JURaSSiC: accuracy of clinician vs risk score prediction of ischemic stroke outcomes

Derivation and Validation of a Scoring System for Predicting Poor Outcome After Posterior Circulation Ischemic Stroke in China

BACKGROUND AND OBJECTIVES

Guidelines for posterior circulation ischemic stroke (PCIS) treatment are lacking and outcome prediction is crucial for patients and clinicians. We aimed to develop and validate a prognostic score to predict the poor outcome for patients with PCIS.

METHODS

The score was developed from a prospective derivation cohort named the Third China National Stroke Registry (August 2015-March 2018) and validated in a spatiotemporal independent validation cohort (December 2017-March 2023) in China. Patients with PCIS with acute infarctions defined as hyperintense lesions on diffusion-weighted imaging were included in this study. The poor outcome was measured as modified Rankin scale (mRS) score 3-6 at 3 months after PCIS. Multivariable logistic regression analysis was used to identify predictors for poor outcome. The prognostic score, namely PCIS Outcome Score (PCISOS), was developed by assigning points to variables based on their relative β-coefficients in the logistic model.

RESULTS

The PCISOS was derived from 3,294 patients (median age 62 [interquartile range (IQR) 55-70] years; 2,250 [68.3%] men) and validated in 501 patients (median age 61 [IQR 53-68] years; 404 [80.6%] men). Among them, 384 (11.7%) and 64 (12.8%) had poor outcome 3 months after stroke in respective cohorts. Age, mRS before admission, NIH Stroke Scale on admission, ischemic stroke history, infarction distribution, basilar artery, and posterior cerebral artery stenosis or occlusion were identified as independent predictors for poor outcome and included in PCISOS. This easy-to-use integer scoring system identified a marked risk gradient between 4 risk groups. PCISOS performed better than previous scores, with an excellent discrimination (C statistic) of 0.80 (95% CI 0.77-0.83) in the derivation cohort and 0.81 (95% CI 0.77-0.84) in the validation cohort. Calibration test showed high agreement between the predicted and observed outcomes in both cohorts.

DISCUSSION

PCISOS can be applied for patients with PCIS with acute infarctions to predict functional outcome at 3 months post-PCIS. This simple tool helps clinicians to identify patients with PCIS with higher risk of poor outcome and provides reliable outcome expectations for patients. This information might be used for personalized rehabilitation plan and patient selection for future clinical trials to reduce disability and mortality.

Artificial intelligence to predict individualized outcome of acute ischemic stroke patients: The SIBILLA project

INTRODUCTION

Formulating reliable prognosis for ischemic stroke patients remains a challenging task. We aimed to develop an artificial intelligence model able to formulate in the first 24 h after stroke an individualized prognosis in terms of NIHSS.

PATIENTS AND METHODS

Seven hundred ninety four acute ischemic stroke patients were divided into a training (597) and testing (197) cohort. Clinical and instrumental data were collected in the first 24 h. We evaluated the performance of four machine-learning models (Random Forest, K-Nearest Neighbors, Support Vector Machine, XGBoost) in predicting NIHSS at discharge both in terms of variation between discharge and admission (regressor approach) and in terms of severity class namely NIHSS 0-5, 6-10, 11-20, >20 (classifier approach). We used Shapley Additive exPlanations values to weight features impact on predictions.

RESULTS

XGBoost emerged as the best performing model. The classifier and regressor approaches perform similarly in terms of accuracy (80% vs 75%) and f1-score (79% vs 77%) respectively. However, the regressor has higher precision (85% vs 68%) in predicting prognosis of very severe stroke patients (NIHSS > 20). NIHSS at admission and 24 hours, GCS at 24 hours, heart rate, acute ischemic lesion on CT-scan and TICI score were the most impacting features on the prediction.

DISCUSSION

Our approach, which employs an artificial intelligence based-tool, inherently able to continuously learn and improve its performance, could improve care pathway and support stroke physicians in the communication with patients and caregivers.

CONCLUSION

XGBoost reliably predicts individualized outcome in terms of NIHSS at discharge in the first 24 hours after stroke.

Biomarkers to improve functional outcome prediction after ischemic stroke: Results from the SICFAIL, STRAWINSKI, and PREDICT studies

BACKGROUND AND AIMS

Acute ischemic stroke (AIS) outcome prognostication remains challenging despite available prognostic models. We investigated whether a biomarker panel improves the predictive performance of established prognostic scores.

METHODS

We investigated the improvement in discrimination, calibration, and overall performance by adding five biomarkers (procalcitonin, copeptin, cortisol, mid-regional pro-atrial natriuretic peptide (MR-proANP), and N-terminal pro-B-type natriuretic peptide (NT-proBNP)) to the Acute Stroke Registry and Analysis of Lausanne (ASTRAL) and age/NIHSS scores using data from two prospective cohort studies (SICFAIL, PREDICT) and one clinical trial (STRAWINSKI). Poor outcome was defined as mRS > 2 at 12 (SICFAIL, derivation dataset) or 3 months (PREDICT/STRAWINSKI, pooled external validation dataset).

RESULTS

Among 412 SICFAIL participants (median age 70 years, quartiles 59-78; 63% male; median NIHSS score 3, quartiles 1-5), 29% had a poor outcome. Area under the curve of the ASTRAL and age/NIHSS were 0.76 (95% CI 0.71-0.81) and 0.77 (95% CI 0.73-0.82), respectively. Copeptin (0.79, 95% CI 0.74-0.84), NT-proBNP (0.80, 95% CI 0.76-0.84), and MR-proANP (0.79, 95% CI 0.75-0.84) significantly improved ASTRAL score's discrimination, calibration, and overall performance. Copeptin improved age/NIHSS model's discrimination, copeptin, MR-proANP, and NT-proBNP improved its calibration and overall performance. In the validation dataset (450 patients, median age 73 years, quartiles 66-81; 54% men; median NIHSS score 8, quartiles 3-14), copeptin was independently associated with various definitions of poor outcome and also mortality. Copeptin did not increase model's discrimination but it did improve calibration and overall model performance.

DISCUSSION

Copeptin, NT-proBNP, and MR-proANP improved modest but consistently the predictive performance of established prognostic scores in patients with mild AIS. Copeptin was most consistently associated with poor outcome in patients with moderate to severe AIS, although its added prognostic value was less obvious.

Barthel Index, SPAN-100, and NIHSS Studies on the Predictive Value of Prognosis in Patients With Thrombolysis

OBJECTIVE

We mainly explore the predictive value of Barthel Index (BI), SPAN-100, and National Institute of Health stroke scale (NIHSS) scores on clinical prognosis and functional outcomes in thrombolytic patients and compare the differences in the predictive values of the above 3 scales so as to provide an effective basis to evaluate the prognosis of thrombolytic patients.

METHODS

Data were collected from 212 patients with the first-onset AIS (acute ischemic stroke). The enrolled patients were treated with recombinant tissue plasminogen activator thrombolytic therapy and were divided into 2 groups according to the modified Rankin Scale (mRS) score at discharge: the prognosis group (mRS≤2 points) and the poor prognosis group (mRS≥3 points). Logistic multivariate analysis was used to analyze the predictors of poor prognosis in patients with thrombolysis. MedCalc software was used to plot receiver operating characteristic (ROC) curves, calculate the area under the ROC curve (AUC), and compare the prediction performance of the 3 scales by the Delong and colleagues' method, and the difference of P <0.05 was statistically significant.

RESULTS

Logistic binary regression multivariate analysis suggested that BI was a predictor of poor prognosis for thrombolytic therapy in patients with AIS. The lower the BI score, the poorer the prognosis. The AUC for BI score was 0.862, 95% CI (0.808-0.906), NIHSS score AUC was 0.665, 95% CI (0.597-0.728), and SPAN-100 score AUC was 0.640, 95% CI (0.572-0.705). AUC comparison of 3 scoring ROC curves suggested statistically significant differences between BI and NIHSS ( PC =0.0000), BI and SPAN-100 ( PC =0.0000); no significant difference was observed between SPAN-100 and NIHSS ( PC =1.7997).

CONCLUSIONS

Simple BI scores have a high prognostic value for thrombolytic therapy in AIS.

Development and internal validation of machine learning-based models and external validation of existing risk scores for outcome prediction in patients with ischaemic stroke

Aims

We developed new machine learning (ML) models and externally validated existing statistical models [ischaemic stroke predictive risk score (iScore) and totalled health risks in vascular events (THRIVE) scores] for predicting the composite of recurrent stroke or all-cause mortality at 90 days and at 3 years after hospitalization for first acute ischaemic stroke (AIS).

