 → O3, each phase building the next. O1 vignettes serve as the name generator in O3; the O2 atlas selects the O3 CLSCs and provides the potential supply. Field stakeholders are involved in every phase: older adults, family caregivers, professionals, managers and partners from all sectors, with one or more patient partners.
An older person does not experience needs one by one: the stairs, groceries, loneliness and forms reinforce each other. Keeping 12 separate domains reproduces the fragmentation the CSBE criticises in home care, where each service answers only its own domain. A profile describes the combination, and therefore the response to organise: whom to mobilise together. It includes one or more morbidities, because they change the nature of the response (half of people aged 65 and over live with several morbidities).
Feasibility first: individual home care data (OEMC, I-CLSC) require legal authorisation and at least 2,000 users; the protocol’s simulation shows that below that the profiles cannot be recovered, and that at 211 (1.4 needs per request) no sample size is enough. Next, the level of decision: a CLSC plans for groups of situations, not for one case. Finally, ethics and transfer: a profile turned into a vignette (Madame Protocole) can be submitted to every team without exposing anyone.
Because the 2026 policy makes the CLSC territory the reference unit: the response must be judged where people live. The map sets met and unmet needs against the resources of the six sectors, measures accessibility on foot and by public transit (E2SFCA) and distinguishes supply that is absent, inaccessible or saturated. It also selects, transparently, the 3 or 4 contrasting CLSCs for O3. R in Positron and QGIS make every indicator recalculable.
A resource available on the map is not necessarily mobilised. Only network analysis shows who actually works with the home care team, at what level of collaboration and where the missing links are. Published examples: Contandriopoulos et al. (2016, Healthcare Policy) mapped collaborations among Canadian chronic disease prevention organisations; Richter et al. (2026, BMC Health Services Research) used an egocentric network to analyse the coordination of health promotion in municipalities; Lemieux-Charles et al. (2005, The Gerontologist) studied community dementia care networks in Ontario.
Egocentric network: the ego is the home care team of each CLSC. For each profile (vignette), a questionnaire asks to which organisations the team referred an older person, or with which it collaborated, over the past 12 months, with the sector, level of collaboration (1 to 4), frequency and delay. The SNA (size, diversity, share of level 3 or 4 ties) is then contextualised through interviews with the actors, who retrace the pathway of each profile: who intervenes, in what order, and where it breaks.
Conceptual illustration of the thesis
In my thesis, network analysis starts from the home care team of each CLSC. For each vignette, the team names the organisations to which it has referred an older adult, or with which it has collaborated, over the past 12 months; line thickness indicates the level of collaboration, from 1 to 4. This declared network is compared with 211 referrals and with the atlas supply, then explained in interviews. Illustrative network, not observed.
This section explains, step by step, how the mapping addresses Objective 2: which data are brought together, which variables link them, how R computes the indicators and how QGIS turns them into maps. The results shown are examples by need domain, computed on October 17, 2026. They will be redone by needs profile once the Objective 1 review is complete.
Resources for older adults are known through separate directories: a list of GMF, a list of pharmacies, a directory of community organizations. None of them says, for a given territory, whether an older adult can find, close to home, both a health resource and the social resources they need.
Objective 2 therefore links three things on a single map: where older adults live, where the resources of the six sectors are located and which needs remain unmet. The final product is an atlas by CLSC territory, a needs × resources × territory matrix and a gap index.
The choice of sectors to map does not come from the data available. It comes first from the diagram in the government Policy on frontline care and services: around the CLSC revolve the hospital, the clinic, the pharmacy, the community organization, the municipality, the school and the CPE (Ministère de la Santé et des Services sociaux [MSSS], 2026a, p. 42).
The same page assigns the CLSC four roles: navigator, coordinator, intersectoral leader and co-facilitator (MSSS, 2026a, p. 42). To play these roles, it must know which partners exist in the territory, and where.
The literature refined the breakdown. The intersectoral framework of the RUISSS de l’Université de Montréal sorts the partners of a continuum for older adults into six sectors (RUISSS de l’Université de Montréal, 2022, p. 33). It adopts the four levels of collaboration of Bilodeau and colleagues: information exchange, cooperation, coordination, integration (Bilodeau et al., 2019; RUISSS de l’Université de Montréal, 2022, p. 33).
Resource types sorted by sector in the project settings (35 types). Assignment to a sector follows a written rule, not the source.
An open breakdown. These six sectors are a deductive starting point, drawn from the Policy and the literature. The network analysis of Objective 3, by contrast, is inductive: home care support teams freely name the organizations they work with. It may therefore reveal unforeseen sectors, such as private businesses or other actors. These will then be added to the atlas.
All the data are public or already in hand. They were inventoried dataset by dataset (73 datasets) and entered in a register of sources, with their year and producer (Bernier, 2026b). No individual-level user data are used.
| Data (source, year) | Variables used | What it measures | What it is used for in Objective 2 |
|---|---|---|---|
| CLSC, RLS, RTS and RSS territoriesMSSS, boundaries as of April 1, 2026 | CLSC_code, CLSC_nom, RTS_code | The official boundaries of the 8 CLSC territories of the CCSMTL | Unit of analysis and decision: each indicator is reported by CLSC |
| Blocks and dissemination areasStatistics Canada and ISQ, 2021 | IDIDU, ADIDU, pop_2021, area | The smallest areas for which the population is known | Starting points for walking trips, and building blocks to reconstitute each CLSC |
| Census profile by dissemination areaStatistics Canada, 2021 | People aged 65, 75 and 85 and over, low income among those 65 and over (MFR-ApI), living alone, neither French nor English, core housing need | How many older adults live in each block, and their social need factors | The denominator (people served) and the potential needs index |
| Population by CLSC 2001-2051MSSS and ISQ | Aged 60, 65 and 75 and over on July 1, 2024 | The recent population of each CLSC | Rates per 1,000 older adults (density, 211 needs) |
| Network and frontline resourcesMSSS, 2025-2026: M02 file, GMF, clinics, nurse practitioners, pharmacies, RPA, authorized beds | type_code, location, capacite (registered patients, places, beds) | Where health services are located, and their capacity when published | The health pillar of the response, and weighting accessibility by capacity |
| Directory of organizations211 Grand Montréal, 2026 | id_211, linked categories, territory served, “home care support” flag | Community organizations and what they offer | The social pillar: each category is linked to a need domain |
| Municipal and neighbourhood resourcesCity of Montréal, SPVM, Ministère de la Famille, MEQ, 2024-2026 | Type, address or location | Public places, food assistance, grocery stores, CPE, schools, police, fire stations, non-market housing | The parapublic, community, municipality and housing sectors |
| OpenStreetMap2026 | Street and sidewalk network; rehabilitation, assistive devices, places of worship, post offices | The actual walking route, and a few resources absent from official sources | Walking times, and a supplementary inventory only |
| Public transit schedules (GTFS)STM, 2026 | Stops, routes, schedules | Travel time by bus and metro | Access within 15 and 30 minutes by public transit |
| Deprivation indicesINSPQ, 2021 | Material and social deprivation quintiles by dissemination area | Neighbourhood disadvantage | Identifying priority areas and validating the needs index |
| 211 dashboard211 Grand Montréal, 2024-2025 (2025-2026 as sensitivity analysis) | Needs expressed and unmet, reason, domain, age group (60-69, 70 and over), borough, FSA (RTA) | The needs actually expressed, and those left unanswered | Locating met or unmet needs, and naming the gap (supply absent, inaccessible, saturated, ineligible) |
| Boroughs and forward sortation areasCity of Montréal, 2026; Statistics Canada, 2021 | arr_code, RTA | The areas for which 211 publishes its data | The bridge between 211 needs and CLSC territories |
| Address pointsCity of Montréal, 2026 | Number, street, coordinates | The exact location of an address | Geocoding published lists that lack coordinates |
The complete register (file, format, year, producer, role) accompanies the scripts. The same sources, extended to the island, the Metropolitan Community and Québec, are used for comparisons outside the CCSMTL.
211 receives requests from the public and classifies them using the AIRS taxonomy. Its public dashboard gives, by borough and age group, the needs expressed and those left unmet, with their reason (211 Grand Montréal, 2026).
Example: in Verdun, in 2024-2025, people aged 60 and over expressed 602 needs to 211; 32 remained unmet, or 5.3% (Bernier, 2026b). Personal, family and community support and housing come first.
Each record in the 211 directory gives an address, service categories and a target clientele. The project keeps nonprofit organizations and cooperatives, places each record on the map, then links each of its categories to a need domain.
This linkage was coded twice, independently. Agreement is very high: Cohen’s κ = 0.95 across 163 categories, with 7 disagreements resolved in writing (Bernier, 2026b).
The sources come from six sectors and do not speak the same language: a GMF has an MSSS code, an organization a 211 number, a grocery store a City permit. To compare them, we need variables that exist, or can be built without strong assumptions, in every source. The dataset-by-dataset check retains six (Bernier, 2026a).
The place where the service is delivered, in NAD83 / MTM zone 8 coordinates (EPSG 32188), with the nature of the point and the geocoding precision.
In 211, 414 of the 842 geolocated records in the territory share their point with another record (shared head office).
The 2026 CLSC, the dissemination area, the borough and the FSA, all computed from the location, by point in polygon.
Never read from the original fields: codes and spellings differ. 27 dissemination areas (6.6% of the population) straddle two CLSC.
One of the six sectors: public network, frontline care, parapublic, community, municipality, housing. Assigned by a written rule.
Dual cases resolved in writing, for example CHSLD (sectors 1 and 6).
The domain the resource addresses, linked to the 211 AIRS taxonomy by a crosswalk table, with double coding.
No crosswalk table existed: it was built for the thesis.
The extraction date, or the row’s update date when there is one. We keep both.
Supply is a snapshot from October 2026, needs cover 2024-2025: the gap is displayed everywhere.
A key assigned after deduplication: same service type within 100 m, or same official identifier. The list of sources for each service is kept.
Matching by name is nearly impossible: 59 of 79 pharmacies are found within 100 m, 1 of 79 by name.
In practice, location does all the work. Each resource is a point. The software looks up which block (IDIDU), which dissemination area (ADIDU), which CLSC (CLSC_code) and which borough (arr_code) the point falls in.
These codes then serve as keys: the census is joined by dissemination area, 211 by borough or FSA. The domain (domaine_code) links resources to needs, and the type (type_code) links them to their sector.
| Rule | Bias avoided | What justifies it in our data |
|---|---|---|
| Across sectors, compare deduplicated counts of services and distances | Incompatible units | Capacity exists only in a few datasets, in beds, places, housing units, registered patients or dollars |
| One homogeneous source per indicator, the same for all 8 CLSC; OpenStreetMap as a supplement only | Uneven completeness | OpenStreetMap finds 59 of 79 pharmacies, 11 of 15 RPA, no occupational therapy |
| Deduplicate: same type within 100 m | Double counting | Hospitals and CHSLD extracted from M02, pharmacies duplicated in two MSSS layers |
| Cross supply and needs among those aged 60 and over, as rates per 1,000, always with the counts | Age that cannot be crossed, size effect | 211 has no cut-off at age 65 |
| Keep needs at the borough (or FSA) level as long as the postal code is missing | Ecological fallacy | 14.8% of requests without an FSA; cells under 5 suppressed |
| Display the date gap, and redo the analysis with the 2025-2026 211 data | Temporal mismatch | Supply from October 2026, needs from 2024-2025 |
Source: Bernier (2026a). Accessibility is measured finely, by block, on the pedestrian network and the STM schedule, with a 3 km buffer, then aggregated to the CLSC. This limits the modifiable areal unit problem and the edge effect (Apparicio et al., 2017; Tao et al., 2025).
CLSC territories, blocks, dissemination areas and the census are read into R, in a single projection in metres. Each file is verified by its fingerprint.
Each block receives its CLSC and its estimated number of people aged 65 and over (block population × share of those 65 and over in its dissemination area). The borough → CLSC and FSA → CLSC crosswalks are built for 211.
Resources from the six sectors are placed on the map, deduplicated, and sorted by type, sector and domain.
For each block: walking time at 3 km/h on the OpenStreetMap network to each resource (10 min ≈ 500 m, 20 min ≈ 1 km), public transit time (15 and 30 min), capacity-weighted accessibility (E2SFCA; Luo & Qi, 2009; Ngui & Apparicio, 2011), intersectoral response score.
Potential needs index (census), met or unmet needs (211), then the needs × resources × territory matrix and the gap index: supply absent, inaccessible, saturated or ineligible.
R exports a GeoPackage. QGIS produces the maps and an atlas, one page per CLSC. After Objective 1, everything is rerun by needs profile.
The central indicator of Objective 2 is the intersectoral response score. It answers a simple question: from home, within a 10-minute walk, can an older adult find a health resource and social resources? And for how many of the 12 social domains?
08_analyse_spatiale.R, lines 43 to 45, 124 to 137 and 178 to 180, run in Positron.
score_intersectoriel_500m.Here is the script: from the blocks layer, it takes the columns prox_m10_LOG, prox_m10_ALIM, prox_m10_TRANSP and the 9 other social domain columns, plus prox_m10_SANTE. Each equals 1 if a resource in the domain is within a 10-minute walk of the block, 0 otherwise. These columns themselves come from three pieces of data: the location of the resources (inventory of the six sectors), their domain (211 AIRS taxonomy) and the OpenStreetMap pedestrian network.
Here is the map it produces in QGIS: each inhabited block is coloured by score_intersectoriel_500m, from 0 to 12. The darker, the weaker the response. Average for the 43,366 people aged 65 and over in the CCSMTL: 7.4 out of 12 at 500 m, 10.0 out of 12 at 1 km (Bernier, 2026b).
classe_reponse equals 1 for a strong response, 3 for a weak one: this is the class that is then crossed with needs.9 of 12 social domains within 10 minutes · pharmacy at 6 minutes: health = 1
Score = 9 × 1 = 9
11 of 12 social domains within 10 minutes · no health resource: health = 0
Score = 11 × 0 = 0
Within a 10-minute walk, almost all older adults in the CCSMTL have a food resource (93.3%). Only 37.1% have an information and referral resource, and 12.8% a transportation resource. A health resource is within reach for 81.4% of them (Bernier, 2026b).
These differences between domains are exactly what the Objective 1 profiles will help refine: a person who combines isolation, reduced mobility and administrative procedures depends precisely on the least present domains.
The average score at 500 m varies widely between territories: 4.3 out of 12 in Verdun, 10.9 in Centre-Ville. Crossed with the potential needs index, it reveals gap areas: blocks where needs are in the highest third and the response in the lowest third.
In total, 172 blocks are in a gap. 6,913 people aged 65 and over live there, or 15.9%. The share reaches 36.8% in Saint-Paul and 0% in Centre-Ville (Bernier, 2026b).
