Deployment and validation of the CLL treatment infection model adjoined to an EHR system.

Research algorithms are seldom externally validated or integrated into clinical practice, leaving unknown challenges in deployment. In such efforts, one needs to address challenges related to data harmonization, the performance of an algorithm in unforeseen missingness, automation and monitoring of predictions, and legal frameworks. We here describe the deployment of a high-dimensional data-driven decision support model into an EHR and derive practical guidelines informed by this deployment that includes the necessary processes, stakeholders and design requirements for a successful deployment. For this, we describe our deployment of the chronic lymphocytic leukemia (CLL) treatment infection model (CLL-TIM) as a stand-alone platform adjoined to an EPIC-based Danish Electronic Health Record (EHR), with the presentation of personalized predictions in a clinical context. CLL-TIM is an 84-variable data-driven prognostic model utilizing 7-year medical patient records and predicts the 2-year risk composite outcome of infection and/or treatment post-CLL diagnosis. As an independent validation cohort for this deployment, we used a retrospective population-based cohort of patients diagnosed with CLL from 2018 onwards (n = 1480). Unexpectedly high levels of missingness for key CLL-TIM variables were exhibited upon deployment. High dimensionality, with the handling of missingness, and predictive confidence were critical design elements that enabled trustworthy predictions and thus serves as a priority for prognostic models seeking deployment in new EHRs. Our setup for deployment, including automation and monitoring into EHR that meets Medical Device Regulations, may be used as step-by-step guidelines for others aiming at designing and deploying research algorithms into clinical practice.

Protocol for a randomised controlled trial comparing warfarin with no oral anticoagulation in patients with atrial fibrillation on chronic dialysis: the Danish Warfarin-Dialysis (DANWARD) trial.

INTRODUCTION: Atrial fibrillation is highly prevalent in patients on chronic dialysis. It is unclear whether anticoagulant therapy for stroke prevention is beneficial in these patients. Vitamin K-antagonists (VKA) remain the predominant anticoagulant choice. Importantly, anticoagulation remains inconsistently used and a possible benefit remains untested in randomised clinical trials comparing oral anticoagulation with no treatment in patients on chronic dialysis. The Danish Warfarin-Dialysis (DANWARD) trial aims to investigate the safety and efficacy of VKAs in patients with atrial fibrillation on chronic dialysis. The hypothesis is that VKA treatment compared with no treatment is associated with stroke risk reduction and overall benefit. METHODS AND ANALYSIS: The DANWARD trial is an investigator-initiated trial at 13 Danish dialysis centres. In an open-label randomised clinical trial study design, a total of 718 patients with atrial fibrillation on chronic dialysis will be randomised in a 1:1 ratio to receive either standard dose VKA targeting an international normalised ratio of 2.0-3.0 or no oral anticoagulation. Principal analyses will compare the risk of a primary efficacy endpoint, stroke or transient ischaemic attack and a primary safety endpoint, major bleeding, in patients allocated to VKA treatment and no treatment, respectively. The first patient was randomised in October 2019. Patients will be followed until 1 year after the inclusion of the last patient. ETHICS AND DISSEMINATION: The study protocol was approved by the Regional Research Ethics Committee (journal number H-18050839) and the Danish Medicines Agency (case number 2018101877). The trial is conducted in accordance with the Helsinki declaration and standards of Good Clinical Practice. Study results will be disseminated to participating sites, at research conferences and in peer-reviewed journals. TRIAL REGISTRATION NUMBERS: NCT03862859, EUDRA-CT 2018-000484-86 and CTIS ID 2022-502500-75-00.

Maintenance of indwelling urinary catheters with a novel polyhexanide-based solution: user experience.

BACKGROUND: Catheter-associated urinary tract infection (CAUTI) can significantly affect patients' quality of life and increase healthcare costs. AIMS: This study aimed to capture patients' and nurses' experience of catheter maintenance using a polyhexanide-based solution (PS) in everyday practice. METHODS: Retrospective analysis of data was collected for a product evaluation. PS was used twice a week for five weeks. FINDINGS: The study included 42 patients, 30 (71%) men and 12 women (29%). After five weeks of rinsing catheters with PS, nine patients reported no or decreased frequency of CAUTI, eight a better quality of life, eight reduced blockage, seven a decrease in odour and five fewer catheter changes. Three patients reported no benefit from PS use. Nurses reported that fewer visits were needed and consumption of disposables was lower. CONCLUSIONS: User experiences suggest that, as a novel means of catheter maintenance, PS has the potential to reduce catheter-associated complications such as CAUTI, improve quality of life and reduce healthcare costs.

Apolipoprotein M in patients with chronic kidney disease.

BACKGROUND AND AIMS: Plasma apolipoprotein M (APOM) is bound to HDL-particles and has anti-atherogenic effects. The present study explored whether plasma APOM is reduced in patients with chronic kidney disease (CKD), and associated with cardiovascular disease (CVD). In addition, we tested the hypothesis that the excretion of APOM into the urine is increased in patients with kidney disease. METHODS: Plasma samples were collected from a cohort of patients with CKD stages 1 to 5D (N = 409) and controls (N = 35). Urine was collected from 47 subjects. Plasma APOM was measured with sandwich ELISA and urine APOM with competitive ELISA. RESULTS: Plasma APOM levels were reduced in patients with CKD stages 3-5D as compared to patients with CKD stages 1 + 2 and controls (p < 0.01). CKD patients with known CVD displayed even further reduction in plasma APOM levels than CKD patients without known CVD (p < 0.001). Fast-phase liquid chromatography showed that plasma APOM was primarily associated with HDL-cholesterol (HDL-C) across CKD stages. Accordingly, when plasma APOM values were corrected for HDL-C, a significant difference only persisted between patients with CKD stage 3 and stages 1 + 2 (p < 0.05), and the difference between CKD patients with and without known CVD disappeared. Urine APOM/creatinine ratio was not significantly increased in patients with kidney disease. CONCLUSIONS: The results show that the difference in plasma APOM levels observed between patients with mild and advanced CKD may mainly be due to differences in plasma HDL-C. Whether APOM plays a role in human uremic atherogenesis warrants further experimental studies.

Predicting apple (cv. Elshof) postharvest dry matter and soluble solids content with near infrared spectroscopy.

BACKGROUND: Fruit dry matter (DM) and soluble solids content (SSC) are primarily composed of carbohydrate and are standard parameters for assessing quality. Near infrared spectroscopy provides potential for non-destructive fruit quality analysis but the collinearity between DM and SSC is an issue for prediction. Shorter wavelength spectra have been used for the prediction of fruit DM and SSC, but radiation between 1000 and 2500 nm may be suitable for distinguishing between the two forms of carbohydrate. RESULTS: Spectra and DM and SSC samples were taken for a total of 450 'Elshof' apples 30, 58 and 93 days after harvest. Regression models were built using the interval partial least squares method. Prediction models for DM and SSC for each day yielded R² values between 0.63 and 0.86 and residual predictive deviations (RPDs) between 1.7 and 2.7 for DM, and R² = 0.76-0.85 and RPDs = 2.2-2.6 for SSC. CONCLUSION: Model RPD values were not high enough for general quantitative predictions, although they compare well to previous work. Certain factors affected model success, including changes in fruit physiology over time and the range of reference data. The complexity of absorbance spectra for DM and SSC plus their strong correlation suggests that prediction models cannot easily distinguish between soluble and non-soluble forms of carbohydrate.