Methods and results

In adults hospitalized with AIS from January 2005 to November 2016, with follow-up until November 2019, we developed three ML models [random forest (RF), support vector machine (SVM), and extreme gradient boosting (XGBOOST)] and externally validated the iScore and THRIVE scores for predicting the composite outcomes after AIS hospitalization, using data from 721 patients and 90 potential predictor variables. At 90 days and 3 years, 11 and 34% of patients, respectively, reached the composite outcome. For the 90-day prediction, the area under the receiver operating characteristic curve (AUC) was 0.779 for RF, 0.771 for SVM, 0.772 for XGBOOST, 0.720 for iScore, and 0.664 for THRIVE. For 3-year prediction, the AUC was 0.743 for RF, 0.777 for SVM, 0.773 for XGBOOST, 0.710 for iScore, and 0.675 for THRIVE.

Conclusion

The study provided three ML-based predictive models that achieved good discrimination and clinical usefulness in outcome prediction after AIS and broadened the application of the iScore and THRIVE scoring system for long-term outcome prediction. Our findings warrant comparative analyses of ML and existing statistical method-based risk prediction tools for outcome prediction after AIS in new data sets.

Recent Trends in the Withdrawal of Life-Sustaining Treatment in Patients with Acute Cerebrovascular Disease : 2017-2021

OBJECTIVE

The Act on Life-Sustaining Treatment (LST) decisions for end-of-life patients has been effective since February 2018. An increasing number of patients and their families want to withhold or withdraw from LST when medical futility is expected. This study aimed to investigate the status of the Act on LST decisions for patients with acute cerebrovascular disease at a single hospital.

METHODS

Between January 2017 and December 2021, 227 patients with acute cerebrovascular diseases, including hemorrhagic stroke (n=184) and ischemic stroke (n=43), died at the hospital. The study period was divided into the periods before and after the Act.

RESULTS

The duration of hospitalization decreased after the Act was implemented compared to before (15.9±16.1 vs. 11.2±18.6 days, p=0.127). The rate of obtaining consent for the LST plan tended to increase after the Act (139/183 [76.0%] vs. 27/44 [61.4%], p=0.077). Notably, none of the patients made an LST decision independently. Ventilator withdrawal was more frequently performed after the Act than before (52/183 [28.4%] vs. 0/44 [0%], p<0.001). Conversely, the rate of organ donation decreased after the Act was implemented (5/183 [2.7%] vs. 6/44 [13.6%], p=0.008). Refusal to undergo surgery was more common after the Act was implemented than before (87/149 [58.4%] vs. 15/41 [36.6%], p=0.021) among the 190 patients who required surgery.

CONCLUSION

After the Act on LST decisions was implemented, the rate of LST withdrawal increased in patients with acute cerebrovascular disease. However, the decision to withdraw LST was made by the patient's family rather than the patient themselves. After the execution of the Act, we also observed an increased rate of refusal to undergo surgery and a decreased rate of organ donation. The Act on LST decisions may reduce unnecessary treatments that prolong end-of-life processes without a curative effect. However, the widespread application of this law may also reduce beneficial treatments and contribute to a decline in organ donation.

The Charlotte Large artery occlusion Endovascular therapy Outcome Score predicts independent outcome after thrombectomy

BACKGROUND AND PURPOSE

Predicting functional outcomes after endovascular thrombectomy (EVT) is of interest to patients and families as they navigate hospital and post-acute care decision-making. We evaluated the prognostic ability of several scales to predict good neurological function after EVT.

METHODS

We retrospectively analyzed records from a health system's code stroke registry, including consecutive successful thrombectomy patients from August 2020 to February 2023 presenting with an anterior circulation large vessel occlusion who were evaluated with pre-EVT CT perfusion. Primary and secondary outcomes were 90-day modified Rankin Scale (mRS) scores 0-2 and 0-1, respectively. Logistic regression was performed to evaluate the ability of each scale to predict the outcomes. Scales were compared by calculating the area under the curve (AUC).

RESULTS

A total of 465 patients (mean age 68.1 [±14.9] years, median National Institutes of Health Stroke Scale [NIHSS] 16 [11-21]) met inclusion criteria. In the logistic regression, the Charlotte Large artery occlusion Endovascular therapy Outcome Score (CLEOS), Totaled Health Risks in Vascular Events, Houston Intra-Arterial Therapy-2, Pittsburgh Response to Endovascular therapy, and Stroke Prognostication using Age and NIHSS were significant in predicting the primary and secondary outcomes. CLEOS was superior to all other scales in predicting 90-day mRS 0-2 (AUC .75, 95% confidence interval [CI] .70-.80) and mRS 0-1 (AUC .74, 95% CI .69-.78). Twenty of 22 patients (90.9%) with CLEOS <315 had 90-day mRS 0-2.

CONCLUSIONS

CLEOS predicts independent and excellent neurological function after anterior circulation EVT.

PREDICT-TBI: Comparison of Physician Predictions with the IMPACT Model to Predict 6-Month Functional Outcome in Traumatic Brain Injury

BACKGROUND

Predicting functional outcome in critically ill patients with traumatic brain injury (TBI) strongly influences end-of-life decisions and information for surrogate decision makers. Despite well-validated prognostic models, clinicians most often rely on their subjective perception of prognosis. In this study, we aimed to compare physicians' predictions with the International Mission on Prognosis and Analysis of Clinical Trials in TBI (IMPACT) prognostic model for predicting an unfavorable functional outcome at 6 months after moderate or severe TBI.

METHODS

PREDICT-TBI is a prospective study of patients with moderate to severe TBI. Patients were admitted to a neurocritical care unit and were excluded if they died or had withdrawal of life-sustaining treatments within the first 24 h. In a paired study design, we compared the accuracy of physician prediction on day 1 with the prediction of the IMPACT model as two diagnostic tests in predicting unfavorable outcome 6 months after TBI. Unfavorable outcome was assessed by the Glasgow Outcome Scale from 1 to 3 by using a structured telephone interview. The primary end point was the difference between the discrimination ability of the physician and the IMPACT model assessed by the area under the curve.

RESULTS

Of the 93 patients with inclusion and exclusion criteria, 80 patients reached the primary end point. At 6 months, 29 patients (36%) had unfavorable outcome. A total of 31 clinicians participated in the study. Physicians' predictions showed an area under the curve of 0.79 (95% confidence interval 0.68-0.89), against 0.80 (95% confidence interval 0.69-0.91) for the laboratory IMPACT model, with no statistical difference (p = 0.88). Both approaches were well calibrated. Agreement between physicians was moderate (κ = 0.56). Lack of experience was not associated with prediction accuracy (p = 0.58).

CONCLUSIONS

Predictions made by physicians for functional outcome were overall moderately accurate, and no statistical difference was found with the IMPACT models, possibly due to a lack of power. The significant variability between physician assessments suggests prediction could be improved through peer reviewing, with the support of the IMPACT models, to provide a realistic expectation of outcome to families and guide discussions about end-of-life decisions.

Cognitive bias during clinical decision-making and its influence on patient outcomes in the emergency department: A scoping review

BACKGROUND

An integral part of clinical practice is decision-making. Yet there is widespread acceptance that there is evidence of cognitive bias within clinical practice among nurses and physicians. However, how cognitive bias among emergency nurses and physicians' decision-making influences patient outcomes remains unclear.

AIM

The aim of this review was to systematically synthesise research exploring the emergency nurses' and physicians' cognitive bias in decision-making and its influence on patient outcomes.

METHODS

This scoping review was guided by the PRISMA Extension for Scoping Reviews. The databases searched included CINAHL, MEDLINE, Web of Science and PubMed. No date limits were applied. The Patterns, Advances, Gaps, Evidence for practice and Research recommendation (PAGER) framework was used to guide the discussion.

RESULTS

The review included 18 articles, consisting of 10 primary studies (nine quantitative and one qualitative) and eight literature reviews. Of the 18 articles, nine investigated physicians, five articles examined nurses, and four both physicians and nurses with sample sizes ranging from 13 to 3547. Six primary studies were cross-sectional and five used hypothetical scenarios, and one real-world assessment. Three were experimental studies. Twenty-nine cognitive biases were identified with Implicit bias (n = 12) most frequently explored, followed by outcome bias (n = 4). Results were inconclusive regarding the influence of biases on treatment decisions and patient outcomes. Four key themes were identified; (i) cognitive biases among emergency clinicians; (ii) measurement of cognitive bias; (iii) influence of cognitive bias on clinical decision-making; and (iv) association between emergency clinicians' cognitive bias and patient outcomes.