The matrix crosses, for each CLSC territory and each domain, the population, supply, accessibility, the joint health and social response, and the 211 needs. A written rule, applied in order, gives each cell a final class. Of the 112 current cells, 5 flag a gap: supply absent, inaccessible or saturated.
Needs × resources × territory matrix, final class per cell, 8 CLSC × 14 domains. 211 needs for 2024-2025, distributed from boroughs to CLSC by the share of those aged 60 and over (assumption), except in Verdun where the correspondence is exact. Calculation: Bernier (2026b), script 15.
The atlas brings all of this together in one page per CLSC territory: the resource map, the territory’s indicators and the local response by domain. It is a tool that managers and local coordination tables can read without any software.
The atlas also serves Objective 3: it transparently proposes the contrasting CLSC territories in which to conduct the network analysis, and provides the potential supply to compare with the network actually mobilized.
Today, each map covers one domain at a time. Once the scoping review has established the profiles of co-occurring needs, the same code will be rerun with a different list: instead of “housing,” we will measure, for example, “housing and income and administrative procedures” within walking distance, with a health resource. The variables, sources and steps stay the same. Only the combination of domains changes.
The map and the network complement each other. The map answers “what exists, and where?” The network answers “who works with whom?” The interviews answer “why?” Where the three do not overlap, we find resources that are available but never mobilized. These gaps become the interview questions.
A thesis is feasible when three conditions are met: the data exist and are accessible, the methods are mastered, and the timeline leaves enough time to carry them out. For this project, the first two are already met; the timeline, presented at the end, shows the work done and the work that remains.
Feasibility is secured
The thesis does not rely on any individual-level database from the health network. It works with profiles, drawing on public or aggregated sources, and on original data collection for Objective 3. Each source has a known status.
The literature search is complete: the records have been deduplicated in Covidence. To strengthen the review, the extraction carried out with the librarian will be redone with several reviewers. Each study will be screened by two independent people, first on title and abstract, then on full text, after a pilot test on 25 records; data will be extracted in duplicate (Bernier, 2026).
This approach follows the standards for scoping reviews: the JBI guidance for conduct (Peters et al., 2020) and PRISMA-ScR for reporting (Tricco et al., 2018). Two independent readings limit the risk of wrongly excluding a study.
211 receives requests for help from individuals, their families and caregivers, and practitioners. Its public dashboard already describes, by borough, the needs expressed, the unmet needs and their reasons. For people aged 70 and over in the territory, it records about 3,700 requests from 2021 to 2026, and 691 unmet needs among people aged 60 and over, all with a coded reason (Bernier, 2026).
The agreement with 211 gives access to a finer extraction. Because its data are more complete in the CMM, access is easier and the comparison with the network reported by SAD teams, in O3, becomes more precise.
For Objective 2, the longest part of the work is done. Seven R scripts, run in Positron, import the territories, link each block to its CLSC, inventory the resources, measure access on foot and by public transit, then export the layers to QGIS. Every decision is recorded in a log, and every file read receives a fingerprint: everything can be recomputed identically (Bernier, 2026).
While waiting for the O1 profiles, the examples are produced by need domain. The mapping will be redone by profile, with the same scripts.
The O2 examples already give an order of magnitude: 164 blocks show a gap between potential needs and nearby response, and 15.4 % of people aged 65 and over live there (Bernier, 2026). This figure will be recalculated by profile; it shows that the method works on real data from the territory.
For O2, the project depends on only one external source: the census, which provides the number of older adults per block and their living conditions. The 2026 Census will be released in stages, from February to December 2027: population in February, age in May, households and income in July, general health status in October (Statistique Canada, 2026).
The decision rule is simple. If the O2 analysis can wait until summer 2027, it will use the 2026 Census. Otherwise, it will use the 2021 Census, already in hand, supplemented by the 2026 population projections by CLSC territory (Bernier, 2026). In both cases, the thesis timeline is maintained: the uncertainty affects how current the denominators are, not feasibility.
The third objective is the only one that requires new data collection. It runs over 12 months in two phases: first an ego-centred network analysis, which maps the network, then interviews that explain the observed interactions. This is an explanatory sequential mixed-methods design: the network figures raise the questions, and the interviews answer them (Creswell & Plano Clark, 2017; Fetters et al., 2013).
The question put to the teams starts from a fictional person, built from the O1 profiles: to which organisations have you referred her, or with which have you collaborated, over the past 12 months? For each tie, the questionnaire records the level of collaboration, from information exchange (1) to integration (4) (Bilodeau et al., 2019).
Using a vignette protects privacy: no real person is described, and all teams respond to the same situation.
The project aims for an 80 % response rate, in at least two contrasting CLSCs. The threshold is high because a network is sensitive to missing data: an organisation that does not respond makes its own ties disappear, and the network looks sparser than it is. To reach it, several respondents are approached per team and follow-ups are made by sector (Bernier, 2026). The 80 % rate is a planning target; recruitment and non-response will be monitored.
With a single CLSC, it is impossible to know whether a dense network reflects the territory (a more advantaged population, many resources nearby) or the way the team is organised. By comparing territories that differ in deprivation, resource density and urbanity, we can distinguish what stems from the territory from what stems from the organisation. This is the logic of diverse cases: selecting cases that span the variation, so that the comparison is instructive and the results transferable (Seawright & Gerring, 2008; Yin, 2018).
The CIUSSS’s 8 CLSC territories are highly unequal. The share of people aged 65 and over with low income nearly doubles from one end to the other, from 21.3 % in Verdun to 41.8 % in Pointe-Saint-Charles. Centre-Sud has the most renters (74.2 %) and the highest core housing need (11.7 %); Verdun has the largest number of older adults (12,535) (Statistique Canada, 2022; Bernier, 2026).
The current proposal retains Verdun, Saint-Henri, Saint-Louis-du-Parc and Centre-Sud, two per local service network. Pointe-Saint-Charles is excluded because its SAD falls under a community clinic. The final choice will be revisited with the O2 atlas (Bernier, 2026).
For the network analysis, discussions on a partnership with the CIUSSS du Centre-Sud are under way, because the institution has stakes in the subject. The primary care policy also makes the CLSC territory the reference unit and notes that the supply “often remains little known” (Ministère de la Santé et des Services sociaux [MSSS], 2026, p. 25).
The network analysis with Centre-Sud adds what the data alone cannot tell us: contextualisation, and an understanding of the mechanisms behind possible responses and of why they are currently organised as they are. Combined with the 211 data, which are more complete in the CMM, it allows a precise comparison of the reported network, the observed network and the mapped supply.
Field actors are involved in every phase, and not only as respondents. The network analysis gives them a voice: it is the teams and their partners who describe the ties that exist and those that are missing. The results are presented to them for validation before being published.
This list is not closed: through its inductive approach, the network analysis may reveal other actors and sectors, including private ones, that no one had anticipated.
The choice of the O3 CLSCs will be presented to the SAD teams before data collection: a session with the teams of the 7 CLSCs of the CIUSSS, then a wave with organisations and interviews in the selected CLSCs (Bernier, 2026). The O1 vignettes, the O2 atlas and the O3 network map all serve as discussion tools that everyone can understand.
Before any data collection, the project will obtain approval from a research ethics board and will specify consent, data protection and dissemination (Conseil de recherches en sciences humaines et al., 2022).
The doctorate is planned over 16 quarters, from April 2025 to March 2029. After a year and a half, the courses and the preparation of the protocol are wrapping up, the O1 literature search is complete and the first O2 maps have been produced. The light part of each bar shows what remains to be done; the solid part, what is done.
The sequence follows the design: O1 provides the vignettes, O2 selects the CLSCs, then O3 runs for 12 months, from the third quarter of 2027 to the second quarter of 2028. This single data collection aligns with component 2 of the RUISSS, planned for 2027-2028 (Bernier, 2026). The 2026 Census does not lengthen the timeline: if its data arrive too late for O2, the 2021 Census takes over.
In summary, feasibility is secured because the data are accessible: the agreement with 211 has been obtained and the rest is online. For O2, the data, the coding and the mastery of the software are already in hand; the project depends only on the census, and would use the 2021 Census if necessary. The review extraction, carried out with a librarian, will be redone with several reviewers. For the network analysis, discussions on a partnership with Centre-Sud are under way: they add contextualisation and an understanding of the mechanisms behind possible responses and of why they are currently organised as they are.
This thesis does not aim for one more article on older adults’ needs. It aims for three tools that a home care support (SAD) team, a community organization, a municipality or a management team can open and use: a typology of need profiles, an atlas of resources by CLSC territory, and a map of the networks that can respond. This chapter explains what each objective produces, who it serves, and what the method makes it possible to estimate for the whole of Quebec.
Status of the figures. The mapping results (O2) are real: they come from the R scripts and QGIS maps produced on October 17, 2026 (Bernier, 2026a). The profiles (O1) and the networks (O3) are still to come: the examples that illustrate them are fictional. The impacts are anticipated; they are not presented as proof that an intervention works. Each estimate gives its calculation and its source.
An older adult with one or more morbidities and several social needs does not always enter through the CLSC. She arrives through the emergency room after a fall, through her pharmacy, through a meals-on-wheels service, through 211 or through her exhausted daughter. The primary care policy names these entry points and asks that none be the wrong one (MSSS, 2026b, p. 27). It also notes that the supply “often remains little known” (p. 25).
A response can therefore come from anywhere: the public network, a community organization, housing, transportation, a respite or day-centre service, a school or a childcare service that frees up a family caregiver. The thesis’s impacts target all these actors, and not only health.
Each objective delivers a final product, and each product feeds the next. Needs (O1) tell us which resources to look for (O2); the atlas (O2) tells us where they are and chooses the territories to compare; the networks (O3) tell us who already works together and which link is missing.
The typology draws on a scoping review conducted according to JBI and reported according to PRISMA-ScR. On October 17, 2026, four databases yielded 60,957 records (MEDLINE 16,942, Embase 21,401, APA PsycInfo 4,820, PubMed 17,794). After the 2006 cut-off and the removal of 25,412 duplicates, 26,609 studies remain to be screened by two independent reviewers (Bernier, 2026b).
Direct impact: a synthesis of profiles already described elsewhere, such as those of met and unmet needs in dementia (Janssen et al., 2020) or of social needs that cluster together (Kreuter et al., 2021), applied to older adults receiving care at home.
Madame Protocole, 72, lives alone on the third floor in Verdun. She has diabetes and mild memory problems; stairs, grocery shopping, appointments, loneliness and paperwork are difficult for her; her daughter is exhausted. This character is fictional.
The same vignette is put to all teams and all organizations. It serves two purposes: identifying the resources to map (O2) and asking “to whom do you refer this person?” (O3). It can then be used to train new practitioners and in intersectoral case discussions.
The inventory counts more than 7,000 sites located in the eight CLSC territories, of 35 types and six sectors: public network, primary care, parapublic, community, municipal, housing (Bernier, 2026a). Each territory has its page: map, indicators, health and social response by domain, and from 375 to 894 potential partners.
The gap index crosses the needs expressed to 211 with the supply. For 2024-2025, people aged 60 and over in the four boroughs that cover the CCSMTL expressed 3,954 needs to 211, of which 140 went unmet (3.5 %). Only 3 of the 112 CLSC × domain cells have enough signal to draw a conclusion: absent supply in material assistance in Saint-Paul and in transportation in Pointe-Saint-Charles, saturated supply in personal and family support in Verdun. Hence the request for a detailed extraction from 211.
For the same profile, the network changes from one territory to another. In the fictional example, the Verdun SAD team mobilizes a meals-on-wheels service, a volunteer action centre and a pharmacy, but no housing resource; in Centre-Sud, the missing link is family caregiving.
Four indicators per profile and per CLSC make this measurable: coverage of domains by a partner at level 2 or higher, the share of level 3 or 4 ties, resources that are available but not mobilized (compared with the O2 atlas) and structural holes. An effective partnership should reach level 3 or 4 (RUISSS de l’UdeM, 2022, pp. 16-17; Bilodeau et al., 2019).
Even before the profiles, the CCSMTL atlas provides operational benchmarks. They concern the 43,366 people aged 65 and over counted in 2021 in the 1,532 inhabited blocks of the territory (Statistique Canada, 2022; Bernier, 2026a).
In the heart of Montréal, food, personal support and recreation are almost everywhere within a ten-minute walk. Transportation is within walking distance for only 12.8 % of people aged 65 and over; information and referral, for 37.1 %.
Estimate. 43,366 × (1 − 0.128) ≈ 37,815 people aged 65 and over have no transportation resource within a 10-minute walk, and 43,366 × (1 − 0.371) ≈ 27,277 have no information and referral service at that distance (Bernier, 2026a).
Impact. For a person who can no longer take the bus alone, accompanied transportation is the link on which all the others depend: grocery shopping, appointments, day centre. The atlas shows in which blocks to provide it first.
A block is a priority when three conditions combine: a high density of people aged 65 and over, a weak response within 500 m and a deprived neighbourhood (quintile 4 or 5 of the INSPQ). In the CCSMTL, 150 of 1,532 blocks qualify; 7,958 people aged 65 and over live there, or 18.3 % (Bernier, 2026a).
The distribution is very uneven: 35.8 % of older adults in Saint-Paul live in a priority block, compared with 0 % in Centre-Ville. In Verdun, 61.6 % of people aged 65 and over have a weak response within 500 m, that is, at most 6 of 12 social domains (12,455 × 0.616 ≈ 7,672 people).
Impact. A management team with a budget for a new neighbourhood service knows where it reaches the most vulnerable older adults per dollar invested.
The bivariate map crosses an index of potential needs drawn from the census (low income, living alone, advanced age, language, housing) with the weakness of the response within 500 m. The 172 blocks where both fall in the unfavourable tercile are home to 6,913 people aged 65 and over, or 15.9 % of those in the CCSMTL (Bernier, 2026a).
These terciles are internal to the territory: they place blocks relative to one another. They do not say that a block is poorly served in absolute terms; they say where to look first.
Impact. A neighbourhood round table or a community organizer can bring this map to a consultation and start from a common basis, rather than from impressions.
The audiences below cover all the sectors that the primary care policy places around the CLSC territory: public network, clinics and pharmacies, community organizations, municipality, CPE and schools (MSSS, 2026b, p. 42). For each: a precise operational impact, a quantified benchmark where one exists, and the product that supports it.
Since March 2026, the CLSC territory has been “the reference unit” of primary care (MSSS, 2026b, p. 15). The CLSC plays four roles there: navigator, coordinator, intersectoral leader and co-facilitator (p. 42). It must respond to all the needs of its population, “whether or not they are expressed” (MSSS, 2026c, p. 19).
