Search

Abstract Practices for controlling intracranial pressure (ICP) in traumatic brain injury (TBI) patients admitted to the intensive care unit (ICU) vary considerably between centres. To help understand the rational basis for such variance in care, this study aims to identify the patient-level predictors of changes in ICP management. We extracted all heterogeneous data (2008 pre-ICU and ICU variables) collected from a prospective cohort (n = 844, 51 ICUs) of ICP-monitored TBI patients in the Collaborative European NeuroTrauma Effectiveness Research in TBI study. We developed the TILTomorrow modelling strategy, which leverages recurrent neural networks to map a token-embedded time series representation of all variables (including missing values) to an ordinal, dynamic prediction of the following day’s five-category therapy intensity level (TIL(Basic)) score. With 20 repeats of fivefold cross-validation, we trained TILTomorrow on different variable sets and applied the TimeSHAP (temporal extension of SHapley Additive exPlanations) algorithm to estimate variable contributions towards predictions of next-day changes in TIL(Basic). Based on Somers’ D xy , the full range of variables explained 68% (95% CI 65–72%) of the ordinal variation in next-day changes in TIL(Basic) on day one and up to 51% (95% CI 45–56%) thereafter, when changes in TIL(Basic) became less frequent. Up to 81% (95% CI 78–85%) of this explanation could be derived from non-treatment variables (i.e., markers of pathophysiology and injury severity), but the prior trajectory of ICU management significantly improved prediction of future de-escalations in ICP-targeted treatment. Whilst there was no significant difference in the predictive discriminability (i.e., area under receiver operating characteristic curve) between next-day escalations (0.80 [95% CI 0.77–0.84]) and de-escalations (0.79 [95% CI 0.76–0.82]) in TIL(Basic) after day two, we found specific predictor effects to be more robust with de-escalations. The most important predictors of day-to-day changes in ICP management included preceding treatments, age, space-occupying lesions, ICP, metabolic derangements, and neurological function. Serial protein biomarkers were also important and may serve a useful role in the clinical armamentarium for assessing therapeutic needs. Approximately half of the ordinal variation in day-to-day changes in TIL(Basic) after day two remained unexplained, underscoring the significant contribution of unmeasured factors or clinicians’ personal preferences in ICP treatment. At the same time, specific dynamic markers of pathophysiology associated strongly with changes in treatment intensity and, upon mechanistic investigation, may improve the timing and personalised targeting of future care.

Abstract Purpose The use of statins for the primary prevention of cardiovascular diseases (CVD) is associated with various beneficial outcomes, alongside certain undesirable effects. This study aims to determine optimal risk thresholds above which statin therapy yields a net benefit, considering both the positive effects and potential adverse effects, as well as their probabilities and patient preferences. Methods Quantitative benefit-harm balance modeling was applied to the Iranian general population aged 40 to 75 years with no history of CVD. The analysis utilized data from prior studies, including statin effect estimates for different outcomes from a meta-analysis, patient preferences obtained from an Iranian survey, and baseline incidence rates of adverse outcomes sourced from the Global Burden of Disease study for Iran. Outcomes were defined as angina, myocardial infarction, fatal coronary heart disease, fatal or non-fatal stroke, and heart failure. Benefit-harm balance indices were calculated for various combinations of age, sex, and 10-year CVD risk. Results Statin therapy was found to be advantageous at a lower 10-year CVD risk threshold in men (18–23%) compared to women (24–28%). Furthermore, individuals aged 40–45 years exhibited a lower risk threshold (18% in men, 24% in women) than those aged 70–75 years (23% in men, 28% in women). Conclusion The desirable 10-year risk thresholds for statin prescription in the primary prevention of CVD vary by age and gender, ranging from 18 to 28%, encompassing a spectrum of outcomes from angina to CVD mortality. These results suggest hard-CVD risk thresholds of 7.5% to 10% for both sexes.

Summary: Background: Postoperative infections significantly impact patient outcomes and costs, exacerbated by late diagnoses, yet early reliable predictors are scarce. Existing artificial intelligence (AI) models for postoperative infection prediction often lack external validation or perform poorly in local settings when validated. We aimed to develop locally valid models as part of the PERISCOPE AI system to enable early detection, safer discharge, and more timely treatment of patients. Methods: We developed and validated XGBoost models to predict postoperative infections within 7 and 30 days of surgery. Using retrospective pre-operative and intra-operative electronic health record data from 2014 to 2023 across various surgical specialities, the models were developed at Hospital A and validated and updated at Hospitals B and C in the Netherlands and Belgium. Model performance was evaluated before and after updating using the two most recent years of data as temporal validation datasets. Main outcome measures were model discrimination (area under the receiver operating characteristic curve (AUROC)), calibration (slope, intercept, and plots), and clinical utility (decision curve analysis with net benefit). Findings: The study included 253,010 surgical procedures with 23,903 infections within 30-days. Discriminative performance, calibration properties, and clinical utility significantly improved after updating. Final AUROCs after updating for Hospitals A, B, and C were 0.82 (95% confidence interval (CI) 0.81–0.83), 0.82 (95% CI 0.81–0.83), and 0.91 (95% CI 0.90–0.91) respectively for 30-day predictions on the temporal validation datasets (2022–2023). Calibration plots demonstrated adequate correspondence between observed outcomes and predicted risk. All local models were deemed clinically useful as the net benefit was higher than default strategies (treat all and treat none) over a wide range of clinically relevant decision thresholds. Interpretation: PERISCOPE can accurately predict overall postoperative infections within 7- and 30-days post-surgery. The robust performance implies potential for improving clinical care in diverse clinical target populations. This study supports the need for approaches to local updating of AI models to account for domain shifts in patient populations and data distributions across different clinical settings. Funding: This study was funded by a REACT EU grant from European Regional Development Fund (ERDF) and Kansen voor West.

Summary: Background: Post-traumatic stress disorder (PTSD) and depression are common after mild traumatic brain injury (mTBI), but their biological drivers are uncertain. We therefore explored whether polygenic risk scores (PRS) derived for PTSD and major depressive disorder (MDD) are associated with the development of cognate TBI-related phenotypes. Methods: Meta-analyses were conducted using data from two multicenter, prospective observational cohort studies of patients with mTBI: the CENTER-TBI study (ClinicalTrials.gov ID NCT02210221) in Europe (December 2014–December 2017) and the TRACK-TBI study in the US (March 2014–July 2018). In both cohorts, the most common causes of injury were road traffic accidents and falls. Primary outcomes, specifically probable PTSD and depression, were defined at 6 months post-injury using scores ≥33 on the PTSD Checklist-5 and ≥15 on the Patient Health Questionnaire-9, respectively. We calculated PTSD-PRS and MDD-PRS for patients aged ≥17 years who had a Glasgow Coma Scale score of 13–15 upon hospital arrival and assessed their association with PTSD and depression following TBI. We also evaluated the transferability of the findings in a cohort of African Americans. Findings: Overall, 11.8% (219/1869) and 6.7% (124/1869) patients were classified as having probable PTSD and depression, respectively. The PTSD-PRS was significantly associated with higher adjusted odds of PTSD in both cohorts, with a pooled odds ratio (OR) of 1.55 [95% confidence interval (CI) 1.30–1.84, p

Summary: Background: The inflammatory response in patients with traumatic brain injury (TBI) offers opportunities for stratification and intervention. Previous unselected approaches to immunomodulation in patients with TBI have not improved patient outcomes. Methods: Serum and plasma samples from two prospective, multi-centre observational studies of patients with TBI were used to discover (Collaborative European NeuroTrauma Effectiveness Research [CENTER-TBI], Europe) and validate (Transforming Research and Clinical Knowledge in Traumatic Brain Injury [TRACK-TBI] Pilot, USA) individual variations in the immune response using a multiplex panel of 30 inflammatory mediators. Mediators that were associated with unfavourable outcomes (Glasgow outcome score-extended [GOS-E] ≤ 4) were used for hierarchical clustering to identify patients with similar signatures. Findings: Two clusters were identified in both the discovery and validation cohorts, termed early-inflammatory and pauci-inflammatory. The early-inflammatory phenotype had higher concentrations of interleukin-6 (IL-6), IL-15, and monocyte chemoattractant protein 1 (MCP1). Patients with the early-inflammatory phenotype were older and more likely to have an unfavourable GOS-E at 6 months. There were no differences in the baseline injury severity scores between patients in each phenotype. A combined IL-15 and MCP1 signature identified patients with the early-inflammatory phenotype in both cohorts. Inflammatory processes mediated outcomes in older patients with moderate-severe TBI. Interpretation: Our findings offer a precision medicine approach for future clinical trials of immunomodulation in patients with TBI, by using inflammatory signatures to stratify patients. Funding: CENTER-TBI study was supported by the European Union 7th Framework Programme. TRACK-TBI is supported by the National Institute of Neurological Disorders and Stroke.

