Long-term prognosis for patients admitted to a medical intermediate care unit. / Langtidsprognose for pasienter innlagt i en medisinsk overvåkningsavdeling.

BACKGROUND: The intermediate care unit at Akershus University Hospital treats patients with incipient or manifest organ failure. Selecting patients who might benefit from treatment in an intermediate care unit is challenging. Few data are available on long-term survival of patients treated in medical intermediate care units and on how assumed favourable and unfavourable prognostic factors predict long-term survival in this population. MATERIAL AND METHOD: Comorbidity, reason for admission and whether an infection was a direct or contributory reason for the admission were prospectively registered for patients in the unit in 2014 and 2016. We registered mortality up to six years after the admission and conducted a logistic regression analysis with three-year survival as the outcome variable. RESULTS: Of the 2 170 included patients, 153 (7 %) died in the intermediate care unit. Of the 2 017 patients who were discharged alive from the intermediate care unit, 55 % were still alive three years later, including 28 % of older patients aged over 80 years and 23 % of patients with cancer. Age, malignancy, other comorbidity and infection were predictors of mortality. INTERPRETATION: Many patient groups in an intermediate care unit have a poor long-term prognosis. However, people older than 80 years, cancer patients or patients with another serious comorbidity may live long after their stay in an intermediate care unit, and the fact of belonging to these groups should not be an independent reason for withholding treatment.

Neural classification of Norwegian radiology reports: using NLP to detect findings in CT-scans of children.

BACKGROUND: With a motivation of quality assurance, machine learning techniques were trained to classify Norwegian radiology reports of paediatric CT examinations according to their description of abnormal findings. METHODS: 13.506 reports from CT-scans of children, 1000 reports from CT scan of adults and 1000 reports from X-ray examination of adults were classified as positive or negative by a radiologist, according to the presence of abnormal findings. Inter-rater reliability was evaluated by comparison with a clinician's classifications of 500 reports. Test-retest reliability of the radiologist was performed on the same 500 reports. A convolutional neural network model (CNN), a bidirectional recurrent neural network model (bi-LSTM) and a support vector machine model (SVM) were trained on a random selection of the children's data set. Models were evaluated on the remaining CT-children reports and the adult data sets. RESULTS: Test-retest reliability: Cohen's Kappa = 0.86 and F1 = 0.919. Inter-rater reliability: Kappa = 0.80 and F1 = 0.885. Model performances on the Children-CT data were as follows. CNN: (AUC = 0.981, F1 = 0.930), bi-LSTM: (AUC = 0.978, F1 = 0.927), SVM: (AUC = 0.975, F1 = 0.912). On the adult data sets, the models had AUC around 0.95 and F1 around 0.91. CONCLUSIONS: The models performed close to perfectly on its defined domain, and also performed convincingly on reports pertaining to a different patient group and a different modality. The models were deemed suitable for classifying radiology reports for future quality assurance purposes, where the fraction of the examinations with abnormal findings for different sub-groups of patients is a parameter of interest.