Clinician judgment vs formal scales for predicting intracerebral hemorrhage outcomes

Timing of neuroprognostication in the ICU

PURPOSE OF REVIEW

Neuroprognostication after acute brain injury (ABI) is complex. In this review, we examine the threats to accurate neuroprognostication, discuss strategies to mitigate the self-fulfilling prophecy, and how to approach the indeterminate prognosis.

RECENT FINDINGS

The goal of neuroprognostication is to provide a timely and accurate prediction of a patient's neurologic outcome so treatment can proceed in accordance with a patient's values and preferences. Neuroprognostication should be delayed until at least 72 h after injury and/or only when the necessary prognostic data is available to avoid early withdraw life-sustaining treatment on patients who may otherwise survive with a good outcome. Clinicians should be aware of the limitations of available predictors and prognostic models, the role of flawed heuristics and the self-fulfilling prophecy, and the influence of surrogate decision-maker bias on end-of-life decisions.

SUMMARY

The approach to neuroprognostication after ABI should be systematic, use highly reliable multimodal data, and involve experts to minimize the risk of erroneous prediction and perpetuating the self-fulfilling prophecy. Even when such standards are rigorously upheld, the prognosis may be indeterminate. In such cases, clinicians should engage in shared decision-making with surrogates and consider the use of a time-limited trial.

The use of cloud based machine learning to predict outcome in intracerebral haemorrhage without explicit programming expertise

Machine Learning (ML) techniques require novel computer programming skills along with clinical domain knowledge to produce a useful model. We demonstrate the use of a cloud-based ML tool that does not require any programming expertise to develop, validate and deploy a prognostic model for Intracerebral Haemorrhage (ICH). The data of patients admitted with Spontaneous Intracerebral haemorrhage from January 2015 to December 2019 was accessed from our prospectively maintained hospital stroke registry. 80% of the dataset was used for training, 10% for validation, and 10% for testing. Seventeen input variables were used to predict the dichotomized outcomes (Good outcome mRS 0-3/ Bad outcome mRS 4-6), using machine learning (ML) and logistic regression (LR) models. The two different approaches were evaluated using Area Under the Curve (AUC) for Receiver Operating Characteristic (ROC), Precision recall and accuracy. Our data set comprised of a cohort of 1000 patients. The data was split 8:1 for training & testing respectively. The AUC ROC of the ML model was 0.86 with an accuracy of 75.7%. With LR AUC ROC was 0.74 with an accuracy of 73.8%. Feature importance chart showed that Glasgow coma score (GCS) at presentation had the highest relative importance, followed by hematoma volume and age in both approaches. Machine learning models perform better when compared to logistic regression. Models can be developed by clinicians possessing domain expertise and no programming experience using cloud based tools. The models so developed lend themselves to be incorporated into clinical workflow.

A Survey of Surrogates and Health Care Professionals Indicates Support of Cognitive Motor Dissociation-Assisted Prognostication

BACKGROUND

Prognostication of patients with acute disorders of consciousness is imprecise but more accurate technology-supported predictions, such as cognitive motor dissociation (CMD), are emerging. CMD refers to the detection of willful brain activation following motor commands using functional magnetic resonance imaging or machine learning-supported analysis of the electroencephalogram in clinically unresponsive patients. CMD is associated with long-term recovery, but acceptance by surrogates and health care professionals is uncertain. The objective of this study was to determine receptiveness for CMD to inform goals of care (GoC) decisions and research participation among health care professionals and surrogates of behaviorally unresponsive patients.

METHODS

This was a two-center study of surrogates of and health care professionals caring for unconscious patients with severe neurological injury who were enrolled in two prospective US-based studies. Participants completed a 13-item survey to assess demographics, religiosity, minimal acceptable level of recovery, enthusiasm for research participation, and receptiveness for CMD to support GoC decisions.

RESULTS

Completed surveys were obtained from 196 participants (133 health care professionals and 63 surrogates). Across all respondents, 93% indicated that they would want their loved one or the patient they cared for to participate in a research study that supports recovery of consciousness if CMD were detected, compared to 58% if CMD were not detected. Health care professionals were more likely than surrogates to change GoC with a positive (78% vs. 59%, p = 0.005) or negative (83% vs. 59%, p = 0.0002) CMD result. Participants who reported religion was the most important part of their life were least likely to change GoC with or without CMD. Participants who identified as Black (odds ratio [OR] 0.12, 95% confidence interval [CI] 0.04-0.36) or Hispanic/Latino (OR 0.39, 95% CI 0.2-0.75) and those for whom religion was the most important part of their life (OR 0.18, 95% CI 0.05-0.64) were more likely to accept a lower minimum level of recovery.

CONCLUSIONS

Technology-supported prognostication and enthusiasm for clinical trial participation was supported across a diverse spectrum of health care professionals and surrogate decision-makers. Education for surrogates and health care professionals should accompany integration of technology-supported prognostication.

Recovery of consciousness after acute brain injury: a narrative review

BACKGROUND

Disorders of consciousness (DoC) are frequently encountered in both, acute and chronic brain injuries. In many countries, early withdrawal of life-sustaining treatments is common practice for these patients even though the accuracy of predicting recovery is debated and delayed recovery can be seen. In this review, we will discuss theoretical concepts of consciousness and pathophysiology, explore effective strategies for management, and discuss the accurate prediction of long-term clinical outcomes. We will also address research challenges.

MAIN TEXT

DoC are characterized by alterations in arousal and/or content, being classified as coma, unresponsive wakefulness syndrome/vegetative state, minimally conscious state, and confusional state. Patients with willful modulation of brain activity detectable by functional MRI or EEG but not by behavioral examination is a state also known as covert consciousness or cognitive motor dissociation. This state may be as common as every 4th or 5th patient without behavioral evidence of verbal command following and has been identified as an independent predictor of long-term functional recovery. Underlying mechanisms are uncertain but intact arousal and thalamocortical projections maybe be essential. Insights into the mechanisms underlying DoC will be of major importance as these will provide a framework to conceptualize treatment approaches, including medical, mechanical, or electoral brain stimulation.

CONCLUSIONS

We are beginning to gain insights into the underlying mechanisms of DoC, identifying novel advanced prognostication tools to improve the accuracy of recovery predictions, and are starting to conceptualize targeted treatments to support the recovery of DoC patients. It is essential to determine how these advancements can be implemented and benefit DoC patients across a range of clinical settings and global societal systems. The Curing Coma Campaign has highlighted major gaps knowledge and provides a roadmap to advance the field of coma science with the goal to support the recovery of patients with DoC.

Benefits of Neurosurgical Teleconsults in the Management of Intracerebral Hemorrhage: Transfers and Transportation Cost Reduction

BACKGROUND

Our study explores the efficacy and economic benefits of neurosurgical teleconsultations in managing intracerebral hemorrhage (ICH), focusing on reducing unnecessary patient transfers and associated costs.

METHODS

We conducted a cost-savings analysis at our institution of a previously published pilot study involving a cohort of patients with ICH who were potential candidates for airlift to our tertiary care center but instead received neurosurgical consultation via teleconsultation to avoid the transfer. Data on patient demographics, distances, and costs were collected and analyzed to assess the economic impact of teleconsultations.

RESULTS

The cohort comprised 14 patients; we noted significant cost savings from avoiding interhospital transfers, ranging from $84,346.52 to $120,495.03 per patient. Teleconsultations facilitated immediate, collaborative decision-making between healthcare providers at community hospitals and a tertiary care center, reducing the need for expensive air transportation and unnecessary hospital transfers.

CONCLUSIONS

Neurosurgical teleconsultations offer a cost-effective alternative to traditional patient transfer methods for ICH management, providing substantial economic benefits while maintaining high physician and patient-family satisfaction levels. This study underscores the potential of our teleneurosurgery program to significantly reduce costs by reducing unnecessary financial burdens on patients' families and healthcare systems.

Impact of Aphasia on Brain Activation to Motor Commands in Patients with Acute Intracerebral Hemorrhage

BACKGROUND

Brain activation to motor commands is seen in 15% of clinically unresponsive patients with acute brain injury. This state called cognitive motor dissociation (CMD) is detectable by electroencephalogram (EEG) or functional magnetic resonance imaging, predicts long-term recovery, and is recommended by recent guidelines to support prognostication. However, false negative CMD results are a particular concern, and occult aphasia in clinically unresponsive patients may be a major factor. This study aimed to quantify the impact of aphasia on CMD testing.

METHODS

We prospectively studied 61 intensive care unit patients admitted with acute primary intracerebral hemorrhage (ICH) who had behavioral evidence of command following or were able to mimic motor commands. All patients underwent an EEG-based motor command paradigm used to detect CMD and comprehensive aphasia assessments. Logistic regression was used to identify predictors of brain activation, including aphasia types and associations with recovery of independence (Glasgow Outcome Scale-Extended score ≥ 4).