Summary: Background: Traumatic brain injury is conventionally categorised as mild, moderate, or severe on the Glasgow Coma Scale (GCS). Recently developed biomarkers can provide more objective and nuanced measures of the extent of brain injury. Methods: Exposure–response relationships were investigated in 2479 patients aged ≥16 enrolled in the Collaborative European NeuroTrauma Effectiveness Research in Traumatic Brain Injury (CENTER-TBI) prospective observational cohort study. Neurofilament protein-light (NFL), ubiquitin carboxy-terminal hydrolase L1 (UCH-L1), and glial fibrillary acidic protein (GFAP) were assayed from serum sampled in the first 24 h; concentrations were divided into quintiles within GCS severity groups. Relationships with the Glasgow Outcome Scale-Extended were examined using modified Poisson regression including age, sex, major extracranial injury, time to sample, and log biomarker concentration as covariates. Findings: Within severity groups there were associations between biomarkers and outcomes after adjustment for covariates: GCS 13–15 and negative CT imaging (relative risks [RRs] from 1.28 to 3.72), GCS 13–15 and positive CT (1.21–2.81), GCS 9–12 (1.16–2.02), GCS 3–8 (1.09–1.94). RRs were associated with clinically important differences in expectations of prognosis. In patients with GCS 3 (RRs 1.51–1.80) percentages of unfavourable outcome were 37–51% in the lowest quintiles of biomarker levels and reached 90–94% in the highest quintiles. Similarly, for GCS 15 (RRs 1.83–3.79), the percentages were 2–4% and 19–28% in the lowest and highest biomarker quintiles, respectively. Interpretation: Conventional TBI severity classification is inadequate and underestimates heterogeneity of brain injury and associated outcomes. The adoption of circulating biomarkers can add to clinical assessment of injury severity. Funding: European Union 7th Framework program (EC grant 602150), Hannelore Kohl Stiftung, One Mind, Integra LifeSciences, Neuro-Trauma Sciences, NIHR Rosetrees Trust.

Abstract Community-acquired Pneumonia (CAP) guidelines generally recommend to admit patients with moderate-to-severe CAP and start treatment with intravenous antibiotics. This study aims to explore the clinical outcomes of oral antibiotics in patients with moderate-to-severe CAP. We performed a nested cohort study of an observational study including all adult patients presenting to the emergency department of the Haga Teaching Hospital, the Netherlands, between April 2019 and May 2020, who had a blood culture drawn. We conducted propensity score matching with logistic and linear regression analysis to compare patients with moderate-to-severe CAP (Pneumonia Severity Index class III–V) treated with oral antibiotics to patients treated with intravenous antibiotics. Outcomes were 30-day mortality, intensive care unit admission, readmission, length of stay (LOS) and length of antibiotic treatment. Of the original 314 patients, 71 orally treated patients were matched with 102 intravenously treated patients. The mean age was 73 years and 58% were male. We found no significant differences in outcomes between the oral and intravenous group, except for an increased LOS of + 2.6 days (95% confidence interval 1.2–4.0, p value

Background: Parkinson’s disease (PD) is the second most common neurodegenerative disease, primarily affecting the middle-aged to elderly population. Among its nonmotor symptoms, cognitive decline (CD) is a precursor to dementia and represents a critical target for early risk assessment and diagnosis. Accurate CD prediction is crucial for timely intervention and tailored management of at-risk patients. This study used machine learning (ML) techniques to predict the CD risk over five-year in early-stage PD. Methods: Data from the Early Parkinson’s Disease Longitudinal Singapore (2014 to 2018) was used to predict CD defined as a one-unit annual decrease or a one-unit decline in Montreal Cognitive Assessment over two consecutive years. Four ML methods—AutoScore, Random Forest, K-Nearest Neighbors and Neural Network—were applied using baseline demographics, clinical assessments and blood biomarkers. Results: Variable selection identified key predictors of CD, including education year, diastolic lying blood pressure, diastolic standing blood pressure, systolic lying blood pressure, Hoehn and Yahr scale, body mass index, phosphorylated tau at threonine 181, total tau, Neurofilament light chain and suppression of tumorigenicity 2. Random Forest was the most effective, achieving an AUC of 0.93 (95% CI: 0.89, 0.97), using 10-fold cross-validation. Conclusions: Here, we demonstrate that ML-based models can identify early-stage PD patients at high risk for CD, supporting targeted interventions and improved PD management.

OBJECTIVES:. To develop and validate a prediction model for 1-year mortality in patients with a hematologic malignancy acutely admitted to the ICU. DESIGN:. A retrospective cohort study. SETTING:. Five university hospitals in the Netherlands between 2002 and 2015. PATIENTS:. A total of 1097 consecutive patients with a hematologic malignancy were acutely admitted to the ICU for at least 24 h. INTERVENTIONS:. None. MEASUREMENTS AND MAIN RESULTS:. We created a 13-variable model from 22 potential predictors. Key predictors included active disease, age, previous hematopoietic stem cell transplantation, mechanical ventilation, lowest platelet count, acute kidney injury, maximum heart rate, and type of malignancy. A bootstrap procedure reduced overfitting and improved the model’s generalizability. This involved estimating the optimism in the initial model and shrinking the regression coefficients accordingly in the final model. We assessed performance using internal–external cross-validation by center and compared it with the Acute Physiology and Chronic Health Evaluation II model. Additionally, we evaluated clinical usefulness through decision curve analysis. The overall 1-year mortality rate observed in the study was 62% (95% CI, 59–65). Our 13-variable prediction model demonstrated acceptable calibration and discrimination at internal–external validation across centers (C-statistic 0.70; 95% CI, 0.63–0.77), outperforming the Acute Physiology and Chronic Health Evaluation II model (C-statistic 0.61; 95% CI, 0.57–0.65). Decision curve analysis indicated overall net benefit within a clinically relevant threshold probability range of 60–100% predicted 1-year mortality. CONCLUSIONS:. Our newly developed 13-variable prediction model predicts 1-year mortality in hematologic malignancy patients admitted to the ICU more accurately than the Acute Physiology and Chronic Health Evaluation II model. This model may aid in shared decision-making regarding the continuation of ICU care and end-of-life considerations.

Abstract Rationale Aspirin is typically discontinued in cranial and spinal surgery because of the increased risk of hemorrhagic complications, but comes together with the risk of resulting in an increase of cardiac and neurologic thrombotic perioperative events. Objective The aim of this study is to investigate the non-inferiority of perioperative continuation of aspirin patients undergoing low complex lumbar spinal surgery, compared with the current policy of perioperative discontinuation of aspirin. Study design A randomized controlled trial with two parallel groups of 277 cases (554 in total). Study population Patients undergoing low complex lumbar spinal surgery and using aspirin. All patients are aged >18 years. Intervention Peri-operative continuation of aspirin. Study outcomes Primary study outcome: composite of the following bleeding complications: Neurological deterioration as a result of hemorrhage in the surgical area with cauda and/or nerve root compression. Post-surgical anemia with hemoglobin level lower than 5 mmol/l, requiring transfusion. Subcutaneous hematoma leading to wound leakage and pain higher than NRS=7. Major and/or minor hemorrhage in any other body system according to the definition of the International Society on Thrombosis and Haemostasis bleeding scale. Secondary study outcomes: Each of the individual components of the primary outcome Absolute mean difference in operative blood loss between the study arms Thrombo-embolic-related complications: Myocardial infarction Venous thromboembolism Stroke Arterial thromboembolism Further study outcomes Anticoagulant treatment satisfaction by the Anti-Clot Treatment Scale (ACTS) and general health by the Patient-Reported Outcomes Measurement Information System (PROMIS Global-10) in the pre- and postoperative phase. Nature and extent of the burden and risks associated with participation, benefit, and group relatedness Participation in this study imposes no additional risk to patients. Currently, there is no consensus on whether or not aspirin should be discontinued before cranial or spinal surgery. Currently, aspirin is typically discontinued in cranial and spinal surgery, because of a potential increased risk of hemorrhagic complication. An argument not based on a clinical trial. However, this policy might delay surgical procedures or carry the risk of resulting in an increase in cardiac and neurologic thrombotic perioperative events. It is unclear if the possibility of an increase in hemorrhage-related complications outweighs the risk of an increase in cardiac and neurologic thrombotic perioperative events. Furthermore, the Data Safety Monitoring Board (DSMB) will be asked for safety analysis by monitoring the study. There are no further disadvantages to participating in this study. Outcome measurements are recorded during admission and regular outpatient visits, and thus, do not require additional visits to the hospital.

Abstract Existing methods to characterise the evolving condition of traumatic brain injury (TBI) patients in the intensive care unit (ICU) do not capture the context necessary for individualising treatment. Here, we integrate all heterogenous data stored in medical records (1166 pre-ICU and ICU variables) to model the individualised contribution of clinical course to 6-month functional outcome on the Glasgow Outcome Scale -Extended (GOSE). On a prospective cohort (n = 1550, 65 centres) of TBI patients, we train recurrent neural network models to map a token-embedded time series representation of all variables (including missing values) to an ordinal GOSE prognosis every 2 h. The full range of variables explains up to 52% (95% CI: 50–54%) of the ordinal variance in functional outcome. Up to 91% (95% CI: 90–91%) of this explanation is derived from pre-ICU and admission information (i.e., static variables). Information collected in the ICU (i.e., dynamic variables) increases explanation (by up to 5% [95% CI: 4–6%]), though not enough to counter poorer overall performance in longer-stay (>5.75 days) patients. Highest-contributing variables include physician-based prognoses, CT features, and markers of neurological function. Whilst static information currently accounts for the majority of functional outcome explanation after TBI, data-driven analysis highlights investigative avenues to improve the dynamic characterisation of longer-stay patients. Moreover, our modelling strategy proves useful for converting large patient records into interpretable time series with missing data integration and minimal processing.