These roles require knowing, territory by territory, which needs accumulate, which resources exist and who already works together. This is exactly what the thesis’s three products deliver, at the CLSC scale.
A response that starts from their combination of needs, and not from one service at a time. The profile says which domains to look at together; the atlas says what exists within walking distance of their homes.
In the CCSMTL, 87.2 % of people aged 65 and over have no transportation resource within a 10-minute walk: accompanied transportation becomes an explicit priority of the intervention plan (Bernier, 2026a).
Relieving the caregiver is part of the response. The “strained caregiving” profile makes exhaustion visible; the atlas identifies blocks with no respite or day centre nearby; the network shows whether caregiving is a link to be created.
Quebec plans 15 million additional hours of respite by 2031, “distributed equitably across Quebec”, and aims for 38 % of caregivers to know about respite services (Gouvernement du Québec, 2026, pp. 1-2). The atlas says where to place them and whom to inform. Day centres reduce the burden on caregivers of people with dementia; the evidence is thinner for other forms of respite (Vandepitte et al., 2016).
Knowing whom to refer a given profile to in a given CLSC. For each domain, the atlas lists the potential partners, those who come to the home and those who receive people on site.
From 375 potential partners in Saint-Paul to 894 in Centre-Sud (Bernier, 2026a). Social work currently coordinates case by case (Couturier et al., 2022); the list by profile spares it from starting from scratch with each situation.
Spotting a social profile behind a consultation, a prescription or a discharge after a fall, and knowing whom to refer the person to in her territory.
60.0 % of CCSMTL residents aged 65 and over have a pharmacy within a 10-minute walk, compared with 24.0 % for a GMF (Bernier, 2026a). The pharmacy is therefore a neighbourhood entry point for screening, as the Policy provides (MSSS, 2026b, p. 27).
Making their contribution visible, which remains impossible to quantify today (CSBE, 2023a, p. 19), and documenting when a service is saturated rather than absent. Organizations have limited capacity to absorb referrals from the network (Hogg-Graham et al., 2021).
The gap index already classifies personal and family support in Verdun as “saturated supply”: 113 services exist, but needs remain unmet (Bernier, 2026a). It is an argument for funding, not a finding of absence.
Locating vulnerable older adults in relation to housing, transportation, air-conditioned sites and community sites. Age-friendly municipalities (MADA) are active partners in these efforts (Garon et al., 2014).
41.6 % of CCSMTL residents aged 65 and over live within a 10-minute walk of a seniors’ HLM and 42.3 % of a municipal cultural or community site (Bernier, 2026a): ready-made places for an activity, a service point or the departure point of accompanied transportation.
Two roles. Relieving the family caregiver who has young children, and supporting young caregivers in their school path (Gouvernement du Québec, 2026, p. 2). Hosting intergenerational activities, which are promising against the isolation of older adults even though the evidence remains to be consolidated (Peters et al., 2021).
The CCSMTL inventory counts 173 CPEs and daycares and 114 schools (Bernier, 2026a). Crossed with the priority blocks, they indicate where an intergenerational pairing or a daycare place would free up a caregiver.
Prioritizing a territorial budget, choosing where to open a service and which corridor to formalize. Carrying out the four roles that the Policy assigns to the CLSC with benchmarks specific to its territory.
Saint-Paul and Verdun together account for 5,142 of the 7,958 older adults in CCSMTL priority blocks (1,806 + 3,336), or 65 % (Bernier, 2026a). A population-proportional allocation would underfund them.
Having comparable indicators for the 166 CLSC territories, updated when the sources change, in order to monitor the primary care policy, Mieux chez soi and the rollout of the AIPPA.
The 17,883 people waiting for SAD (MSSS, 2026a, p. 6) represent on average 108 people per territory (17,883 ÷ 166); the atlas says where, within each territory, the non-public response is thinnest.
| Audience | Typology and vignettes (O1) | Atlas and gap index (O2) | Networks and links (O3) |
|---|---|---|---|
| Older adults and family caregivers | recognize themselves in a profile, name their needs | resources within walking distance | know who will do what, and in what order |
| SAD teams, social work | screening in a few questions | partners by domain and by CLSC | corridors to create first |
| GMFs, pharmacies, hospital | screening for a social profile | patient’s neighbourhood supply | whom to refer to at discharge |
| Community organizations | common language with the network | saturated or absent supply, evidence for funding | actual place in the network |
| Municipalities, housing, transportation | situations experienced by older adults | priority blocks, available sites | role in the pathways |
| Schools, childcare services | caregiving parent, young caregiver | blocks for a pairing | intergenerational link |
| Managers, MSSS, Santé Québec | needs to plan for | budget priorities by territory | transferable response models |
Each page of the analytical atlas brings together, for a CLSC territory, the neighbourhood response block by block, the priority areas and the clusters of weak response. A team finds its working map there; an intersectoral committee, the agenda for its next meeting.
The Verdun example shows a very populous territory (12,455 people aged 65 and over in 2021) where the response within 500 m is weak for 61.6 % of them (Bernier, 2026a).
The method is not specific to the CCSMTL. The same R code has already produced nine maps and about forty indicators for the 166 CLSC territories of Quebec, grouped into 97 local networks and 18 regions, where 1,908,944 people aged 65 and over lived in 2024 (Bernier, 2026a; MSSS & ISQ, 2025). The CCSMTL accounts for only 2.4 % of them (46,509 ÷ 1,908,944). The figures below are real results of this provincial calculation, except for the last row of the table, which is a transposition flagged as such.
Weighting each territory by its population aged 65 and over, 8.9 % of Quebec’s older adults live more than 10 km, as the crow flies, from the nearest CLSC facility; 100,147 live more than 25 km from an emergency room. In rural areas, the share rises to 27.7 % (Bernier, 2026a).
Impact. Where the CLSC is far away, the response depends more on the pharmacy, the local organization, the municipality and family and friends. The response model cannot be the same as in Montréal; the atlas says in which territories to adapt it.
The co-availability index measures the share of people aged 65 and over who have, within the same radius (5 km in cities, 25 km in rural areas), a GMF or a CLSC, a private seniors’ residence or a CHSLD, and a pharmacy, all three. In 33 CLSC territories, fewer than 80 % of older adults have all three; 261,732 people aged 65 and over live there (Bernier, 2026a).
Spatial analysis (LISA) brings out 12 territories in a “low-low” cluster, where 46,071 older adults live in an underserved territory surrounded by underserved territories. They are located in Saguenay–Lac-Saint-Jean, Abitibi-Témiscamingue, the Côte-Nord, Nord-du-Québec, Nunavik and the Terres-Cries-de-la-Baie-James.
Impact. A cluster cannot be fixed territory by territory: it calls for a regional response, for example a mobile service or a corridor between neighbouring CLSCs.
Everywhere in Quebec, the method uses public sources: MSSS territories, the census, MSSS-ISQ projections, and directories of facilities, GMFs, pharmacies and RPAs. The provincial inventory already counts 510 CLSC facilities, 604 GMFs, 1,883 pharmacies and 1,594 RPAs located in the 166 territories (Bernier, 2026a).
Two limits are known. 211 covers at least 80 % of older adults in only 64 territories; elsewhere, community supply must come from other regional directories. And respite and day-centre services do not yet form a distinct layer: they must be added from the lists of institutions and caregiver organizations.
| Estimate | Value | Calculation | Source |
|---|---|---|---|
| Territories where the method applies as is | 166 | all CLSC territories; complete indicators verified by the script | Bernier, 2026a |
| Older adults concerned | 1,908,944 | aged 65 and over, 2024 | MSSS & ISQ, 2025 |
| Older adults more than 10 km from a CLSC | 170,183 | sum, over the 166 territories, of the share more than 10 km away × people aged 65 and over | Bernier, 2026a |
| Older adults in a territory where fewer than 80 % have health, housing and pharmacy together | 261,732 | sum of people aged 65 and over in the 33 territories concerned | Bernier, 2026a |
| Older adults in quintile 5 of material or social deprivation | 666,314 | weighted share of 34.9 % × 1,908,944 | Bernier, 2026a; INSPQ, n.d. |
| SAD users per territory, on average | ≈ 2,384 | 395,783 ÷ 166 (simple average; territories differ greatly) | MSSS, 2026a, p. 6 |
| Additional respite hours to distribute per territory, on average, by 2031 | ≈ 90,361 h | 15,000,000 ÷ 166; the atlas serves to replace this average with a distribution based on needs | Gouvernement du Québec, 2026, p. 2 |
| SAD hours required but not provided by the public sector, to be covered by caregivers, community organizations or the private sector | 209.3 M h | 234.7 M − 25.4 M (2023) | CSBE, 2024, p. 11 |
| Transposition, not measured: older adults in metropolitan territories who would have a weak response within 500 m if the CCSMTL share applied there | ≈ 372,000 | 28.6 % × 1,299,846; order of magnitude only, to be replaced by the actual calculation once community directories from other regions are integrated | Bernier, 2026a |
Read these estimates with caution. Distances are straight-line at the provincial level and along the street network in the CCSMTL. A nearby resource is not necessarily accessible (eligibility criteria, waiting lists, language). The 211 data describe requests, not people. These limits are documented in the reproducible report, and each calculation can be redone identically (Bernier, 2026a).
A tool is useful if it answers the questions of those who use it. Field actors and patient partners are therefore involved at every step, not only at the end. The approach draws on the local intersectoral action described in Quebec, where effects depend on the quality of the partnership process (Bilodeau et al., 2019).
In O3, three networks are compared for each CLSC studied: the network reported by the SAD teams, the network observed in 211 referrals and the potential supply from the atlas. A resource present in the atlas but absent from both networks is available and not mobilized.
These gaps become the questions for the interviews with the actors: why is this partner not called, and what would it take for it to be? The answer belongs to the field; the thesis provides the data to ask the question.
In summary. The thesis delivers three tools that answer three field questions: which needs go together, where are the resources that respond to them, and who needs to work with whom. They serve older adults and their families, all the organizations that can respond, and those who decide where to invest, first in the 8 CCSMTL territories and then, by the same method, in the 166 CLSC territories of Quebec.
The trajectory models, maps and interview explanations are presented to home care teams, to the AIPPA intersectoral liaison committee and to component 2 of the RUISSS (2027-2028).
Define with the settings the coordination roles, the partners concerned, the corridors to create, the resources and the obstacles. The proposals must be examined for their relevance, acceptability and local feasibility.
An atlas for each CLSC, summaries by profile, feedback sessions and scientific communications will support the sharing of results. If implementation is envisaged, its procedures, indicators and follow-up conditions will be addressed through a separate process.
One or more patient partners, older adults or family caregivers with experience of home care, will be recruited to take part in the project: drafting the vignettes, reading the maps, interpreting the networks and sharing the results.
See the project posterAccessible version: 5 languages, easy-to-read French, colour-blind friendly, low vision, audioCo-holder, GSK Research Chair in optimal management of chronic diseases
CRCHUM profileCo-holder, GSK Research Chair in optimal management of chronic diseases
CRCHUM profile01.2 · Research areas
Six research areas connecting questions with the work and experience that inform them. Each card opens a specific part of the site.
Understand how knowledge, policy instruments and public debate shape health decisions.
Study the intersectoral response to profiles of co-occurring social needs among older adults receiving home care.
Examine support for ageing and the conditions that make choice of living arrangements effective.
Analyse how legal norms and planning organise provision and make rights effective.
Connect service transformation, evaluation and practical implementation conditions within organisations.
Compare system configurations, mechanisms and people’s experiences across settings.
01.3 · Publications
Scientific presentations, defended theses and manuscripts in progress are presented here with their original titles, authors and respective statuses.
Manuscripts, posters, presentations and academic work: four clear ways to explore research.
Submitted papers and manuscripts being prepared alongside the thesis.
Health systems and primary care
Submitted to BJGP OpenÉric Bernier (LL.M, PhD student), Nadia Sourial (PhD), Isabelle Vedel (PhD), Géraldine Layani (MD, M.Sc), Aurélien Algoet (MD), Matthieu Calafiore (MD, PhD)
Epistemology and knowledge translation
Manuscript drafted · Work suspendedBernier, É.
Research impact is usually pictured as a forward pathway, from knowledge to policy and practice. The article develops public health as a reflexive science of governance and specifies how consequences in practice and the situated knowledge of affected populations can acquire standing to require reconsideration of upstream categories, indicators and rules. The concept is translated into a documentary audit, tested on four institutional chains.
Governance and the public sphere
Manuscript in preparationBernier, É., Aïssi, F., Diene Faye, C., Boukari Abdo, S.
Health law
Manuscript in preparationBernier, É. · legal scholarship article
Bernier, É., Sourial, N., Layani, G. · Groupe francophone en soins primaires (GFSP), Montréal · poster
Bernier, É., Sourial, N., Layani, G. · CAHSPR, Ottawa
Bernier, É., Sourial, N., Layani, G. · Quebec Network for Research on Aging (RQRV), Sherbrooke · poster and 180-second talk
No oral presentations are listed at present.
Master of Laws · Université Paris Panthéon-Assas · supervisor: Benoît Apollis
Master’s in health care facility management · Université de Bretagne Occidentale · supervisor: Michelle-Hentic Giliberto
State diploma in special needs education · Académie de Rennes
The doctoral project is described separately under “My PhD”. View my PhD
01.4 · Research stance
Situated, critical and reflexive research
My research stance concerns how problems are framed, analytical categories are constructed and knowledge is produced that remains open to discussion and revision. It brings together a relational understanding of health situations, the examination of institutions and researcher reflexivity.
I understand people as singular beings embedded in relationships, histories and the social, material and living environments on which they depend. Each person may have their own aspirations; opportunities to fulfil them also arise through living conditions, available resources and relationships. This tension between individuality and interdependence shapes my understanding of health.
My place as a researcher is within that same world. My experience with people with disabilities, law, management and public health shapes the problems I notice. I treat this position as a starting point to examine: making my choices visible, welcoming disagreement and allowing inquiry to change my categories.
Reflexivity involves examining how the research question, selected categories, included actors and conditions of inquiry shape the knowledge produced. I make these choices, competing interpretations and the limits of conclusions explicit so that the reasoning can be discussed.
I distinguish what inquiry can observe, the interpretations proposed and the purposes I defend. Robust knowledge requires examinable evidence, explicit reasoning and opportunities to discuss disagreements. Concepts and measures are constructed to study realities whose consequences affect lives.
The question decides the method: integrated mixed methods.
A stance that combines a positivist reading (measuring, estimating, testing hypotheses) with a constructivist one (understanding the meaning actors give to their practices). The question determines the method.
Public health, law, management, political science and social intervention: perspectives that inform one another to analyse the same system.
Dewey’s pragmatism connects inquiry with encountered problems and their consequences.
