Enhancing Patient Understanding of Laboratory Test Results: Systematic Review of Presentation Formats and Their Impact on Perception, Decision, Action, and Memory.

BACKGROUND: Direct access of patients to their web-based patient portal, including laboratory test results, has become increasingly common. Numeric laboratory results can be challenging to interpret for patients, which may lead to anxiety, confusion, and unnecessary doctor consultations. Laboratory results can be presented in different formats, but there is limited evidence regarding how these presentation formats impact patients' processing of the information. OBJECTIVE: This study aims to synthesize the evidence on effective formats for presenting numeric laboratory test results with a focus on outcomes related to patients' information processing, including affective perception, perceived magnitude, cognitive perception, perception of communication, decision, action, and memory. METHODS: The search was conducted in 3 databases (PubMed, Web of Science, and Embase) from inception until May 31, 2023. We included quantitative, qualitative, and mixed methods articles describing or comparing formats for presenting diagnostic laboratory test results to patients. Two reviewers independently extracted and synthesized the characteristics of the articles and presentation formats used. The quality of the included articles was assessed by 2 independent reviewers using the Mixed Methods Appraisal Tool. RESULTS: A total of 18 studies were included, which were heterogeneous in terms of study design and primary outcomes used. The quality of the articles ranged from poor to excellent. Most studies (n=16, 89%) used mock test results. The most frequently used presentation formats were numerical values with reference ranges (n=12), horizontal line bars with colored blocks (n=12), or a combination of horizontal line bars with numerical values (n=8). All studies examined perception as an outcome, while action and memory were studied in 1 and 3 articles, respectively. In general, participants' satisfaction and usability were the highest when test results were presented using horizontal line bars with colored blocks. Adding reference ranges or personalized information (eg, goal ranges) further increased participants' perception. Additionally, horizontal line bars significantly decreased participants' tendency to search for information or to contact their physician, compared with numerical values with reference ranges. CONCLUSIONS: In this review, we synthesized available evidence on effective presentation formats for laboratory test results. The use of horizontal line bars with reference ranges or personalized goal ranges increased participants' cognitive perception and perception of communication while decreasing participants' tendency to contact their physicians. Action and memory were less frequently studied, so no conclusion could be drawn about a single preferred format regarding these outcomes. Therefore, the use of horizontal line bars with reference ranges or personalized goal ranges is recommended to enhance patients' information processing of laboratory test results. Further research should focus on real-life settings and diverse presentation formats in combination with outcomes related to patients' information processing.

Deep embedded clustering generalisability and adaptation for integrating mixed datatypes: two critical care cohorts.

We validated a Deep Embedded Clustering (DEC) model and its adaptation for integrating mixed datatypes (in this study, numerical and categorical variables). Deep Embedded Clustering (DEC) is a promising technique capable of managing extensive sets of variables and non-linear relationships. Nevertheless, DEC cannot adequately handle mixed datatypes. Therefore, we adapted DEC by replacing the autoencoder with an X-shaped variational autoencoder (XVAE) and optimising hyperparameters for cluster stability. We call this model "X-DEC". We compared DEC and X-DEC by reproducing a previous study that used DEC to identify clusters in a population of intensive care patients. We assessed internal validity based on cluster stability on the development dataset. Since generalisability of clustering models has insufficiently been validated on external populations, we assessed external validity by investigating cluster generalisability onto an external validation dataset. We concluded that both DEC and X-DEC resulted in clinically recognisable and generalisable clusters, but X-DEC produced much more stable clusters.

Machine learning for risk stratification in the emergency department (MARS-ED) study protocol for a randomized controlled pilot trial on the implementation of a prediction model based on machine learning technology predicting 31-day mortality in the emergency department.