RESULTS

Of 61 patients, 50 completed aphasia and the EEG-based motor command paradigm. A total of 72% (n = 36) were diagnosed with aphasia. Patients with impaired comprehension (i.e., receptive or global aphasia) were less likely to show brain activation than those with intact comprehension (odds ratio [OR] 0.23 [95% confidence interval 0.05-0.89], p = 0.04). Brain activation was independently associated with Glasgow Outcome Scale-Extended ≥ 4 by 12 months (OR 2.4 [95% confidence interval 1.2-5.0], p = 0.01) accounting for the Functional Outcome in Patients with Primary ICH score (OR1.3 [95% confidence interval 1.0-1.8], p = 0.01).

CONCLUSIONS

Brain activation to motor commands is four times less likely for patients with primary ICH with impaired comprehension. False negative results due to occult receptive aphasia need to be considered when interpreting CMD testing. Early detection of brain activation may help predict long-term recovery in conscious patients with ICH.

CT-based deep learning model for predicting hospital discharge outcome in spontaneous intracerebral hemorrhage

OBJECTIVES

To predict the functional outcome of patients with intracerebral hemorrhage (ICH) using deep learning models based on computed tomography (CT) images.

METHODS

A retrospective, bi-center study of ICH patients was conducted. Firstly, a custom 3D convolutional model was built for predicting the functional outcome of ICH patients based on CT scans from randomly selected ICH patients in H training dataset collected from H hospital. Secondly, clinical data and radiological features were collected at admission and the Extreme Gradient Boosting (XGBoost) algorithm was used to establish a second model, named the XGBoost model. Finally, the Convolution model and XGBoost model were fused to build the third "Fusion model." Favorable outcome was defined as modified Rankin Scale score of 0-3 at discharge. The prognostic predictive accuracy of the three models was evaluated using an H test dataset and an external Y dataset, and compared with the performance of ICH score and ICH grading scale (ICH-GS).

RESULTS

A total of 604 patients with ICH were included in this study, of which 450 patients were in the H training dataset, 50 patients in the H test dataset, and 104 patients in the Y dataset. In the Y dataset, the areas under the curve (AUCs) of the Convolution model, XGBoost model, and Fusion model were 0.829, 0.871, and 0.905, respectively. The Fusion model prognostic performance exceeded that of ICH score and ICH-GS (p = 0.043 and p = 0.045, respectively).

CONCLUSIONS

Deep learning models have good accuracy for predicting functional outcome of patients with spontaneous intracerebral hemorrhage.

CLINICAL RELEVANCE STATEMENT

The proposed deep learning Fusion model may assist clinicians in predicting functional outcome and developing treatment strategies, thereby improving the survival and quality of life of patients with spontaneous intracerebral hemorrhage.

KEY POINTS

• Integrating clinical presentations, CT images, and radiological features to establish deep learning model for functional outcome prediction of patients with intracerebral hemorrhage. • Deep learning applied to CT images provides great help in prognosing functional outcome of intracerebral hemorrhage patients. • The developed deep learning model performs better than clinical prognostic scores in predicting functional outcome of patients with intracerebral hemorrhage.

Prognostication in Neurocritical Care

OBJECTIVE

This article synthesizes the current literature on prognostication in neurocritical care, identifies existing challenges, and proposes future research directions to reduce variability and enhance scientific and patient-centered approaches to neuroprognostication.

LATEST DEVELOPMENTS

Patients with severe acute brain injury often lack the capacity to make their own medical decisions, leaving surrogate decision makers responsible for life-or-death choices. These decisions heavily rely on clinicians' prognostication, which is still considered an art because of the previous lack of specific guidelines. Consequently, there is significant variability in neuroprognostication practices. This article examines various aspects of neuroprognostication. It explores the cognitive approach to prognostication, highlights the use of statistical modeling such as Bayesian models and machine learning, emphasizes the importance of clinician-family communication during prognostic disclosures, and proposes shared decision making for more patient-centered care.

ESSENTIAL POINTS

This article identifies ongoing challenges in the field and emphasizes the need for future research to ameliorate variability in neuroprognostication. By focusing on scientific methodologies and patient-centered approaches, this research aims to provide guidance and tools that may enhance neuroprognostication in neurocritical care.

Clinicians' approach to predicting post-cardiac arrest outcomes for patients enrolled in a United States clinical trial

PURPOSE

Perceived poor prognosis can lead to withdrawal of life-sustaining therapies (WLST) in patients who might otherwise recover. We characterized clinicians' approach to post-arrest prognostication in a multicenter clinical trial.

METHODS

Semi-structured interviews were conducted with clinicians who treated a comatose post-cardiac arrest patient enrolled in the Influence of Cooling Duration on Efficacy in Cardiac Arrest Patients (ICECAP) trial (NCT04217551). Two authors independently analyzed each interview using inductive and deductive coding. The clinician reported how they arrived at a prognosis for the specific patient. We summarized the frequency with which clinicians reported using objective diagnostics to formulate their prognosis, and compared the reported approaches to established guidelines. Each respondent provided demographic information and described local neuroprognostication practices.

RESULTS

We interviewed 30 clinicians at 19 US hospitals. Most claimed adherence to local hospital neuroprognostication protocols (n = 19). Prognostication led to WLST for perceived poor neurological prognosis in 15/30 patients, of whom most showed inconsistencies with guidelines or trial recommendations, respectively. In 10/15 WLST cases, clinicians reported relying on multimodal testing. A prevalent theme was the use of "clinical gestalt," defined as prognosticating based on a patient's overall appearance or a subjective impression in the absence of objective data. Many clinicians (21/30) reported using clinical gestalt for initial prognostication, with 9/21 expressing high confidence initially.

CONCLUSION

Clinicians in our study state they follow neuroprognostication guidelines in general but often do not do so in actual practice. They reported clinical gestalt frequently informed early, highly confident prognostic judgments, and few objective tests changed initial impressions. Subjective prognostication may undermine well-designed trials.

Guidelines for Neuroprognostication in Critically Ill Adults with Intracerebral Hemorrhage

BACKGROUND

The objective of this document is to provide recommendations on the formal reliability of major clinical predictors often associated with intracerebral hemorrhage (ICH) neuroprognostication.

METHODS

A narrative systematic review was completed using the Grading of Recommendations Assessment, Development, and Evaluation methodology and the Population, Intervention, Comparator, Outcome, Timing, Setting questions. Predictors, which included both individual clinical variables and prediction models, were selected based on clinical relevance and attention in the literature. Following construction of the evidence profile and summary of findings, recommendations were based on Grading of Recommendations Assessment, Development, and Evaluation criteria. Good practice statements addressed essential principles of neuroprognostication that could not be framed in the Population, Intervention, Comparator, Outcome, Timing, Setting format.

RESULTS

Six candidate clinical variables and two clinical grading scales (the original ICH score and maximally treated ICH score) were selected for recommendation creation. A total of 347 articles out of 10,751 articles screened met our eligibility criteria. Consensus statements of good practice included deferring neuroprognostication-aside from the most clinically devastated patients-for at least the first 48-72 h of intensive care unit admission; understanding what outcomes would have been most valued by the patient; and counseling of patients and surrogates whose ultimate neurological recovery may occur over a variable period of time. Although many clinical variables and grading scales are associated with ICH poor outcome, no clinical variable alone or sole clinical grading scale was suggested by the panel as currently being reliable by itself for use in counseling patients with ICH and their surrogates, regarding functional outcome at 3 months and beyond or 30-day mortality.

CONCLUSIONS

These guidelines provide recommendations on the formal reliability of predictors of poor outcome in the context of counseling patients with ICH and surrogates and suggest broad principles of neuroprognostication. Clinicians formulating their judgments of prognosis for patients with ICH should avoid anchoring bias based solely on any one clinical variable or published clinical grading scale.

Coma Prognostication After Acute Brain Injury: A Review

Importance

Among the most impactful neurologic assessments is that of neuroprognostication, defined here as the prediction of neurologic recovery from disorders of consciousness caused by severe, acute brain injury. Across a range of brain injury etiologies, these determinations often dictate whether life-sustaining treatment is continued or withdrawn; thus, they have major implications for morbidity, mortality, and health care costs. Neuroprognostication relies on a diverse array of tests, including behavioral, radiologic, physiological, and serologic markers, that evaluate the brain's functional and structural integrity.