I favour precise questions, knowledge open to challenge and revisable conclusions. A finding’s usefulness also concerns what it helps us understand and its effects in a situation.
Foucault examines processes through which people become objects of knowledge and subjects.
I question how I define a need, population or complex situation. My categories must be confronted with the experiences they describe and what they render less visible.
Fricker examines injustices affecting the credibility of speech and the resources available to understand experience.
I take seriously the conditions enabling people to tell their story, be heard and contribute to interpretation. Participation involves influence over questions and conclusions.
Institutions enable protection, cooperation and access to resources. Their rules and categories also distribute opportunities for action. My stance is to examine these effects, their justifications and opportunities to challenge them in concrete settings.
Power operates through people’s possibilities for action; it extends beyond orders imposed from above.
In services, I question access rules, classifications and professional roles: what do they enable, who can challenge them and what room remains for self-determination?
Fassin connects the proclaimed value of life with the unequal treatment of lives.
I examine how apparently ordinary decisions allocate attention, protection, time and resources. Average improvements must also be considered in relation to those they reach and those they leave at a distance.
Inequality can affect the possibility of being recognised as a person who knows.
I examine patients’, relatives’ and community actors’ roles in defining problems. Working with people requires opportunities to reshape the discussion and express disagreement.
Autonomy takes meaning in concrete situations, through supportive relationships and genuinely accessible possibilities. This orientation extends my support experience: recognising aspirations, discussing constraints and improving the conditions that make choices practicable.
For Hegel, self-consciousness develops through relations with others, involving recognition and conflict.
I treat tensions between protection and autonomy, common rules and singular situations as research questions. Understanding an organisation requires following its history, relationships and transformations.
Sartrean existentialism connects responsibility with the choices through which an existence takes shape.
I draw from this a requirement to explain my choices. In public health research, I confront it with social constraints and available resources, understanding situated responsibility rather than attributing difficulties solely to individuals.
The capability approach examines real freedoms to lead lives that people have reason to value.
I examine what people can actually do: access services, understand decisions, participate and pursue their aspirations. Available resources and their conditions of use help analyse gaps between formal rights and real opportunities.
My research seeks knowledge that can enter collective deliberation. Recommendations and innovations involve societal choices: who benefits, what burdens a change creates and how those affected can discuss its effects.
Camus examines revolt affirming common dignity while questioning the limits of action and its justifications.
I wish to combine resistance to injustice with attention to those affected by change. A desirable purpose also requires examining the means employed and their consequences.
Dewey connects the formation of a public with shared consequences of activities and their collective handling.
I draw attention to affected actors, deliberation and feedback from practice to knowledge. Findings must be open to discussion, contextualisation and revision when their consequences require it.
Ricœur connects the aspiration to a fulfilled life, attention to others and institutional justice.
I relate the quality of support relationships to the organisational and political conditions enabling them. My responsibility concerns both the person encountered and the collective frameworks my work helps illuminate.
These works open questions and inform connections. Their authors defend different positions: I bring them into dialogue to clarify my own responsibility, without presenting a unified doctrine. First-person statements express this research orientation.
01.5 · Research methods
From question to observation: each question calls for its own data, analytical method and interpretation. I choose methods according to what needs to be understood, drawing on several disciplines.
Designing a study, collecting data and choosing an analytical strategy are decisions linked to the research question. The approaches below outline their possibilities, requirements and limitations.
Integrated quantitative and qualitative analyses: network analysis, regression, thematic analysis, document analysis and mapping, brought together in joint displays.
Data analysis
Collaboration and facilitation
Office tools
Social network analysis describes the relationships between actors rather than their attributes alone. A sociocentric approach examines all ties within a bounded population; an egocentric approach starts from a respondent (ego) and the contacts this person names (alters). In my thesis, the network is ego-centred on the home care (SAD) team of each CLSC (local community service centre). For each vignette, the team names the organisations to which it has referred an older adult, or with which it has collaborated, over the past 12 months. It rates the level of collaboration, from 1 (information exchange or referral) to 4 (integration). This declared network is compared with 211 referrals and with the resource atlas, then explained in interviews.
Mapping and geographic information systems (GIS) link data to their location in order to describe how services, needs and populations are distributed across a territory. Measures of potential accessibility, such as the two-step floating catchment area (2SFCA) method, relate service supply, demand and the distance between them. I acquired the foundations of this approach through a MOOC from the University of Geneva. In my thesis, mapping is objective O2: an atlas of resources for each CLSC territory, E2SFCA accessibility on foot and by public transit, and an index of the gap between needs and supply.
Reproducible programming means carrying out every step of an analysis, from importing data to producing figures, in scripts that can be rerun and checked. I code in R and Python in Positron, Posit’s development environment, and chain processing steps on the command line in the shell. In R, the tidyverse structures data preparation; igraph and statnet build networks and compute centrality, density and cohesion; sf handles spatial data. Python complements this pipeline for data preparation and automation. Everything links network data, territorial data and figures within a single documented project.
Thematic analysis identifies, organizes and interprets patterns of meaning in a corpus of interviews or documents, through an iterative process that moves from initial coding to the definition of themes. I carry out this coding in NVivo, which documents the coding tree, memos and successive groupings. In my thesis, explanatory interviews are centred on the network map of each home care team. The Framework Method combines deductive and inductive codes in a participants × themes matrix to compare sectors, network positions and profiles. Double coding of 20% of the material, a reflexive journal and feedback to participants are planned before the quantitative and qualitative strands are integrated.
I conducted a qualitative analysis of Quebec physicians’ critical discourse during the debates on Bill 2. The purposive corpus comprises 42 press documents and press releases (Eureka database, May to December 2025), coded chronologically until the code structure stabilised. The analysis combines Fairclough’s three-dimensional model (the text, practices of production and circulation, the sociopolitical context), framing theory and the attribution of narrative roles: victims, villains and heroes. Every quotation is traced back to its source, and the analysis separates what the texts say from their sociological interpretation.
A case study investigates a situated phenomenon in depth: an organisation, territory, policy or initiative. Designs may involve one case, multiple cases or embedded units; their strength lies in connecting the phenomenon with its context.
Interviews, particularly semi-structured ones, explore the meaning of experiences, decisions and practices. An open guide organises topics while allowing follow-up questions, unexpected issues and contradictions.
A questionnaire survey translates a research question into comparable measures. It connects the target population, sampling strategy, question selection and administration mode, whether online, by telephone or in person.
Secondary analysis reuses existing surveys, cohorts, administrative or institutional data to address a defined question. It requires understanding why, from whom and how the data were originally produced.
Causal inference concerns the effect of an intervention or exposure relative to a defined alternative. It distinguishes causal questions from observed associations and connects population, intervention, comparison, timing and outcome.
Comparison connects governance, financing, resources, service organisation and outcomes. Carsten Anckar’s work clarifies two case-selection approaches: most similar systems design (MSSD) and most different systems design (MDSD).
01.6 · Documentary synthesis · 14 October 2026
This section brings together assisted programming, tool development and reflection on research, learning and public health. It draws on scientific work on the benefits, limitations and social transformations of these technologies.
A conversational assistant proposes an answer; an agent can chain file reading, modifications and tool execution. This changes work organisation: permitted actions must be specified, a record retained and the resulting solution assessed.
Define users, necessary data, expected features and accessibility constraints. Breaking changes into understandable units makes it possible to discuss a solution and resume work without losing its logic.
A tool’s consistency depends on its files, dependencies, data structures and previous decisions. Locally plausible code may not fit the rest of the project: the assistant needs this context.
Read an error, reproduce the problem, explain its cause and correct the relevant part. Maintainability matters alongside immediate functionality: explicit names, documentation, versions and the ability to undo a change.
Match access rights, actions and decision points to the stakes. An agent acting on files, data or services requires more supervision than a text suggestion.
Examine units, missing values, transformations and assumptions. Known examples, targeted checks and reproducibility help assess both program behaviour and the scientific meaning of its outputs.
Productivity depends on the task, experience and tool. METR’s 2025 study and 2026 methodological update encourage distinguishing perceived speed, quality and measured gains.
Generated line counts are a poor indicator. My interest is in useful, understandable, accessible tools that can evolve.
Practical workshop
Generative AI assistants can support programming, tool development and iterative revision. I distinguish executable code, a valid statistical analysis and an answer to the right research question. Specify the estimand, population, timing and missing-data assumptions before interpreting a result. Human responsibility includes confidentiality, equity and the ability to challenge an output.
Download the worksheet ↓Purpose, permitted data and expected behaviour.
Small, understandable changes.
Inspect code, calculations and assumptions separately.
Document contributions and human decisions.
Fictional situation · proposed approach
A tool must explain a research finding to managers and the public. The level of delegation depends on whether the task concerns code, analysis or learning.
Specify inputs, outputs and edge cases. Review the changes and use synthetic data when appropriate.
A specification, version history and an independently calculated reference example.
Write the question and assumptions first. Request alternatives and distinguish planned analysis from exploration.
An analysis log covering population, variables, estimand, missing data and sensitivity analyses.
Start personally, discuss with a peer, use AI as a second critique and justify accepted and rejected suggestions.
Successive versions, a reflective note and an appropriate transfer task without assistance.
Experimental results, conceptual arguments and governance guidance answer different questions. They inform decisions without supporting universal effectiveness claims. Documentary cut-off: 15 October 2026.
Peer and AI Review + Reflection (PAIRR) brings together peer feedback, AI feedback and reflection. Sperber and colleagues’ model (2025) structures writing through critical comparison of feedback and learner responsibility.
Develop an initial analysis and clarify its objectives before seeking feedback.
Give and receive criteria-based peer feedback; discuss understanding, arguments and areas for improvement.
Request feedback on the same draft and rubric in an authorised environment; check the relevance of suggestions.
Compare human and AI feedback. Justify advice accepted, questioned or rejected, and identify what the comparison taught you.
Revise the work while retaining a record of decisions and limitations. The learner remains responsible for the final choices and their justification.
Proposed public health activity: a short text on coordination around a CLSC, peer and AI feedback, followed by a reflective note. Minnillo and colleagues’ 2026 preprint explores critical AI literacy and integrity in writing courses; it does not validate the approach in public health.
AI now enters question formulation, literature exploration, programming, analysis and dissemination. The scientific challenge is to connect these capabilities with explicit methods, plural forms of knowledge and responsibility that remains human.
In my thesis, the relationship between older adults’ need profiles, local resources and intersectoral networks guides methodological choices. AI can help explore a formulation or prepare code; it does not determine the research object, network boundary or validity of an indicator.
For planned use with qualitative materials, distinguish suggested categories, documented coding and interpretation. Excerpts, counterexamples and memos must remain examinable; any data transfer must respect authorisations and confidentiality.
Hao and colleagues (2026) associate AI adoption with individual advantages and a collective narrowing of research topics. This observational analysis raises questions about diversity without establishing universal causation.
AlphaFold 3 predicts structures of biomolecular complexes. This advance illustrates a different use from conversational assistance: scientific contribution depends on a defined task and appropriate evaluation, followed by comparison with experimental knowledge.
Read cited publications, check references and distinguish facts from proposals. ICMJE recommendations retain human responsibility and require authors to describe AI use in submitted work.
Retain scripts, versions, parameters and decisions. Test code on controlled examples; discuss statistical choices and interpretations beyond whether a script runs.
AI cannot fulfil an author’s responsibilities. Manuscript confidentiality, copyright and transparency also concern peer review and editorial processes.
Messeri and Crockett’s perspective (2024) examines illusions of understanding and uniformity in scientific questions. My concern is to explain an analysis and preserve alternative approaches.
In my thesis, AI is a subject for reflection; this section does not replace the protocol or methodological validation of the network analysis ego-centred on CLSC home care teams.
Producing a good answer with an assistant and developing a lasting skill are distinct goals. Teachers organise activities, clarify permitted uses and make reasoning observable.
A university physics trial finds benefits from a pedagogically designed tutor. A secondary-school mathematics trial documents learning risks without safeguards. These settings do not support a universal conclusion.
Combine unassisted attempts, gradual hints, discussion of errors and oral justification. Also assess the ability to transfer reasoning to a new problem.
Tool access and learners’ rights matter. ASPIRE (2026) proposes a framework for public health education; a pedagogical framework must be distinguished from generalisable evidence of effectiveness.
My educational ambition is critical AI literacy: learning to formulate, verify, challenge and decide when assistance is useful.
Work by Pranav Rajpurkar, Eric Topol and other teams demonstrates the breadth of possibilities in medicine. Public health also asks about service organisation, population effects, relationships of care and the distribution of benefits.
Goh and colleagues’ trial (2024) found no significant improvement in physicians’ diagnostic reasoning with the assistant. Model performance alone therefore does not predict the performance of human–AI work.
Goh et al. (2024) · Rajpurkar, Chen, Banerjee & Topol (2022)
Studies by Tu (2025) and Liévin (2026) assess diagnostic dialogue and management in simulated consultations. They open possibilities without demonstrating clinical benefits for patients in routine practice.
Moore and colleagues (2025) document stigma and inappropriate responses in the systems studied. Their findings question using these models as substitutes for therapeutic relationships and professional responsibility.
Obermeyer and colleagues (2019) show that a cost indicator can misrepresent health needs. This lesson extends beyond generative AI: the chosen target and previous access to care shape outcomes.
Generative models require validation suited to the use and human oversight. Generating information alone does not demonstrate improved health.
WHO’s July 2026 report examines inequities, benefit sharing and power relationships in AI-related research, with particular relevance to low- and middle-income countries.
Who defines the problem? Which languages and populations are represented? Who can challenge a recommendation? How can innovation, professional autonomy and participation be connected?
My perspective is to assess innovation also through its effects on social relationships, resource distribution and the well-being of people and other living beings. The suggested readings illuminate these public health goals.
Pickett et al. (2026) · DeSalvo et al. (2017) · Webster & Neal (2024) · WHO (2024)
A public health approach compares expected benefits, risks, costs and alternatives, including a solution without AI.
Bioethics links innovation to human dignity, autonomy, vulnerability, justice and social responsibility. It asks about AI’s purposes as well as its performance.
People should be able to understand uses that affect them and retain choices. Assistance should support dialogue rather than make decisions opaque.
What concrete benefits are sought? Which errors may cause harm? Who can identify a problem and stop or correct the system?
Examine access, biases and the distribution of benefits and burdens. Social responsibility in health extends beyond optimising a tool.
Limit data to authorised uses, choose appropriate environments and identify human accountability. Pseudonymised data can remain sensitive.
My guiding principle is to discuss innovation with the people and institutions concerned, considering their rights, capabilities and collective consequences.
AI is also economic and political infrastructure. Its development concerns work, resources, languages, service access and our collective capacity to choose desirable transformations.
