Two rare and severe disorders of insulin action, namely insulin autoimmune syndrome (IAS) and type B insulin resistance (TB-IR), are caused by pathogenic antibodies against insulin or the insulin receptor, respectively. These may arise in isolation or may complicate management of pre-existing diabetes mellitus, and milder forms of the conditions are often suspected in patients with insulin-treated diabetes and labile glycaemic control. Antibody depletion can effectively treat either condition in many cases. This research aimed to target major limitations of existing diagnostics, specifically, that anti-insulin antibody (IA) testing alone does not establish whether antibodies alter insulin action to a clinically-significant degree, and that no clinically-accredited diagnostic test for TB-IR currently exists. An initial collaborative study examined the ability of a panel of commercial insulin assays to quantify ten different insulin preparations. Significant variability in performance of assays against animal-derived and insulin analogues was seen, with certain insulin analogues not detected at all, with important implications for the use of insulin immunoassays in insulin-treated patients. A suite of techniques for investigation of the clinical significance of IAs were then developed and assessed. In a study of five widely-used insulin immunoassays, dilution of IAS plasma led to increased insulin recovery, and polyethylene glycol (PEG) precipitation of IAS plasma decreased insulin recovery in the majority of assays. Gel filtration chromatography (GFC) discriminated high molecular weight and monomeric insulin, while ex vivo addition of exogenous insulin to plasma increased sensitivity of insulin immunocomplex detection. An observational study was performed of 7 patients, all ultimately diagnosed with IAS. IAs were measured using radioligand-binding assay and enzyme-linked immunosorbent assay (ELISA). Method comparison showed results to differ in rank order and relative magnitude. For one patient whose screening IA result was not grossly elevated using either IA assay method, immunosubtraction studies were consistent with the presence of an IgA, a class of antibody under-/not detected in the IA assays studied. Competitive radioligand-binding studies demonstrated IAs to have a range of affinities. 4 patients treated with individualised regimens of immunosuppressive therapy varied in clinical response, and 3 were managed conservatively. Plasma insulin and C-peptide measurements made using mass spectrometry demonstrated under-estimation of insulin and over-estimation of C-peptide concentration using immunoassay in IAS. An observational laboratory and clinical study was also undertaken of 30 insulin-treated patients with diabetes and unexplained labile glycaemia. IA, and plasma insulin before and after PEG precipitation, were determined. Three groups were identified: the first were ‘negative’ for actionable IA; the second had demonstrable IAs of potential significance that warrant further study; and the third included 3 patients for whom immunomodulation therapy was indicated, with 1 other patient showing marked improvement of glycaemic control with close supervision and manipulation of insulin. Finally, anti-insulin receptor antibodies were detected using a newly developed ELISA utilising Chinese hamster ovary-expressed myc-tagged wild-type human insulin receptor. ‘Proof of principle’ was demonstrated for the new assay, with clear scope established for future diagnostic development. The ability to robustly prove, or conversely to rule out, the presence, of insulin–antibody complexes and/or anti-insulin receptor antibodies is invaluable in the investigation of patients with insulin resistance and/or unexplained labile glycaemia, and may decisively alter care pathways. Knowledge gained by this research has advanced understanding of the limitations of current laboratory diagnostics, and has thereby aided clinical-decision making for affected patients.