CONCLUSIONS

This review identified that cognitive biases were present among emergency nurses and physicians during clinical decision-making, but it remains unclear how cognitive bias influences patient outcomes. Further research examining emergency clinicians' cognitive bias is required.

RELEVANCE TO CLINICAL PRACTICE

Awareness of emergency clinicians' own cognitive biases may result to the provision of equity in care.

NO PATIENT OR PUBLIC CONTRIBUTION IN THIS REVIEW

We intend to disseminate the results through publication in a peer-reviewed journals and conference presentations.

Development and Validation of a Postprocedural Model to Predict Outcome After Endovascular Treatment for Ischemic Stroke

Importance

Outcome prediction after endovascular treatment (EVT) for ischemic stroke is important to patients, family members, and physicians.

Objective

To develop and validate a model based on preprocedural and postprocedural characteristics to predict functional outcome for individual patients after EVT.

Design, Setting, and Participants

A prediction model was developed using individual patient data from 7 randomized clinical trials, performed between December 2010 and December 2014. The model was developed within the Highly Effective Reperfusion Evaluated in Multiple Endovascular Stroke Trials (HERMES) collaboration and external validation in data from the Dutch Multicenter Randomized Clinical Trial of Endovascular Treatment for Acute Ischemic Stroke in the Netherlands (MR CLEAN) Registry of patients treated in clinical practice between March 2014 and November 2017. Participants included patients from multiple centers throughout different countries in Europe, North America, East Asia, and Oceania (derivation cohort), and multiple centers in the Netherlands (validation cohort). Included were adult patients with a history of ischemic stroke from an intracranial large vessel occlusion in the anterior circulation who underwent EVT within 12 hours of symptom onset or last seen well. Data were last analyzed in July 2022.

Main Outcome(s) and Measure(s)

A total of 19 variables were assessed by multivariable ordinal regression to predict functional outcome (modified Rankin Scale [mRS] score) 90 days after EVT. Variables were routinely available 1 day after EVT. Akaike information criterion (AIC) was used to optimize model fit vs model complexity. Probabilities for functional independence (mRS 0-2) and survival (mRS 0-5) were derived from the ordinal model. Model performance was expressed with discrimination (C statistic) and calibration.

Results

A total of 781 patients (median [IQR] age, 67 [57-76] years; 414 men [53%]) constituted the derivation cohort, and 3260 patients (median [IQR] age, 72 [61-80] years; 1684 men [52%]) composed the validation cohort. Nine variables were included in the model: age, baseline National Institutes of Health Stroke Scale (NIHSS) score, prestroke mRS score, history of diabetes, occlusion location, collateral score, reperfusion grade, NIHSS score at 24 hours, and symptomatic intracranial hemorrhage 24 hours after EVT. External validation in the MR CLEAN Registry showed excellent discriminative ability for functional independence (C statistic, 0.91; 95% CI, 0.90-0.92) and survival (0.89; 95% CI, 0.88-0.90). The proportion of functional independence in the MR CLEAN Registry was systematically higher than predicted by the model (41% vs 34%), whereas observed and predicted survival were similar (72% vs 75%). The model was updated and implemented for clinical use.

Conclusion and relevance

The prognostic tool MR PREDICTS@24H can be applied 1 day after EVT to accurately predict functional outcome for individual patients at 90 days and to provide reliable outcome expectations and personalize follow-up and rehabilitation plans. It will need further validation and updating for contemporary patients.

Heuristics and biases in medical decision-making under uncertainty: The case of neuropronostication for consciousness disorders

Neuropronostication for consciousness disorders can be very complex and prone to high uncertainty. Despite notable advancements in the development of dedicated scales and physiological markers using innovative paradigms, these technical progressions are often overshadowed by factors intrinsic to the medical environment. Beyond the scarcity of objective data guiding medical decisions, factors like time pressure, fatigue, multitasking, and emotional load can drive clinicians to rely more on heuristic-based clinical reasoning. Such an approach, albeit beneficial under certain circumstances, may lead to systematic error judgments and impair medical decisions, especially in complex and uncertain environments. After a brief review of the main theoretical frameworks, this paper explores the influence of clinicians' cognitive biases on clinical reasoning and decision-making in the challenging context of neuroprognostication for consciousness disorders. The discussion further revolves around developing and implementing various strategies designed to mitigate these biases and their impact, aiming to enhance the quality of care and the patient safety.

Developing and Validating a Prediction Model For Death or Critical Illness in Hospitalized Adults, an Opportunity for Human-Computer Collaboration

Hospital early warning systems that use machine learning (ML) to predict clinical deterioration are increasingly being used to aid clinical decision-making. However, it is not known how ML predictions complement physician and nurse judgment. Our objective was to train and validate a ML model to predict patient deterioration and compare model predictions with real-world physician and nurse predictions.

DESIGN

Retrospective and prospective cohort study.

SETTING

Academic tertiary care hospital.

PATIENTS

Adult general internal medicine hospitalizations.

MEASUREMENTS AND MAIN RESULTS

We developed and validated a neural network model to predict in-hospital death and ICU admission in 23,528 hospitalizations between April 2011 and April 2019. We then compared model predictions with 3,374 prospectively collected predictions from nurses, residents, and attending physicians about their own patients in 960 hospitalizations between April 30, and August 28, 2019. ML model predictions achieved clinician-level accuracy for predicting ICU admission or death (ML median F1 score 0.32 [interquartile range (IQR) 0.30-0.34], AUC 0.77 [IQ 0.76-0.78]; clinicians median F1-score 0.33 [IQR 0.30-0.35], AUC 0.64 [IQR 0.63-0.66]). ML predictions were more accurate than clinicians for ICU admission. Of all ICU admissions and deaths, 36% occurred in hospitalizations where the model and clinicians disagreed. Combining human and model predictions detected 49% of clinical deterioration events, improving sensitivity by 16% compared with clinicians alone and 24% compared with the model alone while maintaining a positive predictive value of 33%, thus keeping false alarms at a clinically acceptable level.

CONCLUSIONS

ML models can complement clinician judgment to predict clinical deterioration in hospital. These findings demonstrate important opportunities for human-computer collaboration to improve prognostication and personalized medicine in hospital.

IScore, a useful prognostic tool for patients with acute ischemic stroke treated with intravenous thrombolysis: a validation study

BACKGROUND

 Stroke is one of the major causes of disability and mortality worldwide. Up to 30% of individuals who experience stroke die within 30 days, and more than 50% of those who survive will have some degree of disability. There are some predetermining factors based on admission data that could be used to objectively assess the odds of poor outcomes, including the Ischemic Stroke Predictive Risk Score (IScore).

OBJECTIVE

 To analyze and validate the IScore in patients undergoing intravenous thrombolysis for stroke and compare the results of this predictor with actual death and disability outcomes.

METHODS

 In a retrospective study, data were collected from a database housed at the Stroke Unit of the Teaching Hospital of Universidade Federal do Paraná, Southern Brazil. The IScore was applied to admission data from 239 patients, and the results were compared with actual outcomes (death and disability) within 30 days and 1 year after the stroke event. Data analysis was performed using an analysis of the receiver operating characteristic (ROC) curve to determine the sensitivity and specificity of the IScore in the study population.

RESULTS

 The IScore demonstrated moderate sensitivity and high specificity in patients with stroke who underwent thrombolysis when evaluated after 30 days and 1 year of the event.

CONCLUSIONS

 The IScore can be applied to in stroke patients undergoing thrombolysis; therefore, it may be used as an objective prognostic tool to guide clinical decision-making. Understanding the prognosis of patients in the acute phase can assist clinicians in making the best therapeutic decisions and enable better end-of-life care.

Evolution of Making Clinical Predictions in Neurosurgery

Prediction of clinical outcomes is an essential task for every physician. Physicians may base their clinical prediction of an individual patient on their intuition and on scientific material such as studies presenting population risks and studies reporting on risk factors (prognostic factors). A relatively new and more informative approach for making clinical predictions relies on the use of statistical models that simultaneously consider multiple predictors that provide an estimate of the patient's absolute risk of an outcome. There is a growing body of literature in the neurosurgical field reporting on clinical prediction models. These tools have high potential in supporting (not replacing) neurosurgeons with their prediction of a patient's outcome. If used sensibly, these tools pave the way for more informed decision-making with or for individual patients. Patients and their significant others want to know their risk of the anticipated outcome, how it is derived, and the uncertainty associated with it. Learning from these prediction models and communicating the output to others has become an increasingly important skill neurosurgeons have to master. This article describes the evolution of making clinical predictions in neurosurgery, synopsizes key phases for the generation of a useful clinical prediction model, and addresses some considerations when deploying and communicating the results of a prediction model. The paper is illustrated with multiple examples from the neurosurgical literature, including predicting arachnoid cyst rupture, predicting rebleeding in patients suffering from aneurysmal subarachnoid hemorrhage, and predicting survival in glioblastoma patients.