ILO and NASK (2025) examine task exposure to generative AI. Potential exposure does not mean actual job elimination: organisation, training and the distribution of gains remain decisive.
Who controls models, data and infrastructure? I examine institutional dependencies, the ability to change providers and affected people’s participation in adoption decisions.
Data and tool availability can favour particular languages, disciplines and populations. Automatic translation does not replace attention to contexts, lived experience and disagreements.
The assessment should include energy, resources and expected benefits. Connecting AI with collective well-being means asking which uses deserve these resources and how benefits and burdens should be shared.
Equity and agency are my guiding thread: designing and discussing transformations with people, while making decisions open to challenge.
I bring together trials, simulation studies, observational analyses and synthesis reports. They address different questions: technical capability, actual benefit, work transformation or governance. Their status is indicated in the references.
The international report led by Yoshua Bengio synthesises capabilities, risks and safeguards. It encourages distinguishing documented harms, emerging risks and uncertainty, rather than treating all projections as established facts.
The Singapore consensus proposes safety research priorities. This collective preprint contributes to discussion of evaluation, oversight and international cooperation; it does not settle scientific disagreements.
Hao, Xu, Li and Evans extend questions raised by Messeri and Crockett about scientific uniformity. My concern is preserving field experience, criticism and problems that resist standardisation.
METR observed a slowdown among the developers studied in 2025; its 2026 update highlights selection effects. This cannot establish a single effect for every developer and model.
AMIE’s simulated management studies strengthen interest in health agents. The debate now concerns real-world evaluation, accountability, inequalities and organisational conditions of use.
Selection current to 14 October 2026. Personal proposals, empirical evidence and institutional recommendations are distinguished; earlier readings remain together in the bibliography.
A selection of reading on AI, research, learning, bioethics, equity and the well-being of people and the planet.
02.1 · Implementation
Moving knowledge between people, practice settings and research, then examining what it actually helps to change.
03 · Implementation
Moving knowledge between people, practice settings and research, then examining what it actually helps to change.
In my thesis, knowledge mobilisation accompanies the research: it is neither an implementation study nor a policy evaluation.
Completed trainingMobilizeU, knowledge mobilisation · Research Impact Canada and York University, 2026 · Learn more
Five stages
Everything starts with listening: identifying with the communities the needs and decisions to be informed, then specifying who participates, when and with what influence. Transfer plans often neglect fine-grained knowledge of the public (Tchameni Ngamo et al., 2016); integrated knowledge application, on the other hand, involves users of the results from the formulation of the question (Jull et al., 2017). Respect, relevance, reciprocity and responsibility guide this first encounter (Kirkness & ; Barnhardt, 1991).
Research knowledge does not circulate on its own: it meets experiential and contextual knowledge that needs to be put into dialogue rather than opposed (Breton, 2025). The exchange of knowledge is a social process, made of back-and-forth interactions, and not a simple transmission (Ward et al., 2012). Adapting the format, language and content to the intended uses involves clear language summaries (Stoll et al., 2022) and, often, a knowledge broker who makes the link (Lomas, 2007).
Consulting gathers opinions; co-constructing is developing together; deliberating is examining arguments and making decisions. Deliberative dialogue transforms findings into discussed options by the concerned people (Boyko et al., 2012). Yet it is still necessary to document the real influence of contributions and to recognise the costs of co-production for all parties (Oliver et al., 2019), in order to avoid token participation (Gainforth et al., 2021).
A promising practice does not implement itself: proving its effectiveness does not guarantee its adoption (Bauer & ; Kirchner, 2020). One must examine the context, resources and obstacles using frameworks such as CFIR (Damschroder et al., 2022), then choose and describe the support strategies in a shared vocabulary (Powell et al., 2015). In Quebec, the implementation of a citizen support approach for the elderly illuminates the favourable factors, obstacles and essential conditions (Gobeil et al., 2023).
A disseminated tool is a product; its adoption is an implementation outcome; changes for services and people are other outcomes, to be documented separately (Proctor et al., 2011). Rather than attributing these changes to the research alone, we follow its plausible contribution by examining other influences (Mayne, 2012 ; Morton, 2015). French-speaking frameworks help to make these outcomes visible (Gervais et al., 2016).
These five stages draw on the knowledge-to-action model (Graham et al., 2006) and Ward’s framework (2017).
Each area links the scientific literature to my practice.
Mobilising is to circulate knowledge from research into settings, but also from settings and people into research. The knowledge-to-action model distinguishes the creation of knowledge and the cycle of its implementation: identifying a problem, adapting knowledge to the context, identifying barriers, supporting and monitoring its use (Graham et al., 2006). Five questions organise the approach: what to transfer, to whom, by whom, how and with what effect, and one mobilises first sets of evidence rather than an isolated study (Grimshaw et al., 2012). Knowledge brokerage gives a face to this linking work (Munerol et al., 2013).
A framework makes choices explicit: why mobilise, which knowledges, with whom and how (Ward, 2017). Theories, models and frameworks are grouped into families according to what one seeks to do: guide a process, explain what facilitates or hinders, or evaluate (Nilsen, 2015). Some place facilitation at the heart of implementation (Harvey & ; Kitson, 2016); others cover the entire cycle of knowledge application (Esmail et al., 2020). Choosing them according to the question, and discussing rather than ticking them off as a list, makes the approach more fruitful (Kislov et al., 2019).
A mobilisation plan specifies the audiences, objectives, messages, activities, resources, timeline and indicators (IRSC, 2012). It is built with partners from the outset of the project: designing for dissemination and sustainability means thinking early about the fit between findings and their adoption context (Kwan et al., 2022). Real plans describe well the partners and strategies, but often forget evaluation and resources (Tchameni Ngamo et al., 2016). To start: know your audiences, surround yourself, involve people with lived experience and draw lessons from each trial (Diany et al., 2025 ; Dancause, 2023).
Co-production and integrated application of knowledge involve knowledge users at every stage, from question to dissemination (Jull et al., 2017). Partnership research approaches share common principles, strategies and outcomes, which a shared vocabulary allows to describe and compare (Hoekstra et al., 2020). Engagement is conceptualised across three levels, organisation, values and practices (Boaz et al., 2018), and rests on real interdependence and co-creation (Skipper & ; Pepler, 2020). It requires time and resources: its costs must be recognised (Oliver et al., 2019).
An inclusive mobilisation questions the place of each: who is heard, who participates, who benefits (Georgalakis, 2023). Linking knowledge and action for equity requires redistributing power and addressing the structural causes of inequalities (Plamondon et al., 2019). Experiential knowledge is not a complement: it transforms the questions, methods and interpretation (Breton, 2025). In Indigenous health, promising practices respect the ways of sharing knowledge specific to communities (Morton Ninomiya et al., 2022); in the field of disability, co-constructed principles avoid token participation (Gainforth et al., 2021).
The format follows the aim: co-create (deliberative dialogue, world café, workshop), disseminate (exhibition, webinar, panel discussion) or implement (training, project meeting). Accessibility is considered in all its forms, physical, sensory, financial, cultural and linguistic: audiences must feel included, not just invited. The Montreal model recognises the experiential knowledge of patients as complementary to professional knowledge (Pomey et al., 2015), and many frameworks support the involvement of patients and the public in research (Greenhalgh et al., 2019). Clear language summaries (Stoll et al., 2022) and co-built visual art (Courcy et al., 2022) open research to other audiences.
Strategic communication and knowledge application overlap without merging: one supports visibility, the other use (Barwick et al., 2014). The knowledge broker works in between, building links between researchers and decision-makers rather than disseminating products (Lomas, 2007). In public policy, the relevance of data and organisational factors weigh heavily (Oliver et al., 2026), whereas advice given to researchers rarely rests on evidence and silences the costs of engagement (Oliver & ; Cairney, 2019). A brokerage intervention in a regional health agency shows that skills progress more quickly than influence on decisions (Fillol et al., 2026).
Evaluating impact is to follow the contribution of research to changes without attributing them solely to it: productions, scope, appropriation, documented uses, then outcomes for services and people (Morton, 2015). Contribution analysis offers a reasoned approach when strict causal attribution is out of reach (Mayne, 2012). French-speaking frameworks make visible the outcomes of transfer (Gervais et al., 2016), and a Quebec review lists the models used to evaluate mobilisation strategies (Ziam et al., 2024). There remains an open and debated question: to what extent is transfer evaluation possible (Plouffe, 2015)?
Ward’s framework structures the plan around four questions: clarify the what, the why and the who before choosing the how, the formats and the channels.
(Ward, 2017)
The knowledge-to-action model frames the production, tailoring, communication and monitoring of knowledge, without claiming to secure lasting uptake.
(Graham et al., 2006)
Respect, relevance, reciprocity and responsibility guide partnerships with patients and practice settings, drawing on the CIHR patient engagement framework.
(Kirkness & Barnhardt, 1991; CIHR, 2014; Bowen & Graham, 2013)
A dialogue with home care teams and their partners turns trajectory models and maps into contextualised guidance, with feedback to participants. Involvement is reported using GRIPP2.
(Boyko et al., 2012; Lavis et al., 2009; Staniszewska et al., 2017)
Four indicators (output, reach, uptake, documented use) and a log make it possible to assess the plausible contribution of the research without attributing later changes to it.
(Morton, 2015)
Examining the context and the determinants that help or hinder implementation.
(Damschroder et al., 2022)
Distinguishing acceptability, adoption, appropriateness, feasibility, fidelity, cost, penetration and sustainability.
(Proctor et al., 2011)
Impact frameworks mostly map the forward path, from research to practice; my work also examines the way back.
(Greenhalgh et al., 2016; Brownson et al., 2026)
Linking outputs to uses, then to documented changes, while examining other possible influences.
(Morton, 2015; Bayley & Phipps, 2019)
Co-production mechanisms depend on infrastructure, role clarity and power sharing; their costs matter too.
(Graham et al., 2019; Kothari et al., 2025; Oliver et al., 2019)
Theorising rather than applying frameworks as checklists.
(Kislov et al., 2019)
Designing and evaluating public participation processes that leave a record of their influence.
(Abelson et al., 2003; OECD, 2022)
Proving that an innovation is effective is not enough to ensure that it will be adopted.
(Bauer & Kirchner, 2020)
The ERIC compilation offers a common vocabulary of 73 strategies for describing, selecting and reporting the strategies used.
(Powell et al., 2015)
Implementation becomes a phase in its own right, with context, programme theory and stakeholders at the centre.
(Skivington et al., 2021)
In Quebec, the implementation of personalised citizen support for older adults’ social participation sheds light on the conditions that matter.
(Gobeil et al., 2023)
02.2 · Sharing knowledge
Sixteen concepts connecting people, knowledge production and the transformation of systems. Each is defined, related to my approach and supported by scientific or institutional references in APA format.
Being an agent in one’s own life: choosing, deciding and pursuing personally meaningful goals. This requires real opportunities for choice, recognition of preferences and support adapted to the person’s capabilities and circumstances.
In my approach
In educational support and reflection on where people live, supporting their preferences and their ability to act.
A useful distinction. Wehmeyer’s functional model describes self-determined behaviour; Ryan and Deci’s theory addresses motivation and the needs for autonomy, competence and relatedness.
Developing a person’s or community’s real ability to act on what matters to them. This process connects skills, resources, rights, recognition and access to decision-making spaces.
In my approach
Designing tools that preserve users’ choices and examining the social and organisational barriers that restrict those choices.
A useful distinction. Empowerment has individual, collective and structural dimensions; access to resources matters alongside personal dispositions.
Involving people in the questions and decisions that concern them, with the means to contribute: understandable information, time, accessibility and recognition. Their influence may extend to identifying needs, considering options, making choices and evaluation.
In my approach
In partnership work, specifying who participates, when, and how their contributions influence decisions.
A useful distinction. Consultation, collaboration and shared decision-making involve different degrees of influence. Making that influence explicit helps establish the scope of participation.
Creating fairer conditions for accessing resources, participation and health, taking needs and barriers into account. Equity addresses avoidable and unjust disadvantages that hinder people’s ability to realise their capabilities and aspirations.
In my approach
In my values and public health reflections, considering the social conditions and support that make genuine choice possible.
A useful distinction. Different forms of support may be needed to enable everyone to exercise their capabilities and pursue their choices.
Developing support and action with the person: connecting what they want to do, what they can do and the support available. Professionals contribute a framework, resources and expertise that make participation possible.
In my approach
In specialised educational support and learning tools, developing approaches that people can understand, adjust and make their own.
A useful distinction. Support retains a professional framework and adapts to the person’s autonomy, responsibilities and situational constraints.
Developing a project, tool or knowledge together across several stages: framing the problem, choosing options, producing, interpreting and evaluating. Cooperation benefits from explicit roles and decision-making arrangements.
In my approach
A guiding principle for developing tools and preparing research workshops around the needs of the people concerned.
A useful distinction. Its quality depends on shared power, recognition of contributions and the resources needed to participate.
Knowledge developed through lived experience and making sense of it: illness, disability, support, everyday life or service use. It offers situated insight into how arrangements are used, the difficulties encountered and their practical effects.
In my approach
Recognising the expertise of people concerned alongside professional and scientific knowledge in care, teaching and research.
A useful distinction. An individual experience illuminates a context; collective discussion and combining knowledge sources help clarify its scope and limits.
In my ongoing work, this term describes experience and the consequences of action feeding back into knowledge frameworks: practice can prompt reconsideration of categories, indicators and decision-making rules.
In my approach
The manuscript presented under Research areas explores this movement from practice back towards knowledge and public health governance.
A useful distinction. Here it is a research proposition under development, distinct from simply disseminating or returning findings. The references below illuminate its theoretical foundations.
Understanding phenomena through interactions between people, organisations, institutions and environments. Interdependence, feedback loops and delays help explain how a local action may produce effects at several levels.
In my approach
In my thesis, connecting social needs, networks around CLSC home care and the rules that organise different actors’ responses.
A useful distinction. A systems perspective leads to hypotheses about mechanisms and then to documenting relationships and anticipated or observed effects.
Organising cooperation between different sectors around shared needs and determinants: health, housing, social services, education or community action. Common goals, explicit roles and coordination arrangements support this response.
In my approach
A central concept in my thesis: studying how several sectors respond to the co-occurring social needs of older adults receiving home care from CLSCs.
A useful distinction. It involves negotiating priorities, responsibilities and resources between sectors, beyond professionals simply working near one another.
Connecting professional and organisational interventions into a continuous, understandable pathway centred on people’s needs. Integration operates at clinical, professional, organisational and system levels.
In my approach
Connecting local services, transitions between organisations and the conditions supporting independent living at home.
A useful distinction. Shared supports, including information, goals and coordination arrangements, connect these different levels of integration.
