Insights from explainable AI in oesophageal cancer team decisions.

Background: Clinician-led quality control into oncological decision-making is crucial for optimising patient care. Explainable artificial intelligence (XAI) techniques provide data-driven approaches to unravel how clinical variables influence this decision-making. We applied global XAI techniques to examine the impact of key clinical decision-drivers when mapped by a machine learning (ML) model, on the likelihood of receiving different oesophageal cancer (OC) treatment modalities by the multidisciplinary team (MDT).

Evaluation of thyroid function monitoring in people treated with lithium: Advice based on real-world data.

Introduction: Blood test monitoring is essential for the management of lithium treatment and National Institute for Health and Care Excellence guidance recommends 6-monthly serum testing of thyroid function. We examined conformity to these guidelines and the impact of monitoring outside these intervals.

Methods: We extracted serum lithium and thyroid hormone results at one centre between January 2009 and December 2020.

Can we check serum lithium levels less often without compromising patient safety?

Background: Lithium is viewed as the first-line long-term treatment for prevention of relapse in people with bipolar disorder.

Aims: This study examined factors associated with the likelihood of maintaining serum lithium levels within the recommended range and explored whether the monitoring interval could be extended in some cases.

Method: We included 46 555 lithium rest requests in 3371 individuals over 7 years from three UK centres.

Serum lithium test requesting across three UK regions: an evaluation of adherence to monitoring guidelines.

Background: Bipolar disorder is the fourth most common mental health condition, affecting ~ 1% of UK adults. Lithium is an effective treatment for prevention of relapse and hospital admission, and is widely recommended as a first-line treatment. We previously showed in other areas that laboratory testing patterns are variable with sub-optimal conformity to guidance.