Palliative care approaches to acute stroke in the hospital setting

Stroke is a prevalent neurologic condition that portends a high risk of morbidity and mortality such that patients impacted by stroke and their caregivers can benefit from palliative care at the time of diagnosis and throughout the disease trajectory. Clinicians who care for stroke patients should be adept at establishing rapport with patients and caregivers, delivering serious news, responding to emotions, discussing prognosis, and establishing goals of care efficiently in an acute stroke setting. Aggressive stroke care can be integrated with a palliative approach to care that involves aligning the available treatment options with a patient's values and goals of care. Reassessing the goals throughout the hospitalization provides an opportunity for continued shared decision-making about the intensity of poststroke interventions. The palliative needs for stroke patients may increase over time depending on the severity of disease, poststroke complications, stroke-related symptoms, and treatment intensity preferences. If the decision is made to transition the focus of care to comfort, the support of an interdisciplinary palliative care or hospice team can be beneficial to the patient, family members, and surrogate decision makers.

Ethical Considerations in Surgical Decompression for Stroke

Stroke is a major cause of morbidity and mortality. Neurosurgical decompression is often considered for the treatment of malignant infarcts and intraparenchymal hemorrhages, but this treatment can be frought with ethical dilemmas. In this article, the authors outline the primary principles of bioethics and their application to stroke care, provide an overview of key ethical issues and special situations in the neurosurgical management of stroke, and highlight methods to improve ethical decision-making for patients with stroke. Understanding these ethical principles is essential for stroke care teams to deliver appropriate, timely, and ethical care to patients with stroke.

Prognostication of ICU Patients by Providers with and without Neurocritical Care Training

BACKGROUND

Predictions of functional outcome in neurocritical care (NCC) patients impact care decisions. This study compared the predictive values (PVs) of good and poor functional outcome among health care providers with and without NCC training.

METHODS

Consecutive patients who were intubated for  ≥ 72 h with primary neurological illness or neurological complications were prospectively enrolled and followed for 6-month functional outcome. Medical intensive care unit (MICU) attendings, NCC attendings, residents (RES), and nurses (RN) predicted 6-month functional outcome on the modified Rankin scale (mRS). The primary objective was to compare these four groups' PVs of a good (mRS score 0-3) and a poor (mRS score 4-6) outcome prediction.

RESULTS

Two hundred eighty-nine patients were enrolled. One hundred seventy-six had mRS scores predicted by a provider from each group and were included in the primary outcome analysis. At 6 months, 54 (31%) patients had good outcome and 122 (69%) had poor outcome. Compared with other providers, NCC attendings expected better outcomes (p < 0.001). Consequently, the PV of a poor outcome prediction by NCC attendings was higher (96% [95% confidence interval [CI] 89-99%]) than that by MICU attendings (88% [95% CI 80-93%]), RES (82% [95% CI 74-88%]), and RN (85% [95% CI 77-91%]) (p = 0.047, 0.002, and 0.012, respectively). When patients who had withdrawal of life-sustaining therapy (n = 67) were excluded, NCC attendings remained better at predicting poor outcome (NCC 90% [95% CI 75-97%] vs. MICU 73% [95% CI 59-84%], p = 0.064). The PV of a good outcome prediction was similar among groups (MICU 65% [95% CI 52-76%], NCC 63% [95% CI 51-73%], RES 71% [95% CI 55-84%], and RN 64% [95% CI 50-76%]).

CONCLUSIONS

Neurointensivists expected better outcomes than other providers and were better at predicting poor functional outcomes. The PV of a good outcome prediction was modest among all providers.

Prognostication in Acute Neurological Emergencies

BACKGROUND

For patients with acute, serious neurological conditions presenting to the emergency department (ED), prognostication is typically based on clinical experience, scoring systems and patient co-morbidities. Because estimating a poor prognosis influences caregiver decisions to withdraw life-sustaining therapy, we investigated the consistency of prognostication across a spectrum of neurology physicians.

METHODS

Five acute neurological presentations (2 with large hemispheric infarction; 1 with brainstem infarction, 1 with lobar hemorrhage, and 1 with hypoxic-ischemic encephalopathy) were selected for a department-wide prognostication simulation exercise. All had presented to our tertiary care hospital's ED, where a poor outcome was predicted by the ED neurology team within 24 hours of onset. Relevant clinical, laboratory and imaging data available before ED prognostication were presented on a web-based platform to 120 providers blinded to the actual outcome. The provider was requested to rank-order, from most to least likely, the predicted 90-day modified Rankin Scale (mRS) score. To determine the accuracy of individual outcome predictions we compared the patient's the actual 90-day mRS score to highest ranked predicted mRS score. Additionally, the group's "weighted" outcomes, accounting for the entire spectrum of mRS scores ranked by all respondents, were compared to the actual outcome for each case. Consistency was compared between pre-specified provider roles: neurology trainees versus faculty; non-vascular versus vascular faculty.

RESULTS

Responses ranged from 106-110 per case. Individual predictions were highly variable, with predictions matching the actual mRS scores in as low as 2% of respondents in one case and 95% in another case. However, as a group, the weighted outcome matched the actual mRS score in 3 of 5 cases (60%). There was no significant difference between subgroups based on expertise (stroke/neurocritical care versus other) or experience (faculty versus trainee) in 4 of 5 cases.

CONCLUSION

Acute neuro-prognostication is highly variable and often inaccurate among neurology providers. Significant differences are not attributable to experience or subspecialty expertise. The mean outcome prediction from group of providers ("the wisdom of the crowd") may be superior to that of individual providers.

The intersection of prognostication and code status in patients with severe brain injury

Accurately estimating the prognosis of brain injury patients can be difficult, especially early in their course. Prognostication is important because it largely determines the care level we provide, from aggressive treatment for patients we predict could have a good outcome to withdrawal of treatment for those we expect will have a poor outcome. Accurate prognostication is required for ethical decision-making. However, several studies have shown that prognostication is frequently inaccurate and variable. Overly optimistic prognostication can lead to false hope and futile care. Overly pessimistic prognostication can lead to therapeutic nihilism. Overlapping is the powerful effect that cognitive biases, in particular code status, can play in shaping our perceptions and the care level we provide. The presence of Do Not Resuscitate orders has been shown to be associated with increased mortality. Based on a comprehensive search of peer-reviewed journals using a wide range of key terms, including prognostication, critical illness, brain injury, cognitive bias, and code status, the following is a review of prognostic accuracy and the effect of code status on outcome. Because withdrawal of treatment is the most common cause of death in the ICU, a clearer understanding of this intersection of prognostication and code status is needed.

Applying Machine Learning in Distributed Data Networks for Pharmacoepidemiologic and Pharmacovigilance Studies: Opportunities, Challenges, and Considerations

Increasing availability of electronic health databases capturing real-world experiences with medical products has garnered much interest in their use for pharmacoepidemiologic and pharmacovigilance studies. The traditional practice of having numerous groups use single databases to accomplish similar tasks and address common questions about medical products can be made more efficient through well-coordinated multi-database studies, greatly facilitated through distributed data network (DDN) architectures. Access to larger amounts of electronic health data within DDNs has created a growing interest in using data-adaptive machine learning (ML) techniques that can automatically model complex associations in high-dimensional data with minimal human guidance. However, the siloed storage and diverse nature of the databases in DDNs create unique challenges for using ML. In this paper, we discuss opportunities, challenges, and considerations for applying ML in DDNs for pharmacoepidemiologic and pharmacovigilance studies. We first discuss major types of activities performed by DDNs and how ML may be used. Next, we discuss practical data-related factors influencing how DDNs work in practice. We then combine these discussions and jointly consider how opportunities for ML are affected by practical data-related factors for DDNs, leading to several challenges. We present different approaches for addressing these challenges and highlight efforts that real-world DDNs have taken or are currently taking to help mitigate them. Despite these challenges, the time is ripe for the emerging interest to use ML in DDNs, and the utility of these data-adaptive modeling techniques in pharmacoepidemiologic and pharmacovigilance studies will likely continue to increase in the coming years.