BACKGROUND: Many prediction models have been developed to help identify emergency department (ED) patients at high risk of poor outcome. However, these models often underperform in clinical practice and their actual clinical impact has hardly ever been evaluated. We aim to perform a clinical trial to investigate the clinical impact of a prediction model based on machine learning (ML) technology. METHODS: The study is a prospective, randomized, open-label, non-inferiority pilot clinical trial. We will investigate the clinical impact of a prediction model based on ML technology, the RISKINDEX, which has been developed to predict the risk of 31-day mortality based on the results of laboratory tests and demographic characteristics. In previous studies, the RISKINDEX was shown to outperform internal medicine specialists and to have high discriminatory performance. Adults patients (18 years or older) will be recruited in the ED. All participants will be randomly assigned to the control group or the intervention group in a 1:1 ratio. Participants in the control group will receive care as usual in which the study team asks the attending physicians questions about their clinical intuition. Participants in the intervention group will also receive care as usual, but in addition to asking the clinical impression questions, the study team presents the RISKINDEX to the attending physician in order to assess the extent to which clinical treatment is influenced by the results. DISCUSSION: This pilot clinical trial investigates the clinical impact and implementation of an ML based prediction model in the ED. By assessing the clinical impact and prognostic accuracy of the RISKINDEX, this study aims to contribute valuable insights to optimize patient care and inform future research in the field of ML based clinical prediction models. TRIAL REGISTRATION: ClinicalTrials.gov NCT05497830. Machine Learning for Risk Stratification in the Emergency Department (MARS-ED). Registered on August 11, 2022. URL: https://clinicaltrials.gov/study/NCT05497830 .

Lower Limits for Reporting High-Sensitivity Cardiac Troponin Assays and Impact of Analytical Performance on Patient Misclassification.

BACKGROUND: Cardiac troponin measurements are indispensable for the diagnosis of myocardial infarction and provide useful information for long-term risk prediction of cardiovascular disease. Accelerated diagnostic pathways prevent unnecessary hospital admission, but require reporting cardiac troponin concentrations at low concentrations that are sometimes below the limit of quantification. Whether analytical imprecision at these concentrations contributes to misclassification of patients is debated. CONTENT: The International Federation of Clinical Chemistry Committee on Clinical Application of Cardiac Bio-Markers (IFCC C-CB) provides evidence-based educational statements on analytical and clinical aspects of cardiac biomarkers. This mini-review discusses how the reporting of low concentrations of cardiac troponins impacts on whether or not assays are classified as high-sensitivity and how analytical performance at low concentrations influences the utility of troponins in accelerated diagnostic pathways. Practical suggestions are made for laboratories regarding analytical quality assessment of cardiac troponin results at low cutoffs, with a particular focus on accelerated diagnostic pathways. The review also discusses how future use of cardiac troponins for long-term prediction or management of cardiovascular disease may require improvements in analytical quality. SUMMARY: Clinical guidelines recommend using cardiac troponin concentrations as low as the limit of detection of the assay to guide patient care. Laboratories, manufacturers, researchers, and external quality assessment providers should extend analytical performance monitoring of cardiac troponin assays to include the concentration ranges applicable in these pathways.

Explainable Machine Learning Models for Rapid Risk Stratification in the Emergency Department: A Multicenter Study.

BACKGROUND: Risk stratification of patients presenting to the emergency department (ED) is important for appropriate triage. Diagnostic laboratory tests are an essential part of the workup and risk stratification of these patients. Using machine learning, the prognostic power and clinical value of these tests can be amplified greatly. In this study, we applied machine learning to develop an accurate and explainable clinical decision support tool model that predicts the likelihood of 31-day mortality in ED patients (the RISKINDEX). This tool was developed and evaluated in four Dutch hospitals. METHODS: Machine learning models included patient characteristics and available laboratory data collected within the first 2 h after ED presentation, and were trained using 5 years of data from consecutive ED patients from the Maastricht University Medical Center (Maastricht), Meander Medical Center (Amersfoort), and Zuyderland Medical Center (Sittard and Heerlen). A sixth year of data was used to evaluate the models using area under the receiver-operating-characteristic curve (AUROC) and calibration curves. The Shapley additive explanations (SHAP) algorithm was used to obtain explainable machine learning models. RESULTS: The present study included 266 327 patients with 7.1 million laboratory results available. Models show high diagnostic performance with AUROCs of 0.94, 0.98, 0.88, and 0.90 for Maastricht, Amersfoort, Sittard and Heerlen, respectively. The SHAP algorithm was utilized to visualize patient characteristics and laboratory data patterns that underlie individual RISKINDEX predictions. CONCLUSIONS: Our clinical decision support tool has excellent diagnostic performance in predicting 31-day mortality in ED patients. Follow-up studies will assess whether implementation of these algorithms can improve clinically relevant end points.

