70% total loss + 30% loss beyond placebo. We convert each percentage to points using the scales below, combine the points, then round to a whole number. Headline percentages use one decimal place; calculations and study details retain the source values. Values between anchors receive proportionate points; for example, 22.5% total loss earns 88.5 points. Loss of 30% or more earns 100; no loss or weight gain earns zero for that component.
Percentage loss → points out of 100| Loss | Total-loss points | Beyond-placebo points |
|---|
| 0% | 0 | 0 |
|---|
| 5% | 20 | 25 |
|---|
| 10% | 40 | 50 |
|---|
| 15% | 62 | 72 |
|---|
| 20% | 82 | 90 |
|---|
| 25% | 95 | 98 |
|---|
| 30% | 100 | 100 |
|---|
One matched result per program. Treatment and placebo must refer to the same trial, dose, population, week and analysis. A published adjusted difference takes precedence over subtracting rounded means. Placebo weight gain increases the difference. Total loss and a placebo comparison are both required; a missing value never becomes zero.
Which result? We prefer treatment-policy analyses, which include treatment stops; next, efficacy analyses assuming continued treatment; then analyses whose treatment-stop assumptions are not fully specified. Within that basis, we prefer the highest trial phase and longest reported endpoint. The reviewed dose is identified in study details. We do not select a newer early trial simply because it has a bigger headline result.
Time is shown, not projected. A 13-week result is not extrapolated to a year. The “68–84 wk data” group includes adults without diabetes and accounts for people who stopped treatment. “Other trial data” covers shorter studies, longer extensions, less specified populations or different analyses. Results from diabetes-only, enriched, withdrawal, add-on or active-comparator settings need their own comparison group and do not receive a score in this general obesity ranking.
What the score means. This is a Saddl3-designed index of weight loss, not a validated clinical scale or an overall treatment recommendation. Its two inputs overlap: both contain the treatment arm’s weight loss. The placebo component accounts partly for the trial’s background response; it does not make separate trials equivalent. A one-point difference or rank is not evidence of statistical superiority. Whole-number ties share the same score and are ordered alphabetically.
Safety, treatment discontinuation, body composition, dosing, access and cardiovascular outcomes require separate judgment. Trial phase and publication type describe the evidence rather than changing the score. Confidence intervals and adverse-event discontinuation are shown where verified; missing details are labeled. A program awaiting a score may have public results that have not yet been verified here.