Organising change in a particular setting by addressing adoption conditions, adaptation, available resources and evaluation. The CFIR helps examine factors that enable or impede implementation.
In my approach
For proposals arising from research, examining roles, resources and local constraints before considering deployment.
A useful distinction. A determinant framework helps explain context; a theory, a process model and an evaluation approach serve different functions.
A process of adaptation and rebuilding in the face of adversity, supported by personal, relational, social and institutional resources. Its form depends on people’s histories, environments and opportunities for support.
In my approach
In my own path, turning hardships into resources while recognising the collective conditions that make rebuilding possible.
A useful distinction. Resilience varies with circumstances and accessible resources. It does not justify leaving individuals to compensate for inequalities on their own.
Examining how one’s position, experience, values and choices shape questions, relationships and interpretations. This makes influences open to discussion and allows choices to be revised as work develops.
In my approach
In my research stance, the reflective journal planned in the protocol and reflection on the use of AI, particularly in PAIRR.
A useful distinction. A reflective journal, critical discussion and a record of decisions give this process a concrete form.
Examining sources and arguments, distinguishing facts, hypotheses and interpretations, assessing evidence quality and adjusting judgement. It combines reasoning skills with a willingness to question one’s own beliefs.
In my approach
In scientific reading, teaching and AI use, comparing explanations, their supporting evidence and their limitations.
A useful distinction. A critical judgement is reasoned and also considers evidence that challenges initial hypotheses.
Sharing, adapting and co-creating knowledge to support its use in practice and decisions. The process connects knowledge production, users’ needs, ownership, implementation and learning from observed effects.
In my approach
Adapting courses, public science communication resources and tools to their intended uses, then attending to what is actually used.
A useful distinction. Goals and formats vary across professionals, students, citizens and decision-makers. Sharing knowledge is part of an ongoing relationship.
02.3 · Teaching
My academic ambition extends a question that runs through my work as a specialised educator: how can we create conditions in which a person develops their capabilities, exercises judgment and participates in collective action? In public health, this question connects command of knowledge, attention to human experience and responsibility for decisions. I want it to guide my teaching.
University teaching involves more than organising content for students to reproduce. It provides access to demanding knowledge, its history and the approaches through which it can be questioned. I want students to explain how a conclusion was reached, what it supports and what remains uncertain. A concept becomes educationally meaningful when it changes how a situation is understood: identifying a social determinant, making an organisational choice visible or explaining why an intervention affects people differently.
Drawing on Biesta’s distinction between qualification, socialisation and subjectification, I see three connected responsibilities: developing knowledge and capabilities, enabling entry into a community of practice and supporting people in finding their own position within it. Educating competent professionals also means enabling them to question their field’s norms, explain a disagreement and take responsibility for the consequences of their actions.
My role would be to organise progression, identify expectations and make the criteria for sound reasoning explicit. Student participation rests on this structure. Productive dialogue requires knowledge, preparation and rules that allow arguments to be challenged while respecting people. I want to explain my teaching choices, welcome objections and acknowledge a limit or an error. Academic authority gains legitimacy through responsibility and the possibility of discussing its reasons.
Freire and hooks offer ways of thinking about a relationship in which students’ experience matters and learning engages their relationship with the world. I take from them a responsibility to make room for people’s voices while examining their claims. Lived experience can reveal a decisive question; an empirical proposition then requires appropriate sources and methods. Respecting a person and examining the validity of a statement are part of the same educational task.
Sources informing this reflection : Biesta (2009) · Freire (2018) · hooks (1994)
In supporting people with disabilities, I have assessed needs, jointly developed individualised plans and coordinated interventions. Adapted holidays and respite programmes also brought me into contact with different rhythms, communication modes and practical conditions for participation. This experience shapes my attention to what someone seeks to achieve, the resources they draw on and the obstacles that restrict their possibilities.
It also invites me to examine the learning environment: an implicit instruction, an inaccessible resource, a participation rule or a timetable can create barriers. The Human Development Model–Disability Creation Process and Universal Design for Learning offer ways of understanding the interaction between a person and their environment. At university, this attention must respect students as adults, their privacy and their ability to define their own needs. It calls for a pedagogical relationship suited to that setting.
The bridge I want to build rests on self-determination: supporting the capacity to choose and act by providing access to the means to do so. Working with students then involves understandable objectives, feedback, meaningful pedagogical choices and support that can diminish as autonomy develops. The group offers a further resource: learning to seek help, offer it and recognise others’ contributions.
Grounding in support work
Translation into my teaching project
Identify prior knowledge, constraints and barriers to participation without reducing a student to a difficulty.
Grounding in support work
Translation into my teaching project
Clarify expected learning and offer choices of topic or approach where these fit the course requirements.
Grounding in support work
Translation into my teaching project
Provide worked examples, intermediate stages and feedback; gradually withdraw support that is no longer needed.
Grounding in support work
Translation into my teaching project
Connect peer learning, mentoring, teaching teams and university services while respecting roles and confidentiality.
Sources informing this reflection : RIPPH · MDH-PPH · CAST (2024) · UDL 3.0 · CAST (2024) · Strategy development
My four teaching assistant appointments at Université de Montréal provide an initial foundation for this ambition. Student support and the marking of written and oral examinations, documented for ASA 6754 in winter 2026, focus attention on how someone understands a question, uses their knowledge and makes their reasoning visible. These responsibilities take place within the framework established by the course instructors.
I want to develop feedback that helps students rework their understanding: identify a specific achievement, name a priority difficulty and suggest an actionable next step. When a student describes an organisation accurately but explains little about its effects, feedback can invite them to explain the mechanism, consider another hypothesis and examine who is affected. The aim is for a comment to become a resource for subsequent reasoning. Carless and Boud emphasise that using feedback requires learning to make judgments, manage one’s response and take action.
Coordinating the doctoral peer mentoring programme extends this attention to learning conditions. Entering a programme involves understanding its expectations, resources and sometimes implicit conventions. My ambition is to help create an environment in which seeking support is compatible with intellectual rigour and experience circulates among peers. The approach described here guides my development; it is not an evaluation of a course design whose effects I have already measured.
Université de Montréal.
Health system and population health, Université de Montréal: student support, grading of written and oral exams.
ESPUM, Université de Montréal.
Sources informing this reflection : Carless & Boud (2018)
I aim to design teaching in which knowledge is put to work on public health situations rather than merely recalled. This is not about opposing lectures to discussion or mistaking participation for learning: each activity should build an identifiable capability, draw on verifiable sources and make reasoning visible.
Proposed approaches, not outcomes from an evaluation of my teaching.
Clarify concepts, available data and uncertainties.
Bring methods, stakeholders and competing explanations into dialogue.
Develop a justified decision and reconsider it in another context.
These stages form a loop: new findings can lead to a revised initial question.
A health problem, its context and affected populations.
Hypotheses, mechanisms and missing information.
Data, literature and appropriate methods.
Options, disagreements and explicit criteria.
Decision, limitations and opportunities to revise.
The format depends on intended learning, group composition, available resources and real constraints.
| Format | Student activity | Observable evidence | Consideration |
|---|---|---|---|
| Interactive lecture | Answer conceptual questions and defend an objection. | Annotated concept map or short justification. | Interaction must involve more than checking attendance. |
| Case study | Analyse a contextualised situation and compare responses. | Decision brief stating assumptions and limitations. | A simplified case is not the whole reality. |
| Data workshop | Explore an indicator, its sources and possible biases. | Captioned chart with critical commentary. | Distinguish association, causation and missing data. |
| Stakeholder simulation | Defend options with explicit roles and constraints. | Trade-off matrix and reasoned collective decision. | Do not claim to speak for the people concerned. |
| Peer review | Assess a draft using shared criteria. | Revised version with reasons for changes. | Teach students how to give and use feedback. |
Select a field to see how the same approach leads to different problems, activities and outputs. Cases are fictional.
A city must adapt its heatwave prevention plan amid unevenly distributed risks.
Compare vulnerability maps, costs, needs and stakeholder proposals; discuss trade-offs without reducing equity to one metric.
A policy brief comparing two options, their consequences and missing evidence.
A primary care team wants to improve continuity after repeated emergency visits.
Map a fictional care journey, identify gaps across organisations and test two models against feasibility constraints.
A coordination diagram and reasoned change proposal with monitoring indicators.
An international comparison finds a surprising association between reported experience and avoidable admissions.
Examine units of analysis, coding, confounding, data sources and ecological inference limits.
A sensitivity-analysis plan and cautious interpretation without a causal claim.
A meta-analysis of 225 undergraduate STEM studies found better outcomes on average with active learning than with traditional lecturing. This finding concerns those disciplines; it neither measures my teaching nor guarantees transfer to every public health course.
225studies reviewed
≈ 6 %approximate average improvement in examination scores
≈ 1,5 ×approximate relative likelihood of failure with traditional lecturing vs active learning
Relative index, active learning set to 1; these are not absolute failure rates. Freeman et al. (2014). Freeman et al. (2014)
Align learning outcomes, learning activities and assessment; distinguish teaching from what students can demonstrate.
Vary engagement, representation and action and expression while keeping meaningful shared goals.
Give learners an opportunity to interpret feedback, make a judgment and improve their work.
These formats describe a direction I would like to pursue. Actual choices depend on course objectives, the teaching team and students. Their usefulness should be examined through assessable outputs and documented feedback.
These studies differ in populations and outcomes. I want to use them to design, evaluate and adjust teaching, rather than as universal prescriptions.
In undergraduate STEM, active learning was associated with higher average examination scores and fewer failures than conventional lectures.
A pooled result from varied formats and settings.
In STEM courses, average examination achievement gaps narrowed by 33% and passing gaps by 45% among studied groups.
Defined populations and settings; intensity of active learning matters.
Case-, problem- and team-based formats have been used across courses, with reported learning opportunities and implementation challenges.
Scoping reviews map evidence; they do not estimate one pooled causal effect.
Students learned more in active classes even when they felt they learned less; perceived effort does not measure achievement.
A specific course; neither direction nor size necessarily transfers.
Three poolable studies found no significant difference on tests; satisfaction and reported behaviours were more favourable.
Do not conflate satisfaction, measured learning and practice transfer.
Illustrative mapping, not a reproduction of the official frameworks. Informed by Canada's 2025 core competencies and the WHO–ASPHER European framework.
Examine an indicator, its population and limitations.
Critical reading of surveillance data.
Compare options, equity, resources and uncertainties.
Reasoned policy brief.
Identify responsibilities, connections and coordination risks.
Stakeholder map and deliberation scenario.
Adapt information to its audience without distorting evidence.
Accessible summary for a community organisation.
Make trade-offs and affected perspectives explicit.
Decision log explaining limitations.
One possible progression that does not stop at satisfaction or immediate recall.
Diagnostic question and initial individual reasoning.
Guided discussion, source comparison and joint output.
Justified revision, feedback and transfer task.
Retrieval activity and unfamiliar problem under comparable criteria.
Group work alone does not demonstrate learning; individuals need to explain and apply.
Enjoyment or difficulty describes experience but does not alone establish competence.
Cohort size, accessibility, preparation workload and partner availability affect feasibility.
Learning inequalities are also shaped by social and institutional conditions.
Freeman et al. (2014) · Theobald et al. (2020) · Mansoor et al. (2026) · Deslauriers et al. (2019) · Mensour et al. (2025) · CCN en santé publique (2025) · OMS Europe & ASPHER (2021) · Roediger & Karpicke (2006) · Cepeda et al. (2006) · Wallar & Papadopoulos (2015)
I want to start from the capabilities being developed and use them to design activities and assessment. If the aim is to analyse a health policy, students need opportunities to practise before an examination: define the problem, make assumptions explicit, interpret data and discuss consequences. Constructive alignment gives these activities and their criteria a shared direction.
Active learning interests me when it involves an identifiable intellectual task: explaining, comparing, solving, arguing or revising. Freeman and colleagues’ meta-analysis concerns undergraduate STEM teaching; it supports these approaches within that scope while leaving their adaptation to other contexts to be examined. Retrieval questions, spaced revisiting of key ideas and explanatory correction can also help consolidate knowledge.
I also want assessment to be understandable and usable. Distinguishing practice from assessments that certify achievement gives error and feedback a clear role. An explicit rubric can address problem definition, source quality, argument coherence, consideration of alternatives and attention to consequences. Collective work can include an individual justification to make each person’s learning visible.
A public health learning sequence I would like to develop
A proposed teaching design: use a fictional case of co-occurring social needs in primary care to examine how a CLSC and its partners might organise a response. This illustrates an educational ambition and presents no thesis findings.
An initial individual analysis distinguishes needs, missing information and hypotheses about barriers to a response.
Groups compare the perspectives of patients, professionals, community organisations and decision-makers, without claiming to speak on their behalf.
Compare scenarios using evidence and sources; distinguish established findings from hypotheses and value choices.
Produce a recommendation, explain the alternatives set aside and discuss feasibility, equity and organisational consequences.
Compare the work with the criteria, peers’ questions and teaching team feedback; select a revision priority.
Submit a reasoned revision and a short reflection on changes made, remaining uncertainties and knowledge that needs further development.
Sources informing this reflection : Eberly Center · Learning objectives · Freeman et al. (2014) · Dunlosky et al. (2013) · Eberly Center · Assessment · Carless & Boud (2018)
Practical workshop
My teaching ambition connects demanding knowledge with participation and autonomy. Start with a capability, organise practice and ask for evidence of reasoning. An accessible environment offers different ways to reach clear objectives. Individual preparation, peer discussion and a revised personal explanation help distinguish participation from learning.
Download the worksheet ↓Explain, interpret or decide.
Practise with appropriate support.
Justify choices and examine alternatives.
Use feedback on a new attempt.
Fictional situation · proposed approach
A team must organise a response to co-occurring social needs in primary care. A fictional vignette, data and different actors’ perspectives support three possible learning goals.
Compare individual and organisational explanations, define the concepts and examine what each reveals.
An annotated concept map, a personal explanation and a counterexample.
Choose the denominator, calculate a measure and distinguish association, causal hypotheses and value judgements.
An annotated calculation, plain-language interpretation and an explicit limitation.
Compare options, affected people, displaced burdens and uncertainty. Deliberate, then revise individually.
A decision brief with criteria, alternatives, a major objection and reasons to reconsider.
The trials by Kestin and Bastani concern different settings and instructional designs. They support examining learning without assistance. The medical never-skilling concern discussed in 2026 remains a research agenda; direct evidence is limited.
The rise of generative AI assistants raises a pedagogical question: what can a person actually do when part of the work can be delegated? I want students to learn to choose which tasks to assign to a tool, explain its limits, check its sources and critically examine its outputs. The quality of an assisted text or programme is insufficient to establish its author’s understanding.