Precision medicine in stroke: towards personalized outcome predictions using artificial intelligence

Stroke ranks among the leading causes for morbidity and mortality worldwide. New and continuously improving treatment options such as thrombolysis and thrombectomy have revolutionized acute stroke treatment in recent years. Following modern rhythms, the next revolution might well be the strategic use of the steadily increasing amounts of patient-related data for generating models enabling individualized outcome predictions. Milestones have already been achieved in several health care domains, as big data and artificial intelligence have entered everyday life.

The aim of this review is to synoptically illustrate and discuss how artificial intelligence approaches may help to compute single-patient predictions in stroke outcome research in the acute, subacute and chronic stage. We will present approaches considering demographic, clinical and electrophysiological data, as well as data originating from various imaging modalities and combinations thereof. We will outline their advantages, disadvantages, their potential pitfalls and the promises they hold with a special focus on a clinical audience. Throughout the review we will highlight methodological aspects of novel machine-learning approaches as they are particularly crucial to realize precision medicine. We will finally provide an outlook on how artificial intelligence approaches might contribute to enhancing favourable outcomes after stroke.

Predictive Analytics in Clinical Practice: Advantages and Disadvantages

Predictive analytics are increasingly reported by clinicians. These tools aim to improve patient outcomes in terms of quality, safety, and efficiency. However, deploying predictive analytics in clinical practice remains challenging today. We highlight several advantages and disadvantages of the application of predictive analytics in clinical practice. To flourish and reach its potential, predictive analytics need data that is of adequate quantity and quality, ideally tailored to clinical scenarios in equipoise regarding optimal management. Adequate reporting of predictive analytic tools is incumbent for uptake into clinical workflows. At least for now, the clinicians' knowledge, experience, and vigilance remain imperative for applying predictive analytics in clinical practice.

Machine Intelligence in Clinical Neuroscience: Taming the Unchained Prometheus

The democratization of machine learning (ML) through availability of open-source learning libraries, the availability of datasets in the "big data" era, increasing computing power even on mobile devices, and online training resources have both led to an explosion in applications and publications of ML in the clinical neurosciences, but has also enabled a dangerous amount of flawed analyses and cardinal methodological errors committed by benevolent authors. While powerful ML methods are nowadays available to almost anyone and can be applied after just few minutes of familiarizing oneself with these methods, that does not imply that one has mastered these techniques. This textbook for clinicians aims to demystify ML by illustrating its methodological foundations, as well as some specific applications throughout clinical neuroscience, and its limitations. While our mind can recognize, abstract, and deal with the many uncertainties in clinical practice, algorithms cannot. Algorithms must remain tools of our own mind, tools that we should be able to master, control, and apply to our advantage in an adjunctive manner. Our hope is that this book inspires and instructs physician-scientists to continue to develop the seeds that have been planted for machine intelligence in clinical neuroscience, not forgetting their inherent limitations.

Prognostic Scores for Large Vessel Occlusion Strokes

PURPOSE OF THE REVIEW

Endovascular thrombectomy (EVT) for large vessel occlusion strokes (LVOS) presents several treatment challenges. We provide a summary of existing tools for patient selection (pre-EVT tools) and for prognostication of long-term outcomes following reperfusion therapy (post-EVT tools).

RECENT FINDINGS

Recently published randomized trials demonstrated superiority of EVT over medical therapy alone for LVOS. Uniform patient selection paradigms based on demographic, clinical, and radiographic variables are not completely standardized, leading to variability in patient selection for EVT for LVOS. Post-EVT, an accurate assessment of long-term prognosis is critical in the decision-making process.

SUMMARY

Prognostic scores can serve as useful adjuncts to facilitate clinical decision-making during early management of patients with ischemic stroke, particularly those with LVOS. The acute management of LVOS comprises rapid clinical assessment, triage, and cerebrovascular imaging, followed by evaluation for candidacy for thrombolysis and EVT. Pre-EVT prognostic tools that accurately predict the likelihood of benefit from EVT may guide reliable, efficient, and cost-effective patient selection. Following EVT, severe stroke deficits and subacute poststroke complications that portend a poor prognosis may warrant invasive therapies. Clinical decisions regarding these treatment options involve careful discussions between providers and patient families, and are also based on prognosis provided by the treating clinician. Reliable post-EVT prognostic tools can facilitate this by providing accurate and objective prognostic information. Several prognostic tools have been developed and validated in the literature, some of which may be applicable in the pre-EVT and post-EVT settings, although clinical utility and application varies. Validation in contemporary datasets as well as implementation and impact studies are needed before these scales can be used to guide clinical decisions for individual patients.

End-of-Life Care Decision-Making in Stroke

Stroke is one of the leading causes of death and long-term disability in the United States. Though advances in interventions have improved patient survival after stroke, prognostication of long-term functional outcomes remains challenging, thereby complicating discussions of treatment goals. Stroke patients who require intensive care unit care often do not have the capacity themselves to participate in decision making processes, a fact that further complicates potential end-of-life care discussions after the immediate post-stroke period. Establishing clear, consistent communication with surrogates through shared decision-making represents best practice, as these surrogates face decisions regarding artificial nutrition, tracheostomy, code status changes, and withdrawal or withholding of life-sustaining therapies. Throughout decision-making, clinicians must be aware of a myriad of factors affecting both provider recommendations and surrogate concerns, such as cognitive biases. While decision aids have the potential to better frame these conversations within intensive care units, aids specific to goals-of-care decisions for stroke patients are currently lacking. This mini review highlights the difficulties in decision-making for critically ill ischemic stroke and intracerebral hemorrhage patients, beginning with limitations in current validated clinical scales and clinician subjectivity in prognostication. We outline processes for identifying patient preferences when possible and make recommendations for collaborating closely with surrogate decision-makers on end-of-life care decisions.

Cognitive biases, environmental, patient and personal factors associated with critical care decision making: A scoping review

PURPOSE

Cognitive biases and factors affecting decision making in critical care can potentially lead to life-threatening errors. We aimed to examine the existing evidence on the influence of cognitive biases and factors on decision making in critical care.

MATERIALS AND METHODS

We conducted a scoping review by searching MEDLINE for articles from 2004 to November 2020. We included studies conducted in physicians that described cognitive biases or factors associated with decision making. During the study process we decided on the method to summarize the evidence, and based on the obtained studies a descriptive summary of findings was the best fit.

RESULTS

Thirty heterogenous studies were included. Four main biases or factors were observed, e.g. cognitive biases, personal factors, environmental factors, and patient factors. Six (20%) studies reported biases associated with decision making comprising omission-, status quo-, implicit-, explicit-, outcome-, and overconfidence bias. Nineteen (63%) studies described personal factors, twenty-two (73%) studies described environmental factors, and sixteen (53%) studies described patient factors.

CONCLUSIONS

The current evidence on cognitive biases and factors is heterogenous, but shows they influence clinical decision. Future studies should investigate the prevalence of cognitive biases and factors in clinical practice and their impact on clinical outcomes.

Actigraphic Measurement of the Upper Limbs for the Prediction of Ischemic Stroke Prognosis: An Observational Study

Background: It is often challenging to formulate a reliable prognosis for patients with acute ischemic stroke. The most accepted prognostic factors may not be sufficient to predict the recovery process. In this view, describing the evolution of motor deficits over time via sensors might be useful for strengthening the prognostic model. Our aim was to assess whether an actigraphic-based parameter (Asymmetry Rate Index for the 24 h period (AR2_24 h)) obtained in the acute stroke phase could be a predictor of a 90 d prognosis. Methods: In this observational study, we recorded and analyzed the 24 h upper limb movement asymmetry of 20 consecutive patients with acute ischemic stroke during their stay in a stroke unit. We recorded the motor activity of both arms using two programmable actigraphic systems positioned on patients’ wrists. We clinically evaluated the stroke patients by NIHSS in the acute phase and then assessed them across 90 days using the modified Rankin Scale (mRS). Results: We found that the AR2_24 h parameter positively correlates with the 90 d mRS (r = 0.69, p < 0.001). Moreover, we found that an AR2_24 h > 32% predicts a poorer outcome (90 d mRS > 2), with sensitivity = 100% and specificity = 89%. Conclusions: Sensor-based parameters might provide useful information for predicting ischemic stroke prognosis in the acute phase.