Head-to-head comparison of 19 prediction models for short-term outcome in medical patients in the emergency department: a retrospective study.

INTRODUCTION: Prediction models for identifying emergency department (ED) patients at high risk of poor outcome are often not externally validated. We aimed to perform a head-to-head comparison of the discriminatory performance of several prediction models in a large cohort of ED patients. METHODS: In this retrospective study, we selected prediction models that aim to predict poor outcome and we included adult medical ED patients. Primary outcome was 31-day mortality, secondary outcomes were 1-day mortality, 7-day mortality, and a composite endpoint of 31-day mortality and admission to intensive care unit (ICU).The discriminatory performance of the prediction models was assessed using an area under the receiver operating characteristic curve (AUC). Finally, the prediction models with the highest performance to predict 31-day mortality were selected to further examine calibration and appropriate clinical cut-off points. RESULTS: We included 19 prediction models and applied these to 2185 ED patients. Thirty-one-day mortality was 10.6% (231 patients), 1-day mortality was 1.4%, 7-day mortality was 4.4%, and 331 patients (15.1%) met the composite endpoint. The RISE UP and COPE score showed similar and very good discriminatory performance for 31-day mortality (AUC 0.86), 1-day mortality (AUC 0.87), 7-day mortality (AUC 0.86) and for the composite endpoint (AUC 0.81). Both scores were well calibrated. Almost no patients with RISE UP and COPE scores below 5% had an adverse outcome, while those with scores above 20% were at high risk of adverse outcome. Some of the other prediction models (i.e. APACHE II, NEWS, WPSS, MEWS, EWS and SOFA) showed significantly higher discriminatory performance for 1-day and 7-day mortality than for 31-day mortality. CONCLUSIONS: Head-to-head validation of 19 prediction models in medical ED patients showed that the RISE UP and COPE score outperformed other models regarding 31-day mortality.

Branched-chain ketoacid co-ingestion with protein lowers amino acid oxidation during hemodialysis: A randomized controlled cross-over trial.

BACKGROUND & AIMS: Hemodialysis removes amino acids from the circulation, thereby stimulating muscle proteolysis. Protein ingestion during hemodialysis can compensate for amino acid removal but may also increase uremic toxin production. Branched-chain ketoacid (BCKA) co-ingestion may provide an additional anabolic stimulus without adding to uremic toxin accumulation. In the present study we assessed the impact of BCKA co-ingestion with protein on forearm amino acid balance and amino acid oxidation during hemodialysis. METHODS: Nine patients (age: 73 ± 10 y) on chronic hemodialysis participated in this crossover trial. During two 4-h hemodialysis sessions, patients ingested 18 g protein with (PRO + BCKA) or without (PRO) 9 g BCKAs in a randomized order. Test beverages were labeled with L-[ring-13C6]-phenylalanine and provided throughout the last 3 h of hemodialysis as 18 equal sips consumed with 10-min intervals. Arterial and venous plasma as well as breath samples were collected frequently throughout hemodialysis. RESULTS: Arterial plasma total amino acid (TAA) concentrations during PRO and PRO + BCKA treatments were significantly lower after 1 h of hemodialysis (2.6 ± 0.3 and 2.6 ± 0.3 mmol/L, respectively) when compared to pre-hemodialysis concentrations (4.2 ± 1.0 and 4.0 ± 0.5 mmol/L, respectively; time effect: P < 0.001). Arterial plasma TAA concentrations increased throughout test beverage ingestion (time effect: P = 0.027) without differences between treatments (time∗treatment: P = 0.62). Forearm arteriovenous TAA balance during test beverage ingestion did not differ between timepoints (time effect: P = 0.31) or treatments (time∗treatment: P = 0.34). Whole-body phenylalanine oxidation was 33 ± 16% lower during PRO + BCKA when compared to PRO treatments (P < 0.001). CONCLUSIONS: BCKA co-ingestion with protein during hemodialysis does not improve forearm net protein balance but lowers amino acid oxidation.

