Oxford · Technology & Innovation in Digital Health
Candidacy · Summer 2026

Can a video game cure depression?

SPARX and the social shaping of digital mental health

A sociotechnical reading by Luke Fullagar — Founder, Ideal Nurture, here attending a sociology of technology and innovation in health course at the University of Oxford. This is the first assignment in the course deliverables — meant to be a group-work process, but I had to leave early for my son, so I decided to build a fresh assignment in full from first principles, for myself, as a group of one: to honour the team I couldn't deliver to, to respect the gravity of the unique opportunity to learn here offered by the faculty, and to honour myself, my son, and the future of whoever may be touched by what I learn here properly. The talk, the full analysis, a seven-domain NASSS deployment, and the process behind it are all rendered below.

▶  Watch the talk Downloads

In 2012 the headlines said yes. This is what happens when you take that claim apart.

SPARX, the social shaping of technology, and the NASSS implementation framework.  ·  Trouble playing? Watch on YouTube ↗

The throughline

A technology's effects are not properties of the software. They are produced — or lost — in the social and organisational world it enters.

How it was built

From evidence to argument

Twelve paper-notes, grouped into four argument strands, feed six Moment notes — which converge on a single throughline and the four-section synthesis behind the talk.

5
Lens
  • Pinch & Bijker 1984
  • Wyatt 2008
  • Borup 2006
  • Lupton 2014
  • Sætra 2024
social shaping of technology
1
Headline
  • Merry 2012
the SPARX RCT, re-read
4
Co-design
  • Arnstein 1969
  • Bate & Robert 2006
  • Lucassen 2014
  • Lucassen 2015
graded on two ladders
2
NASSS
  • Greenhalgh 2017
  • Berg 2001
implementation & complexity
↓
Six Moments
2.0 · Lens 3.0 · Headline 4.0 · Co-design 4.1 · Co-design (deep) 5.0 · NASSS 5.1 · NASSS (deep)
↓
The throughline

SPARX's effects are not properties of the software — they are produced, or fail to be, in the social and organisational world it enters.

↓
1 · Lens & overview
social shaping + Lupton
2 · Headline teardown
the RCT, re-read
3 · Co-design
graded on two ladders
4 · NASSS
complexity & scale-up

Colour = argument strand. Each paper-note links up to its Moment(s); backlinks make the chain two-way in Obsidian.

The instrument, deployed

SPARX through all seven NASSS domains

Each domain rated for complexity the way the eNASSS-CAT toolkit asks — simple, complicated, or complex. Slide through the deployment. The rule: the more domains that are complex, the less likely a technology is to be adopted, sustained, and scaled.

Domain 1 · The Condition

Adolescent depression

What is the nature of the illness — its co-morbidities and socio-cultural character?

Not bounded or simple. Highly co-morbid, culturally variable in expression and help-seeking, driven by structural forces no app touches. Scoped for mild-to-moderate; the real world sent ~47% sicker than the design target. A condition this socially embedded resists the tool's framing.

Complex
Domain 2 · The Technology

"Just a game"

Its material properties, dependability, and the data it generates.

Deceptively simple — but the material substrate bit hard: desktop-only at launch, slow downloads, device incompatibility, an awkward sign-up. Each a documented reason for non-completion. The "simplicity" is exactly what made it easy to under-resource.

Complicated → complex
Domain 3 · The Value Proposition

Value for whom?

Supply-side value to the funder vs demand-side value to the user.

The supply-side case was strong — scalable, cheap, evidence-based, politically attractive. The demand-side case broke: does a fifteen-year-old, alone at home, experience this as worth their evening, week after week? The 7.4% four-module figure is demand-side value collapsing.

Complex
Domain 4 · The Adopter System

Becoming a digital health citizen

Will staff, patients and carers adapt their identities and routines around it?

It asks an adolescent to become a self-managing "digital health citizen" (Lupton); clinicians to absorb new referral and monitoring work; parents to scaffold sustained use. That identity and role work is demanding, unpaid, and precisely what the completion collapse measures failing to occur. The heart of the failure.

Highly complex
Domain 5 · The Organisation(s)

Who does the weaving?

Organisational capacity, readiness, and changes to routine.

To land, an organisation must build it into a workflow — who recommends, who monitors, who catches deterioration, who runs the safety-net. Health IT succeeds only insofar as it is woven into everyday care (Berg). Unguided national rollout removed the human who would do the weaving.

Complex
Domain 6 · The Wider System

Game, or medical device?

Political, regulatory, professional and socio-cultural context.

Government funding was a real, rare advantage. But framing it as a game rather than a medical device quietly sidesteps the adverse-event obligations of an unguided mental-health intervention. Professional buy-in varies; the data-justice question — who owns the use data — sits unresolved.

Complex
Domain 7 · Embedding & Adaptation

Where digital therapeutics die

The scope to adapt locally and sustain as contexts change.

The field's deepest tension: evidence-based medicine demands a fixed, high-fidelity intervention; digital health demands continuous updating to stay usable. Those imperatives are at war. SPARX's successful variants prove flexibility is possible — but the mundane ongoing maintenance and funding are what deployment could not sustain.

Highly complex
The complexity tally

The verdict

1 · Condition Complex
2 · Technology Complicated→complex
3 · Value proposition Complex
4 · Adopter system Highly complex
5 · Organisation Complex
6 · Wider system Complex
7 · Embedding over time Highly complex

Six of seven domains sit in the complex band, two acutely. SPARX didn't fail despite being effective and well-designed — it failed in exactly the way the framework predicts a good tool fails when the world around it is complex.

Drag, swipe, or use the arrows · 8 panels

Take it with you

The deliverables

Everything behind the talk, in citable form.

PDF The analysis

The 1,000-word write-up

The throughline, argued section by section · APA-7 PDF The talk

The presentation deck

The slides used in the video XLS The instrument

eNASSS-CAT — SPARX

The complexity toolkit, filled in across all seven domains ▶ The presentation

Watch the video presentation

The full talk, on YouTube · ~25 min
Behind the analysis

The process — how this was actually made

Before the talk there were long days and nights of reading, note-making and argument-building in Obsidian — the early phases, the literature reviews, and the Moment-by-Moment drilldowns that became the four-section synthesis. This is the working that the polished deliverables sit on top of — and that forms the corpus of the way this process of discernment is being learned as well, not just the content.

Case study identificationsLiterature reviewThe four strandsSix MomentsDrilldownsSynthesis
Enter the deep-dive →
Go deeper

Around the work

The presentation

The video presentation

The full talk as delivered — the sociotechnical reading of SPARX, start to finish, on YouTube.

Watch on YouTube → The working

The process deep-dive

The early phases, the literature review, the six Moments and the four-section synthesis behind the talk.

Enter the deep-dive → The chapter

The Oxford chapter

Where this analysis sits in the wider journey — finding that it could be done, and turning to whether, and how, it should.

Back to Oxford →
Luke Fullagar
Luke Fullagar
Founder · Ideal Nurture

Built from the conviction that care lives in the world a technology enters — not in the code.

Luke Fullagar is the founder of Ideal Nurture, building therapeutic technology grounded in attachment science. This analysis was prepared for the University of Oxford's Technology & Innovation in Digital Health programme — a study in why effective, well-designed tools still succeed or fail in the social and organisational systems around them.

Private by design. Clinically grounded. Emotionally literate.