Habit Formation

4 claims that cleared review. Each one shows what has to be true for it to hold and how strongly it is supported — including the places the sources contradict each other.

In adults at high risk of type 2 diabetes, an intensive lifestyle intervention achieving roughly 7% weight loss is more effective at preventing progression to diabetes than medication.

Losing about 7% of body weight through diet and exercise changes cuts the risk of developing type 2 diabetes more than medication does, in people at high risk.

Only when
  • specific figure: 7% weight loss through lifestyle intervention reduced type 2 diabetes onset by 58% in the Diabetes Prevention Program (DPP) / DPPOS
  • lifestyle intervention also offered longer-term and broader health benefits beyond diabetes prevention alone, compared with medication
  • see medication_interaction--bd64 (same underlying DPP/DPPOS trial) for the metformin comparison arm and 10-year weight-durability finding
Who

adults at high risk of type 2 diabetes (prediabetes / insulin resistance)

Moderate confidence

Both members cite the same landmark trial (the Diabetes Prevention Program) rather than being independent replications of the finding -- this is a single-study result reported by two independent secondary voices, not two separate studies confirming each other. Confidence held at moderate rather than higher on that basis, despite the underlying evidence being RCT-level.

2 independent people · 2 sources · strongest evidence: rct

Sources
  • Comparative effectiveness of different exercise modality on glycaemic control and lipid profile for prediabetes: systematic review and network meta-analysis. · Frontiers in endocrinology · 2025 · doi:10.3389/fendo.2025.1518871

Real-world carbohydrate intake on a ketogenic diet tends to drift substantially upward over 12 months even among trial participants who started very low-carbohydrate, indicating adherence erodes over time without active reinforcement.

Even people who start a keto diet very strictly tend to drift back toward higher carb intake over about a year if nothing actively keeps them on track. Planning regular check-ins helps counter that normal drift.

Only when
  • A TO Z Weight Loss Study: keto-diet arm rose from <10 g/day carbohydrate at start to 34% of total energy intake from carbohydrate at 12 months
  • DIRECT trial: keto-diet arm rose from 20 g/day carbohydrate at start to 40% of total energy intake from carbohydrate at 12 months
  • both drawn from an umbrella review's discussion of adherence in named randomized trials
Who

adults enrolled in randomized trials assigned to a ketogenic-diet arm

Moderate confidence

Drawn from an umbrella review citing two named RCTs, but attributed to a single first-author voice (the review's collective authorial byline) rather than two independently confirming sources; RCT-tier evidence keeps it above low.

1 independent person · 2 sources · strongest evidence: rct

Sources
  • Effects of ketogenic diet on health outcomes: an umbrella review of meta-analyses of randomized clinical trials. · BMC medicine · 2023 · doi:10.1186/s12916-023-02874-y

In a 12-month digital habit-formation app for adults at increased risk of type 2 diabetes, engagement splits into distinct trajectories: roughly half sustain some use through the year, but nearly half drop to near-zero use by 6-7 months.

About half of people using a habit-tracking app like this stop engaging by 6-7 months. That's normal, not a personal failure — a check-in around that point can help you get back on track.

Only when
  • 53.1% of participants (1022/1926) continued using the app throughout the full 12-month intervention
  • four engagement trajectories identified: terminated use, dropping to near zero by 6-7 months (46.9%, 904/1926); weekly use (38%, 731/1926); twice-weekly use (10.8%, 208/1926); daily use (4.3%, 83/1926)
Who

adults at increased risk of type 2 diabetes using a digital habit-formation app

Moderate confidence

Observational data from a single first-author cohort study (n=1926), not replicated by an independent group; large sample size and clear trajectory analysis support moderate rather than low confidence.

1 independent person · 2 sources · strongest evidence: observational

Sources
  • Digitally Supported Lifestyle Intervention to Prevent Type 2 Diabetes Through Healthy Habits: Secondary Analysis of Long-Term User Engagement Trajectories in a Randomized Controlled Trial. · Journal of Medical Internet Research · 2022 · doi:10.2196/31530

Stopping use of a habit-tracking app does not automatically mean someone has abandoned the habit, but in one digital lifestyle-intervention study, people who stopped had selected fewer habits and showed no improvement in diabetes risk factors, suggesting most who quit had disengaged rather than internalized the habit.

If someone stops using a habit-tracking app, it doesn't automatically mean they gave up -- but in one study, most people who quit really had disengaged, not just outgrown the app. Worth checking in rather than assuming.

Only when
  • General possibility (author assertion): app termination can reflect either genuine disengagement or successful internalization of the habit without further need for the app
  • Study-specific finding (observational, this study's terminated-usage group): terminators had selected fewer habits than continued users and showed no improvement in type 2 diabetes risk factors, which favors the disengagement explanation over internalization in this population
Who

adults at increased risk of type 2 diabetes using a digital habit-formation app

Moderate confidence

Single voice (lavikainen-p); the stronger, study-specific observational finding (c3) outranks the accompanying author-assertion caveat (c2) and sets the evidence floor, but one study from one source caps confidence below high.

1 independent person · 2 sources · strongest evidence: observational

Sources
  • Digitally Supported Lifestyle Intervention to Prevent Type 2 Diabetes Through Healthy Habits: Secondary Analysis of Long-Term User Engagement Trajectories in a Randomized Controlled Trial. · Journal of Medical Internet Research · 2022 · doi:10.2196/31530

This is a summary of published research, not medical advice. It cannot account for your medications, conditions or history. Talk to a clinician before acting on any of it.