Observations

Prognostic markers, such as the neurologic examination, electroencephalography, and conventional computed tomography and magnetic resonance imaging (MRI), have been foundational in assessing a patient's current level of consciousness and capacity for recovery. Emerging techniques, such as functional MRI, diffusion MRI, and advanced forms of electroencephalography, provide new ways of evaluating the brain, leading to evolving schemes for characterizing neurologic function and novel methods for predicting recovery.

Conclusions and Relevance

Neuroprognostic markers are rapidly evolving as new ways of assessing the brain's structural and functional integrity after brain injury are discovered. Many of these techniques remain in development, and further research is needed to optimize their prognostic utility. However, even as such efforts are underway, a series of promising findings coupled with the imperfect predictive value of conventional prognostic markers and the high stakes of these assessments have prompted clinical guidelines to endorse emerging techniques for neuroprognostication. Thus, clinicians have been thrust into an uncertain predicament in which emerging techniques are not yet perfected but too promising to ignore. This review illustrates the current, and likely future, landscapes of prognostic markers. No matter how much prognostic markers evolve and improve, these assessments must be approached with humility and individualized to reflect each patient's values.

Doctors and nurses subjective predictions of 6-month outcome compared to actual 6-month outcome for adult patients with spontaneous intracerebral haemorrhage (ICH) in neurocritical care: An observational study

Background

Acute spontaneous intracerebral haemorrhage is a devastating form of stroke. Prognostication after ICH may be influenced by clinicians' subjective opinions.

Purpose

To evaluate subjective predictions of 6-month outcome by clinicians' for ICH patients in a neurocritical care using the modified Rankin Scale (mRS) and compare these to actual 6-month outcome.

Method

We included clinicians' predictions of 6-month outcome in the first 48 h for 52 adults with ICH and compared to actual 6-month outcome using descriptive statistics and multilevel binomial logistic regression.

Results

35/52 patients (66%) had a poor 6-month outcome (mRS 4-6); 19/52 (36%) had died. 324 predictions were included. For good (mRS 0-3) versus poor (mRS 4-6), outcome, accuracy of predictions was 68% and exact agreement 29%. mRS 6 and mRS 4 received the most correct predictions. Comparing job roles, predictions of death were underestimated, by doctors (12%) and nurses (13%) compared with actual mortality (36%). Predictions of vital status showed no significant difference between doctors and nurses: OR = 1.24 {CI; 0.50-3.05}; (p = 0.64) or good versus poor outcome: OR = 1.65 {CI; 0.98-2.79}; (p = 0.06). When predicted and actual 6-month outcome were compared, job role did not significantly relate to correct predictions of good versus poor outcome: OR = 1.13 {CI;0.67-1.90}; (p = 0.65) or for vital status: OR = 1.11 {CI; 0.47-2.61}; p = 0.81).

Conclusions

Early prognostication is challenging. Doctors and nurses were most likely to correctly predict poor outcome but tended to err on the side of optimism for mortality, suggesting an absence of clinical nihilism in relation to ICH.

Medical versus neurosurgical treatment in ICH patients: a single center experience

BACKGROUND AND AIMS

The effect of surgical treatment for spontaneous intracerebral hemorrhage (ICH) remains uncertain. We conducted an observational retrospective cohort study on supra-centimeter spontaneous ICH treated with either neurosurgical or conservative management. The baseline demographics and risk factors were correlated with in-hospital mortality and 3 and 6-month survival rates stratified by management.

METHODS

We included all patients with evidence of spontaneous ICH > 1 cm detected by CT and admitted between august 2020 and march 2021 to the "SMM" Hospital in Perugia.

RESULTS

Onehundredandtwentytwo patients were included in the study, and 45% (n.55) were surgically treated. The mean age was 71.9 ± 15.3, and 61% (n.75) were males. Intra-hospital mortality ended up being 31% (n.38), 3 months-survival was 63% (n.77) and 6 months-survival was 60% (n.73). From the multivariate analysis of the surgical patients versus medical patient, we observed that the surgical patients were younger (67.5 ± 14.9 vs 75.5 ± 14.7 y; OR 0.87; Cl 95% 0.85-0.94; p 0.001), with greater ICH volume at the onset (61 ± 39.4 cc vs 51 ± 64 cc; OR 1.03; Cl 95% 1.005-1.07; p 0.05), more midline shift (7.61 ± 5.54 mm vs 4.09 ± 5.88 mm; OR 1.37; Cl 95% 1.045-1.79; p 0.023), and a higher ICH score (3 vs 2 mean ICH score; OR 21.12; Cl 95% 2.6-170.6; p 0.004). Intra-hospital mortality in the surgical group and in the conservative treatment group was respectively 33% vs 30%, 3 month-survival was 64% vs 63% and 6 month- survival were 60% in both groups.

CONCLUSIONS

Our patient cohort shows no overall benefit from surgery over conservative treatment, but surgical patients were younger and had larger ICH volume.

Clinical Grading Scales and Neuroprognostication in Acute Brain Injury

Prediction of neurological clinical outcome after acute brain injury is critical because it helps guide discussions with patients and families and informs treatment plans and allocation of resources. Numerous clinical grading scales have been published that aim to support prognostication after acute brain injury. However, the development and validation of clinical scales lack a standardized approach. This in turn makes it difficult for clinicians to rely on prognostic grading scales and to integrate them into clinical practice. In this review, we discuss quality measures of score development and validation and summarize available scales to prognosticate outcomes after acute brain injury. These include scales developed for patients with coma, cardiac arrest, ischemic stroke, nontraumatic intracerebral hemorrhage, subarachnoid hemorrhage, and traumatic brain injury; for each scale, we discuss available validation studies.

Cognitive Biases and Shared Decision Making in Acute Brain Injury

Many patients hospitalized after severe acute brain injury are comatose and require life-sustaining therapies. Some of these patients make favorable recoveries with continued intensive care, while others do not. In addition to providing medical care, clinicians must guide surrogate decision makers through high-stakes, emotionally charged decisions about whether to continue life-sustaining therapies. These consultations require clinicians first to assess a patient's likelihood of recovery given continued life-sustaining therapies (i.e., prognosticate), then to communicate that prediction to surrogates, and, finally, to elicit and interpret the patient's preferences. At each step, both clinicians and surrogates are vulnerable to flawed decision making. Clinicians can be imprecise, biased, and overconfident when prognosticating after brain injury. Surrogates can misperceive the choice and misunderstand or misrepresent a patient's wishes, which may never have been communicated clearly. These biases can undermine the ability to reach choices congruent with patients' preferences through shared decision making (SDM). Decision science has extensively studied these biases. In this article, we apply that research to improving SDM for patients who are comatose after acute brain injury. After introducing SDM and the medical context, we describe principal decision science results as they relate to neurologic prognostication and end-of-life decisions, by both clinicians and surrogates. Based on research regarding general processes that can produce imprecise, biased, and overconfident prognoses, we propose interventions that could improve SDM, supporting clinicians and surrogates in making these challenging decisions.

Prognostic Language in Critical Neurologic Illness: A Multicenter Mixed-Methods Study

BACKGROUND AND OBJECTIVES

There are no evidence-based guidelines for discussing prognosis in critical neurologic illness, but in general, experts recommend that clinicians communicate prognosis using estimates, such as numerical or qualitative expressions of risk. Little is known about how real-world clinicians communicate prognosis in critical neurologic illness. Our primary objective was to characterize prognostic language clinicians used in critical neurologic illness. We additionally explored whether prognostic language differed between prognostic domains (e.g., survival, cognition).

METHODS

We conducted a multicenter cross-sectional mixed-methods study analyzing deidentified transcripts of audio-recorded clinician-family meetings for patients with neurologic illness requiring intensive care (e.g., intracerebral hemorrhage, traumatic brain injury, severe stroke) from 7 US centers. Two coders assigned codes for prognostic language type and domain of prognosis to each clinician prognostic statement. Prognostic language was coded as probabilistic (estimating the likelihood of an outcome occurring, e.g., "80% survival"; "She'll probably survive") or nonprobabilistic (characterizing outcomes without offering likelihood; e.g., "She may not survive"). We applied univariate and multivariate binomial logistic regression to examine independent associations between prognostic language and domain of prognosis.

RESULTS

We analyzed 43 clinician-family meetings for 39 patients with 78 surrogates and 27 clinicians. Clinicians made 512 statements about survival (median 0/meeting [interquartile range (IQR) 0-2]), physical function (median 2 [IQR 0-7]), cognition (median 2 [IQR 0-6]), and overall recovery (median 2 [IQR 1-4]). Most statements were nonprobabilistic (316/512 [62%]); 10 of 512 prognostic statements (2%) offered numeric estimates; and 21% (9/43) of family meetings only contained nonprobabilistic language. Compared with statements about cognition, statements about survival (odds ratio [OR] 2.50, 95% CI 1.01-6.18, p = 0.048) and physical function (OR 3.22, 95% 1.77-5.86, p < 0.001) were more frequently probabilistic. Statements about physical function were less likely to be uncertainty-based than statements about cognition (OR 0.34, 95% CI 0.17-0.66, p = 0.002).