Abstract Background A number of recent papers have proposed methods to calculate confidence intervals and p values for net benefit used in decision curve analysis. These papers are sparse on the rationale for doing so. We aim to assess the relation between sampling variability, inference, and decision-analytic concepts. Methods and results We review the underlying theory of decision analysis. When we are forced into a decision, we should choose the option with the highest expected utility, irrespective of p values or uncertainty. This is in some distinction to traditional hypothesis testing, where a decision such as whether to reject a given hypothesis can be postponed. Application of inference for net benefit would generally be harmful. In particular, insisting that differences in net benefit be statistically significant would dramatically change the criteria by which we consider a prediction model to be of value. We argue instead that uncertainty related to sampling variation for net benefit should be thought of in terms of the value of further research. Decision analysis tells us which decision to make for now, but we may also want to know how much confidence we should have in that decision. If we are insufficiently confident that we are right, further research is warranted. Conclusion Null hypothesis testing or simple consideration of confidence intervals are of questionable value for decision curve analysis, and methods such as value of information analysis or approaches to assess the probability of benefit should be considered instead.

Background: Patients with cancer often have to make complex decisions about treatment, with the options varying in risk profiles and effects on survival and quality of life. Moreover, inefficient care paths make it hard for patients to participate in shared decision-making. Data-driven decision-support tools have the potential to empower patients, support personalized care, improve health outcomes and promote health equity. However, decision-support tools currently seldom consider quality of life or individual preferences, and their use in clinical practice remains limited, partly because they are not well integrated in patients’ care paths. Aim and objectives: The central aim of the 4D PICTURE project is to redesign patients’ care paths and develop and integrate evidence-based decision-support tools to improve decision-making processes in cancer care delivery. This article presents an overview of this international, interdisciplinary project. Design, methods and analysis: In co-creation with patients and other stakeholders, we will develop data-driven decision-support tools for patients with breast cancer, prostate cancer and melanoma. We will support treatment decisions by using large, high-quality datasets with state-of-the-art prognostic algorithms. We will further develop a conversation tool, the Metaphor Menu, using text mining combined with citizen science techniques and linguistics, incorporating large datasets of patient experiences, values and preferences. We will further develop a promising methodology, MetroMapping, to redesign care paths. We will evaluate MetroMapping and these integrated decision-support tools, and ensure their sustainability using the Nonadoption, Abandonment, Scale-Up, Spread, and Sustainability (NASSS) framework. We will explore the generalizability of MetroMapping and the decision-support tools for other types of cancer and across other EU member states. Ethics: Through an embedded ethics approach, we will address social and ethical issues. Discussion: Improved care paths integrating comprehensive decision-support tools have the potential to empower patients, their significant others and healthcare providers in decision-making and improve outcomes. This project will strengthen health care at the system level by improving its resilience and efficiency.

The TRIPOD (Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis) Statement includes a 22-item checklist, which aims to improve the reporting of studies developing, validating, or updating a prediction model, whether for diagnostic or prognostic purposes. The TRIPOD Statement aims to improve the transparency of the reporting of a prediction model study regardless of the study methods used. This explanation and elaboration document describes the rationale; clarifies the meaning of each item; and discusses why transparent reporting is important, with a view to assessing risk of bias and clinical usefulness of the prediction model. Each checklist item of the TRIPOD Statement is explained in detail and accompanied by published examples of good reporting. The document also provides a valuable reference of issues to consider when designing, conducting, and analyzing prediction model studies. To aid the editorial process and help peer reviewers and, ultimately, readers and systematic reviewers of prediction model studies, it is recommended that authors include a completed checklist in their submission. The TRIPOD checklist can also be downloaded from www.tripod-statement.org. This article is the translation in to Russian by Dr. Ruslan Saygitov (ORCID: https://orcid.org/0000-0002-8915-6153) from the original published in [Ann Intern Med. 2015;162:W1-W73. doi: https://doi.org/10.7326/M14-0698].

Abstract Background Baseline outcome risk can be an important determinant of absolute treatment benefit and has been used in guidelines for “personalizing” medical decisions. We compared easily applicable risk-based methods for optimal prediction of individualized treatment effects. Methods We simulated RCT data using diverse assumptions for the average treatment effect, a baseline prognostic index of risk, the shape of its interaction with treatment (none, linear, quadratic or non-monotonic), and the magnitude of treatment-related harms (none or constant independent of the prognostic index). We predicted absolute benefit using: models with a constant relative treatment effect; stratification in quarters of the prognostic index; models including a linear interaction of treatment with the prognostic index; models including an interaction of treatment with a restricted cubic spline transformation of the prognostic index; an adaptive approach using Akaike’s Information Criterion. We evaluated predictive performance using root mean squared error and measures of discrimination and calibration for benefit. Results The linear-interaction model displayed optimal or close-to-optimal performance across many simulation scenarios with moderate sample size (N = 4,250; ~ 785 events). The restricted cubic splines model was optimal for strong non-linear deviations from a constant treatment effect, particularly when sample size was larger (N = 17,000). The adaptive approach also required larger sample sizes. These findings were illustrated in the GUSTO-I trial. Conclusions An interaction between baseline risk and treatment assignment should be considered to improve treatment effect predictions.

Abstract Treatment effects are often anticipated to vary across groups of patients with different baseline risk. The Predictive Approaches to Treatment Effect Heterogeneity (PATH) statement focused on baseline risk as a robust predictor of treatment effect and provided guidance on risk-based assessment of treatment effect heterogeneity in a randomized controlled trial. The aim of this study is to extend this approach to the observational setting using a standardized scalable framework. The proposed framework consists of five steps: (1) definition of the research aim, i.e., the population, the treatment, the comparator and the outcome(s) of interest; (2) identification of relevant databases; (3) development of a prediction model for the outcome(s) of interest; (4) estimation of relative and absolute treatment effect within strata of predicted risk, after adjusting for observed confounding; (5) presentation of the results. We demonstrate our framework by evaluating heterogeneity of the effect of thiazide or thiazide-like diuretics versus angiotensin-converting enzyme inhibitors on three efficacy and nine safety outcomes across three observational databases. We provide a publicly available R software package for applying this framework to any database mapped to the Observational Medical Outcomes Partnership Common Data Model. In our demonstration, patients at low risk of acute myocardial infarction receive negligible absolute benefits for all three efficacy outcomes, though they are more pronounced in the highest risk group, especially for acute myocardial infarction. Our framework allows for the evaluation of differential treatment effects across risk strata, which offers the opportunity to consider the benefit-harm trade-off between alternative treatments.

Abstract Lobular primary breast cancer (PBC) histology has been proposed as a risk factor for contralateral breast cancer (CBC), but results have been inconsistent. We investigated CBC risk and the impact of systemic therapy in lobular versus ductal PBC. Further, CBC characteristics following these histologic subtypes were explored. We selected 74,373 women diagnosed between 2003 and 2010 with stage I‐III invasive PBC from the nationwide Netherlands Cancer Registry. We assessed absolute risk of CBC taking into account competing risks among those with lobular (n = 8903), lobular mixed with other types (n = 3512), versus ductal (n = 62,230) histology. Hazard ratios (HR) for CBC were estimated in a cause‐specific Cox model, adjusting for age at PBC diagnosis, radiotherapy, chemotherapy and/or endocrine therapy. Multivariable HRs for CBC were 1.18 (95% CI: 1.04–1.33) for lobular and 1.37 (95% CI: 1.16–1.63) for lobular mixed versus ductal PBC. Ten‐year cumulative CBC incidences in patients with lobular, lobular mixed versus ductal PBC were 3.2%, 3.6% versus 2.8% when treated with systemic therapy and 6.6%, 7.7% versus 5.6% in patients without systemic therapy, respectively. Metachronous CBCs were diagnosed in a less favourable stage in 19%, 26% and 23% and less favourable differentiation grade in 22%, 33% and 27% than the PBCs of patients with lobular, lobular mixed and ductal PBC, respectively. In conclusion, lobular and lobular mixed PBC histology are associated with modestly increased CBC risk. Personalised CBC risk assessment needs to consider PBC histology, including systemic treatment administration. The impact on prognosis of CBCs with unfavourable characteristics warrants further evaluation.

