A state transition modelling approach, either as a cohort or individual patient-level simulation (IPS), is preferred (recommended). However, when alternative modelling approaches are used, these must be transparent and validated and capture all important differences in costs or outcomes between the intervention and comparator. (required)
Cycle length must not exceed 1 year. Time in each health state including associated costs and utilities will generally be half-cycle corrected. Front-loading of some high-cost items such as admissions or surgical procedures must be done if this is the only way to capture the acute resource use associated with the transition. (required)
Carefully select outcomes (health states and events) for each population group (stratum) guided by the following criteria (required):
direct evidence of a treatment effect (from trials or real-world evidence), or strong evidence of an association of an improved outcome with weight loss and clinical plausibility that the intervention affects the outcome and
evidence that the outcome has a meaningful impact on costs, quality of life or risk of other outcomes (for example, mortality)
their inclusion within a high-quality core outcome set for that stratum, as these indicate outcomes that are important to patients and clinicians.
Capture health states in the model as follows:
type 2 diabetes mellitus (T2DM) status: 'normoglycaemia' or 'non-diabetic hyperglycaemia' or 'T2DM' (required)
atherosclerotic cardiovascular disease (ASCVD) acute events: myocardial infarction 'MI' or 'stroke' (ischaemic stroke or unspecified stroke) or 'stroke after MI' or 'MI after stroke' (required)
ASCVD status: 'no ASCVD' or 'post-myocardial infarction' ('post-MI') or 'post-stroke' or 'post-stroke and post-MI' or 'ASCVD other' or 'post-MI and chronic heart failure' ('post-MI and CHF') (required)
chronic kidney disease (CKD) status: 'no CKD' or 'CKD stage G1A2 to G4' or 'CKD stage G5' (recommended)
weight category: distinct health state ('healthy weight' or 'overweight' or 'obesity class 1' or 'obesity class 2' or 'obesity class 3') or using more granular weight or body mass index (BMI) trajectories (required)
if applicable, line of weight management treatment – from line 1 to line 2 to line 3, etc (required)
a 'dead' state (required).
Strata containing people with T2DM at baseline should have additional health states containing diabetes-related core outcomes. (recommended)
Capture movement between health states as follows:
remission and recurrence of non-diabetic hyperglycaemia (recommended)
progression of T2DM status (from no diabetes to non-diabetic hyperglycaemia to diabetes) (recommended)
progression of ASCVD status
from 'no ASCVD' or 'ASCVD other' to 'post-stroke' or 'post-MI' (recommended)
from 'post-stroke' or 'post-MI' to 'post-stroke and post-MI' (recommended)
from 'post-MI' to 'post-MI and CHF' (recommended)
progression of CKD status (from 'no CKD' to 'CKD stage G1A2 to G4' to 'CKD stage G5') (recommended)
change in weight category (required)
all health states transition to the 'dead' state (required).
Key model transitions are illustrated in figure 1.
|
Definitions: CHF: chronic heart failure; MI: myocardial infarction.
Calculate costs and QALY losses for the following health events:
obstructive sleep apnoea (recommended)
osteoarthritis requiring knee replacement (recommended)
osteoarthritis requiring hip replacement (recommended)
bariatric procedure (unless this is an intervention or comparator in the model), with costs and utility decrements reflecting a weighted average of different procedures (recommended)
treatment and procedure-related adverse effects (non-serious adverse effects with an incidence of more than 5% for each intervention and any serious adverse effects) (required)
discontinuation (required).
The following outcomes should be considered for inclusion as health states in a sensitivity analysis if there is direct evidence of a treatment effect:
metabolic dysfunction-associated steatotic liver disease (MASLD) (recommended)
cancer (breast, colorectal, kidney, liver and womb). (recommended)
For the rationale for these statements, see the rationale and supporting information section on model structure and health states.