e-ISSN: 2722-3558 Sriwijaya Journal of Surgery [SJS] https://sriwijayasurgery.
External Validation of the RASH Score in Surgically Managed Acute Subdural Hematoma: A Critical Appraisal of Prognostic Accuracy and Surgical Factors in a Southeast Asian Cohort Irwansyah1*.
Trijoso Permono2.
Dwiandi Susilo2.
Erial Bahar3 1Surgery Study Program.
Faculty of Medicine.
Universitas Sriwijaya.
Palembang.
Indonesia 2Department of Neurosurgery.
Dr.
Mohammad Hoesin General Hospital.
Palembang.
Indonesia 3Faculty of Medicine.
Universitas Sriwijaya.
Palembang.
Indonesia
ARTICLE
ABSTRACT
INFO
Introduction: Acute subdural hematoma (ASDH) is a lethal form of traumatic brain injury (TBI) with high mortality.
The Richmond Acute Subdural Hematoma (RASH) score is a simple prognostic tool, but its validity in diverse populations is untested.
This study aimed to perform the first external validation of the RASH score in an Indonesian cohort and critically appraise its performance alongside key surgical factors.
Methods: We conducted a retrospective, single-center, diagnostic accuracy study of 67 adult patients who underwent surgery for traumatic ASDH between January 2022 and December 2024 at a tertiary neurosurgical center in Palembang.
Indonesia.
The RASH score was calculated from admission data.
additionally analyzed the type of surgery .
raniotomy vs.
and time from injury to operation.
The primary outcome was in-hospital mortality.
Receiver Operating Characteristic (ROC) curve analysis was used to evaluate the RASH score's predictive performance.
Results: The overall in-hospital mortality rate was 20.
9% .
The RASH score demonstrated excellent discrimination for mortality, with an Area Under the ROC Curve (AUC) of 0.
% CI: 0.
715Ae0.
p<0.
A score of 5 or greater was identified as the optimal cut-off, yielding a sensitivity of 78.
and specificity of 77.
This threshold provided a high Negative Predictive Value (NPV) of 93.
2% but a modest Positive Predictive Value (PPV) of 47.
In bivariate analysis, decompressive craniectomy and longer time to surgery were significantly associated with mortality.
Conclusion: The RASH score is a simple and robust tool for risk stratification in this selected surgical Its high NPV is valuable for identifying patients with a higher likelihood of survival.
However, its utility must be interpreted cautiously due to the significant selection bias inherent in studying only operable patients.
The score should serve as an adjunct to, not a replacement for, comprehensive clinical judgment.
Keywords:
Acute subdural hematoma Mortality Prognosis RASH score Traumatic brain injury *Corresponding author:
Irwansyah E-mail address:
irwansyahbondan@gmail.
All authors have reviewed and approved the Anal version of the manuscript.
https://doi.
org/10.
37275/sjs.
Introduction the dura mater and the arachnoid materAistands as Traumatic brain injury (TBI) represents a global one of the most lethal entities encountered in public health catastrophe and a leading cause of neurosurgical practice.
Typically caused by the mortality and profound long-term disability, aptly rupture of cortical bridging veins due to sudden termed a "silent epidemic".
1 Its incidence continues to acceleration-deceleration forces.
ASDH is frequently rise, particularly in low- and middle-income countries, associated with severe underlying parenchymal brain where traffic accidents and violence are endemic.
injury, including contusions and diffuse axonal Among TBI Consequently, umbrella, acute subdural hematoma (ASDH)Aithe accumulation of blood in the potential space between neurocritical care, and intracranial pressure (ICP) management, the postoperative mortality rate for an influx of calcium into neurons.
This calcium ASDH remains distressingly high, with contemporary overload activates catabolic enzymes, promotes free literature consistently reporting figures between 30% radical formation, and ultimately leads to neuronal .
Neuroinflammation and Oxidative Stress:
persistent and critical need for accurate, early, and The brain mounts an intense inflammatory response reliable risk stratification tools to guide clinical to the injury, involving the activation of microglia and This astrocytes and the infiltration of peripheral immune While initially protective, this response can communication with patients' families in the face of become dysregulated, releasing cytotoxic cytokines devastating injury.
and reactive oxygen species that cause widespread The pathophysiology of ASDH is a devastating two- oxidative damage to lipids, proteins, and DNA, further act process, comprising a primary and a secondary perpetuating cell death.
