Discovery of clinical biomarkers for the prediction of delirium in older people admitted to the acute hospital

Confusion is common in older people admitted to hospital. People with confusion have poor outcomes following admission to hospital; they can experience a reduction in their ability to undertake everyday tasks, longer hospital stays and increased risk of death. For the clinical team, the differential diagnosis of confusion is fraught with difficulty due to the overlapping picture of symptoms. Confusion is generally due to either delirium (a rapid, fluctuating and potentially reversible change in mental state), preexisting dementia, delirium superimposed on dementia or un-diagnosed dementia. Distinguishing between these conditions is critical in order to tailor a management plan and advise on prognosis, and yet currently this diagnostic process relies heavily on clinical history alone, undertaken by an experienced specialist. The unprecedented access to patient medical records makes it now possible to develop algorithms capable of extracting important features indicative of delirium in hospitalized patients. Patient information contained in electronic medical records such as longitudinal laboratory tests and drug prescribing together with information on co-morbidity can be potentially used to aid distinguishing delirium from the symptoms of dementia in hospital settings, thereby increasing the accuracy of clinical diagnoses able to be delivered by emergency care teams. The study will use state-of-the-art data analytics on anonymised routine healthcare patients records, to identify clinical factors which may support the clinical diagnosis of delirium in older people admitted to the acute hospital.