Homeostatic model assessment (HOMA) is a structural model of fasting glucose–insulin feedback, first described by Matthews and colleagues in 1985 to estimate insulin resistance and β-cell function from a single fasting sample. In the original work, HOMA-derived insulin resistance correlated well with euglycaemic clamp and CIGMA measures, but coefficients of variation for individual estimates were around 30%, limiting precision for single-patient assessment.

Subsequent Oxford work produced the HOMA2 computer model and later iHOMA2 tools, which better reflect non-linear physiology and modern insulin assays. These publications emphasise that the simple HOMA-IR equation is an approximation, best suited to epidemiological and cohort studies and to tracking change over time within a person, rather than as a stand-alone diagnostic test.

Reference ranges and “insulin resistance” cut-offs for HOMA-IR are not standardised. Some PCOS-focused guidance for adult women describes a normal range of about 0.7–2.0 with ≥2.0 suggesting increased insulin resistance, while a large Mexican-American cohort proposed <2.60 as normal, 2.60–3.80 as borderline high and >3.80 as high, based on clustering of metabolic risk factors. Local laboratory reference data and population mix therefore matter when interpreting absolute values.

Source / population	Illustrative HOMA-IR ranges	Comment
PCOS guidance (adult women)	≈0.7–2.0 normal; ≥2.0 higher insulin resistance	Used in polycystic ovary syndrome metabolic assessment; not generalisable to all adults.
Cameron County Hispanic Cohort (Mexican-American adults)	<2.60 normal; 2.60–3.80 borderline high; >3.80 high	Cut-offs derived against clustered metabolic risk factors in 1854 adults.

Ethnic differences and assay variation further limit universal cut-offs. In some African-American, Afro-Caribbean and South-Asian cohorts, HOMA-IR correlates less well with clamp-derived insulin sensitivity than other indices, so it is often best treated as a continuous fasting index interpreted alongside fasting plasma glucose, HbA1c, oral glucose tolerance testing and broader cardiometabolic risk assessment.

This implementation is intended for non-pregnant adults with suspected insulin resistance or cardiometabolic risk (for example obesity, metabolic syndrome, NAFLD, PCOS or prediabetes) in whom fasting glucose and insulin have been measured. It is less reliable in type 1 diabetes, advanced β-cell failure, pregnancy and in patients treated with exogenous insulin, where fasting insulin does not reflect endogenous secretion.

The primary source or current guidance provides the model definition, source population and supporting evidence.

HOMA-IR (Homeostatic Model Assessment of Insulin Resistance)