PAIRR (peer support and review, AI assistance, then reflection) offers an approach I would like to adapt. Initial personal reasoning, peer discussion, critique of AI suggestions and justification of revisions can make the process visible. Depending on the objective, an oral explanation or an unassisted task can complement assessment. Use rules should be explicit, account for unequal access and protect personal data and confidential work.
Public health involves debating freedoms, limited resources and unequally distributed risks. Teaching should prepare students to recognise conflicts of values, distinguish scientific uncertainty from political disagreement and listen to people whose experiences differ. I want to develop reasoned deliberation: make reasons accessible, examine decisions’ effects and leave room for challenge. This capacity concerns professional responsibility and democratic life alike.
Climate challenges, inequalities and changing health systems require attention to interdependence. A useful response within one service can transfer a burden to families, another sector or the environment. I want to use comparison and systems analysis to help students identify these transfers, their timescales and the people overlooked in a decision. UNESCO’s report on the futures of education provides a framework for reflecting on these collective responsibilities, rather than a single course model to reproduce.
Educating for tomorrow requires sound knowledge and the capacity to revise it. Barnett helps frame education for a future we cannot fully predict: learning to formulate problems, work with reasoned uncertainty and cooperate with other areas of expertise. My ambition is for students to construct a provisional response, explain what would strengthen it and identify when new information warrants changing it.
This ambition also involves the institution: make support accessible, sustain teaching work and examine conditions that enable or prevent participation. I want to contribute to a university that educates people to act together, care for relationships and question the purposes of what they build.
Sources informing this reflection : Sperber et al. (2025) · UNESCO (2021) · Barnett (2004)
These references support different dimensions of my thinking. Freire, hooks, Biesta and Barnett offer philosophical and theoretical perspectives; UNESCO, CAST and RIPPH provide guiding frameworks. Research on learning and feedback must be read in light of its methods and context. It informs my approach without, by itself, demonstrating the effects of my personal practice.
02.4 · Science outreach
I want science communication to be a space where people can understand a question, examine the reasons behind a claim and join the discussion. My background in specialised education, law, management and public health supports a shared concern: connect rigorous knowledge with lived situations while preserving uncertainty and diverse perspectives. This section brings together a vision, visual explanations, proposed formats and resources for practical use.
Carousel: organise a visual progression
Video: make a mechanism visible
Podcast: bring perspectives into dialogue
Access to information does not automatically provide the means to interpret it. I want to explain how knowledge is produced, how it changes our understanding of a problem and what remains open. My ambition is for people to ask a more precise question, recognise a limitation and place a decision in context. Curiosity can begin the encounter; access to sources and the possibility of discussing reasons give it depth.
My experience as a specialised educator encourages me to start from people’s capabilities, constraints and projects. A family carer, student and community volunteer may approach the same topic with different questions. I want these experiences to inform the point of entry and reveal problems sometimes obscured by institutional language. This dialogue requires time, listening and an explanation of what was taken into account.
Communicating about health involves data, experience and collective choices. I want to distinguish these: testimony illuminates a situated experience; a study addresses a question under particular conditions; a decision also involves values and constraints. Bringing them into dialogue helps explain disagreements. This requires stating my role, sources and any interests involved, then making corrections visible when necessary.
Make knowledge available.
A short explanation introduces co-occurring social needs.
Exchange questions, knowledge and objections.
A conversation brings out difficulties encountered between services.
Share choices about content or process.
People concerned help choose examples and questions to address.
Examine options and their consequences together.
A workshop compares possible responses to an organisational problem.
These approaches serve different aims and can be combined. Inviting comments does not by itself share decision-making: the scope of participation must be explicit.
Explore further : National Academies (2017) · NCCPE · Public engagement · UNESCO (2021) · Open science
Before writing a headline, reconstruct the path between the research question and the proposed conclusion. This diagram is a reading framework, not an automatic ranking of methods.
What problem, population and context?
How were people, data and comparisons selected?
What was observed, with what magnitude and uncertainty?
What can be concluded here? What still needs examining?
State whether it is a protocol, preprint, peer-reviewed article or evidence synthesis. Status provides information about the production process; it does not replace examining the content. For my thesis, the objectives and methods describe ongoing work, with no findings available.
An observed association is insufficient to attribute a causal effect. An interview can deepen understanding of an experience and its meanings without estimating its population frequency. A synthesis inherits limitations of the included studies. The wording should match the question and design.
Explain what is missing, what varies and what could change the conclusion. An estimate, its uncertainty range and its practical relevance address different questions. Van der Bles and colleagues’ experiments shed light on some ways of expressing uncertainty; they do not guarantee trust in every setting.
Explore further : van der Bles et al. (2020) · Schünemann et al. (2024) · Cochrane
An entirely fictional example: an event occurs in 20 out of 100 people in one group and 10 out of 100 in another over the same period. None of these figures comes from my research.
20 out of 100 people
10 out of 100 people
Each dot represents one person; filled dots indicate the event.
20% − 10% = 10 percentage points: 10 fewer events per 100 people.
(20 − 10) ÷ 20 = 50% fewer events relative to group A.
The same denominator, event and duration make the comparison interpretable.
Both presentations describe the same difference. They do not establish why the groups differ. Interpreting a real finding also requires information about recruitment, design, uncertainty and outcomes that matter to people.
Explore further : Schünemann et al. (2024) · Cochrane · CDC · Clear Communication Index
From initial framing to public feedback: six principles for clear, accessible communication that welcomes dialogue.
Specify whom you are addressing, their questions and the intended outcome: understanding, discussion or support for a decision. Then select suitable content and channels.
Identify their questions, experiences and vocabulary. Speak with some intended audience members before finalising the content.
Define what the content should enable: understanding an idea, discussing an issue or informing a decision.
Choose the channel, length and detail according to audience practices, then frame a concrete question.
One audience contains diverse experiences. Information needs and trust cannot be inferred from knowledge levels alone.
Organise content around one main idea. Explain terms, use relevant examples and make key information and sources easy to find.
Make the main message visible at the start and organise explanations around it.
Explain jargon when first used. Include examples or diagrams when they genuinely improve understanding.
Present numbers with their population and period, make sources accessible and specify the next step offered to the audience.
An engaging headline must remain faithful to available knowledge. Simple language can preserve necessary distinctions.
Create opportunities for questions and sharing experience. Plan respectful responses, explicit moderation and feedback that informs future content.
Invite experiences or questions and leave room for responses.
Explain moderation rules, the limits of the response offered and follow-up arrangements.
Gather recurring questions, clarify disagreements and show how exchanges shape subsequent content.
Provide a respectful space and avoid publicly requesting health information or details identifying a person.
Distinguish findings, interpretations and hypotheses; explain uncertainty and the context of numbers. Transparency requires careful wording, without promising an automatic effect on trust.
Specify what was observed, what is interpretation and what remains a hypothesis or planned project.
Explain the population studied, scope of findings and important limitations in accessible language.
Date content, retain sources and make corrections visible when errors or developments require them.
Communicating uncertainty involves choices of language and context; its effect on trust cannot be guaranteed in advance.
Include checked captions, transcripts and descriptions of essential visual information. Design alternatives from the scripting stage to offer several ways to access the content.
Include what will be spoken, shown and provided in text from the scripting stage.
Provide reviewed captions, transcripts and useful descriptions of visual information; ensure clear reading and navigation.
Invite intended users to read or try the content, considering mobile access, languages and listening contexts.
Needs vary by medium. Unreviewed automatic captions or vague alternative text may leave essential information inaccessible.
Distinguish reach, understanding and quality of dialogue. Collect questions and feedback, check understanding and adjust formats to the intended goal.
Distinguish reach, message comprehension, participation and perceived usefulness.
Ask for a brief restatement and remaining questions, instead of relying solely on views or likes.
Identify what needs revision, change the format or explanation and record improvements.
Views describe distribution. Alone, they demonstrate neither understanding, trust nor behaviour change.
Public feedback informs subsequent messages: revisit and adapt this process.
These mockups are editorial proposals to develop. They show how a question can be explored without inventing findings or implying an episode has already been published.
Design example 01
Example: “Why can one situation require several services?” Five cards can make the reasoning visible and create a resource to revisit.
Start from a recognisable situation without exposing a real person.
Explain the coexistence of housing, income or support needs.
Show actors and necessary relationships; explain what the diagram simplifies.
More actors do not automatically mean better coordination.
Ask a question and provide the source and its date.
Keep essential text in the post, describe the diagram, number the cards and avoid conveying information through colour alone.
Explore further : NCCPE (2018) · Social media · CDC · Clear Communication Index · W3C · Images tutorial
Design example 02
Example: explain how information can be lost between two services. The visual should help explain the mechanism while narration names its elements.
Begin with a concrete difficulty and state the idea being explored.
Build a small network progressively, with a key and named actors.
Compare a coordinated information transfer with an interrupted one as illustrative scenarios.
Explain what the example shows, its limits and the sources to consult.
Plan checked captions, a transcript and spoken descriptions of necessary visual information. Music should not obscure speech.
Explore further : CDC · Clear Communication Index · W3C · Accessible media · W3C · Images tutorial
Design example 03
Example: an interview about conditions for cooperation across sectors. Preparation should enable perspectives to be compared and terms explained without staging an artificial consensus.
Introduce participants, their roles and connections to the topic.
Invite a situated account and clarify important terms.
Ask what supports a claim, what is disputed and whose voices are missing.
Distinguish agreement, disagreement and questions to revisit.
Publish a structured transcript, listening markers and episode notes with sources and limitations. Agree on recording and publication arrangements.
Explore further : NCCPE · Public engagement · W3C · Accessible media
Communication quality also depends on the relationship it creates: who can contribute, who remains distant and what happens to contributions.
I want to restate a question before answering, distinguish challenges to findings from criticism of an institution and acknowledge experience without generalising it. Explicit discussion rules should protect people and enable arguments to be examined. Consultation becomes meaningful when its scope is explained and participants hear what happened afterwards.
A useful correction states the supported point, identifies the specific problem and explains the reasons for the correction with an accessible source. I want to avoid humiliation and sensational repetition of a misleading claim. In health, a public reply also needs to acknowledge the limits of an online exchange and preserve confidentiality.
AI tools can help prepare a structure or explore an illustration; responsibility for claims, sources and editing remains with the author. A generated image should be identifiable as an illustration when it could be mistaken for a real scene. Voices, quotations and testimony should not be invented to make content more persuasive. I want to retain provenance and avoid sharing sensitive data with a tool without an appropriate framework.
Gather questions without requesting identifying details.
Group needs, objections and uncertainties.
Adapt the explanation and identify what remains open.
Show revisions, corrections and proposed next steps.
One way to document it
Document consultation counts where such measurement exists.
What we cannot conclude
A visit does not establish reading or understanding.
One way to document it
Invite a short restatement or ask which distinction was retained.
What we cannot conclude
An immediate response is insufficient to establish lasting understanding.
One way to document it
Examine questions, diversity of contributions and reported barriers.
What we cannot conclude
Respondents do not necessarily represent the whole audience.
One way to document it
Ask how the content helped someone find their bearings or pursue an issue.
What we cannot conclude
Reported usefulness does not establish a causal effect on health or decisions.
Usage figures describe a resource’s reach; they do not demonstrate understanding, usefulness or an effect on health.
Explore further : NCCPE · Public engagement · NCCPE · Evaluation · WHO (2025) · Social listening
Short worksheets to reuse in a project, workshop or discussion. They structure the work; they are not validated instruments or quality certifications.
Explore further : CDC · Clear Communication Index · Schünemann et al. (2024) · Cochrane · W3C · Images tutorial
Each resource connects to a specific task. Research articles, reports and guides serve different purposes; their examples need adaptation to audiences and settings.
Research report · EN
Understand science communication goals, audiences and conditions.
Read chapter 2 to clarify what an approach aims to enable.
Scope and limitations. A research agenda, not a recipe guaranteeing acceptance.
National Academies of Sciences, Engineering, and Medicine. (2017). Communicating science effectively: A research agenda. The National Academies Press.
Design tool · EN / ES
Structure messages, numbers and intended actions.
Use the framing questions before drafting, then discuss criteria with reviewers.
Scope and limitations. A clarity checklist does not replace scientific examination or audience feedback.
Centers for Disease Control and Prevention. (n.d.). How to use the Index. The CDC Clear Communication Index.
Practice guide · EN
Prepare public engagement and its conditions.
Clarify participation goals and how contributions will be followed up.
Scope and limitations. Engagement requires resources and local adaptation.
National Co-ordinating Centre for Public Engagement. (n.d.). A quick guide to developing high quality public engagement.
Research article · EN
Experimentally examine some effects of uncertainty wording on trust.
Compare formulations and studied contexts before deriving a rule.
Scope and limitations. Findings do not guarantee identical effects for every topic or audience.
van der Bles, A. M., van der Linden, S., Freeman, A. L. J., & Spiegelhalter, D. J. (2020). The effects of communicating uncertainty on public trust in facts and numbers. Proceedings of the National Academy of Sciences, 117(14), 7672–7683.
Technical guide · EN
Plan captions, transcripts and descriptions of audiovisual information.
Build alternatives into scripts and episode notes.
Scope and limitations. Needs and requirements vary across media.
World Wide Web Consortium, Web Accessibility Initiative. (n.d.). Making audio and video media accessible.
Practice guide · EN
Distinguish outputs, intended outcomes and impacts in public engagement.
Connect an evaluation question to a realistic feedback method.
Scope and limitations. An isolated indicator does not establish causation.
National Co-ordinating Centre for Public Engagement. (2023, August 14). How to evaluate public engagement projects and programmes.
Practice synthesis · EN
Think through public engagement on social media.
Use its questions about strategy and audience relationships.
Scope and limitations. Published in 2018: platforms and practices have changed.
National Co-ordinating Centre for Public Engagement. (2018). What works: Engaging the public through social media.
Methods handbook · EN
Interpret effect sizes and uncertainty.
Use sections on absolute and relative effects when explaining numbers.
Scope and limitations. Focused on intervention syntheses; adapt its use to other designs.
Schünemann, H. J., Vist, G. E., Higgins, J. P. T., Santesso, N., Deeks, J. J., Glasziou, P., Akl, E. A., & Guyatt, G. H. (2024). Chapter 15: Interpreting results and drawing conclusions. In J. P. T. Higgins, J. Thomas, J. Chandler, M. Cumpston, T. Li, M. J. Page, & V. A. Welch (Eds.), Cochrane handbook for systematic reviews of interventions (Version 6.5). Cochrane.
Technical guide · EN
Choose text alternatives according to a visual’s function.
Prepare a full description of a graph or diagram.
Scope and limitations. Descriptions need to convey the information relevant to their context.