Associations between plasma sulfur amino acids and specific fat depots in two independent cohorts: CODAM and The Maastricht Study.

PURPOSE: Sulfur amino acids (SAAs) have been associated with obesity and obesity-related metabolic diseases. We investigated whether plasma SAAs (methionine, total cysteine (tCys), total homocysteine, cystathionine and total glutathione) are related to specific fat depots. METHODS: We examined cross-sectional subsets from the CODAM cohort (n = 470, 61.3% men, median [IQR]: 67 [61, 71] years) and The Maastricht Study (DMS; n = 371, 53.4% men, 63 [55, 68] years), enriched with (pre)diabetic individuals. SAAs were measured in fasting EDTA plasma with LC-MS/MS. Outcomes comprised BMI, skinfolds, waist circumference (WC), dual-energy X-ray absorptiometry (DXA, DMS), body composition, abdominal subcutaneous and visceral adipose tissues (CODAM: ultrasound, DMS: MRI) and liver fat (estimated, in CODAM, or MRI-derived, in DMS, liver fat percentage and fatty liver disease). Associations were examined with linear or logistic regressions adjusted for relevant confounders with z-standardized primary exposures and outcomes. RESULTS: Methionine was associated with all measures of liver fat, e.g., fatty liver disease [CODAM: OR = 1.49 (95% CI 1.19, 1.88); DMS: OR = 1.51 (1.09, 2.14)], but not with other fat depots. tCys was associated with overall obesity, e.g., BMI [CODAM: ß = 0.19 (0.09, 0.28); DMS: ß = 0.24 (0.14, 0.34)]; peripheral adiposity, e.g., biceps and triceps skinfolds [CODAM: ß = 0.15 (0.08, 0.23); DMS: ß = 0.20 (0.12, 0.29)]; and central adiposity, e.g., WC [CODAM: ß = 0.16 (0.08, 0.25); DMS: ß = 0.17 (0.08, 0.27)]. Associations of tCys with VAT and liver fat were inconsistent. Other SAAs were not associated with body fat. CONCLUSION: Plasma concentrations of methionine and tCys showed distinct associations with different fat depots, with similar strengths in the two cohorts.

Intradialytic Protein Ingestion and Exercise do Not Compromise Uremic Toxin Removal Throughout Hemodialysis.

OBJECTIVE: Dietary protein and physical activity interventions are increasingly implemented during hemodialysis to support muscle maintenance in patients with end-stage renal disease (ESRD). Although muscle maintenance is important, adequate removal of uremic toxins throughout hemodialysis is the primary concern for patients. It remains to be established whether intradialytic protein ingestion and/or exercise modulate uremic toxin removal during hemodialysis. METHODS: We recruited 10 patients with ESRD (age: 65 ± 16 y, BMI: 24.2 ± 4.8 kg/m2) on chronic hemodialysis treatment to participate in this randomized cross-over trial. During hemodialysis, patients were assigned to ingest 40 g protein or a nonprotein placebo both at rest (protein [PRO] and placebo [PLA], respectively) and following 30 min of exercise (PRO + exercise [EX] and PLA + EX, respectively). Blood and spent dialysate samples were collected throughout hemodialysis to assess reduction ratios and removal of urea, creatinine, phosphate, cystatin C, and indoxyl sulfate. RESULTS: The reduction ratios of urea and indoxyl sulfate were higher during PLA (76 ± 6% and 46 ± 9%, respectively) and PLA + EX interventions (77 ± 5% and 45 ± 10%, respectively) when compared to PRO (72 ± 4% and 40 ± 8%, respectively) and PRO + EX interventions (73 ± 4% and 43 ± 7%, respectively; protein effect: P = .001 and P = .023, respectively; exercise effect: P = .25 and P = .52, respectively). Nonetheless, protein ingestion resulted in greater urea removal (P = .046) during hemodialysis. Reduction ratios and removal of creatinine, phosphate, and cystatin C during hemodialysis did not differ following intradialytic protein ingestion or exercise (protein effect: P > .05; exercise effect: P>.05). Urea, creatinine, and phosphate removal were greater throughout the period with intradialytic exercise during PLA + EX and PRO + EX interventions when compared to the same period during PLA and PRO interventions (exercise effect: P = .034, P = .039, and P = .022, respectively). CONCLUSION: The removal of uremic toxins is not compromised by protein feeding and/or exercise implementation during hemodialysis in patients with ESRD.