The extent and severity of this injury phase that ultimately dictates patient outcome.
secondary injury cascade are often the primary The primary injury is the instantaneous, direct mechanical damage inflicted at the moment of impact.
outcome and survival.
Therefore, a cornerstone of This includes the tearing of blood vessels, formation of modern neurotrauma care is the prediction and the hematoma, and direct contusion of the brain mitigation of these secondary insults.
4 The hematoma itself acts as a rapidly For decades, prognostication in ASDH has relied on expanding intracranial mass lesion.
Within the fixed a handful of key clinical and radiological parameters.
volume of the skull, this leads to a precipitous rise in The Glasgow Coma Scale (GCS) has been the bedrock ICP, causing compression of adjacent brain tissue and distortion of vital deep-brain structures, including the indispensable measure of the level of consciousness.
Pupillary examination serves as a critical bedside However, the more insidious and often more indicator of brainstem compression, particularly of the damaging phase is the secondary injury.
This is a third cranial nerve.
On computed tomography (CT), complex and self-propagating cascade of deleterious measurements such as hematoma thickness and the biochemical, cellular, and molecular events triggered degree of midline shift (MLS) provide a direct by the primary insult and the subsequent mass effect.
anatomical quantification of the mass effect.
This cascade includes: .
Cerebral Edema: Both While invaluable, each of these factors in isolation cytotoxic .
ellular swelling due to ion pump failur.
provides an incomplete picture.
The GCS is a snapshot and vasogenic .
reakdown of the blood-brain barrie.
edema contribute to a further increase in intracranial intoxication, or intubation.
7 Radiological markers, volume and ICP.
Cerebral Ischemia: Elevated ICP while anatomically precise, do not capture the reduces the cerebral perfusion pressure (CPP), defined dynamic physiological state of the brain.
Relying on as Mean Arterial Pressure (MAP) minus ICP.
When CPP any single factor is insufficient to encapsulate the falls below critical thresholds, cerebral blood flow multifaceted nature of the injury and predict the (CBF) is compromised, leading to widespread ischemic complex interplay of pathophysiological processes that damage and creating a vicious cycle of swelling, determine survival.
8 This has driven the development further ICP elevation, and worsening ischemia.
of integrated clinical prediction scores, which aim to Excitotoxicity: The ischemic and traumatic insult provide a more holistic, objective, and reliable prognosis by combining multiple weighted risk factors neurotransmitters, primarily glutamate.
This over- into a single composite value.
An ideal prognostic tool activates receptors like the NMDA receptor, leading to for the acute setting should be simple to use, rapid to calculate, based on universally available data, and key surgical variables, namely the type of surgery and the timing of intervention.
The novelty of this work lies not only in its unique geographic and ethnic context The Richmond Acute Subdural Hematoma (RASH) but also in its critical appraisal of the score's score was specifically developed to meet this need.
is an intuitive grading scale based on five preoperative framework of surgical decision-making.
real-world variables that are readily available upon patient admission: Age.
GCS severity.
Pupillary response.
Midline shift >5 mm, and post-traumatic Loss of Methods This study was a retrospective, single-center.
Consciousness (LOC).
10 Each of these components reflects a distinct and critical aspect of the injury's Mohammad Hoesin General Hospital in Palembang.
In its original derivation and internal South Sumatra.
Indonesia.
This institution is a validation on a large cohort of over 2,500 patients in national tertiary referral hospital and serves as the the United States, the RASH score demonstrated a primary neurotrauma center for the region.
The study strong, linear correlation with postoperative mortality, offering a rapid and powerful estimate of risk.
The Institutional Review Board and Ethics Committee.
score's greatest strength is its simplicity, allowing for Given the retrospective nature of the data collection immediate calculation in high-pressure emergency from existing medical records, the requirement for individual informed consent was waived.
The study Dr.
was conducted in strict accordance with the ethical However, the utility and generalizability of any principles of the Declaration of Helsinki.
clinical prediction rule are contingent upon its We retrospectively screened the medical and performance in populations different from the one in radiological records of all patients admitted with a which it was developedAia process known as external Factors such as genetic background.
January 1st, 2022, and December 31st, 2024.
A total of healthcare system infrastructure, pre-hospital care 95 patients with traumatic ASDH were identified during this period.
