Seven things a study would need to be the study this thesis proposes.

D1
D1

Healthy adults 50+

Community-dwelling, non-patient adults aged 50 or older.

D2
D2

Six months or longer

The intervention or its follow-up lasts ≥ 6 months, long enough to see past the novelty window.

D3
D3

State and context as input

The system changes what it does based on the person's current situation.

D4
D4

Behaviour theory

A named theory or framework shapes the intervention.

D5
D5

LLM delivers

A large language model generates the user-facing message or dialogue.

D6
D6

JITAI / adaptive timing

Decision points and tailoring variables decide when to intervene.

D7
D7

Closed-loop calibration

The person's response changes how much the system asks next time.

Footnote: D6 and D7 are different claims: when versus how much. A system can read "you dismissed the last three nudges" (D3-F) without changing how hard it pushes next time (D7).
See where the 138 papers sit →