Why digital nudges stop working — and what it would take to make them last.

Engagement over six months: typical digital intervention versus the hypothesised behaviour-science-grounded JITAI Both lines rise in the first weeks. The grey line, a typical digital intervention, decays inside the novelty window (first three months) and is rarely measured after it, shown dotted. The orange line, the hypothesised intervention, stays level to six months. The claim is tested by what happens after month three. Typical digital intervention (literature) Hypothesis: behaviour-science JITAI NOVELTY WINDOW · weeks to ~3 months 0123456 Months since start Engagement / adherence sustained to ≥ 6 months — the test typical: decays — rarely measured past month 3 (dotted)

Motivation: Longer healthspan for the population, through upstream prevention in healthy adults aged 50+.

Hypothesis: Digital nudges grounded in behavioural science can produce sustained lifestyle change — persisting beyond the novelty window — and do so economically, without human coaching hours or cash incentives.

Engagement

Engagement — people keep responding to the nudges (telemetry: prompts answered, dialogue turns, app opens).

Adherence

Adherence — people keep doing the behaviour (wearable data: self-selected weekly activity goals met).

Sustainability

Sustainability — both hold over ≥ 6 months, and the shape of the curve is reported, not just the endpoint.

A PRISMA-ScR scoping review charts every eligible study on seven binary dimensions and reports coverage, co-occurrence (the UpSet plot on the Gap map page) and the empty cells. It cannot test the hypothesis; it shows where the evidence is, and where it is not. The last paragraph of the paper states the missing study.

Framework demonstration — coding is machine-assisted and not yet human reviewed (0 of 138). Counts will change as papers are checked.