The final study cohort was significantly across geographic regions, potentially composed of patients who met specific inclusion and altering a score's predictive accuracy.
To date, the exclusion criteria.
The inclusion criteria were as RASH score has not been validated in a Southeast follows: .
age Ou18 years.
a definitive diagnosis of Asian traumatic ASDH confirmed by a non-contrast head CT management of TBI may differ, representing a critical and .
underwent surgical evacuation of the knowledge gap.
Clinicians in these regions lack the ASDH, evidence to confidently apply this tool in their local The exclusion criteria were: .
ASDH
from non-traumatic causes, such as spontaneous or Therefore, this study was designed to address this post-procedural ASDH The primary aim was to perform the first external polytrauma with hemodynamic instability stemming validation of the RASH score in a cohort of Indonesian from severe thoracic, abdominal, or pelvic injuries that patients undergoing surgical evacuation for traumatic could independently be the primary cause of mortality.
ASDH.
The secondary aims were to determine the and .
Incomplete or missing medical records that score's predictive accuracy for in-hospital mortality, precluded the calculation of the full RASH score or identify an optimal cut-off for risk stratification in this determination of the primary outcome.
population, and explore the prognostic significance of Of the 95 patients initially screened, 28 were not head CT scan as the maximum perpendicular included in the final analysis.
Ten patients were displacement of the septum pellucidum from the managed non-operatively due to small hematoma size anatomical midline.
This was dichotomized as: MLS O5 and stable neurological status.
Eighteen patients were mm .
and MLS > 5 mm .
Loss of Consciousness non-survivable (LOC):
The admission and were managed with palliative care after documented LOC at the time of injury, as reported by discussion with their families, thus were not surgical the patient, family, or emergency responders, was This left 71 patients who underwent recorded as a binary variable: No .
and Yes .
Surgical and outcome variables were defined.
subsequently excluded due to incomplete admission .
Primary Outcome: The primary outcome variable Thus, a final cohort of 67 patients was was in-hospital mortality, defined as death from any included in the analysis.
This selection process is cause during the index hospitalization for the ASDH.
critical to note, as the study cohort inherently .
)Type of Surgical Procedure: The primary surgical represents a filtered subset of all ASDH patients, procedure was categorized as either Craniotomy with specifically those for whom surgery was deemed both hematoma evacuation or Decompressive Craniectomy necessary and potentially beneficial.
(DC) with hematoma evacuation.
Time from Four A standardized data collection form was developed Injury to Operation: This was calculated in hours from to extract all relevant information from the hospital's the documented time of injury to the time of skin electronic and paper-based medical records.
To ensure incision for surgery.
For each of the 67 patients, the data integrity and minimize extraction errors, two total RASH score was calculated by summing the investigators independently extracted the data for all points assigned to each of the five components.
The Any total score ranges from a minimum of 0 to a maximum of 8, with higher scores indicating a greater predicted All patient identifiers were anonymized to maintain strict confidentiality.
risk of mortality.
All statistical analyses were conducted using IBM RASH score components consist of five variables SPSS Statistics for Windows.
Version 28.
0 (Armonk, were collected based on the patient's clinical and NY: IBM Cor.
A two-sided p-value of <0.
05 was radiological status at the time of initial presentation to the emergency department.
Age: Recorded in years Continuous variables were assessed for normality and categorized for scoring: O59 years .
, 60Ae using the Shapiro-Wilk test.
Normally distributed data 79 years .
, and Ou80 years .
were reported as mean and standard deviation Glasgow Coma Scale (GCS) Severity: The initial GCS (SD), while non-normally distributed data (Time to score was recorded and categorized: Mild (GCS 14Ae15.
Surgery.
RASH scor.
were reported as median and 0 point.
Moderate (GCS 9Ae13, 1 poin.
, and Severe interquartile range (IQR).
Categorical variables were (GCS O8, 2 point.
For patients intubated prior to a presented as frequencies and percentages .
, %).
formal GCS assessment, the score was estimated from identify factors associated with in-hospital mortality, motor and eye components or from documented scores by emergency responders before intubation.
characteristics were compared between the survivor Pupillary and non-survivor groups.
The Chi-square test or documented for both pupils and categorized: both FisherAos exact test was used for categorical variables pupils reactive .
, unilateral non-reactive pupil as appropriate.