DISCUSSION

Clinicians preferred not to use estimates (either numeric or qualitative) when discussing critical neurologic illness prognosis, especially when they discussed cognitive outcomes. These findings may inform interventions to improve prognostic communication in critical neurologic illness.

Prediction of poor outcome in stroke patients using radiomics analysis of intraparenchymal and intraventricular hemorrhage and clinical factors

PURPOSE

To build three prognostic models using radiomics analysis of the hemorrhagic lesions, clinical variables, and their combination, to predict the outcome of stroke patients with spontaneous intracerebral hemorrhage (sICH).

MATERIALS AND METHODS

Eighty-three sICH patients were included. Among them, 40 patients (48.2%) had poor prognosis with modified Rankin scale (mRS) of 5 and 6 at discharge, and the prognostic model was built to differentiate mRS ≤ 4 vs. 5 + 6. The region of interest (ROI) of intraparenchymal hemorrhage (IPH) and intraventricular hemorrhage (IVH) were separately segmented. Features were extracted using PyRadiomics, and the support vector machine was applied to select features and build radiomics models based on IPH and IPH + IVH. The clinical models were built using multivariate logistic regression, and then the radiomics scores were combined with clinical variables to build the combined model.

RESULTS

When using IPH, the AUC for radiomics, clinical, and combined model was 0.78, 0.82, and 0.87, respectively. When using IPH + IVH, the AUC was increased to 0.80, 0.84, and 0.90, respectively. The combined model had a significantly improved AUC compared to the radiomics by DeLong test. A clinical prognostic model based on the ICH score of 0-1 only achieved AUC of 0.71.

CONCLUSIONS

The combined model using the radiomics score derived from IPH + IVH and the clinical factors could achieve a high accuracy in prediction of sICH patients with poor outcome, which may be used to assist in making the decision about the optimal care.

Acute Spontaneous Lobar Cerebral Hemorrhages Present a Different Clinical Profile and a More Severe Early Prognosis than Deep Subcortical Intracerebral Hemorrhages-A Hospital-Based Stroke Registry Study

Acute spontaneous intracerebral hemorrhage (ICH) is the most severe stroke subtype, with a high risk of death, dependence, and dementia. Knowledge about the clinical profile and early outcomes of ICH patients with lobar versus deep subcortical brain topography remains limited. In this study, we investigated the effects of ICH topography on demographics, cerebrovascular risk factors, clinical characteristics, and early outcomes in a sample of 298 consecutive acute ICH patients (165 with lobar and 133 with subcortical hemorrhagic stroke) available in a single-center-based stroke registry over 24 years. The multiple logistic regression analysis shows that variables independently associated with lobar ICH were early seizures (OR 6.81, CI 95% 1.27−5.15), chronic liver disease (OR 4.55, 95% CI 1.03−20.15), hemianopia (OR 2.55, 95% CI 1.26−5.15), headaches (OR 1.90, 95% CI 1.90, 95% IC 1.06−3.41), alcohol abuse (>80 gr/day) (OR 0−10, 95% CI 0.02−0,53), hypertension (OR 0,41, 95% CI 0.23−0−70), sensory deficit (OR 0.43, 95% CI 0.25−0.75), and limb weakness (OR: 0.47, 95% CI 0.24−0.93). The in-hospital mortality was 26.7% for lobar and 16.5% for subcortical ICH. The study confirmed that the clinical spectrum, prognosis, and early mortality of patients with ICH depend on the site of bleeding, with a more severe early prognosis in lobar intracerebral hemorrhage.

Reappraisal of Intracerebral Hemorrhages and Intracerebral Hemorrhage Grading Scale Score in Surgically and Medically Managed Cerebellar Intracerebral Hemorrhage

BACKGROUND

As compared with supratentorial intracerebral hemorrhages (ICH), bleeds that occur within the cerebellum require special consideration given the nature of the posterior fossa.

OBJECTIVE

To validate ICH and ICH grading scale (ICH-GS) scores in patients with cerebellar hemorrhage and examine the outcomes of patients managed surgically as compared with those who underwent conservative treatment.

METHODS

This observational multicenter study included 475 patients with cerebellar hemorrhage from 9 different neurosurgical departments in Germany between 2005 and 2021. The prognostic accuracy of ICH and ICH-GS scores were calculated by the area under the curve of the receiver operating characteristic curves. Analyzed outcomes were the in-hospital mortality, mortality at 6 months, in-hospital outcome, and outcome at 6 months.

RESULTS

Of 403 patients, 252 patients (62.5%) underwent surgical treatment and 151 patients (37.5%) conservative treatment. Both ICH and ICH-GS scores demonstrated good prognostic accuracy regarding both overall mortality and functional outcomes. In those patients presenting with severe cerebellar hemorrhages, ie, ICH score >3 and ICH-GS score >11, overall mortality was significantly lower in surgically treated patients. Mortality was significantly higher in those patients managed surgically who presented with ICH scores ≤3; in such patients, improved outcomes were noted when the hematoma was treated conservatively.

CONCLUSION

ICH and ICH scores are useful tools for prediction of survival and outcome in patients with cerebellar ICH. Surgical management may be beneficial for those who present with severe cerebellar ICH as reflected by ICH scores >3, while conservative management seems reasonable in patients with lower ICH scores.

ANAID-ICH nomogram for predicting unfavorable outcome after intracerebral hemorrhage

OBJECTIVE

Diffusion-weighted imaging lesions (DWILs) are associated with unfavorable outcome in intracerebral hemorrhage (ICH). We proposed a novel predictive nomogram incorporating DWILs.

METHODS

A total of 738 patients with primary ICH in a tertiary hospital were prospectively enrolled as a training cohort. DWILs were defined as remote focal hyperintensities on DWI corresponding to low intensities on apparent diffusion coefficient images and remote from the focal hematoma. The outcome of interest was modified Rankin Scale scores of 4-6 at 90 days after onset. Multivariate logistic regression was used to construct a nomogram. Model performance was tested in the training cohort and externally validated with respect to discrimination, calibration, and clinical usefulness in another institute. Additionally, the nomogram was compared with the ICH score in terms of predictive ability.

RESULTS

Overall, 153 (20.73%) and 23 (15.54%) patients developed an unfavorable outcome in the training and validation cohorts, respectively. The multivariate analysis revealed that age, National Institutes of Health Stroke Scale (NIHSS) score, anemia, infratentorial location, presence of DWILs, and prior ICH were associated with unfavorable outcome. Our ANAID-ICH nomogram was constructed according to the aforementioned variables; the area under the receiver operating characteristic curve was 0.842 and 0.831 in the training and validation sets, respectively. With regard to the 90-day outcome, the nomogram showed a significantly higher predictive value than the ICH score in both cohorts.

CONCLUSIONS

The ANAID-ICH nomogram comprising age, NIHSS score, anemia, infratentorial location, presence of DWILs, and prior ICH may facilitate the identification of patients at higher risk for an unfavorable outcome.

Prognosis Predictions by Families, Physicians, and Nurses of Patients with Severe Acute Brain Injury: Agreement and Accuracy

BACKGROUND

Effective shared decision-making relies on some degree of alignment between families and the medical team regarding a patient's likelihood of recovery. Patients with severe acute brain injury (SABI) are often unable to participate in decisions, and therefore family members make decisions on their behalf. The goal of this study was to evaluate agreement between prognostic predictions by families, physicians, and nurses of patients with SABI regarding their likelihood of regaining independence and to measure each group's prediction accuracy.

METHODS

This observational cohort study, conducted from 01/2018 to 07/2020, was based in the neuroscience and medical/cardiac intensive care units of a single center. Patient eligibility included a diagnosis of SABI-specifically stroke, traumatic brain injury, or hypoxic ischemic encephalopathy-and a Glasgow Coma Scale ≤ 12 after hospital day 2. At enrollment, families, physicians, and nurses were asked separately to predict a patient's likelihood of recovering to independence within 6 months on a 0-100 scale, regardless of whether a formal family meeting had occurred. True outcome was based on modified Rankin Scale assessment through a family report or medical chart review. Prognostic agreement was measured by (1) intraclass correlation coefficient; (2) mean group prediction comparisons using paired Student's t-tests; and (3) prevalence of concordance, defined as an absolute difference of less than 20 percentage points between predictions. Accuracy for each group was measured by calculating the area under a receiver operating characteristic curve (C statistic) and compared by using DeLong's test.