Use of artificial intelligence to assess the risk of coronary artery disease without additional (non-invasive) testing: validation in a low-risk to intermediate-risk outpatient clinic cohort.

OBJECTIVES: Predicting the presence or absence of coronary artery disease (CAD) is clinically important. Pretest probability (PTP) and CAD consortium clinical (CAD2) model and risk scores used in the guidelines are not sufficiently accurate as the only guidance for applying invasive testing or discharging a patient. Artificial intelligence without the need of additional non-invasive testing is not yet used in this context, as previous results of the model are promising, but available in high-risk population only. Still, validation in low-risk patients, which is clinically most relevant, is lacking. DESIGN: Retrospective cohort study. SETTING: Secondary outpatient clinic care in one Dutch academic hospital. PARTICIPANTS: We included 696 patients referred from primary care for further testing regarding the presence or absence of CAD. The results were compared with PTP and CAD2 using receiver operating characteristic (ROC) curves (area under the curve (AUC)). CAD was defined by a coronary stenosis >50% in at least one coronary vessel in invasive coronary or CT angiography, or having a coronary event within 6 months. OUTCOME MEASURES: The first cohort validating the memetic pattern-based algorithm (MPA) model developed in two high-risk populations in a low-risk to intermediate-risk cohort to improve risk stratification for non-invasive diagnosis of the presence or absence of CAD. RESULTS: The population contained 49% male, average age was 65.6±12.6 years. 16.2% had CAD. The AUCs of the MPA model, the PTP and the CAD2 were 0.87, 0.80, and 0.82, respectively. Applying the MPA model resulted in possible discharge of 67.7% of the patients with an acceptable CAD rate of 4.2%. CONCLUSIONS: In this low-risk to intermediate-risk population, the MPA model provides a good risk stratification of presence or absence of CAD with a better ROC compared with traditional risk scores. The results are promising but need prospective confirmation.

Metformin and high-sensitivity cardiac troponin I and T trajectories in type 2 diabetes patients: a post-hoc analysis of a randomized controlled trial.

BACKGROUND: Metformin has favorable effects on cardiovascular outcomes in both newly onset and advanced type 2 diabetes, as previously reported findings from the UK Prospective Diabetes Study and the HOME trial have demonstrated. Patients with type 2 diabetes present with chronically elevated circulating cardiac troponin levels, an established predictor of cardiovascular endpoints and prognostic marker of subclinical myocardial injury. It is unknown whether metformin affects cardiac troponin levels. The study aimed to evaluate cardiac troponin I and T trajectories in patients with diabetes treated either with metformin or placebo. METHODS: This study is a post-hoc analysis of a randomized controlled trial (HOME trial) that included 390 patients with advanced type 2 diabetes randomized to 850 mg metformin or placebo up to three times daily concomitant to continued insulin treatment. Cardiac troponin I and T concentrations were measured at baseline and after 4, 17, 30, 43 and 52 months. We evaluated cardiac troponin trajectories by linear mixed-effects modeling, correcting for age, sex, smoking status and history of cardiovascular disease. RESULTS: This study enrolled 390 subjects, of which 196 received metformin and 194 received placebo. In the treatment and placebo groups, mean age was 64 and 59 years; with 50% and 58% of subjects of the female sex, respectively. Despite the previously reported reduction of macrovascular disease risk in this cohort by metformin, linear mixed-effects regression modelling did not reveal evidence for an effect on cardiac troponin I and cardiac troponin T levels [- 8.4% (- 18.6, 3.2), p = 0.150, and - 4.6% (- 12, 3.2), p = 0.242, respectively]. A statistically significant time-treatment interaction was found for troponin T [- 1.6% (- 2.9, - 0.2), p = 0.021] but not troponin I concentrations [- 1.5% (- 4.2, 1.2), p = 0.263]. CONCLUSIONS: In this post-hoc analysis of a 4.3-year randomized controlled trial, metformin did not exert a clinically relevant effect on cardiac troponin I and cardiac troponin T levels when compared to placebo. Cardioprotective effects of the drug observed in clinical studies are not reflected by a reduction in these biomarkers of subclinical myocardial injury. Trial registration ClinicalTrials.gov identifier NCT00375388.