The non-parametric Mann-Whitney U .
, and bilateral non-reactive pupils .
test was used to compare the median RASH scores and .
Midline Shift (MLS): Measured on the initial axial median Time to Surgery between the two outcome Response:
Reactivity groups.
The (GCS 9Ae.
3% had a severe TBI (GCS O.
discriminatory power of the RASH score to predict Pupillary abnormalities were present in 17.
9% of mortality using Receiver Operating Characteristic patients, with a concerning 13.
4% having bilateral (ROC) curve analysis.
The Area Under the ROC Curve non-reactive pupils.
A midline shift exceeding 5 mm (AUC) and its 95% confidence interval (CI) were was observed in 46.
3% of patients, and the vast calculated to quantify the overall predictive accuracy.
AUC
consciousness at the time of injury.
The median time outstanding, 0.
80Ae0.
89 = excellent, 0.
70Ae0.
from injury to the start of surgery was 6.
2 hours (IQR acceptable, and <0.
70 = poor.
5Ae9.
8 hour.
Regarding the surgical procedure, 41 90Ae1.
The optimal cut-off point for the RASH score that .
2%) underwent craniotomy, while 26 best-differentiated survivors from non-survivors was determined using YoudenAos J index (J = Sensitivity decompressive craniectomy.
The overall in-hospital Specificity Ae .
For this optimal threshold, we mortality rate for this surgical cohort was 20.
9% .
of 67 patient.
Positive .
Predictive Value (PPV), and Negative Predictive Value Table 2 presents a comparison of preoperative and (NPV).
To assess the independent predictive value of surgical factors between patients who survived and the RASH score and other risk factors, two separate those who died.
The median RASH score was significantly higher in the non-survivor group (Median constructed to avoid multicollinearity.
Model 1:
= 5.
IQR = 4.
0Ae6.
compared to the survivor group Assessed the RASH score as a continuous variable.
(Median = 3.
IQR = 2.
0Ae4.
, a difference that was Model 2: Assessed individual demographic, clinical, highly statistically significant (Mann-Whitney U = and surgical variables identified as significant or near- significant in the bivariate analysis.
Due to the components, the presence of a midline shift > 5 mm observation of quasi-complete separation in the showed the strongest association with mortality.
bivariate analysis .
here one predictor perfectly Strikingly, all 14 patients .
%) who died had a predicted the outcome in a subgrou.
, a penalized midline shift > 5 mm, compared to only 17 of 53 logistic regression model using Firth's method was survivors .
1%) .
<0.
This finding represents a employed to ensure model stability and generate case of quasi-complete separation in the data.
While reliable odds ratio estimates.
Results were expressed other factors, such as severe GCS .
0% mortalit.
as Adjusted Odds Ratios .
OR) with corresponding and bilateral non-reactive pupils .
4% mortalit.
95% CIs.
were more common in the non-survivor group, these p<0.
Among RASH differences did not achieve statistical significance.
Results likely due to the limited number of events .
ow A final cohort of 67 patients who underwent statistical powe.
in this modest sample size.
Crucially, surgery for traumatic ASDH was included in the the analysis of surgical factors revealed significant The baseline demographic, clinical, and associations with mortality.
A significantly higher surgical characteristics of the cohort are detailed in proportion of non-survivors underwent decompressive Table 1.
The population was predominantly male craniectomy .
4%) compared to survivors .
6%), with a mean age of 52.
4 A 15.
8 years.
The .
=0.
Furthermore, the median time from injury most common mechanism of injury was motor vehicle to surgery was significantly longer in patients who died accidents .
2%), followed by falls .
9%).
The .
5 hour.
compared to those who survived .
neurological status upon presentation was generally hour.
=0.
poor: 46.
3% of patients presented with a moderate TBI Table 1.
Baseline demographics, clinical, and surgical characteristics of the study cohort (N=.
Notes: Data are presented as n (%), mean A standard deviation (SD), or median (Interquartile Rang.
GCS.
Glasgow Coma Scale.
RASH.
Richmond Acute Subdural Hematoma.
Table 2.
Bivariate analysis of preoperative and surgical factors by in-hospital mortality status.
Notes: Values are n (%) or median (Interquartile Rang.
P-values are from Chi-square/Fisher's exact test for categorical variables and Mann-Whitney U test for continuous variables.