RESULTS

Data were collected from 222 patients and families, 45 physicians, and 103 nurses. Complete data on agreement and accuracy were available for 187 and 177 patients, respectively. The intraclass correlation coefficient, in which 1 indicates perfect correlation and 0 indicates no correlation, was 0.49 for physician-family pairs, 0.40 for family-nurse pairs, and 0.66 for physician-nurse pairs. The difference in mean predictions between families and physicians was 23.5 percentage points (p < 0.001), 25.4 between families and nurses (p < 0.001), and 1.9 between physicians and nurses (p = 0.38). Prevalence of concordance was 39.6% for family-physician pairs, 30.0% for family-nurse pairs, and 56.2% for physician-nurse pairs. The C statistic for prediction accuracy was 0.65 for families, 0.82 for physicians, and 0.76 for nurses. The p values for differences in C statistics were < 0.05 for family-physician and family-nurse groups and 0.18 for physician-nurse groups.

CONCLUSIONS

For patients with SABI, agreement in predictions between families, physicians, and nurses regarding likelihood of recovery is poor. Accuracy appears higher for physicians and nurses compared with families, with no significant difference between physicians and nurses.

Top Ten Tips Palliative Care Clinicians Should Know About Disorders of Consciousness: A Focus on Traumatic and Anoxic Brain Injury

Palliative care (PC) teams commonly encounter patients with disorders of consciousness (DOC) following anoxic or traumatic brain injury (TBI). Primary teams may consult PC to help surrogates in making treatment choices for these patients. PC clinicians must understand the complexity of predicting neurologic outcomes, address clinical nihilism, and appropriately guide surrogates in making decisions that are concordant with patients' goals. The purpose of this article was to provide PC providers with a better understanding of caring for patients with DOC, specifically following anoxic or TBI. Many of the tips acknowledge the uncertainty of DOC and provide strategies to help tackle this dilemma.

Prognostication and shared decision making in neurocritical care

Prognostication is crucial in the neurological intensive care unit (neuroICU). Patients with severe acute brain injury (SABI) are unable to make their own decisions because of the insult itself or sedation needs. Surrogate decision makers, usually family members, must make decisions on the patient's behalf. However, many are unprepared for their role as surrogates owing to the sudden and unexpected nature of SABI. Surrogates rely on clinicians in the neuroICU to provide them with an outlook (prognosis) with which to make substituted judgments and decide on treatments and goals of care on behalf of the patient. Therefore, how a prognostic estimate is derived, and then communicated, is extremely important. Prognostication in the neuroICU is highly variable between clinicians and institutions, and evidence based guidelines are lacking. Shared decision making (SDM), where surrogates and clinicians arrive together at an individualized decision based on patient values and preferences, has been proposed as an opportunity to improve clinician-family communication and ensure that patients receive treatments they would choose. This review outlines the importance and current challenges of prognostication in the neuroICU and how prognostication and SDM intersect, based on relevant research and expert opinion.

Herniation World Federation of Neurosurgical Societies Scale Improves Prediction of Outcome in Patients With Poor-Grade Aneurysmal Subarachnoid Hemorrhage

BACKGROUND

Favorable outcomes are seen in up to 50% of patients with World Federation of Neurosurgical Societies (WFNS) grade V aneurysmal subarachnoid hemorrhage. Therefore, the usefulness of the current WFNS grading system for identifying the worst scenarios for clinical studies and for making treatment decisions is limited. We previously modified the WFNS scale by requiring positive signs of brain stem dysfunction to assign grade V. This study aimed to validate the new herniation WFNS grading system in an independent prospective cohort.

METHODS

We conducted an international prospective multicentre study in poor-grade aneurysmal subarachnoid hemorrhage patients comparing the WFNS classification with a modified version-the herniation WFNS scale (hWFNS). Here, only patients who showed positive signs of brain stem dysfunction (posturing, anisocoric, or bilateral dilated pupils) were assigned hWFNS grade V. Outcome was assessed by modified Rankin Scale score 6 months after hemorrhage. The primary end point was the difference in specificity of the WFNS and hWFNS grading with respect to poor outcomes (modified Rankin Scale score 4-6).

RESULTS

Of the 250 patients included, 237 reached the primary end point. Comparing the WFNS and hWFNS scale after neurological resuscitation, the specificity to predict poor outcome increased from 0.19 (WFNS) to 0.93 (hWFNS) (McNemar, P<0.001) whereas the sensitivity decreased from 0.88 to 0.37 (P<0.001), and the positive predictive value from 61.9 to 88.3 (weighted generalized score statistic, P<0.001). For mortality, the specificity increased from 0.19 to 0.93 (McNemar, P<0.001), and the positive predictive value from 52.5 to 86.7 (weighted generalized score statistic, P<0.001).

CONCLUSIONS

The identification of objective positive signs of brain stem dysfunction significantly improves the specificity and positive predictive value with respect to poor outcome in grade V patients. Therefore, a simple modification-presence of brain stem signs is required for grade V-should be added to the WFNS classification.

REGISTRATION

URL: https://clinicaltrials.gov; Unique identifier: NCT02304328.

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.

Risk Factors Associated With Mortality and Neurologic Disability After Intracerebral Hemorrhage in a Racially and Ethnically Diverse Cohort

Introduction

Intracerebral hemorrhage (ICH) is the most severe subtype of stroke. Its mortality rate is high, and most survivors experience significant disability.

Objective

To assess primary patient risk factors associated with mortality and neurologic disability 3 months after ICH in a large, racially and ethnically balanced cohort.

Design, Setting, and Participants

This cohort study included participants from the Ethnic/Racial Variations of Intracerebral Hemorrhage (ERICH) study, which prospectively recruited 1000 non-Hispanic White, 1000 non-Hispanic Black, and 1000 Hispanic patients with spontaneous ICH to study the epidemiological characteristics and genomics associated with ICH. Participants included those with uniform data collection and phenotype definitions, centralized neuroimaging review, and telephone follow-up at 3 months. Analyses were completed in November 2021.

Exposures

Patient demographic and clinical characteristics as well as hospital event and imaging variables were examined, with characteristics meeting P < .20 considered candidates for a multivariate model. Elements included in the ICH score were specifically analyzed.

Main Outcomes and Measures

Individual characteristics were screened for association with 3-month outcome of neurologic disability or mortality, as assessed by a modified Rankin Scale (mRS) score of 4 or greater vs 3 or less under a logistic regression model. A total of 25 characteristics were tested in the final model, which minimized the Akaike information criterion. Analyses were repeated removing individuals who had withdrawal of care.

Results

A total of 2568 patients (mean [SD] age, 62.4 [14.7] years; 1069 [41.6%] women and 1499 [58.4%] men) had a 3-month outcome determination available, including death. The final logistic model had a significantly higher area under the receiver operating characteristics curve (C = 0.88) compared with ICH score alone (C = 0.76; P < .001). Among characteristics associated with neurologic disability and mortality were larger log ICH volume (OR, 2.74; 95% CI, 2.36-3.19; P < .001), older age (OR per 1-year increase, 1.04; 95% CI, 1.02-1.05; P < .001), pre-ICH mRS score (OR, 1.62; 95% CI, 1.41-1.87; P < .001), lobar location (OR, 0.22; 95% CI, 0.16-0.30; P < .001), and presence of infection (OR, 1.85; 95% CI, 1.42-2.41; P < .001).

Conclusions and Relevance

The findings of this cohort study validate ICH score elements and suggest additional baseline and interim patient characteristics were associated with variation in 3-month outcome.

Prognostication in neurology

Prognosticating is central to primary palliative care in neurology. Many neurologic diseases carry a high burden of troubling symptoms, and many individuals consider health states due to neurologic disease worse than death. Many patients and families report high levels of need for information at all disease stages, including information about prognosis. There are many barriers to communicating prognosis including prognostic uncertainty, lack of training and experience, fear of destroying hope, and not enough time. Developing the right mindset, tools, and skills can improve one's ability to formulate and communicate prognosis. Prognosticating is subject to many biases which can dramatically affect the quality of patient care; it is important for providers to recognize and reduce them. Patients and surrogates often do not hear what they are told, and even when they hear correctly, they form their own opinions. With practice and self-reflection, one can improve their prognostic skills, help patients and families create honest roadmaps of the future, and deliver high-quality person-centered care.

