According to the UK’s NHS National Patient Safety Agency*, between 2005 and 2010, 45% of all cases of patient harm caused by a malpositioned nasogastric tube (NGT) were due to misinterpreted CXRs. In 2009 in the UK, complications of a malpositioned NGT became a ‘never event’ – a serious adverse incident which is wholly preventable if clinical safety protocols are in place. Annalise CXR provides the immediate assessment of NGT position to prevent such an incident.

*Patient Safety Alert NPSA/2011/PSA002: Reducing the harm caused by misplaced nasogastric feeding tubes in adults, children and infants. London: National Patient Safety Agency, 2011.

The Annalise CXR algorithm is trained on a broad dataset covering a spectrum of CXR machine equipment and manufacturers, departmental and portable imaging, patient demographics, and inpatient and outpatient groups.

The process of interpreting cases with multiple clinical findings lead to complex associations between findings, known as hidden stratification. This is inherent to any model examining multiple findings, and robust examination of the interplay of findings allows for validation of the model and improved model performance. Annalise CXR is the first solution to analyse the hidden stratification effect of a subset of clinically relevant findings.