Abstract Background Clinical prediction models should be validated before implementation in clinical practice. But is favorable performance at internal validation or one external validation sufficient to claim that a prediction model works well in the intended clinical context? Main body We argue to the contrary because (1) patient populations vary, (2) measurement procedures vary, and (3) populations and measurements change over time. Hence, we have to expect heterogeneity in model performance between locations and settings, and across time. It follows that prediction models are never truly validated. This does not imply that validation is not important. Rather, the current focus on developing new models should shift to a focus on more extensive, well-conducted, and well-reported validation studies of promising models. Conclusion Principled validation strategies are needed to understand and quantify heterogeneity, monitor performance over time, and update prediction models when appropriate. Such strategies will help to ensure that prediction models stay up-to-date and safe to support clinical decision-making.

Abstract Mortality is a frequently reported outcome in clinical studies of acute respiratory distress syndrome (ARDS). However, timing of mortality assessment has not been well characterized. We aimed to identify a crossing-point between cumulative survival and death in the intensive care unit (ICU) of patients with moderate-to-severe ARDS, beyond which the number of survivors would exceed the number of deaths. We hypothesized that this intersection would occur earlier in a successful clinical trial vs. observational studies of moderate/severe ARDS and predict treatment response. We conducted an ancillary study of 1580 patients with moderate-to-severe ARDS managed with lung-protective ventilation to assess the relevance and timing of measuring ICU mortality rates at different time-points during ICU stay. First, we analyzed 1303 patients from four multicenter, observational cohorts enrolling consecutive patients with moderate/severe ARDS. We assessed cumulative ICU survival from the time of moderate/severe ARDS diagnosis to ventilatory support discontinuation within 7-days, 28-days, 60-days, and at ICU discharge. Then, we compared these findings to those of a successful randomized trial of 277 moderate/severe ARDS patients. In the observational cohorts, ICU mortality (487/1303, 37.4%) and 28-day mortality (425/1102, 38.6%) were similar (p = 0.549). Cumulative proportion of ICU survivors and non-survivors crossed at day-7; after day-7, the number of ICU survivors was progressively higher compared to non-survivors. Measures of oxygenation, lung mechanics, and severity scores were different between survivors and non-survivors at each point-in-time (p

Summary: Background: Limited evidence existed on the comparative effectiveness of decompressive craniectomy (DC) versus craniotomy for evacuation of traumatic acute subdural hematoma (ASDH) until the recently published randomised clinical trial RESCUE-ASDH. In this study, that ran concurrently, we aimed to determine current practice patterns and compare outcomes of primary DC versus craniotomy. Methods: We conducted an analysis of centre treatment preference within the prospective, multicentre, observational Collaborative European NeuroTrauma Effectiveness Research in Traumatic Brain Injury (known as CENTER-TBI) and NeuroTraumatology Quality Registry (known as Net-QuRe) studies, which enrolled patients throughout Europe and Israel (2014–2020). We included patients with an ASDH who underwent acute neurosurgical evacuation. Patients with severe pre-existing neurological disorders were excluded. In an instrumental variable analysis, we compared outcomes between centres according to treatment preference, measured by the case-mix adjusted proportion DC per centre. The primary outcome was functional outcome rated by the 6-months Glasgow Outcome Scale Extended, estimated with ordinal regression as a common odds ratio (OR), adjusted for prespecified confounders. Variation in centre preference was quantified with the median odds ratio (MOR). CENTER-TBI is registered with ClinicalTrials.gov, number NCT02210221, and the Resource Identification Portal (Research Resource Identifier SCR_015582). Findings: Between December 19, 2014 and December 17, 2017, 4559 patients with traumatic brain injury were enrolled in CENTER-TBI of whom 336 (7%) underwent acute surgery for ASDH evacuation; 91 (27%) underwent DC and 245 (63%) craniotomy. The proportion primary DC within total acute surgery cases ranged from 6 to 67% with an interquartile range (IQR) of 12–26% among 46 centres; the odds of receiving a DC for prognostically similar patients in one centre versus another randomly selected centre were trebled (adjusted median odds ratio 2.7, p

The generalizability of predictive algorithms is of key relevance to application in clinical practice. We provide an overview of three types of generalizability, based on existing literature: temporal, geographical, and domain generalizability. These generalizability types are linked to their associated goals, methodology, and stakeholders.

Abstract Early detection of severe asthma exacerbations through home monitoring data in patients with stable mild-to-moderate chronic asthma could help to timely adjust medication. We evaluated the potential of machine learning methods compared to a clinical rule and logistic regression to predict severe exacerbations. We used daily home monitoring data from two studies in asthma patients (development: n = 165 and validation: n = 101 patients). Two ML models (XGBoost, one class SVM) and a logistic regression model provided predictions based on peak expiratory flow and asthma symptoms. These models were compared with an asthma action plan rule. Severe exacerbations occurred in 0.2% of all daily measurements in the development (154/92,787 days) and validation cohorts (94/40,185 days). The AUC of the best performing XGBoost was 0.85 (0.82–0.87) and 0.88 (0.86–0.90) for logistic regression in the validation cohort. The XGBoost model provided overly extreme risk estimates, whereas the logistic regression underestimated predicted risks. Sensitivity and specificity were better overall for XGBoost and logistic regression compared to one class SVM and the clinical rule. We conclude that ML models did not beat logistic regression in predicting short-term severe asthma exacerbations based on home monitoring data. Clinical application remains challenging in settings with low event incidence and high false alarm rates with high sensitivity.

The TRIPOD (Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis) Statement includes a 22-item checklist, which aims to improve the reporting of studies developing, validating, or updating a prediction model, whether for diagnostic or prognostic purposes. The TRIPOD Statement aims to improve the transparency of the reporting of a prediction model study regardless of the study methods used. This explanation and elaboration document describes the rationale; clarifies the meaning of each item; and discusses why transparent reporting is important, with a view to assessing risk of bias and clinical usefulness of the prediction model. Each checklist item of the TRIPOD Statement is explained in detail and accompanied by published examples of good reporting. The document also provides a valuable reference of issues to consider when designing, conducting, and analyzing prediction model studies. To aid the editorial process and help peer reviewers and, ultimately, readers and systematic reviewers of prediction model studies, it is recommended that authors include a completed checklist in their submission. The TRIPOD checklist can also be downloaded from www.tripod-statement.org. For members of the TRIPOD Group, see the Appendix. This article is the translation in to Russian by Dr. Ruslan Saygitov (ORCID: 0000-0002-8915-6153) from the original published in [Ann Intern Med. 2015; 162:W1-W73. doi: 10.7326/M14-0698 ].

Abstract Background Prediction of contralateral breast cancer (CBC) risk is challenging due to moderate performances of the known risk factors. We aimed to improve our previous risk prediction model (PredictCBC) by updated follow-up and including additional risk factors. Methods We included data from 207,510 invasive breast cancer patients participating in 23 studies. In total, 8225 CBC events occurred over a median follow-up of 10.2 years. In addition to the previously included risk factors, PredictCBC-2.0 included CHEK2 c.1100delC, a 313 variant polygenic risk score (PRS-313), body mass index (BMI), and parity. Fine and Gray regression was used to fit the model. Calibration and a time-dependent area under the curve (AUC) at 5 and 10 years were assessed to determine the performance of the models. Decision curve analysis was performed to evaluate the net benefit of PredictCBC-2.0 and previous PredictCBC models. Results The discrimination of PredictCBC-2.0 at 10 years was higher than PredictCBC with an AUC of 0.65 (95% prediction intervals (PI) 0.56–0.74) versus 0.63 (95%PI 0.54–0.71). PredictCBC-2.0 was well calibrated with an observed/expected ratio at 10 years of 0.92 (95%PI 0.34–2.54). Decision curve analysis for contralateral preventive mastectomy (CPM) showed the potential clinical utility of PredictCBC-2.0 between thresholds of 4 and 12% 10-year CBC risk for BRCA1/2 mutation carriers and non-carriers. Conclusions Additional genetic information beyond BRCA1/2 germline mutations improved CBC risk prediction and might help tailor clinical decision-making toward CPM or alternative preventive strategies. Identifying patients who benefit from CPM, especially in the general breast cancer population, remains challenging.

Abstract Background While the Glasgow coma scale (GCS) is one of the strongest outcome predictors, the current classification of traumatic brain injury (TBI) as ‘mild’, ‘moderate’ or ‘severe’ based on this fails to capture enormous heterogeneity in pathophysiology and treatment response. We hypothesized that data-driven characterization of TBI could identify distinct endotypes and give mechanistic insights. Methods We developed an unsupervised statistical clustering model based on a mixture of probabilistic graphs for presentation (

Abstract Background Evaluating patients’ experiences is essential when incorporating the patients’ perspective in improving healthcare. Experiences are mainly collected using closed-ended questions, although the value of open-ended questions is widely recognized. Natural language processing (NLP) can automate the analysis of open-ended questions for an efficient approach to patient-centeredness. Methods We developed the Artificial Intelligence Patient-Reported Experience Measures (AI-PREM) tool, consisting of a new, open-ended questionnaire, an NLP pipeline to analyze the answers using sentiment analysis and topic modeling, and a visualization to guide physicians through the results. The questionnaire and NLP pipeline were iteratively developed and validated in a clinical context. Results The final AI-PREM consisted of five open-ended questions about the provided information, personal approach, collaboration between healthcare professionals, organization of care, and other experiences. The AI-PREM was sent to 867 vestibular schwannoma patients, 534 of which responded. The sentiment analysis model attained an F1 score of 0.97 for positive texts and 0.63 for negative texts. There was a 90% overlap between automatically and manually extracted topics. The visualization was hierarchically structured into three stages: the sentiment per question, the topics per sentiment and question, and the original patient responses per topic. Conclusions The AI-PREM tool is a comprehensive method that combines a validated, open-ended questionnaire with a well-performing NLP pipeline and visualization. Thematically organizing and quantifying patient feedback reduces the time invested by healthcare professionals to evaluate and prioritize patient experiences without being confined to the limited answer options of closed-ended questions.