Machine Learning in Action: Stroke Diagnosis and Outcome Prediction

The application of machine learning has rapidly evolved in medicine over the past decade. In stroke, commercially available machine learning algorithms have already been incorporated into clinical application for rapid diagnosis. The creation and advancement of deep learning techniques have greatly improved clinical utilization of machine learning tools and new algorithms continue to emerge with improved accuracy in stroke diagnosis and outcome prediction. Although imaging-based feature recognition and segmentation have significantly facilitated rapid stroke diagnosis and triaging, stroke prognostication is dependent on a multitude of patient specific as well as clinical factors and hence accurate outcome prediction remains challenging. Despite its vital role in stroke diagnosis and prognostication, it is important to recognize that machine learning output is only as good as the input data and the appropriateness of algorithm applied to any specific data set. Additionally, many studies on machine learning tend to be limited by small sample size and hence concerted efforts to collate data could improve evaluation of future machine learning tools in stroke. In the present state, machine learning technology serves as a helpful and efficient tool for rapid clinical decision making while oversight from clinical experts is still required to address specific aspects not accounted for in an automated algorithm. This article provides an overview of machine learning technology and a tabulated review of pertinent machine learning studies related to stroke diagnosis and outcome prediction.

Machine Learning Models Prognosticate Functional Outcomes Better than Clinical Scores in Spontaneous Intracerebral Haemorrhage

OBJECTIVE

This study aims to develop and compare the use of deep neural networks (DNN) and support vector machines (SVM) to clinical prognostic scores for prognosticating 30-day mortality and 90-day poor functional outcome (PFO) in spontaneous intracerebral haemorrhage (SICH).

MATERIALS AND METHODS

We conducted a retrospective cohort study of 297 SICH patients between December 2014 and May 2016. Clinical data was collected from electronic medical records using standardized data collection forms. The machine learning workflow included imputation of missing data, dimensionality reduction, imbalanced-class correction, and evaluation using cross-validation and comparison of accuracy against clinical prognostic scores.

RESULTS

32 (11%) patients had 30-day mortality while 177 (63%) patients had 90-day PFO. For prognosticating 30-day mortality, the class-balanced accuracies for DNN (0.875; 95% CI 0.800-0.950; McNemar's p-value 1.000) and SVM (0.848; 95% CI 0.767-0.930; McNemar's p-value 0.791) were comparable to that of the original ICH score (0.833; 95% CI 0.748-0.918). The c-statistics for DNN (0.895; DeLong's p-value 0.715), and SVM (0.900; DeLong's p-value 0.619), though greater than that of the original ICH score (0.862), were not significantly different. For prognosticating 90-day PFO, the class-balanced accuracies for DNN (0.853; 95% CI 0.772-0.934; McNemar's p-value 0.003) and SVM (0.860; 95% CI 0.781-0.939; McNemar's p-value 0.004) were better than that of the ICH-Grading Scale (0.706; 95% CI 0.600-0.812). The c-statistic for SVM (0.883; DeLong's p-value 0.022) was significantly greater than that of the ICH-Grading Scale (0.778), while the c-statistic for DNN was 0.864 (DeLong's p-value 0.055).

CONCLUSION

We showed that the SVM model performs significantly better than clinical prognostic scores in predicting 90-day PFO in SICH.

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.

SMASH-U aetiological classification: A predictor of long-term functional outcome after intracerebral haemorrhage

BACKGROUND

SMASH-U is a systematic aetiological classification system for intracerebral haemorrhage (ICH) proven to be a predictor of post-ICH haematoma expansion and mortality. However, its role in predicting functional outcome remains elusive. Therefore, we aimed to investigate whether SMASH-U is associated with long-term functional outcome after ICH and improves the accuracy of prediction when added to max-ICH score.

METHODS

Consecutive acute ICH patients from 2012 to 2018 from the neurology department of Tongji Hospital were enrolled. ICH aetiology was classified according to the SMASH-U system. The association of SMASH-U with 12-month functional outcome after ICH and the predictive value were evaluated.

RESULTS

Of 1938 ICH patients, the aetiology of 1295 (66.8%) patients were classified as hypertension, followed by amyloid angiopathy (n = 250, 12.9%), undetermined (n = 159, 8.2%), structural lesions (n = 149, 7.7%), systemic disease (n = 74, 3.8%) and medication (n = 11, 0.6%). The baseline characteristics were different among the six aetiologies. In multivariate analysis, SMASH-U was proven to be a predictor of 12-month unfavourable functional outcome. When adding the SMASH-U system, the predictive performance of max-ICH score was improved (area under the receiver operating characteristic curve from 0.802 to 0.812, p = 0.010) and the predictive accuracy was enhanced (integrated discrimination improvement [IDI]: 1.60%, p < 0.001; continuous net reclassification improvement [NRI]: 28.16%, p < 0.001; categorical NRI: 3.34%, p = 0.004).

CONCLUSIONS

SMASH-U predicted long-term unfavourable functional outcomes after acute ICH and improved the accuracy of prediction when added to max-ICH score. Integrating the aetiology to a score model to predict the post-ICH outcome may be meaningful and worthy of further exploration.

Frequency of Withdrawal of Life-Sustaining Therapy for Perceived Poor Neurologic Prognosis

Supplemental Digital Content is available in the text.

OBJECTIVES:

To measure the frequency of withdrawal of life-sustaining therapy for perceived poor neurologic prognosis among decedents in hospitals of different sizes and teaching statuses.

DESIGN:

We performed a multicenter, retrospective cohort study.

SETTING:

Four large teaching hospitals, four affiliated small teaching hospitals, and nine affiliated nonteaching hospitals in the United States.

PATIENTS:

We included a sample of all adult inpatient decedents between August 2017 and August 2019.

MEASUREMENTS AND MAIN RESULTS:

We reviewed inpatient notes and categorized the immediately preceding circumstances as withdrawal of life-sustaining therapy for perceived poor neurologic prognosis, withdrawal of life-sustaining therapy for nonneurologic reasons, limitations or withholding of life support or resuscitation, cardiac death despite full treatment, or brain death. Of 2,100 patients, median age was 71 years (interquartile range, 60–81 yr), median hospital length of stay was 5 days (interquartile range, 2–11 d), and 1,326 (63%) were treated at four large teaching hospitals. Withdrawal of life-sustaining therapy for perceived poor neurologic prognosis occurred in 516 patients (25%) and was the sole contributing factor to death in 331 (15%). Withdrawal of life-sustaining therapy for perceived poor neurologic prognosis was common in all hospitals: 30% of deaths at large teaching hospitals, 19% of deaths in small teaching hospitals, and 15% of deaths at nonteaching hospitals. Withdrawal of life-sustaining therapy for perceived poor neurologic prognosis happened frequently across all hospital units. Withdrawal of life-sustaining therapy for perceived poor neurologic prognosis contributed to one in 12 deaths in patients without a primary neurologic diagnosis. After accounting for patient and hospital characteristics, significant between-hospital variability in the odds of withdrawal of life-sustaining therapy for perceived poor neurologic prognosis persisted.

CONCLUSIONS:

A quarter of inpatient deaths in this cohort occurred after withdrawal of life-sustaining therapy for perceived poor neurologic prognosis. The rate of withdrawal of life-sustaining therapy for perceived poor neurologic prognosis occurred commonly in all type of hospital settings. We observed significant unexplained variation in the odds of withdrawal of life-sustaining therapy for perceived poor neurologic prognosis across participating hospitals.

The Intracerebral Hemorrhage Score Overestimates Mortality in Young Adults

OBJECTIVE

To determine whether the intracerebral hemorrhage (ICH) score is accurate in predicting 30-day mortality in young adults, we calculated the ICH score for 156 young adults (aged 18-45) with primary spontaneous ICH and compared predicted to observed 30-day mortality rates.

METHODS

We retrospectively reviewed all patients aged 18-45 consecutively presenting to the University of Iowa from 2009 to 2019 with ICH. We calculated the ICH score and recorded its individual subcomponents for each patient. Poisson regression was used to test the association of ICH score components with 30-day mortality.

RESULTS

We identified 156 patients who met the inclusion criteria; mean± standard deviation (SD) age was 35±8 years. The 30-day mortality rate was 15% (n=24). The ICH score was predictive of 30-day mortality for each unit increase (p= 0.04 for trend), but the observed mortality rates for each ICH score varied considerably from the original ICH score predictions. Most notably, the 30-day mortality rates for ICH scores of 1, 2, and 3 are predicted to be 13%, 26%, and 72% respectively, but were observed in our population to be 0%, 3%, and 41%. An ICH volume of >30cc [relative risk (RR) 28, 95% confidence intervals (CI) 3-315, p=0.01] and a GCS score of <5 (RR 13, 95% CI 0.1-1176, p=0.01) were independently associated with 30-day mortality.

CONCLUSIONS

The ICH score tends to overestimate mortality in young adults. ICH volume and GCS score are the most relevant items in predicting mortality at 30 days in young adults.