Fructose Intake From Fruit Juice and Sugar-Sweetened Beverages Is Associated With Higher Intrahepatic Lipid Content: The Maastricht Study.

OBJECTIVE: Epidemiological evidence regarding the relationship between fructose intake and intrahepatic lipid (IHL) content is inconclusive. We, therefore, assessed the relationship between different sources of fructose and IHL at the population level. RESEARCH DESIGN AND METHODS: We used cross-sectional data from The Maastricht Study, a population-based cohort study (n = 3,981; mean ± SD age: 60 ± 9 years; 50% women). We assessed the relationship between fructose intake (assessed with a food-frequency questionnaire)-total and derived from fruit, fruit juice, and sugar-sweetened beverages (SSB)-and IHL (quantified with 3T Dixon MRI) with adjustment for age, sex, type 2 diabetes, education, smoking status, physical activity, and intakes of total energy, alcohol, saturated fat, protein, vitamin E, and dietary fiber. RESULTS: Energy-adjusted total fructose intake and energy-adjusted fructose from fruit were not associated with IHL in the fully adjusted models (P = 0.647 and P = 0.767). In contrast, energy-adjusted intake of fructose from fruit juice and SSB was associated with higher IHL in the fully adjusted models (P = 0.019 and P = 0.009). Individuals in the highest tertile of energy-adjusted intake of fructose from fruit juice and SSB had a 1.04-fold (95% CI 0.99; 1.11) and 1.09-fold (95% CI 1.03; 1.16) higher IHL, respectively, in comparison with the lowest tertile in the fully adjusted models. Finally, the association for fructose from fruit juice was stronger in individuals with type 2 diabetes (P for interaction = 0.071). CONCLUSIONS: Fructose from fruit juice and SSB is independently associated with higher IHL. These cross-sectional findings contribute to current knowledge in support of measures to reduce the intake of fructose-containing beverages as a means to prevent nonalcoholic fatty liver disease at the population level.

Biological variation of cardiac myosin-binding protein C in healthy individuals.

OBJECTIVES: Cardiac myosin-binding protein C (cMyC) is a novel biomarker of myocardial injury, with a promising role in the triage and risk stratification of patients presenting with acute cardiac disease. In this study, we assess the weekly biological variation of cMyC, to examine its potential in monitoring chronic myocardial injury, and to suggest analytical quality specification for routine use of the test in clinical practice. METHODS: Thirty healthy volunteers were included. Non-fasting samples were obtained once a week for ten consecutive weeks. Samples were tested in duplicate on the Erenna® platform by EMD Millipore Corporation. Outlying measurements and subjects were identified and excluded systematically, and homogeneity of analytical and within-subject variances was achieved before calculating the biological variability (CVI and CVG), reference change values (RCV) and index of individuality (II). RESULTS: Mean age was 38 (range, 21-64) years, and 16 participants were women (53%). The biological variation, RCV and II with 95% confidence interval (CI) were: CVA (%) 19.5 (17.8-21.6), CVI (%) 17.8 (14.8-21.0), CVG (%) 66.9 (50.4-109.9), RCV (%) 106.7 (96.6-120.1)/-51.6 (-54.6 to -49.1) and II 0.42 (0.29-0.56). There was a trend for women to have lower CVG. The calculated RCVs were comparable between genders. CONCLUSIONS: cMyC exhibits acceptable RCV and low II suggesting that it could be suitable for disease monitoring, risk stratification and prognostication if measured serially. Analytical quality specifications based on biological variation are similar to those for cardiac troponin and should be achievable at clinically relevant concentrations.

Sex Differences in Cardiac Troponin I and T and the Prediction of Cardiovascular Events in the General Population.