Abstract Machine learning is increasingly being used to predict clinical outcomes. Most comparisons of different methods have been based on empirical analyses in specific datasets. We used Monte Carlo simulations to determine when machine learning methods perform better than statistical learning methods in a specific setting. We evaluated six learning methods: stochastic gradient boosting machines using trees as the base learners, random forests, artificial neural networks, the lasso, ridge regression, and linear regression estimated using ordinary least squares (OLS). Our simulations were informed by empirical analyses in patients with acute myocardial infarction (AMI) and congestive heart failure (CHF) and used six data-generating processes, each based on one of the six learning methods, to simulate continuous outcomes in the derivation and validation samples. The outcome was systolic blood pressure at hospital discharge, a continuous outcome. We applied the six learning methods in each of the simulated derivation samples and evaluated performance in the simulated validation samples. The primary observation was that neural networks tended to result in estimates with worse predictive accuracy than the other five methods in both disease samples and across all six data-generating processes. Boosted trees and OLS regression tended to perform well across a range of scenarios.

Abstract Background Prediction modeling studies often have methodological limitations, which may compromise model performance in new patients and settings. We aimed to examine the relation between methodological quality of model development studies and their performance at external validation. Methods We systematically searched for externally validated multivariable prediction models that predict functional outcome following moderate or severe traumatic brain injury. Risk of bias and applicability of development studies was assessed with the Prediction model Risk Of Bias Assessment Tool (PROBAST). Each model was rated for its presentation with sufficient detail to be used in practice. Model performance was described in terms of discrimination (AUC), and calibration. Delta AUC (dAUC) was calculated to quantify the percentage change in discrimination between development and validation for all models. Generalized estimation equations (GEE) were used to examine the relation between methodological quality and dAUC while controlling for clustering. Results We included 54 publications, presenting ten development studies of 18 prediction models, and 52 external validation studies, including 245 unique validations. Two development studies (four models) were found to have low risk of bias (RoB). The other eight publications (14 models) showed high or unclear RoB. The median dAUC was positive in low RoB models (dAUC 8%, [IQR − 4% to 21%]) and negative in high RoB models (dAUC − 18%, [IQR − 43% to 2%]). The GEE showed a larger average negative change in discrimination for high RoB models (− 32% (95% CI: − 48 to − 15) and unclear RoB models (− 13% (95% CI: − 16 to − 10)) compared to that seen in low RoB models. Conclusion Lower methodological quality at model development associates with poorer model performance at external validation. Our findings emphasize the importance of adherence to methodological principles and reporting guidelines in prediction modeling studies.

Traumatic brain injury is associated with changes to the metabolome. Here the authors show that acute traumatic brain injury has distinctive serum metabolic patterns which may suggest protective changes of systemic lipid metabolism aiming to maintain lipid homeostasis in the brain.

Abstract Background The rapidly increasing number of elderly (≥ 65 years old) with TBI is accompanied by substantial medical and economic consequences. An ASDH is the most common injury in elderly with TBI and the surgical versus conservative treatment of this patient group remains an important clinical dilemma. Current BTF guidelines are not based on high-quality evidence and compliance is low, allowing for large international treatment variation. The RESET-ASDH trial is an international multicenter RCT on the (cost-)effectiveness of early neurosurgical hematoma evacuation versus initial conservative treatment in elderly with a t-ASDH Methods In total, 300 patients will be recruited from 17 Belgian and Dutch trauma centers. Patients ≥ 65 years with at first presentation a GCS ≥ 9 and a t-ASDH > 10 mm or a t-ASDH 5 mm, or a GCS

Background The Pneumonia Severity Index (PSI) and the CURB-65 score assess disease severity in patients with community-acquired pneumonia (CAP). We compared the clinical performance of both prognostic scores according to clinical outcomes and admission rates. Methods A nationwide retrospective cohort study was conducted using claims data from adult CAP patients presenting to the emergency department (ED) in 2018 and 2019. Dutch hospitals were divided into three categories: “CURB-65 hospitals” (n=25), “PSI hospitals” (n=19) and hospitals using both (“no-consensus hospitals”, n=15). Main outcomes were hospital admission rates, intensive care unit admissions, length of hospital stay, delayed admissions, readmissions and all-cause 30-day mortality. Multilevel logistic and Poisson regression analysis were used to adjust for potential confounders. Findings Of 50 984 included CAP patients, 21 157 were treated in CURB-65 hospitals, 17 279 in PSI hospitals and 12 548 in no-consensus hospitals. The 30-day mortality was significantly lower in CURB-65 hospitals versus PSI hospitals (8.6% and 9.7%, adjusted odds ratio (aOR) 0.89, 95% CI: 0.83–0.96, p=0.003). Other clinical outcomes were similar between CURB-65 hospitals and PSI hospitals. No-consensus hospitals had higher admission rates compared to the CURB-65 and PSI hospitals combined (78.4% and 81.5%, aOR 0.78, 95% CI: 0.62–0.99). Interpretation In this study, using the CURB-65 in CAP patients at the ED is associated with similar and possibly even better clinical outcomes compared to using the PSI. After confirmation in prospective studies, the CURB-65 may be recommended over the use of the PSI since it is associated with lower 30-day mortality and is more user-friendly.

Abstract Background Obtaining accurate estimates of the risk of COVID-19-related death in the general population is challenging in the context of changing levels of circulating infection. Methods We propose a modelling approach to predict 28-day COVID-19-related death which explicitly accounts for COVID-19 infection prevalence using a series of sub-studies from new landmark times incorporating time-updating proxy measures of COVID-19 infection prevalence. This was compared with an approach ignoring infection prevalence. The target population was adults registered at a general practice in England in March 2020. The outcome was 28-day COVID-19-related death. Predictors included demographic characteristics and comorbidities. Three proxies of local infection prevalence were used: model-based estimates, rate of COVID-19-related attendances in emergency care, and rate of suspected COVID-19 cases in primary care. We used data within the TPP SystmOne electronic health record system linked to Office for National Statistics mortality data, using the OpenSAFELY platform, working on behalf of NHS England. Prediction models were developed in case-cohort samples with a 100-day follow-up. Validation was undertaken in 28-day cohorts from the target population. We considered predictive performance (discrimination and calibration) in geographical and temporal subsets of data not used in developing the risk prediction models. Simple models were contrasted to models including a full range of predictors. Results Prediction models were developed on 11,972,947 individuals, of whom 7999 experienced COVID-19-related death. All models discriminated well between individuals who did and did not experience the outcome, including simple models adjusting only for basic demographics and number of comorbidities: C-statistics 0.92–0.94. However, absolute risk estimates were substantially miscalibrated when infection prevalence was not explicitly modelled. Conclusions Our proposed models allow absolute risk estimation in the context of changing infection prevalence but predictive performance is sensitive to the proxy for infection prevalence. Simple models can provide excellent discrimination and may simplify implementation of risk prediction tools.

Aim: BRCA1/2 mutation carriers with primary breast cancer (PBC) are at high risk of contralateral breast cancer (CBC). In a nationwide cohort, we investigated the effects of chemotherapeutic agents given for PBC on CBC risk separately in BRCA1 and BRCA2 mutation carriers. Patients and methods: BRCA1 or BRCA2 mutation carriers with an invasive PBC diagnosis from 1990 to 2017 were selected from a Dutch cohort. We estimated cumulative CBC incidence using competing risks analysis. Hazard ratios (HR) for the effect of neo-adjuvant or adjuvant chemotherapy and different chemotherapeutic agents on CBC risk were estimated using Cox regression. Results: We included 1090 BRCA1 and 568 BRCA2 mutation carriers; median follow-up was 8.9 and 8.4 years, respectively. Ten-year cumulative CBC incidence for treatment with and without chemotherapy was 6.7% [95%CI: 5.1–8.6] and 16.7% [95%CI: 10.8–23.7] in BRCA1 and 4.8% [95%CI: 2.7–7.8] and 16.0% [95%CI: 9.3–24.4] in BRCA2 mutation carriers, respectively. Chemotherapy was associated with reduced CBC risk in BRCA1 (multivariable HR: 0.46, 95%CI: 0.29–0.74); a similar trend was observed in BRCA2 mutation carriers (HR: 0.63, 95%CI: 0.29–1.39). In BRCA1, risk reduction was most pronounced in the first 5 years (HR: 0.32, 95%CI: 0.17–0.61). Anthracyclines and the combination of anthracyclines with taxanes were associated with substantial CBC risk reduction in BRCA1 carriers (HR: 0.34, 95%CI: 0.17–0.68 and HR: 0.22, 95%CI: 0.08–0.62, respectively). Conclusion: Risk-reducing effects of chemotherapy are substantial for at least 5 years and may be used in personalised CBC risk prediction in any case for BRCA1 mutation carriers.