Setting performance benchmarks for stroke care delivery: Which quality indicators should be prioritized in quality improvement; an analysis in 500,331 stroke admissions

BACKGROUND AND AIM

Benchmarking is a management approach for implementing best medical practices at the lowest cost. The objectives of this study were to set achievable performance benchmarks for individual quality indicators to determine the predicted quality achievement related to better adherence, and to select optimal quality indicators for improving the quality of acute ischemic stroke care.

METHODS

We analyzed data on 500,331 patients diagnosed with acute ischemic stroke who were treated at 518 hospitals in China from January 2011 to May 2017. The primary outcome was independence (modified Rankin Scale score ≤2) at discharge. Data-driven achievable benchmarking used the "pared-mean" approach to set objective performance targets. Hierarchical logistic regression models were employed to evaluate the process-outcome association, as well as the predicted quality improvement if all hospitals were to operate at the benchmark level.

RESULTS

Of the overall population, 64.01% were independent patients at discharge. The performance benchmarks were >90% for most of the quality indicators. After adjusting for patient-level and hospital-level characteristics and unifying hospital performance to the benchmark level, the quality indicators with high increase in both overall adherence rate and independence rate were thrombolytic therapy, anticoagulant therapy, venous thrombosis prophylaxis.

CONCLUSIONS

Performance targets for three acute treatments, including thrombolytic therapy, anticoagulant therapy, venous thrombosis prophylaxis, could best motivate improvements in both overall adherence rate and independence rate at discharge. The finding suggests that the above three types of acute treatment should be given priority to improve the quality of acute ischemic stroke care.

Prognostic nomogram for patients with minor stroke and transient ischaemic attack

BACKGROUND

Ischaemic stroke and transient ischaemic attack (TIA) share a common cause. We aim to develop and validate a concise prognostic nomogram for patients with minor stroke and TIA.

METHODS

A total of 994 patients with minor stroke and TIA were included. They were split into a derivation (n=746) and validation (n=248) cohort. The modified Rankin Scale (mRS) scores 3 months after onset were used to assess the prognosis as unfavourable outcome (mRS≥2) or favourable outcome (mRS<2).

RESULT

The final model included seven independent predictors: gender, age, baseline National Institute of Health Stroke Scale (NIHSS), hypertension, diabetes mellitus, white blood cell and serum uric acid. The Harrell's concordance index (C-index) of the nomogram for predicting the outcome was 0.775 (95% CI 0.735 to 0.814), which was confirmed by the validation cohort (C-index=0.787 (95% CI 0.722 to 0.853)). The calibration curve showed that the nomogram-based predictions were consistent with actual observation in both derivation cohort and validation cohort.

CONCLUSION

The proposed nomogram showed favourable predictive accuracy for minor stroke and TIA. This has the potential to contribute to clinical decision-making.

A grading scale for surgically treated patients with spontaneous supratentorial intracerebral hemorrhage: the Surgical Swedish ICH Score

OBJECTIVE

The authors aimed to develop the first clinical grading scale for patients with surgically treated spontaneous supratentorial intracerebral hemorrhage (ICH).

METHODS

A nationwide multicenter study including 401 ICH patients surgically treated by craniotomy and evacuation of a spontaneous supratentorial ICH was conducted between January 1, 2011, and December 31, 2015. All neurosurgical centers in Sweden were included. All medical records and neuroimaging studies were retrospectively reviewed. Independent predictors of 30-day mortality were identified by logistic regression. A risk stratification scale (the Surgical Swedish ICH [SwICH] Score) was developed using weighting of independent predictors based on strength of association.

RESULTS

Factors independently associated with 30-day mortality were Glasgow Coma Scale (GCS) score (p = 0.00015), ICH volume ≥ 50 mL (p = 0.031), patient age ≥ 75 years (p = 0.0056), prior myocardial infarction (MI) (p = 0.00081), and type 2 diabetes (p = 0.0093). The Surgical SwICH Score was the sum of individual points assigned as follows: GCS score 15-13 (0 points), 12-5 (1 point), 4-3 (2 points); age ≥ 75 years (1 point); ICH volume ≥ 50 mL (1 point); type 2 diabetes (1 point); prior MI (1 point). Each increase in the Surgical SwICH Score was associated with a progressively increased 30-day mortality (p = 0.0002). No patient with a Surgical SwICH Score of 0 died, whereas the 30-day mortality rates for patients with Surgical SwICH Scores of 1, 2, 3, and 4 were 5%, 12%, 31%, and 58%, respectively.

CONCLUSIONS

The Surgical SwICH Score is a predictor of 30-day mortality in patients treated surgically for spontaneous supratentorial ICH. External validation is needed to assess the predictive value as well as the generalizability of the Surgical SwICH Score.

Identifying Performance Outliers for Stroke Care Based on Composite Score of Process Indicators: an Observational Study in China

BACKGROUND

Variability in the quality of stroke care is widespread. Identifying performance-based outlier hospitals based on quality indicators (QIs) has become a common practice.

OBJECTIVES

To develop a tool for identifying performance-based outlier hospitals based on risk-adjusted adherence rates of process indicators.

DESIGN

Hospitals were classified into five-level outliers based on the observed-to-expected ratio and P value. The composite quality score was derived by summation of the points for each indicator for each hospital, and associations between outlier status and outcomes were determined.

PARTICIPANTS

Patients diagnosed with acute ischemic stroke, January 1, 2011-May 31, 2017.

INTERVENTION

N/A MAIN OUTCOME MEASURES: Independence at discharge (the modified Rankin Scale = 0-2).

KEY RESULTS

A total of 501,132 patients from 519 hospitals were identified. From 0.39 to 19.65% of hospitals were identified as high outliers according to various QIs. Composite quality scores ranged from - 20 to 16. Providers that were high outliers based on QI2, QI8, QI9, and QI11 had higher independent rates. For composite quality score, each point increase corresponded to an 8% increase in the odds of independent rate.

CONCLUSION

Nationwide variation in the quality of acute stroke care exists at the hospital level. Variability in the quality of stroke care can be captured by our proposed quality score. Applying this quality score as a benchmarking tool could provide audit-level feedback to policymakers and hospitals to aid quality improvement.

Fundamentals of Clinical Prediction Modeling for the Neurosurgeon

Clinical prediction models in neurosurgery are increasingly reported. These models aim to provide an evidence-based approach to the estimation of the probability of a neurosurgical outcome by combining 2 or more prognostic variables. Model development and model reporting are often suboptimal. A basic understanding of the methodology of clinical prediction modeling is needed when interpreting these models. We address basic statistical background, 7 modeling steps, and requirements of these models such that they may fulfill their potential for major impact for our daily clinical practice and for future scientific work.

The impact of disease severity adjustment on hospital standardised mortality ratios: Results from a service-wide analysis of ischaemic stroke admissions using linked pre-hospital, admissions and mortality data

Background

Administrative data are used to examine variation in thirty-day mortality across health services in several jurisdictions. Hospital performance measurement may be error-prone as information about disease severity is not typically available in routinely collected data to incorporate into case-mix adjusted analyses. Using ischaemic stroke as a case study, we tested the extent to which accounting for disease severity impacts on hospital performance assessment.

Methods

We linked all recorded ischaemic stroke admissions between July, 2011 and June, 2014 to death registrations and a measure of stroke severity obtained at first point of patient contact with health services, across New South Wales, Australia’s largest health service jurisdiction. Thirty-day hospital standardised mortality ratios were adjusted for either comorbidities, as is typically done, or for both comorbidities and stroke severity. The impact of stroke severity adjustment on mortality ratios was determined using 95% and 99% control limits applied to funnel plots and by calculating the change in rank order of hospital risk adjusted mortality rates.

Results

The performance of the stroke severity adjusted model was superior to incorporating comorbidity burden alone (c-statistic = 0.82 versus 0.75; N = 17,700 patients, 176 hospitals). Concordance in outlier classification was 89% and 97% when applying 95% or 99% control limits to funnel plots, respectively. The sensitivity rates of outlier detection using comorbidity adjustment compared with gold-standard severity and comorbidity adjustment was 74% and 83% with 95% and 99% control limits, respectively. Corresponding positive predictive values were 74% and 91%. Hospital rank order of risk adjusted mortality rates shifted between 0 to 22 places with severity adjustment (Median = 4.0, Inter-quartile Range = 2–7).

Conclusions

Rankings of mortality rates varied widely depending on whether stroke severity was taken into account. Funnel plots yielded largely concordant results irrespective of severity adjustment and may be sufficiently accurate as a screening tool for assessing hospital performance.

