Severity domains for quantifying the severity of Type 2 diabetes using primary care data and linked hospitalisation records from England

Aims: Despite the increasing prevalence of Type 2 diabetes, whichpresents a significant burden on healthcare resources, there is noagreed validated UK measure to infer Type 2 diabetes severity fromelectronic health records (EHRs). This study aimed to identifyimportant clinical severity domains, recorded in EHRs, whichcould be used to quantify the severity of Type 2 diabetes.Methods: Data from the Clinical Practice Research Datalink(CPRD) and hospitalisation records for people with Type 2diabetes, registered with linked English general practices, to beused to develop a clinical decision algorithm to grade diabetesseverity between 2006 and 2016. At this stage of the study, wehave identified clinically relevant severity domains (main riskfactors for adverse outcomes) to be incorporated in the severityalgorithm.Results: The main severity domains for Type 2 diabetes identifiedthrough a systematic review and expert opinion include age,diabetes duration, HbA1c, hypoglycaemia, microvascular complications,cardiovascular (CV) disease (heart failure, coronary heartdisease, revascularisation interventions), cerebrovascular disease,renal disease, patterns of prescribed antidiabetic and cardiovasculartreatments, hospital admissions (any-cause, diabetes-relatedand CV disease-related admissions).Summary: All identified severity domains, recorded in routinelycollected real-world EHRs, are clinically relevant in our work todefine the severity of Type 2 diabetes. This first step forms a platformon which to develop the severity tool, which will be informative topractitioners, could stratify clinical management of people withType 2 diabetes, support commissioning and public health programmesand inform the methodology of measuring the severity ofother chronic conditions managed in primary care.