BackgroundConsidering the importance of cardiovascular disease (CVD) risk prediction for healthcare systems and the limited information available in the Middle East, we evaluated the SCORE and Globorisk models to predict CVD death in a country of this region.MethodsWe included 24 427 participants (11 187 men) aged 40-80 years from four population-based cohorts in Iran. Updating approaches were used to recalibrate the baseline survival and the overall effect of the predictors of the models. We assessed the models’ discrimination using C-index and then compared the observed with the predicted risk of death using calibration plots. The sensitivity and specificity of the models were estimated at the risk thresholds of 3%, 5%, 7%, and 10%. An agreement between models was assessed using the intra-class correlation coefficient (ICC). We applied decision analysis to provide perception into the consequences of using the models in general practice; for this reason, the clinical usefulness of the models was assessed using the net benefit (NB) and decision curve analysis. The NB is a sensitivity penalized by a weighted false positive (FP) rate in population level.ResultsAfter 154 522 person-years of follow-up, 437 cardiovascular deaths (280 men) occurred. The 10-year observed risks were 4.2% (95% CI: 3.7%-4.8%) in men and 2.1% (1.8-2%.5%) in women. The c-index for SCORE function was 0.784 (0.756-0.812) in men and 0.780 (0.744-0.815) in women. Corresponding values for Globorisk were 0.793 (0.766-0.820) and 0.793 (0.757-0.829). The deviation of the calibration slopes from one reflected a need for recalibration; after which, the predicted-to-observed ratio for both models was 1.02 in men and 0.95 in women. Models showed good agreement (ICC 0.93 in men, and 0.89 in women). Decision curve showed that using both models results in the same clinical usefulness at the risk threshold of 5%, in both men and women; however, at the risk threshold of 10%, Globorisk had better clinical usefulness in women (Difference: 8%, 95% CI: 4%-13%).ConclusionOriginal Globorisk and SCORE models overestimate the CVD risk in Iranian populations resulting in a high number of people who need intervention. Recalibration could adopt these models to precisely predict CVD mortality. Globorisk showed better performance clinically, only among high-risk women.

Abstract Background We investigated whether we could use influenza data to develop prediction models for COVID-19 to increase the speed at which prediction models can reliably be developed and validated early in a pandemic. We developed COVID-19 Estimated Risk (COVER) scores that quantify a patient’s risk of hospital admission with pneumonia (COVER-H), hospitalization with pneumonia requiring intensive services or death (COVER-I), or fatality (COVER-F) in the 30-days following COVID-19 diagnosis using historical data from patients with influenza or flu-like symptoms and tested this in COVID-19 patients. Methods We analyzed a federated network of electronic medical records and administrative claims data from 14 data sources and 6 countries containing data collected on or before 4/27/2020. We used a 2-step process to develop 3 scores using historical data from patients with influenza or flu-like symptoms any time prior to 2020. The first step was to create a data-driven model using LASSO regularized logistic regression, the covariates of which were used to develop aggregate covariates for the second step where the COVER scores were developed using a smaller set of features. These 3 COVER scores were then externally validated on patients with 1) influenza or flu-like symptoms and 2) confirmed or suspected COVID-19 diagnosis across 5 databases from South Korea, Spain, and the United States. Outcomes included i) hospitalization with pneumonia, ii) hospitalization with pneumonia requiring intensive services or death, and iii) death in the 30 days after index date. Results Overall, 44,507 COVID-19 patients were included for model validation. We identified 7 predictors (history of cancer, chronic obstructive pulmonary disease, diabetes, heart disease, hypertension, hyperlipidemia, kidney disease) which combined with age and sex discriminated which patients would experience any of our three outcomes. The models achieved good performance in influenza and COVID-19 cohorts. For COVID-19 the AUC ranges were, COVER-H: 0.69–0.81, COVER-I: 0.73–0.91, and COVER-F: 0.72–0.90. Calibration varied across the validations with some of the COVID-19 validations being less well calibrated than the influenza validations. Conclusions This research demonstrated the utility of using a proxy disease to develop a prediction model. The 3 COVER models with 9-predictors that were developed using influenza data perform well for COVID-19 patients for predicting hospitalization, intensive services, and fatality. The scores showed good discriminatory performance which transferred well to the COVID-19 population. There was some miscalibration in the COVID-19 validations, which is potentially due to the difference in symptom severity between the two diseases. A possible solution for this is to recalibrate the models in each location before use.

Abstract While the opportunities of ML and AI in healthcare are promising, the growth of complex data-driven prediction models requires careful quality and applicability assessment before they are applied and disseminated in daily practice. This scoping review aimed to identify actionable guidance for those closely involved in AI-based prediction model (AIPM) development, evaluation and implementation including software engineers, data scientists, and healthcare professionals and to identify potential gaps in this guidance. We performed a scoping review of the relevant literature providing guidance or quality criteria regarding the development, evaluation, and implementation of AIPMs using a comprehensive multi-stage screening strategy. PubMed, Web of Science, and the ACM Digital Library were searched, and AI experts were consulted. Topics were extracted from the identified literature and summarized across the six phases at the core of this review: (1) data preparation, (2) AIPM development, (3) AIPM validation, (4) software development, (5) AIPM impact assessment, and (6) AIPM implementation into daily healthcare practice. From 2683 unique hits, 72 relevant guidance documents were identified. Substantial guidance was found for data preparation, AIPM development and AIPM validation (phases 1–3), while later phases clearly have received less attention (software development, impact assessment and implementation) in the scientific literature. The six phases of the AIPM development, evaluation and implementation cycle provide a framework for responsible introduction of AI-based prediction models in healthcare. Additional domain and technology specific research may be necessary and more practical experience with implementing AIPMs is needed to support further guidance.

The aims of this study are to describe the use of pharmaceutical venous thromboembolism (pVTE) prophylaxis in patients with traumatic brain injury (TBI) in Europe and study the association of pVTE prophylaxis with outcome. We included 2006 patients ?18 years of age admitted to the intensive care unit from the CENTER-TBI study. VTE events were recorded based on clinical symptoms. Variation between 54 centers in pVTE prophylaxis use was assessed with a multi-variate random-effect model and quantified with the median odds ratio (MOR). The association between pVTE prophylaxis and outcome (Glasgow Outcome Scale-Extended at 6 months) was assessed at center level with an instrumental variable analysis and at patient level with a multi-variate proportional odds regression analysis and a propensity-matched analysis. A time-dependent Cox survival regression analysis was conducted to determine the effect of pVTE prophylaxis on survival during hospital stay. The association between VTE prophylaxis and computed tomography (CT) progression was assessed with a logistic regression analysis. Overall, 56 patients (2%) had a VTE during hospital stay. The majority, 1279 patients (64%), received pVTE prophylaxis, with substantial between-center variation (MOR, 2.7; p?0.001). A moderate association with improved outcome was found at center level (odds ratio [OR], 1.2 [0.7?2.1]) and patient level (multi-variate adjusted OR, 1.4 [1.1?1.7], and propensity adjusted OR, 1.5 [1.1?2.0]), with similar results in subgroup analyses. Survival was higher with the use of pVTE prophylaxis (p?

Abstract Background To evaluate the cost‐effectiveness of prophylactic hysterectomy (PH) in women with Lynch syndrome (LS). Methods We developed a microsimulation model incorporating the natural history for the development of hyperplasia with and without atypia into endometrial cancer (EC) based on the MISCAN‐framework. We simulated women identified as first‐degree relatives (FDR) with LS of colorectal cancer patients after universal testing for LS. We estimated costs and benefits of offering this cohort PH, accounting for reduced quality of life after PH and for having EC. Three minimum ages (30/35/40) and three maximum ages (70/75/80) were compared to no PH. Results In the absence of PH, the estimated number of EC cases was 300 per 1,000 women with LS. Total associated costs for treatment of EC were $5.9 million. Offering PH to FDRs aged 40–80 years was considered optimal. This strategy reduced the number of endometrial cancer cases to 5.4 (−98%), resulting in 516 quality‐adjusted life years (QALY) gained and increasing the costs (treatment of endometrial cancer and PH) to $15.0 million (+154%) per 1,000 women. PH from earlier ages was more costly and resulted in fewer QALYs, although this finding was sensitive to disutility for PH. Conclusions Offering PH to 40‐ to 80‐year‐old women with LS is expected to add 0.5 QALY per person at acceptable costs. Women may decide to have PH at a younger age, depending on their individual disutility for PH and premature menopause.