External Validation of the PREMISE Score in the Athens Stroke Registry

BACKGROUND

A simple score was proposed recently for Predicting Early Mortality from Ischemic Stroke (PREMISE) derived from the Austrian Stroke Unit Registry. This score could be useful in clinical practice and research. However, its generalizability is uncertain, as it was validated internally only.

AIMS

We aimed to validate the PREMISE score externally.

METHODS

The analysis was performed in the Athens Stroke Registry. The PREMISE score was calculated as described in the original publication. The outcome was death within 7 days after stroke. Logistic regression analysis was used to estimate the relative death risk in different strata of the PREMISE score using the lowest values of the score (ie, 0-4) as the reference category. We assessed the score's calibration by the Hosmer-Lemeshow goodness-of-fit test and its discriminatory power by calculating the area under the receiver operating characteristics curve (AUC).

RESULTS

In 2608 consecutive patients (median age 71 years, 38.8% women) with acute ischemic stroke treated in the stroke unit, mortality increased with increasing PREMISE score from .1% (95% confidence intervals [95% CI]: 0%-.2%) in patients with a score of 0-4 to 28.2% (95% CI: 14.1%-42.3%) in patients with a score of ≥10. The risk for death was more than 6 times higher in patients with a PREMISE score of ≥10 compared to patients with 0-4 points (odds ratio [OR]:6.21, 95% CI:4.13-8.29). Τhe PREMISE score showed excellent calibration (Hosmer-Lemeshow χ²: .01, P= .99) and good discriminatory power (AUC .873, 95% CI: .844-.901).

CONCLUSIONS

The present study confirms the prognostic accuracy of the PREMISE score in an independent cohort of patients with acute ischemic stroke treated in the stroke unit.

A novel biomarker-based prognostic score in acute ischemic stroke: The CoRisk score

OBJECTIVES

To derive and externally validate a copeptin-based parsimonious score to predict unfavorable outcome 3 months after an acute ischemic stroke (AIS).

METHODS

The derivation cohort consisted of patients with AIS enrolled prospectively at the University Hospital Basel, Switzerland. The validation cohort was prospectively enrolled after the derivation cohort at the University Hospital of Bern and University Hospital Basel, Switzerland, as well as Frankfurt a.M., Germany. The score components were copeptin levels, age, NIH Stroke Scale, and recanalization therapy (CoRisk score). Copeptin levels were measured in plasma drawn within 24 hours of AIS and before any recanalization therapy. The primary outcome of disability and death at 3 months was defined as modified Rankin Scale score of 3 to 6.

RESULTS

Overall, 1,102 patients were included in the analysis; the derivation cohort contributed 319 patients, and the validation cohort contributed 783. An unfavorable outcome was observed among 436 patients (40%). For the 3-month prediction of disability and death, the CoRisk score was well calibrated in the validation cohort, for which the area under the receiver operating characteristic curve was 0.819 (95% confidence interval [CI] 0.787-0.849). The calibrated CoRisk score correctly classified 75% of patients (95% CI 72-78). The net reclassification index between the calibrated CoRisk scores with and without copeptin was 46% (95% CI 32-60).

CONCLUSIONS

The biomarker-based CoRisk score for the prediction of disability and death was externally validated, was well calibrated, and performed better than the same score without copeptin.

CLINICALTRIALSGOV IDENTIFIER

NCT00390962 (derivation cohort) and NCT00878813 (validation cohort).

A Review of Pre-Intervention Prognostic Scores for Early Prognostication and Patient Selection in Endovascular Management of Large Vessel Occlusion Stroke

BACKGROUND

Endovascular therapy (ET) has emerged as a highly effective treatment for acute large vessel occlusion stroke (LVOS). Tools that facilitate optimal patient selection of patients for ET are needed in order to maximize therapeutic benefit in a cost-effective manner. Several pre-intervention prognostic scores for prediction of outcomes in LVOS patients and patient selection for ET have been developed and validated, but their clinical use has been limited. Here, we review existing pre-intervention prognostic scores, compare their prognostic accuracies and levels of validation and identify gaps in current knowledge.

SUMMARY

We have reviewed published literature pertinent to development, validation, and implementation of pre-intervention prognostic scores for LVOS. Using receiver operating characteristic curve analysis, the prognostic accuracies of validated pre-interventional scores (Pittsburgh Response to Endovascular therapy [PRE], Totaled Health Risks in Vascular Events [THRIVE], Houston Intra-Arterial Therapy-2 (HIAT-2), Stroke Prognostication using Age and NIHSS [SPAN-100]) were compared in published work. Pre-intervention scores predicted functional out comes at 3 months with moderate prognostic accuracies (area under the receiver operator characteristic curve range 0.68-0.73). Using successful reperfusion (mTICI 2B/3) as the therapeutic objective of ET and 3-month modified Rankin Score 0-2 as good clinical outcome, patients most likely to clinically benefit from endovascular reperfusion can be identified using the PRE and HIAT-2 scores. Scores that incorporate collateral imaging or perfusion-based estimation of core and penumbra have not been published. Existing scores are predominantly limited to anterior circulation LVOS, and implementation studies of pre-interventional scores are lacking.

KEY MESSAGES

Pre-intervention prognostic scores can serve as useful adjuncts for patient selection in ET for acute LVOS. Pre-intervention scores including HIAT-2, THRIVE, SPAN-100, and PRE have comparable moderate prognostic accuracies for good 3-month outcomes and can identify patients who derive maximal benefit from successful reperfusion. Improvements in prognostic accuracy may be achieved by incorporating variables such as collateral status and perfusion imaging data. Implementation and impact studies using pre-intervention scores are needed to guide clinical application.

Predictive accuracy of physicians' estimates of outcome after severe stroke

Introduction

End-of-life decisions after stroke should be guided by accurate estimates of the patient’s prognosis. We assessed the accuracy of physicians’ estimates regarding mortality, functional outcome, and quality of life in patients with severe stroke.

Methods

Treating physicians predicted mortality, functional outcome (modified Rankin scale (mRS)), and quality of life (visual analogue scale (VAS)) at six months in patients with major disabling stroke who had a Barthel Index ≤6 (of 20) at day four. Unfavorable functional outcome was defined as mRS >3, non-satisfactory quality of life as VAS <60. Patients were followed-up at six months after stroke. We compared physicians’ estimates with actual outcomes.

Results

Sixty patients were included, with a mean age of 72 years. Of fifteen patients who were predicted to die, one actually survived at six months (positive predictive value (PPV), 0.93; 95% CI, 0.66–0.99). Of thirty patients who survived, one was predicted to die (false positive rate (FPR), 0.03; 95%CI 0.00–0.20). Of forty-six patients who were predicted to have an unfavorable outcome, four had a favorable outcome (PPV, 0.93; 95% CI, 0.81–0.98; FPR, 0.30; 95% CI; 0.08–0.65). Prediction of non-satisfactory quality of life was less accurate (PPV, 0.63; 95% CI, 0.26–0.90; FPR, 0.18; 95% CI 0.05–0.44).

Conclusions

In patients with severe stroke, treating physicians’ estimation of the risk of mortality or unfavorable functional outcome at six months is relatively inaccurate. Prediction of quality of life is even more imprecise.

Communication, Leadership, and Decision-Making in the Neuro-ICU

The Neuro-ICU is a multidisciplinary location that presents peculiar challenges and opportunities for patients with life-threatening neurological disease. Communication skills are essential in supporting caregivers and other embedded providers (e.g., neurosurgeons, advanced practice providers, nurses, pharmacists), through leadership. Limitations to prognostication complicate how decisions are made on behalf of non-communicative patients. Cognitive dysfunction and durable reductions in health-related quality of life are difficult to predict, and the diagnosis of brain death may be challenging and confounded by medications and comorbidities. The Neuro-ICU team, as well as utilization of additional consultants, can be structured to optimize care. Future research should explore how to further improve the composition, communication and interactions of the Neuro-ICU team to maximize outcomes, minimize caregiver burden, and promote collegiality.

Early transition to comfort measures only in acute stroke patients: Analysis from the Get With The Guidelines-Stroke registry

BACKGROUND

Death after acute stroke often occurs after forgoing life-sustaining interventions. We sought to determine the patient and hospital characteristics associated with an early decision to transition to comfort measures only (CMO) after ischemic stroke (IS), intracerebral hemorrhage (ICH), and subarachnoid hemorrhage (SAH) in the Get With The Guidelines-Stroke registry.