World Wide Web Consortium, Web Accessibility Initiative. (n.d.). Images tutorial.
Institutional guidance · EN
Consider social listening ethics and people’s rights.
Examine purpose, proportionality and data protection before collecting information.
Scope and limitations. Developed for public health emergencies.
World Health Organization. (2025). Social listening in infodemic management for public health emergencies: Guidance on ethical considerations.
International recommendation · multilingual
Place knowledge sharing within openness and participation.
Connect access to knowledge with dialogue with society.
Scope and limitations. An international normative framework, not an effectiveness measure for a format.
UNESCO. (2021). UNESCO recommendation on open science.
Seven languages, including Inuktitut and Innu-aimun (preliminary versions), an easy-to-read version, and versions adapted for colour blindness and low vision.
03.1 · Biography
Special needs educator, jurist, entrepreneur and PhD student in public health, I built my path at the meeting point of different worlds: supporting people, managing organisations, law, public policy and research.
Behind this diversity lies one ambition: to understand the mechanisms that shape individual and collective trajectories, in order to contribute to fairer, more supportive societies that recognise everyone’s capacity to act.
My journey, personal and professional, feeds one conviction: human realities cannot be separated from the social, organisational and political systems in which they are embedded.
It follows an upward movement: from supporting people towards transforming health and social service organisations and policies, with a clear focus on supporting public decision-making.
Human realities cannot be separated from the social, organisational and political systems in which they are embedded.
My path begins in France, in a working-class setting, marked in particular by time in the child welfare system and by leaving school at sixteen.
Far from a conventional academic route, I gradually built my trajectory through work, through exploring different settings and through returning to study.
These experiences, like my neurodivergence, nurtured my independent thinking, my adaptability and my interest in the mechanisms of social reproduction, emancipation and solidarity. Today they inform how I look at institutions and at the diversity of human experience.
My interest in music led me to a bachelor’s degree in musicology, alongside my first experiences in youth and leisure programmes and then managing adapted stays for people with disabilities.
For all their differences, these worlds let me explore human relationships in several dimensions: individual expression, collective life, support and responsibility towards others.
Over eleven years of commitment alongside human vulnerability, these experiences gradually placed the ethics of care at the centre, together with the recognition of individual capacities and the importance of support that fosters autonomy rather than dependence. They also showed me the strength of solidarity.
After several years in the field, I wanted to give my professional life a new direction, guided by a reflection on social usefulness and on the place I wished to hold in society.
I therefore returned to study special needs education to deepen the theoretical, ethical and methodological foundations of support in the social and medico-social sectors, refocusing my practice on self-determination and on working with people.
My professional practice thesis, on how educational practices influence self-determination, was a turning point in my questioning: how can people’s freedom of choice be fostered when the conditions of their support are also determined by how institutions work?
This question led me to widen my reflection to organisations themselves.
I completed training in the management of health and medico-social facilities, as well as a university diploma in entrepreneurship at IAE Brest, obtained in 2024.
There I deepened issues of management, funding, coordination, human resources and change management, and observed that the facilities’ difficulties, from funding to partnership work and recruitment, were largely conditioned by health policy.
Above all, I sought to understand the conditions that allow professionals to turn their intentions into effective practice, and organisations to reconcile their constraints with people’s needs.
Studying organisations led me to explore the mechanisms that steer their decisions: legal frameworks, public policy, planning, funding arrangements and governance.
I pursued this reflection at Université Paris-Panthéon-Assas, in public, health and social law, to move beyond management issues and adopt a multilevel approach to the health system and to the legal instruments of national solidarity, then as a project manager at France’s national agency for performance support (ANAP).
This experience allowed me to work at the interface between national directions and operational realities, turning those directions into concrete arrangements, in close contact with the challenges faced by managers, professionals and users of the health system.
It strengthened my interest in the difficulties of implementing public policy, especially when the responses needed require crossing the boundaries between organisations and sectors. Consolidated by a research thesis, this immersion revealed the difficulties of implementing integrated care and service models, and the need for an intersectoral and interdisciplinary approach to siloed practices. I then reached out to several researchers in Québec, presenting my findings and a project to contrast health systems, in search of answers to the difficulties of implementing ageing-in-place policies.
Today, my PhD in public health at Université de Montréal, supervised by Drs. N. Sourial and G. Layani, continues this journey.
My work focuses on the organisation of health and social service systems, populations’ social needs, intersectoral collaboration and the conditions for implementing integrated responses, particularly in home support.
My approach is resolutely interdisciplinary and pragmatic. It seeks to bring together lived experience, scientific knowledge and decision-making realities, to better understand the phenomena studied and to contribute to changing practice and public policy.
I pay particular attention to people losing autonomy and those far from services, to partnership research that recognises patients as producers of knowledge, to knowledge translation and to the use of knowledge by people, professionals, organisations and decision-makers.
This drive to create, experiment and propose also extends to entrepreneurship, notably through the development of OWNNIGHT, a project under way.
I believe that academic, professional and experiential knowledge gain from being confronted rather than ranked.
My path has let me discover different layers of society and of the health system, hold various positions within them and observe how they interact. This plurality of perspectives is now one of the resources of my scientific approach.
For me, research is an instrument of understanding, but also of human emancipation, social justice and support for collective decision-making. Its value lies as much in the rigour of the knowledge it produces as in its capacity to feed debate, inform choices and support their implementation.
It is aligned with the United Nations Sustainable Development Goals: improving the well-being of older adults, reducing inequalities in ageing at home and strengthening inclusive, accountable institutions. In the long run, my ambition is to become an independent researcher, recognised for combining evidence, dialogue among stakeholders and outreach to the general public.
Intellectual curiosity, resilience, autonomy, critical thinking, co-construction and the boldness to undertake are values that run through my journey.
I wish to pursue this exploration at the interface of disciplines and settings, convinced that today’s public health challenges require bringing people, knowledge and institutions into dialogue.
Ad astra per aspera: through hardships to the stars.
03.2 · Background
From frontline practice to health policy
From special needs education to the PhD: the degrees, positions and training that mark the path.
French degrees evaluated by the MIFI (Québec) and WES Canada.
An Android app for sleep well-being and sleep education: a personalised plan, lessons and guided sessions.
École de santé publique de l’Université de Montréal (ESPUM), Canada. Health systems, organisations and policy (SOPS), second year; research trainee at the Centre de recherche du CHUM (CRCHUM), Canada.
Cumulative GPA: 4.1
Université de Montréal.
Health system and population health, Université de Montréal: student support, grading of written and oral exams.
Public law, social law and health law · Université Paris Panthéon-Assas, with special distinction from the jury.
Agence nationale d’appui à la performance (ANAP), Paris (France). Co-leading the expert network, hosting webinars, collecting and analysing data, and facilitating communities of practice.
Université de Bretagne Occidentale (UBO), France, with special distinction from the jury.
Fondation Ildys, Landerneau (France). EHPAD (établissement d’hébergement pour personnes âgées dépendantes), a French residential care home for dependent older adults, comparable to a Quebec CHSLD (long-term care centre); internship with the management team.
IAE de Brest (Institut d’administration des entreprises), France, with special distinction from the jury.
Académie de Rennes.
Residential home, Adapei 35, Saint-Malo: assessing needs, co-designing individual plans, coordinating a multidisciplinary team.
Adapted holidays and respite stays for people with disabilities, Brest (France): co-designing educational programmes, providing socio-educational support and supervising teams, budgets and logistics.
03.3 · Background
Trained in health and social law at Université Paris-Panthéon-Assas (master’s degree, special mention from the jury), I analyse the legal environment of facilities, professionals and health policies, from people’s rights to litigation.
03.4 · Background
Trained in the management of medico-social organisations and personal services at Université de Bretagne Occidentale (master’s degree, special mention from the jury), I have practised day-to-day management as head of adapted holiday programmes, then during a management internship in a nursing home.
03.5 · Background
State-certified (Académie de Rennes, 2023), I supported people with disabilities, including autistic people, in a residential home and on adapted holiday programmes, with one conviction: supporting their self-determination.
A position’s classification depends on its actual duties and the population supported. In residential disability services, duties may combine day-to-day social support with adapted educational activities; these two occupational profiles clarify those dimensions.
Psychosocial support, access to resources, assistance in daily life and social participation. This category includes specialised educational support roles in social and community services and positions in group homes.
View the official occupational profileAdapted educational support for people with disabilities, focusing on communication, learning, social and practical skills, independence and rehabilitation.
View the official occupational profileDigital productions
Two digital projects developed from concrete needs: making knowledge useful, accessible and applicable in everyday life.
The science of your sleep, in your hands.
Entrepreneurial project in progress · AndroidAn Android app for sleep well-being and education, in French and English. It offers an individualised plan that adapts to each answer, short lessons, guided sessions, tools to calm the mind, an evening ritual and stories read aloud.
Empowering people to act on their own sleep: assessing it themselves, understanding what happens at night and learning to take care of it, through sleep education grounded in cognitive behavioural therapy for insomnia (CBT-I).
Support based on CBT-I, the first-line treatment for insomnia, remains hard to access, and many apps merely measure sleep without helping people understand it, while exploiting personal data. OWNNIGHT focuses on empowerment: people learn, assess themselves and take an active role in their own sleep. Everything stays on the phone, encrypted, with no account.
Episphère and Statosphère: understanding epidemiology and statistics at your own pace.
Computer, tablet and phone · offlineTwo sister learning apps, one for epidemiology and the other for statistics. They work offline on computer, tablet and phone: step-by-step lessons, annotated slides, summary sheets, multiple-choice questions with feedback, flashcards, worked exercises and progress tracking, with a built-in tutor, Sigma.
Helping people understand methods rather than memorise them: each formula is explained by what it does, then read step by step, and everyone organises their revision around their own needs.
Quantitative methods often intimidate people from other disciplines, and resources are scattered, often in English or highly technical. The Spheres bring everything together in one place, in French, with an accessible progression and plain-language explanations.
Learning software
Two independent learning environments for reading, calculating, practising and learning from errors. Lessons and quizzes are currently in French.
Descriptive statistics, probability, inference, regression models, causal graphs, experiments and R practice.
27 chapters1,544 quizzes with feedback
Launch in your browser ↗Frequency and association measures, study designs, bias, standardisation, screening, trials, survival and evaluation.
20 chapters2,100 quizzes with feedback
Launch in your browser ↗These public editions are separate from the personal applications. They include concepts, lessons, formulas, flashcards and quizzes; calculation exercises use new fictional data. Progress stays locally in this browser.
03.7 · Values
Six values that guide my work, from practice to research.
MottoAd astra per asperaThrough hardships to the stars
Exploring other disciplines and ways of thinking, comparing ideas and questioning assumptions, data and institutions, starting with one’s own beliefs.
Turning hardships into resources to move forward.
Thinking for oneself, and supporting everyone’s capacity to choose their own life.
Doing with people, not for them: impactful research carried out with and for the people it concerns.
Daring to propose, innovate, create and implement new responses.
Enabling everyone, as far as possible, to fully develop their capabilities and realise their own aspirations.
04.1 · Engagement
Taking part in student and scientific life, in continuity with the path so far.
Elected.
ESPUM, Université de Montréal.
ESPUM, Université de Montréal.
Department of Health Management, Evaluation and Policy, ESPUM.
France.
France.
04.2 · Engagement
Member.
Student member.
04.3 · Engagement
Early-career research day, Groupe francophone en soins primaires.
Excellence scholarship, ranked 3rd.
Graduate and Postdoctoral Studies, Université de Montréal.
Academic excellence.
04.4 · Engagement
Dinard fire and rescue station, France.
Association de la fondation étudiante pour la ville (AFEV), France: coordinating student volunteers and supporting a young person through her school and cultural journey.
04.5 · Engagement
05.1 · Learning software
Two independent learning environments for reading, calculating, practising and learning from errors. Lessons and quizzes are currently in French.
Descriptive statistics, probability, inference, regression models, causal graphs, experiments and R practice.
27 chapters1,544 quizzes with feedback
Launch in your browser ↗Frequency and association measures, study designs, bias, standardisation, screening, trials, survival and evaluation.
20 chapters2,100 quizzes with feedback
Launch in your browser ↗These public editions are separate from the personal applications. They include concepts, lessons, formulas, flashcards and quizzes; calculation exercises use new fictional data. Progress stays locally in this browser.
The library of knowledge
All references cited on this website, collected in APA format and alphabetical order.
1,003 references · APA
1003 references
No references match your search.
05.2 · Resources
An annotated selection to explore knowledge mobilisation and implementation science further, in French and English.
In English
In French
In English
In English
In French
In English
In French
In English
In French
In English
In English
In French
In English
In French
In English
In English
In French
In English
In French
In English
In French
In English
In French
In English
In French
In English
In French
In English
In French
In English
In French
In English
In French
In English
In French
In English
In French
In English
In French
In English
05.3 · Resources
In French
In French
In English
05.4 · Resources
In French
In English
05.5 · Resources
In French
A short presentation that traces where knowledge brokering comes from and how the role has evolved, as an introduction to the work of a knowledge broker.
Open the sourceReviews the competencies a person needs to work in knowledge brokering, a practical complement to the readings on the broker’s role.
Open the sourceA concrete case of research conducted with a citizen researcher on family peer support in mental health, showing what participatory science changes in the way a project is run.
Open the sourcePresents, in eight minutes, the CCNMO model, which shows how research evidence, local context, population preferences and resources come together in a public health decision.
Open the sourceCitizen–researcher pairs describe how they share roles in a research project launched by a citizen’s question, without either of them losing their own expertise.
Open the sourceA patient who became partnership lead at a research centre explains to a general audience what the patient-partner movement is and why patients’ lived experience should inform research and care.
Open the sourceIn English
An introductory lecture that distinguishes dissemination research, research dissemination, implementation research and pragmatic research, and presents the core concepts of the field.
Open the source05.6 · Resources
In French
A nursing researcher explains how she built a bridge between research and practice, notably through communities of practice with nurses.
Open the sourceIn English
The deputy director for implementation science at the National Cancer Institute reflects on the history of the field, the collective nature of science and what makes a research team effective.
Open the sourceAn accessible introduction in which a researcher presents implementation science as the study of planned behaviour change within organisational constraints, illustrated with clinical examples.
Open the sourceWhat if you could help Éric defend his PhD research proposal? Explore a 3D campus, face his supervisors, solve real methodological challenges, manage your working hours and survive Reviewer 2. Your choices have consequences, including a real game over.
Site tutorial video
This site reflects that ambition: a hall, six doors and rooms to explore. Here is how to visit it.
Contact
Have an idea, an interesting reading, an experience to share or a project to explore? I would be glad to hear from you.
School of Public Health, Université de Montréal · 7101 avenue du Parc, Montréal, Québec