BACKGROUND: Cardiac troponin concentrations differ in women and men, but how this influences risk prediction and whether a sex-specific approach is required is unclear. We evaluated whether sex influences the predictive ability of cardiac troponin I and T for cardiovascular events in the general population. METHODS: High-sensitivity cardiac troponin (hs-cTn) I and T were measured in the Generation Scotland Scottish Family Health Study of randomly selected volunteers drawn from the general population between 2006 and 2011. Cox-regression models evaluated associations between hs-cTnI and hs-cTnT and the primary outcome of cardiovascular death, myocardial infarction, or stroke. RESULTS: In 19 501 (58% women, mean age 47 years) participants, the primary outcome occurred in 2.7% (306/11 375) of women and 5.1% (411/8126) of men during the median follow-up period of 7.9 (IQR, 7.1-9.2) years. Cardiac troponin I and T concentrations were lower in women than men (P < 0.001 for both), and both were more strongly associated with cardiovascular events in women than men. For example, at a hs-cTnI concentration of 10 ng/L, the hazard ratio relative to the limit of blank was 9.7 (95% CI 7.6-12.4) and 5.6 (95% CI 4.7-6.6) for women and men, respectively. The hazard ratio for hs-cTnT at a concentration of 10 ng/L relative to the limit of blank was 3.7 (95% CI 3.1-4.3) and 2.2 (95% CI 2.0-2.5) for women and men, respectively. CONCLUSIONS: Cardiac troponin concentrations differ in women and men and are stronger predictors of cardiovascular events in women. Sex-specific approaches are required to provide equivalent risk prediction.

Machine learning-based glucose prediction with use of continuous glucose and physical activity monitoring data: The Maastricht Study.

BACKGROUND: Closed-loop insulin delivery systems, which integrate continuous glucose monitoring (CGM) and algorithms that continuously guide insulin dosing, have been shown to improve glycaemic control. The ability to predict future glucose values can further optimize such devices. In this study, we used machine learning to train models in predicting future glucose levels based on prior CGM and accelerometry data. METHODS: We used data from The Maastricht Study, an observational population-based cohort that comprises individuals with normal glucose metabolism, prediabetes, or type 2 diabetes. We included individuals who underwent >48h of CGM (n = 851), most of whom (n = 540) simultaneously wore an accelerometer to assess physical activity. A random subset of individuals was used to train models in predicting glucose levels at 15- and 60-minute intervals based on either CGM data or both CGM and accelerometer data. In the remaining individuals, model performance was evaluated with root-mean-square error (RMSE), Spearman's correlation coefficient (rho) and surveillance error grid. For a proof-of-concept translation, CGM-based prediction models were optimized and validated with the use of data from individuals with type 1 diabetes (OhioT1DM Dataset, n = 6). RESULTS: Models trained with CGM data were able to accurately predict glucose values at 15 (RMSE: 0.19mmol/L; rho: 0.96) and 60 minutes (RMSE: 0.59mmol/L, rho: 0.72). Model performance was comparable in individuals with type 2 diabetes. Incorporation of accelerometer data only slightly improved prediction. The error grid results indicated that model predictions were clinically safe (15 min: >99%, 60 min >98%). Our prediction models translated well to individuals with type 1 diabetes, which is reflected by high accuracy (RMSEs for 15 and 60 minutes of 0.43 and 1.73 mmol/L, respectively) and clinical safety (15 min: >99%, 60 min: >91%). CONCLUSIONS: Machine learning-based models are able to accurately and safely predict glucose values at 15- and 60-minute intervals based on CGM data only. Future research should further optimize the models for implementation in closed-loop insulin delivery systems.

Sex differences in investigations and outcomes among patients with type 2 myocardial infarction.