Abstract Background The novel coronavirus SARS-19 produces ‘COVID-19’ in patients with symptoms. COVID-19 patients admitted to the hospital require early assessment and care including isolation. The National Early Warning Score (NEWS) and its updated version NEWS2 is a simple physiological scoring system used in hospitals, which may be useful in the early identification of COVID-19 patients. We investigate the performance of multiple enhanced NEWS2 models in predicting the risk of COVID-19. Methods Our cohort included unplanned adult medical admissions discharged over 3 months (11 March 2020 to 13 June 2020 ) from two hospitals (YH for model development; SH for external model validation). We used logistic regression to build multiple prediction models for the risk of COVID-19 using the first electronically recorded NEWS2 within ± 24 hours of admission. Model M0’ included NEWS2; model M1’ included NEWS2 + age + sex, and model M2’ extends model M1’ with subcomponents of NEWS2 (including diastolic blood pressure + oxygen flow rate + oxygen scale). Model performance was evaluated according to discrimination (c statistic), calibration (graphically), and clinical usefulness at NEWS2 ≥ 5. Results The prevalence of COVID-19 was higher in SH (11.0 %=277/2520) than YH (8.7 %=343/3924) with a higher first NEWS2 scores ( SH 3.2 vs YH 2.8) but similar in-hospital mortality (SH 8.4 % vs YH 8.2 %). The c-statistics for predicting the risk of COVID-19 for models M0’,M1’,M2’ in the development dataset were: M0’: 0.71 (95 %CI 0.68–0.74); M1’: 0.67 (95 %CI 0.64–0.70) and M2’: 0.78 (95 %CI 0.75–0.80)). For the validation datasets the c-statistics were: M0’ 0.65 (95 %CI 0.61–0.68); M1’: 0.67 (95 %CI 0.64–0.70) and M2’: 0.72 (95 %CI 0.69–0.75) ). The calibration slope was similar across all models but Model M2’ had the highest sensitivity (M0’ 44 % (95 %CI 38-50 %); M1’ 53 % (95 %CI 47-59 %) and M2’: 57 % (95 %CI 51-63 %)) and specificity (M0’ 75 % (95 %CI 73-77 %); M1’ 72 % (95 %CI 70-74 %) and M2’: 76 % (95 %CI 74-78 %)) for the validation dataset at NEWS2 ≥ 5. Conclusions Model M2’ appears to be reasonably accurate for predicting the risk of COVID-19. It may be clinically useful as an early warning system at the time of admission especially to triage large numbers of unplanned hospital admissions.

BACKGROUND:. We previously reported the added value of 24-hour lactate concentration alone and in combination with 24-hour lactate clearance and lactate concentration at admission for the prediction of inhospital mortality in critically ill patients with sepsis. We aimed to validate this finding. DERIVATION COHORT:. The derivation cohort from Leiden, The Netherlands, consisted of 451 critically ill patients with sepsis. VALIDATION COHORT:. The validation cohort consisted of 4,440 critically ill adult patients with sepsis from the Medical Information Mart for Intensive Care cohort admitted to the ICU of Beth Israel Deaconness Medical Center, Boston, MA, between January 2006 and 2018. PREDICTION MODEL:. Predictors of mortality were: age, chronic comorbidities, length of stay pre-ICU, Glasgow Coma Scale, and Acute Physiology Score. Lactate concentration at 24-hour alone, in combination with 24-hour lactate clearance and in combination with lactate concentration at admission, was added to assess improvement of the prediction model. The outcome was inhospital mortality. RESULTS:. Inhospital mortality occurred in 160 patients (36%) in the derivation cohort and in 2,347 patients (53%) in the validation cohort. The Acute Physiology and Chronic Health Evaluation (APACHE) IV model had a moderate discriminative performance (recalibrated C-statistic, 0.62; 95% CI, 0.60–0.63). Addition of 24-hour lactate concentration increased the recalibrated C-statistic to 0.64 (95% CI, 0.62–0.66). The model with 24-hour lactate concentration and lactate concentration at admission showed the best fit as depicted by the smallest Akaike Information Criterion in both the derivation and validation data. CONCLUSION:. The 24-hour lactate concentration and lactate concentration at admission contribute modestly to prediction of inhospital mortality in critically ill patients with sepsis. Future updates and possible modification of APACHE IV should consider the incorporation of lactate concentration at baseline and at 24 hours.

Abstract Background Prehospital triage protocols typically try to select patients with Injury Severity Score (ISS) above 15 for direct transportation to a Level-1 trauma center. However, ISS does not necessarily discriminate between patients who benefit from immediate care at Level-1 trauma centers. The aim of this study was to assess which patients benefit from direct transportation to Level-1 trauma centers. Methods We used the American National Trauma Data Bank (NTDB), a retrospective observational cohort. All adult patients (ISS > 3) between 2015 and 2016 were included. Patients who were self-presenting or had isolated limb injury were excluded. We used logistic regression to assess the association of direct transportation to Level-1 trauma centers with in-hospital mortality adjusted for clinically relevant confounders. We used this model to define benefit as predicted probability of mortality associated with transportation to a non-Level-1 trauma center minus predicted probability associated with transportation to a Level-1 trauma center. We used a threshold of 1% as absolute benefit. Potential interaction terms with transportation to Level-1 trauma centers were included in a penalized logistic regression model to study which patients benefit. Results We included 388,845 trauma patients from 232 Level-1 centers and 429 Level-2/3 centers. A small beneficial effect was found for direct transportation to Level-1 trauma centers (adjusted Odds Ratio: 0.96, 95% Confidence Interval: 0.92–0.99) which disappeared when comparing Level-1 and 2 versus Level-3 trauma centers. In the risk approach, predicted benefit ranged between 0 and 1%. When allowing for interactions, 7% of the patients (n = 27,753) had more than 1% absolute benefit from direct transportation to Level-1 trauma centers. These patients had higher AIS Head and Thorax scores, lower GCS and lower SBP. A quarter of the patients with ISS > 15 were predicted to benefit from transportation to Level-1 centers (n = 26,522, 22%). Conclusions Benefit of transportation to a Level-1 trauma centers is quite heterogeneous across patients and the difference between Level-1 and Level-2 trauma centers is small. In particular, patients with head injury and signs of shock may benefit from care in a Level-1 trauma center. Future prehospital triage models should incorporate more complete risk profiles.

Abstract Background The Surgery As Needed for Oesophageal cancer (SANO) trial compares active surveillance with standard oesophagectomy for patients with a clinically complete response (cCR) to neoadjuvant chemoradiotherapy. The last patient with a clinically complete response is expected to be included in May 2021. The purpose of this update is to present all amendments to the SANO trial protocol as approved by the Institutional Research Board (IRB) before accrual is completed. Design The SANO trial protocol has been published ( https://doi.org/10.1186/s12885-018-4034-1 ). In this ongoing, phase-III, non-inferiority, stepped-wedge, cluster randomised controlled trial, patients with cCR (i.e. after neoadjuvant chemoradiotherapy no evidence of residual disease in two consecutive clinical response evaluations [CREs]) undergo either active surveillance or standard oesophagectomy. In the active surveillance arm, CREs are repeated every 3 months in the first year, every 4 months in the second year, every 6 months in the third year, and yearly in the fourth and fifth year. In this arm, oesophagectomy is offered only to patients in whom locoregional regrowth is highly suspected or proven, without distant metastases. The primary endpoint is overall survival. Update Amendments to the study design involve the first cluster in the stepped-wedge design being partially randomised as well and continued accrual of patients at baseline until the predetermined number of patients with cCR is reached. Eligibility criteria have been amended, stating that patients who underwent endoscopic treatment prior to neoadjuvant chemoradiotherapy cannot be included and that patients who have highly suspected residual tumour without histological proof can be included. Amendments to the study procedures include that patients proceed to the second CRE if at the first CRE the outcome of the pathological assessment is uncertain and that patients with a non-passable stenosis at endoscopy are not considered cCR. The sample size was recalculated following new insights on response rates (34% instead of 50%) and survival (expected 2-year overall survival of 75% calculated from the moment of reaching cCR instead of 3-year overall survival of 67% calculated from diagnosis). This reduced the number of required patients with cCR from 264 to 224, but increased the required inclusions from 480 to approximately 740 patients at baseline. Conclusion Substantial amendments were made prior to closure of enrolment of the SANO trial. These amendments do not affect the outcomes of the trial compared to the original protocol. The first results are expected late 2023. If active surveillance plus surgery as needed after neoadjuvant chemoradiotherapy for oesophageal cancer leads to non-inferior overall survival compared to standard oesophagectomy, active surveillance can be implemented as a standard of care.

Abstract Background The Framingham hypertension risk score is a well-known and simple model for predicting hypertension in adults. In the current study, we aimed to assess the predictive ability of this model in a Middle Eastern population. Methods We studied 5423 participants, aged 20–69 years, without hypertension, who participated in two consecutive examination cycles of the Tehran Lipid and Glucose Study (TLGS). We assessed discrimination based on Harrell’s concordance statistic (c-index) and calibration (graphical comparison of predicted vs. observed). We evaluated the original, recalibrated (for intercept and slope), and revised (for beta coefficients) models. Results Over the 3-year follow-up period, 319 participants developed hypertension. The Framingham hypertension risk score performed well in discriminating between individuals who developed hypertension and those who did not (c-index = 0.81, 95% CI: 0.79–0.83). Initially, there was a systematic underestimation of the original risk score (events predicted), which was readily corrected by a simple model revision. Conclusions The revised Framingham hypertension risk score can be used as a screening tool in public health and clinical practice to facilitate the targeting of preventive interventions in high-risk Middle Eastern people.