METHODS

We identified patients with IS, ICH, or SAH between November 2009 and September 2013 who met study criteria. Early CMO was defined as the withdrawal of life-sustaining treatments and interventions by hospital day 0 or 1. Using multivariable logistic regression, we identified patient and hospital factors associated with an early (by hospital day 0 or 1) CMO order.

RESULTS

Among 963,525 patients from 1,675 hospitals, 54,794 (5.6%) had an early CMO order (IS: 3.0%; ICH: 19.4%; SAH: 13.1%). Early CMO use varied widely by hospital (range 0.6%-37.6% overall) and declined over time (from 6.1% in 2009 to 5.4% in 2013; p < 0.001). In multivariable analysis, older age, female sex, white race, Medicaid and self-pay/no insurance, arrival by ambulance, arrival off-hours, baseline nonambulatory status, and stroke type were independently associated with early CMO use (vs no early CMO). The correlation between hospital-level risk-adjusted mortality and the use of early CMO was stronger for SAH (r = 0.52) and ICH (r = 0.50) than AIS (r = 0.15) patients.

CONCLUSIONS

Early CMO was utilized in about 5% of stroke patients, being more common in ICH and SAH than IS. Early CMO use varies widely between hospitals and is influenced by patient and hospital characteristics.

Simple prediction scores predict good and devastating outcomes after stroke more accurately than physicians

Introduction

physicians are often asked to prognosticate soon after a patient presents with stroke. This study aimed to compare two outcome prediction scores (Five Simple Variables [FSV] score and the PLAN [Preadmission comorbidities, Level of consciousness, Age, and focal Neurologic deficit]) with informal prediction by physicians.

Methods

demographic and clinical variables were prospectively collected from consecutive patients hospitalised with acute ischaemic or haemorrhagic stroke (2012-13). In-person or telephone follow-up at 6 months established vital and functional status (modified Rankin score [mRS]). Area under the receiver operating curves (AUC) was used to establish prediction score performance.

Results

five hundred and seventy-five patients were included; 46% female, median age 76 years, 88% ischaemic stroke. Six months after stroke, 47% of patients had a good outcome (alive and independent, mRS 0-2) and 26% a devastating outcome (dead or severely dependent, mRS 5-6). The FSV and PLAN scores were superior to physician prediction (AUCs of 0.823-0.863 versus 0.773-0.805, P < 0.0001) for good and devastating outcomes. The FSV score was superior to the PLAN score for predicting good outcomes and vice versa for devastating outcomes (P < 0.001). Outcome prediction was more accurate for those with later presentations (>24 hours from onset).

Conclusion

the FSV and PLAN scores are validated in this population for outcome prediction after both ischaemic and haemorrhagic stroke. The FSV score is the least complex of all developed scores and can assist outcome prediction by physicians.

Family discussions on life-sustaining interventions in neurocritical care

Approximately 20% of all deaths in the USA occur in the intensive care unit (ICU) and the majority of ICU deaths involves decision of de-escalation of life-sustaining interventions. Life-sustaining interventions may include intubation and mechanical ventilation, artificial nutrition and hydration, antibiotic treatment, brain surgery, or vasoactive support. Decision making about goals of care can be defined as an end-of-life communication and the decision-making process between a clinician and a patient (or a surrogate decision maker if the patient is incapable) in an institutional setting to establish a plan of care. This process includes deciding whether to use life-sustaining treatments. Therefore, family discussion is a critical element in the decision-making process throughout the patient's stay in the neurocritical care unit. A large part of care in the neurosciences intensive care unit is discussion of proportionality of care. This chapter provides a stepwise approach to hold these conferences and discusses ways to do it effectively.

Cognitive biases associated with medical decisions: a systematic review

Background

Cognitive biases and personality traits (aversion to risk or ambiguity) may lead to diagnostic inaccuracies and medical errors resulting in mismanagement or inadequate utilization of resources. We conducted a systematic review with four objectives: 1) to identify the most common cognitive biases, 2) to evaluate the influence of cognitive biases on diagnostic accuracy or management errors, 3) to determine their impact on patient outcomes, and 4) to identify literature gaps.

Methods

We searched MEDLINE and the Cochrane Library databases for relevant articles on cognitive biases from 1980 to May 2015. We included studies conducted in physicians that evaluated at least one cognitive factor using case-vignettes or real scenarios and reported an associated outcome written in English. Data quality was assessed by the Newcastle-Ottawa scale. Among 114 publications, 20 studies comprising 6810 physicians met the inclusion criteria. Nineteen cognitive biases were identified.

Results

All studies found at least one cognitive bias or personality trait to affect physicians. Overconfidence, lower tolerance to risk, the anchoring effect, and information and availability biases were associated with diagnostic inaccuracies in 36.5 to 77 % of case-scenarios. Five out of seven (71.4 %) studies showed an association between cognitive biases and therapeutic or management errors. Of two (10 %) studies evaluating the impact of cognitive biases or personality traits on patient outcomes, only one showed that higher tolerance to ambiguity was associated with increased medical complications (9.7 % vs 6.5 %; p = .004). Most studies (60 %) targeted cognitive biases in diagnostic tasks, fewer focused on treatment or management (35 %) and on prognosis (10 %). Literature gaps include potentially relevant biases (e.g. aggregate bias, feedback sanction, hindsight bias) not investigated in the included studies. Moreover, only five (25 %) studies used clinical guidelines as the framework to determine diagnostic or treatment errors. Most studies (n = 12, 60 %) were classified as low quality.

Conclusions

Overconfidence, the anchoring effect, information and availability bias, and tolerance to risk may be associated with diagnostic inaccuracies or suboptimal management. More comprehensive studies are needed to determine the prevalence of cognitive biases and personality traits and their potential impact on physicians’ decisions, medical errors, and patient outcomes.

Electronic supplementary material

The online version of this article (doi:10.1186/s12911-016-0377-1) contains supplementary material, which is available to authorized users.

ASTRAL, DRAGON and SEDAN scores predict stroke outcome more accurately than physicians

BACKGROUND AND PURPOSE

ASTRAL, SEDAN and DRAGON scores are three well-validated scores for stroke outcome prediction. Whether these scores predict stroke outcome more accurately compared with physicians interested in stroke was investigated.

METHODS

Physicians interested in stroke were invited to an online anonymous survey to provide outcome estimates in randomly allocated structured scenarios of recent real-life stroke patients. Their estimates were compared to scores' predictions in the same scenarios. An estimate was considered accurate if it was within 95% confidence intervals of actual outcome.

RESULTS

In all, 244 participants from 32 different countries responded assessing 720 real scenarios and 2636 outcomes. The majority of physicians' estimates were inaccurate (1422/2636, 53.9%). 400 (56.8%) of physicians' estimates about the percentage probability of 3-month modified Rankin score (mRS) > 2 were accurate compared with 609 (86.5%) of ASTRAL score estimates (P < 0.0001). 394 (61.2%) of physicians' estimates about the percentage probability of post-thrombolysis symptomatic intracranial haemorrhage were accurate compared with 583 (90.5%) of SEDAN score estimates (P < 0.0001). 160 (24.8%) of physicians' estimates about post-thrombolysis 3-month percentage probability of mRS 0-2 were accurate compared with 240 (37.3%) DRAGON score estimates (P < 0.0001). 260 (40.4%) of physicians' estimates about the percentage probability of post-thrombolysis mRS 5-6 were accurate compared with 518 (80.4%) DRAGON score estimates (P < 0.0001).

CONCLUSIONS

ASTRAL, DRAGON and SEDAN scores predict outcome of acute ischaemic stroke patients with higher accuracy compared to physicians interested in stroke.

Ischemic stroke outcome: A review of the influence of post-stroke complications within the different scenarios of stroke care

Stroke remains one of the main causes of death and disability worldwide. The challenge of predicting stroke outcome has been traditionally assessed from a general point of view, where baseline non-modifiable factors such as age or stroke severity are considered the most relevant factors. However, after stroke occurrence, some specific complications such as hemorrhagic transformations or post stroke infections, which lead to a poor outcome, could be developed. An early prediction or identification of these circumstances, based on predictive models including clinical information, could be useful for physicians to individualize and improve stroke care. Furthermore, the addition of biological information such as blood biomarkers or genetic polymorphisms over these predictive models could improve their prognostic value. In this review, we focus on describing the different post-stroke complications that have an impact in short and long-term outcome across different time points in its natural history and on the clinical-biological information that might be useful in their prediction.