Proceedings of the First Curing Coma Campaign NIH Symposium: Challenging the Future of Research for Coma and Disorders of Consciousness

Coma and disorders of consciousness (DoC) are highly prevalent and constitute a burden for patients, families, and society worldwide. As part of the Curing Coma Campaign, the Neurocritical Care Society partnered with the National Institutes of Health to organize a symposium bringing together experts from all over the world to develop research targets for DoC. The conference was structured along six domains: (1) defining endotype/phenotypes, (2) biomarkers, (3) proof-of-concept clinical trials, (4) neuroprognostication, (5) long-term recovery, and (6) large datasets. This proceedings paper presents actionable research targets based on the presentations and discussions that occurred at the conference. We summarize the background, main research gaps, overall goals, the panel discussion of the approach, limitations and challenges, and deliverables that were identified.

Prognostication after intracerebral hemorrhage: a review

Background

Approximately half of patients with spontaneous intracerebral hemorrhage (ICH) die within 1 year. Prognostication in this context is of great importance, to guide goals of care discussions, clinical decision-making, and risk stratification. However, available prognostic scores are hardly used in clinical practice. The purpose of this review article is to identify existing outcome prediction scores for spontaneous intracerebral hemorrhage (ICH) discuss their shortcomings, and to suggest how to create and validate more useful scores.

Main text

Through a literature review this article identifies existing ICH outcome prediction models. Using the Essen-ICH-score as an example, we demonstrate a complete score validation including discrimination, calibration and net benefit calculations. Score performance is illustrated in the Erlangen UKER-ICH-cohort (NCT03183167). We identified 19 prediction scores, half of which used mortality as endpoint, the remainder used disability, typically the dichotomized modified Rankin score assessed at variable time points after the index ICH. Complete score validation by our criteria was only available for the max-ICH score. Our validation of the Essen-ICH-score regarding prediction of unfavorable outcome showed good discrimination (area under the curve 0.87), fair calibration (calibration intercept 1.0, slope 0.84), and an overall net benefit of using the score as a decision tool. We discuss methodological pitfalls of prediction scores, e.g. the withdrawal of care (WOC) bias, physiological predictor variables that are often neglected by authors of clinical scores, and incomplete score validation. Future scores need to integrate new predictor variables, patient-reported outcome measures, and reduce the WOC bias. Validation needs to be standardized and thorough. Lastly, we discuss the integration of current ICH scoring systems in clinical practice with the awareness of their shortcomings.

Conclusion

Presently available prognostic scores for ICH do not fulfill essential quality standards. Novel prognostic scores need to be developed to inform the design of research studies and improve clinical care in patients with ICH.

Supplementary Information

The online version contains supplementary material available at 10.1186/s42466-021-00120-5.

Predicting long-term outcomes in acute intracerebral haemorrhage using delayed prognostication scores

Objective

The concept of the ‘self-fulfilling prophecy’ is well established in intracerebral haemorrhage (ICH). The ability to improve prognostication and prediction of long-term outcomes during the first days of hospitalisation is important in guiding conversations around goals of care. We previously demonstrated that incorporating delayed imaging into various prognostication scores for ICH improves the predictive accuracy of 90-day mortality. However, delayed prognostication scores have not been used to predict long-term functional outcomes beyond 90 days.

Design, setting and participants

We analysed data from the ICH Deferoxamine trial to see if delaying the use of prognostication scores to 96 hours after ICH onset will improve performance to predict outcomes at 180 days. 276 patients were included.

Interventions and measurements

We calculated the original ICH score (oICH), modified-ICH score (MICH), max-ICH score and the FUNC score on presentation (baseline), and on day 4 (delayed). Outcomes assessed were mortality and poor functional outcome in survivors (defined as modified Rankin Scale of 4–5) at 180 days. We generated receiver operating characteristic curves, and measured the area under the curve values (AUC) for mortality and functional outcome. We compared baseline and delayed AUCs with non-parametric methods.

Results

At 180 days, 21 of 276 (7.6%) died. Out of the survivors, 54 of 255 had poor functional outcome (21.2%). The oICH, MICH and max-ICH performed significantly better at predicting 180-day mortality when calculated 4 days later compared with their baseline equivalents ((0.74 vs 0.83, p=0.005), (0.73 vs 0.80, p=0.036), (0.74 vs 0.83, p=0.008), respectively). The delayed calculation of these scores did not significantly improve our accuracy for predicting poor functional outcomes.

Conclusion

Delaying the calculation of prognostication scores in acute ICH until day 4 improved prediction of 6-month mortality but not functional outcomes.

Trial registration number

ClinicalTrials.gov Registry (NCT02175225).

Adapting a Traumatic Brain Injury Goals-of-Care Decision Aid for Critically Ill Patients to Intracerebral Hemorrhage and Hemispheric Acute Ischemic Stroke

OBJECTIVES

Families in the neurologic ICU urgently request goals-of-care decision support and shared decision-making tools. We recently developed a goals-of-care decision aid for surrogates of critically ill traumatic brain injury patients using a systematic development process adherent to the International Patient Decision Aid Standards. To widen its applicability, we adapted this decision aid to critically ill patients with intracerebral hemorrhage and large hemispheric acute ischemic stroke.

DESIGN

Prospective observational study.

SETTING

Two academic neurologic ICUs.

SUBJECTS

Twenty family members of patients in the neurologic ICU were recruited from July 2018 to October 2018.

INTERVENTIONS

None.

MEASUREMENTS AND MAIN RESULTS

We reviewed the existing critically ill traumatic brain injury patients decision aid for content and changed: 1) the essential background information, 2) disease-specific terminology to "hemorrhagic stroke" and "ischemic stroke", and 3) disease-specific prognosis tailored to individual patients. We conducted acceptability and usability testing using validated scales. All three decision aids contain information from validated, disease-specific outcome prediction models, as recommended by international decision aid standards, including careful emphasis on their uncertainty. We replaced the individualizable icon arrays graphically depicting probabilities of a traumatic brain injury patient's prognosis with icon arrays visualizing intracerebral hemorrhage and hemispheric acute ischemic stroke prognostic probabilities using high-quality disease-specific data. We selected the Intracerebral Hemorrhage Score with validated 12-month outcomes, and for hemispheric acute ischemic stroke, the 12-month outcomes from landmark hemicraniectomy trials. Twenty family members participated in acceptability and usability testing (n = 11 for the intracerebral hemorrhage decision aid; n = 9 for the acute ischemic stroke decision aid). Median usage time was 22 minutes (interquartile range, 16-26 min). Usability was excellent (median System Usability Scale = 84/100 [interquartile range, 61-93; with > 68 indicating good usability]); 89% of participants graded the decision aid content as good or excellent, and greater than or equal to 90% rated it favorably for information amount, balance, and comprehensibility.

CONCLUSIONS

We successfully adapted goals-of-care decision aids for use in surrogates of critically ill patients with intracerebral hemorrhage and hemispheric acute ischemic stroke and found excellent usability and acceptability. A feasibility trial using these decision aids is currently ongoing to further validate their acceptability and test their feasibility for use in busy neurologic ICUs.

Management of Intracerebral Hemorrhage: JACC Focus Seminar

Intracerebral hemorrhage (ICH) accounts for a disproportionate amount of stroke-related morbidity and mortality. Although chronic hypertension and cerebral amyloid angiopathy are the underlying cerebral vasculopathies accounting for the majority of ICH, there are a broad range of potential causes, and effective management requires accurate identification and treatment of the underlying mechanism of hemorrhage. Magnetic resonance imaging and vascular imaging techniques play a critical role in identifying disease mechanisms. Modern treatment of ICH focuses on rapid stabilization, often requiring urgent treatment of mass effect, aggressive blood pressure reduction and correction of contributing coagulopathies to achieve hemostasis. We discuss management of patients with ICH who continue to require long-term anticoagulation, the interaction of ICH with neurodegenerative diseases, and our approach to prognostication after ICH. We close this review with a discussion of novel medical and surgical approaches to ICH treatment that are being tested in clinical trials.

Treatment of intracerebral hemorrhage: From specific interventions to bundles of care

Intracerebral hemorrhage (ICH) represents a major, global, unmet health need with few treatments. A significant minority of ICH patients present taking an anticoagulant; both vitamin-K antagonists and increasingly direct oral anticoagulants. Anticoagulants are associated with an increased risk of hematoma expansion, and rapid reversal reduces this risk and may improve outcome. Vitamin-K antagonists are reversed with prothrombin complex concentrate, dabigatran with idarucizumab, and anti-Xa agents with PCC or andexanet alfa, where available. Blood pressure lowering may reduce hematoma growth and improve clinical outcomes and careful (avoiding reductions ≥60 mm Hg within 1 h), targeted (as low as 120-130 mm Hg), and sustained (minimizing variability) treatment during the first 24 h may be optimal for achieving better functional outcomes in mild-to-moderate severity acute ICH. Surgery for ICH may include hematoma evacuation and external ventricular drainage to treat hydrocephalus. No large, well-conducted phase III trial of surgery in ICH has so far shown overall benefit, but meta-analyses report an increased likelihood of good functional outcome and lower risk of death with surgery, compared to medical treatment only. Expert supportive care on a stroke unit or critical care unit improves outcomes. Early prognostication is difficult, and early do-not-resuscitate orders or withdrawal of active care should be used judiciously in the first 24-48 h of care. Implementation of acute ICH care can be challenging, and using a care bundle approach, with regular monitoring of data and improvement of care processes can ensure consistent and optimal care for all patients.