OBJECTIVES: Type 2 myocardial infarction (MI) is a heterogenous condition and whether there are differences between women and men is unknown. We evaluated sex differences in clinical characteristics, investigations and outcomes in patients with type 2 MI. METHODS: In the Swedish Web based system for Enhancement and Development of Evidence based care in Heart disease Evaluated According to Recommended Therapies (SWEDEHEART) registry, we compared patients admitted to coronary care units with a diagnosis of type 1 or type 2 MI. Sex-stratified Cox regression models evaluated the association with all-cause death in men and women separately. RESULTS: We included 57 264 (median age 73 years, 65% men) and 6485 (median age 78 years, 50% men) patients with type 1 and type 2 MI, respectively. No differences were observed in the proportion of men and women with type 2 MI who underwent echocardiography and coronary angiography, but women were less likely than men to have left ventricular (LV) impairment and obstructive coronary artery disease (CAD). Compared with type 1 MI, patients with type 2 MI had higher risk of death regardless of sex (men: adjusted HR 1.55 (95% CI 1.44 to 1.67); women: adjusted HR 1.34 (95% CI 1.24 to 1.45)). In those with type 2 MI, the risk of death was lower for women than men (adjusted HR 0.85 (95% CI 0.76 to 0.92) (men, reference)). CONCLUSIONS: Type 2 MI occurred in men and women equally and we found no evidence of sex bias in the selection of patients for cardiac investigations. Patients with type 2 MI had worse outcomes, but women were less likely to have obstructive CAD or severe LV impairment and were more likely to survive than men.

A comparison of machine learning models versus clinical evaluation for mortality prediction in patients with sepsis.

INTRODUCTION: Patients with sepsis who present to an emergency department (ED) have highly variable underlying disease severity, and can be categorized from low to high risk. Development of a risk stratification tool for these patients is important for appropriate triage and early treatment. The aim of this study was to develop machine learning models predicting 31-day mortality in patients presenting to the ED with sepsis and to compare these to internal medicine physicians and clinical risk scores. METHODS: A single-center, retrospective cohort study was conducted amongst 1,344 emergency department patients fulfilling sepsis criteria. Laboratory and clinical data that was available in the first two hours of presentation from these patients were randomly partitioned into a development (n = 1,244) and validation dataset (n = 100). Machine learning models were trained and evaluated on the development dataset and compared to internal medicine physicians and risk scores in the independent validation dataset. The primary outcome was 31-day mortality. RESULTS: A number of 1,344 patients were included of whom 174 (13.0%) died. Machine learning models trained with laboratory or a combination of laboratory + clinical data achieved an area-under-the ROC curve of 0.82 (95% CI: 0.80-0.84) and 0.84 (95% CI: 0.81-0.87) for predicting 31-day mortality, respectively. In the validation set, models outperformed internal medicine physicians and clinical risk scores in sensitivity (92% vs. 72% vs. 78%;p<0.001,all comparisons) while retaining comparable specificity (78% vs. 74% vs. 72%;p>0.02). The model had higher diagnostic accuracy with an area-under-the-ROC curve of 0.85 (95%CI: 0.78-0.92) compared to abbMEDS (0.63,0.54-0.73), mREMS (0.63,0.54-0.72) and internal medicine physicians (0.74,0.65-0.82). CONCLUSION: Machine learning models outperformed internal medicine physicians and clinical risk scores in predicting 31-day mortality. These models are a promising tool to aid in risk stratification of patients presenting to the ED with sepsis.

Decreased serial scores of severe organ failure assessments are associated with survival in mechanically ventilated patients; the prospective Maastricht Intensive Care COVID cohort.

BACKGROUND: The majority of patients with severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) infection are admitted to the Intensive Care Unit (ICU) for mechanical ventilation. The role of multi-organ failure during ICU admission as driver for outcome remains to be investigated yet. DESIGN AND SETTING: Prospective cohort of mechanically ventilated critically ill with SARS-CoV-2 infection. PARTICIPANTS AND METHODS: 94 participants of the MaastrICCht cohort (21% women) had a median length of stay of 16 days (maximum of 77). After division into survivors (n = 59) and non-survivors (n = 35), we analysed 1555 serial SOFA scores using linear mixed-effects models. RESULTS: Survivors improved one SOFA score point more per 5 days (95% CI: 4-8) than non-survivors. Adjustment for age, sex, and chronic lung, renal and liver disease, body-mass index, diabetes mellitus, cardiovascular risk factors, and Acute Physiology and Chronic Health Evaluation II score did not change this result. This association was stronger for women than men (P-interaction = 0.043). CONCLUSIONS: The decrease in SOFA score associated with survival suggests multi-organ failure involvement during mechanical ventilation in patients with SARS-CoV-2. Surviving women appeared to improve faster than surviving men. Serial SOFA scores may unravel an unfavourable trajectory and guide decisions in mechanically ventilated patients with SARS-CoV-2.