Abstract Background We suggest an adaptive sample size calculation method for developing clinical prediction models, in which model performance is monitored sequentially as new data comes in. Methods We illustrate the approach using data for the diagnosis of ovarian cancer (n = 5914, 33% event fraction) and obstructive coronary artery disease (CAD; n = 4888, 44% event fraction). We used logistic regression to develop a prediction model consisting only of a priori selected predictors and assumed linear relations for continuous predictors. We mimicked prospective patient recruitment by developing the model on 100 randomly selected patients, and we used bootstrapping to internally validate the model. We sequentially added 50 random new patients until we reached a sample size of 3000 and re-estimated model performance at each step. We examined the required sample size for satisfying the following stopping rule: obtaining a calibration slope ≥ 0.9 and optimism in the c-statistic (or AUC) < = 0.02 at two consecutive sample sizes. This procedure was repeated 500 times. We also investigated the impact of alternative modeling strategies: modeling nonlinear relations for continuous predictors and correcting for bias on the model estimates (Firth’s correction). Results Better discrimination was achieved in the ovarian cancer data (c-statistic 0.9 with 7 predictors) than in the CAD data (c-statistic 0.7 with 11 predictors). Adequate calibration and limited optimism in discrimination was achieved after a median of 450 patients (interquartile range 450–500) for the ovarian cancer data (22 events per parameter (EPP), 20–24) and 850 patients (750–900) for the CAD data (33 EPP, 30–35). A stricter criterion, requiring AUC optimism < = 0.01, was met with a median of 500 (23 EPP) and 1500 (59 EPP) patients, respectively. These sample sizes were much higher than the well-known 10 EPP rule of thumb and slightly higher than a recently published fixed sample size calculation method by Riley et al. Higher sample sizes were required when nonlinear relationships were modeled, and lower sample sizes when Firth’s correction was used. Conclusions Adaptive sample size determination can be a useful supplement to fixed a priori sample size calculations, because it allows to tailor the sample size to the specific prediction modeling context in a dynamic fashion.

Abstract The number of clinician burnouts is increasing and has been linked to a high administrative burden. Automatic speech recognition (ASR) and natural language processing (NLP) techniques may address this issue by creating the possibility of automating clinical documentation with a “digital scribe”. We reviewed the current status of the digital scribe in development towards clinical practice and present a scope for future research. We performed a literature search of four scientific databases (Medline, Web of Science, ACL, and Arxiv) and requested several companies that offer digital scribes to provide performance data. We included articles that described the use of models on clinical conversational data, either automatically or manually transcribed, to automate clinical documentation. Of 20 included articles, three described ASR models for clinical conversations. The other 17 articles presented models for entity extraction, classification, or summarization of clinical conversations. Two studies examined the system’s clinical validity and usability, while the other 18 studies only assessed their model’s technical validity on the specific NLP task. One company provided performance data. The most promising models use context-sensitive word embeddings in combination with attention-based neural networks. However, the studies on digital scribes only focus on technical validity, while companies offering digital scribes do not publish information on any of the research phases. Future research should focus on more extensive reporting, iteratively studying technical validity and clinical validity and usability, and investigating the clinical utility of digital scribes.

ObjectiveSpine injury is highly prevalent in patients with poly-trauma, but data on the co-occurrence of spine trauma in patients with traumatic brain injury (TBI) are scarce. In this study, we used the Collaborative European NeuroTrauma Effectiveness Research in Traumatic Brain Injury (CENTER-TBI) database to assess the prevalence, characteristics, and outcomes of patients with TBI and a concurrent traumatic spinal injury (TSI).MethodsData from the European multi-center CENTER-TBI study were analyzed. Adult patients with TBI (≥18 years) presenting with a concomitant, isolated TSI of at least serious severity (Abbreviated Injury Scale; AIS ≥3) were included. For outcome analysis, comparison groups of TBI patients with TSI and systemic injuries (non-isolated TSI) and without TSI were created using propensity score matching. Rates of mortality, unfavorable outcomes (Glasgow Outcome Scale Extended; GOSe < 5), and full recovery (GOSe 7–8) of all patients and separately for patients with only mild TBI (mTBI) were compared between groups at 6-month follow-up.ResultsA total of 164 (4%) of the 4,254 CENTER-TBI core study patients suffered from a concomitant isolated TSI. The median age was 53 [interquartile range (IQR): 37–66] years and 71% of patients were men. mTBI was documented in 62% of cases, followed by severe TBI (26%), and spine injuries were mostly cervical (63%) or thoracic (31%). Surgical spine stabilization was performed in 19% of cases and 57% of patients were admitted to the ICU. Mortality at 6 months was 11% and only 36% of patients regained full recovery. There were no significant differences in the 6-month rates of mortality, unfavorable outcomes, or full recovery between TBI patients with and without concomitant isolated TSI. However, concomitant non-isolated TSI was associated with an unfavorable outcome and a higher mortality. In patients with mTBI, a negative association with full recovery could be observed for both concomitant isolated and non-isolated TSI.ConclusionRates of mortality, unfavorable outcomes, and full recovery in TBI patients with and without concomitant, isolated TSIs were comparable after 6 months. However, in patients with mTBI, concomitant TSI was a negative predictor for a full recovery. These findings might indicate that patients with moderate to severe TBI do not necessarily exhibit worse outcomes when having a concomitant TSI, whereas patients with mTBI might be more affected.

Abstract Background There is a growing interest in assessment of the quality of hospital care, based on outcome measures. Many quality of care comparisons rely on binary outcomes, for example mortality rates. Due to low numbers, the observed differences in outcome are partly subject to chance. We aimed to quantify the gain in efficiency by ordinal instead of binary outcome analyses for hospital comparisons. We analyzed patients with traumatic brain injury (TBI) and stroke as examples. Methods We sampled patients from two trials. We simulated ordinal and dichotomous outcomes based on the modified Rankin Scale (stroke) and Glasgow Outcome Scale (TBI) in scenarios with and without true differences between hospitals in outcome. The potential efficiency gain of ordinal outcomes, analyzed with ordinal logistic regression, compared to dichotomous outcomes, analyzed with binary logistic regression was expressed as the possible reduction in sample size while keeping the same statistical power to detect outliers. Results In the IMPACT study (9578 patients in 265 hospitals, mean number of patients per hospital = 36), the analysis of the ordinal scale rather than the dichotomized scale (‘unfavorable outcome’), allowed for up to 32% less patients in the analysis without a loss of power. In the PRACTISE trial (1657 patients in 12 hospitals, mean number of patients per hospital = 138), ordinal analysis allowed for 13% less patients. Compared to mortality, ordinal outcome analyses allowed for up to 37 to 63% less patients. Conclusions Ordinal analyses provide the statistical power of substantially larger studies which have been analyzed with dichotomization of endpoints. We advise to exploit ordinal outcome measures for hospital comparisons, in order to increase efficiency in quality of care measurements. Trial registration We do not report the results of a health care intervention.

Objetivo: Desarrollar un modelo de predicción de riesgo para infección posoperatoria mayor (IPM) a cirugía cardiaca pediátrica y validar el de la Society of Thoracic Surgeons (STS). Materiales y métodos: Se analizó una cohorte retrospectiva de 1025 niños sometidos a cirugía cardiaca con circulación extracorpórea (CEC) del 2000 al 2010. Se empleó un modelo de regresión logística y se validó el modelo. Resultados: De los 1025 pacientes, 59 (5,8%) tuvieron al menos un episodio de IPM (4,8% sepsis, 1% mediastinitis, 0% endocarditis). La mortalidad hospitalaria (63% vs. 13%; p 0,001), al igual que la duración de la ventilación posoperatoria (301,6 vs. 34,3 horas; p 0,001) y la estancia en la unidad de cuidados intensivos (20,9 vs. 5,1 días; p 0,001) fueron mayores en los pacientes con IPM. Los factores predictores fueron: edad, sexo, peso, cardiopatía cianótica, RACHS-1 3-4, clase funcional IV modificada por Ross, estancia hospitalaria previa y antecedente de ventilación mecánica. El modelo tuvo un c-estadístico de 0,80 (intervalo de confianza [IC] al 95%: 0,74-0,86) y es clínicamente útil. El modelo de la STS mostró un c-estadístico de 0,78 (IC 95%: 0,71-0,84) y Hosmer-Lemeshow de 18,2 (p = 0,020). Se realizó una comparación entre ambos modelos empleando una prueba exacta de Fisher. Conclusión: Se desarrolló un modelo para identificar preoperatoriamente a niños con alto riesgo de infección grave después de una cirugía cardiaca con CEC con buen desempeño y calibración. Asimismo, se validó el modelo de la STS con moderada discriminación.