The impact of delirium on withdrawal of life-sustaining treatment after intracerebral hemorrhage

OBJECTIVE

To determine the impact of delirium on withdrawal of life-sustaining treatment (WLST) after intracerebral hemorrhage (ICH) in the context of established predictors of poor outcome, using data from an institutional ICH registry.

METHODS

We performed a single-center cohort study on consecutive patients with ICH admitted over 12 months. ICH features were prospectively adjudicated, and WLST and corresponding hospital day were recorded retrospectively. Patients were categorized using DSM-5 criteria as never delirious, ever delirious (either on admission or later during hospitalization), or persistently comatose. We determined the impact of delirium on WLST using Cox regression models adjusted for demographics and ICH predictors (including Glasgow Coma Scale score), then used logistic regression with receiver operating characteristic curve analysis to compare the accuracy of ICH score-based models with and without delirium category in predicting WLST.

RESULTS

Of 311 patients (mean age 70.6 ± 15.6, median ICH score 1 [interquartile range 1-2]), 50% had delirium. WLST occurred in 26%, and median time to WLST was 1 day (0-6). WLST was more frequent in patients who developed delirium (adjusted hazard ratio 8.9 [95% confidence interval (CI) 2.1-37.6]), with high rates of WLST in both early (occurring ≤24 hours from admission) and later delirium groups. An ICH score-based model was strongly predictive of WLST (area under the curve [AUC] 0.902 [95% CI 0.863-0.941]), and the addition of delirium category further improved the model's accuracy (AUC 0.936 [95% CI 0.909-0.962], p = 0.004).

CONCLUSION

Delirium is associated with WLST after ICH regardless of when it occurs. Further study on the impact of delirium on clinician and surrogate decision-making is warranted.

Wisdom of the expert crowd prediction of response for 3 neurology randomized trials

OBJECTIVE

To explore the accuracy of combined neurology expert forecasts in predicting primary endpoints for trials.

METHODS

We identified one major randomized trial each in stroke, multiple sclerosis (MS), and amyotrophic lateral sclerosis (ALS) that was closing within 6 months. After recruiting a sample of neurology experts for each disease, we elicited forecasts for the primary endpoint outcomes in the trial placebo and treatment arms. Our main outcome was the accuracy of averaged predictions, measured using ordered Brier scores. Scores were compared against an algorithm that offered noncommittal predictions.

RESULTS

Seventy-one neurology experts participated. Combined forecasts of experts were less accurate than a noncommittal prediction algorithm for the stroke trial (pooled Brier score = 0.340, 95% subjective probability interval [sPI] 0.340 to 0.340 vs 0.185 for the uninformed prediction), and approximately as accurate for the MS study (pooled Brier score = 0.107, 95% confidence interval [CI] 0.081 to 0.133 vs 0.098 for the noncommittal prediction) and the ALS study (pooled Brier score = 0.090, 95% CI 0.081 to 0.185 vs 0.090). The 95% sPIs of individual predictions contained actual trial outcomes among 44% of experts. Only 18% showed prediction skill exceeding the noncommittal prediction. Independent experts and coinvestigators achieved similar levels of accuracy.

CONCLUSION

In this first-of-kind exploratory study, averaged expert judgments rarely outperformed noncommittal forecasts. However, experts at least anticipated the possibility of effects observed in trials. Our findings, if replicated in different trial samples, caution against the reliance on simple approaches for combining expert opinion in making research and policy decisions.

Scoping Review and Commentary on Prognostication for Patients with Intracerebral Hemorrhage with Advances in Surgical Techniques

INTRODUCTION

The intracerebral hemorrhage (ICH) score provides an estimate of 30-day mortality for patients with intracerebral hemorrhage in order to guide research protocols and clinical decision making. Several variations of such scoring systems have attempted to optimize its prognostic value. More recently, minimally invasive surgical techniques are increasingly being used with promising results. As more patients become candidates for surgical intervention, there is a need to re-discuss the best methods for predicting outcomes with or without surgical intervention.

METHODS

We systematically performed a scoping review with a comprehensive literature search by two independent reviewers using the PubMed and Cochrane databases for articles pertaining to the "intracerebral hemorrhage score." Relevant articles were selected for analysis and discussion of potential modifications to account for increasing surgical indications.

RESULTS

A total of 64 articles were reviewed in depth and identified 37 clinical grading scales for prognostication of spontaneous intracerebral hemorrhage. The original ICH score remains the most widely used and validated. Various authors proposed modifications for improved prognostic accuracy, though no single scale showed consistent superiority. Most recently, scales to account for advances in surgical techniques have been developed but lack external validation.

CONCLUSION

We provide the most comprehensive review to date of prognostic grading scales for patients with intracerebral hemorrhage. Current prognostic tools for patients with intracerebral hemorrhage remain limited and may overestimate risk of a poor outcome. As minimally invasive surgical techniques are developed, prognostic scales should account for surgical candidacy and outcomes.

Gap Analysis Regarding Prognostication in Neurocritical Care: A Joint Statement from the German Neurocritical Care Society and the Neurocritical Care Society

Background/Objective

Prognostication is a routine part of the delivery of neurocritical care for most patients with acute neurocritical illnesses. Numerous prognostic models exist for many different conditions. However, there are concerns about significant gaps in knowledge regarding optimal methods of prognostication.

Methods

As part of the Arbeitstagung NeuroIntensivMedizin meeting in February 2018 in Würzburg, Germany, a joint session on prognostication was held between the German NeuroIntensive Care Society and the Neurocritical Care Society. The purpose of this session was to provide presentations and open discussion regarding existing prognostic models for eight common neurocritical care conditions (aneurysmal subarachnoid hemorrhage, intracerebral hemorrhage, acute ischemic stroke, traumatic brain injury, traumatic spinal cord injury, status epilepticus, Guillain–Barré Syndrome, and global cerebral ischemia from cardiac arrest). The goal was to develop a qualitative gap analysis regarding prognostication that could help inform a future framework for clinical studies and guidelines.

Results

Prognostic models exist for all of the conditions presented. However, there are significant gaps in prognostication in each condition. Furthermore, several themes emerged that crossed across several or all diseases presented. Specifically, the self-fulfilling prophecy, lack of accounting for medical comorbidities, and absence of integration of in-hospital care parameters were identified as major gaps in most prognostic models.

Conclusions

Prognostication in neurocritical care is important, and current prognostic models are limited. This gap analysis provides a summary assessment of issues that could be addressed in future studies and evidence-based guidelines in order to improve the process of prognostication.

Are providers overconfident in predicting outcome after cardiac arrest?

AIM

To quantify the accuracy of health care providers' predictions of survival and function at hospital discharge in a prospective cohort of patients resuscitated from cardiac arrest. To test whether self-reported confidence in their predictions was associated with increased accuracy and whether this relationship varied across providers.

METHODOLOGY

We presented critical care and neurology providers with clinical vignettes using real data from post-arrest patients. We asked providers to predict survival, function at discharge, and report their confidence in these predictions. We used mixed effects models to explore predictors of confidence, accuracy, and the relationship between the two.

RESULTS

We completed 470 assessments of 62 patients with 65 providers. Of patients, 49 (78%) died and 9 (15%) had functionally favourable survival. Providers accurately predicted survival in 308/470 (66%) assessments. In most errors (146/162, 90%), providers incorrectly predicted survival. Providers accurately predicted function in 349/470 (74%) assessments. In most errors (114/121, 94%), providers incorrectly predicted favourable functional recovery. Providers were confident (median confidence predicting survival 80 [IQR 60-90]; median confidence predicting function 80 [IQR 60-95]). Confidence explained 9% and 18% of variation in accuracy predicting survival and function, respectively. We observed significant between-provider variability in accuracy (median odds ratio (MOR) for predicting survival 2.93, 95%CI 1.94-5.52; MOR for predicting function 5.42, 95%CI 3.01-13.2).

CONCLUSIONS

Providers varied in accuracy predicting post-arrest outcomes and most errors were optimistic. Self-reported confidence explained little variation in accuracy.