Significant p-values are in bold.
The ability of the total RASH score to discriminate metrics: .
Sensitivity: 78.
6% .
of 14 non-survivors between survivors and non-survivors was evaluated were correctly identified as high ris.
Specificity:
4% .
of 53 survivors were correctly identified as low ris.
Positive Predictive Value (PPV): 47.
8% (Of yielding an Area Under the ROC Curve (AUC) of 0.
23 patients with a score Ou5, 11 die.
Negative .
% CI: 0.
715Ae0.
, which was highly statistically Predictive Value (NPV): 93.
2% (Of 44 patients with a significant .
<0.
Analysis of the ROC curve score <5, 41 survive.
The distribution of patients identified an optimal cut-off value based on the according to this cut-off score and their mortality maximal YoudenAos J index.
For clinical practicality, outcome is detailed in Table 3.
The high NPV is this corresponds to a RASH score of 5 or greater to particularly noteworthy, indicating that a score below classify a patient as "high risk.
" At this threshold, the 5 is strongly associated with survival in this cohort.
ROC
(Figure The RASH score demonstrated the following performance Figure 1.
Receiver Operating Characteristic (ROC) Curve for the RASH Score in Predicting In-Hospital Mortality.
The diagonal line represents a test with no discriminatory ability (AUC = 0.
The curve for the RASH score shows excellent separation from the line of no-discrimination.
To further explore the independent predictors of time to surgery, the RASH score remained a powerful mortality, two separate penalized multivariable logistic independent predictor of mortality.
For every one-point regression models were constructed (Table .
In Model increase in the RASH score, the odds of in-hospital 1, the RASH score was analyzed as a continuous mortality increased by nearly threefold (Adjusted OR = After adjusting for the type of surgery and 95, 95% CI: 1.
52Ae5.
74, p=0.
In Model 2, individual clinical and surgical factors were assessed.
This performance is remarkably consistent with that Severe GCS (O.
and the presence of bilateral non- reported in the original U.
cohort (AUC = 0.
reactive pupils remained strong predictors of a fatal lending strong support to its external validity.
Notably, both decompressive craniectomy However, the interpretation of this finding must be and a longer time to surgery were also confirmed as heavily qualified by the clinical and methodological independent predictors of mortality after adjusting for context of our study, particularly the profound other factors.
The wide confidence intervals for many selection bias inherent in evaluating a surgical-only uncertainty due to the modest sample size and low number of events.
The strength of the RASH score lies in its elegant integration of five domains of neurotrauma prognosis, each reflecting a distinct facet of the underlying Discussion pathophysiology (Figure .
13 Our analysis, now This study represents the first external validation of the Richmond Acute Subdural Hematoma (RASH) exploration of these correlates.
Age and Cerebral score in a Southeast Asian population, providing Reserve: Age is a well-established, non-modifiable risk critical new evidence on its utility outside of its original factor in TBI.
The poorer outcomes observed in older derivation cohort.
11 Our principal finding is that the adults are multifactorial.
Physiologically, aging is RASH score serves as a simple and robust predictor of associated with cerebral atrophy, which increases in-hospital mortality for Indonesian patients selected tension on bridging veins, predisposing them to for surgical management of ASDH, demonstrating excellent discriminatory power with an AUC of 0.
Table 3.
2x2 contingency table for the RASH score cut-off of Ou5.
Table 4.
Multivariable logistic regression analysis of predictors for in-hospital mortality.
Abbreviation: aOR, adjusted odds ratio.
CI, confidence interval.
GCS.
Glasgow Coma Scale.
Notes: Reference categories are in parentheses.
AEstimate derived from penalized regression due to quasi-complete separation.
More critically, aging diminishes cerebral reserveAi the severity of the primary impact but also the the brainAos intrinsic capacity to withstand and recover profound effect of raised ICP on the ascending reticular from injury by recruiting alternative neural pathways activating system (ARAS), the brainstem network and mounting effective cellular repair mechanisms.
governing consciousness.
15 Direct compression or This is compounded by immunosenescence, a state of secondary ischemic injury to the ARAS results in a age-related immune dysregulation that can lead to a depressed level of consciousness that is strongly and more pronounced and damaging neuroinflammatory response, and a higher prevalence of comorbidities Response that limit the physiological tolerance to the extreme examination is arguably the most critical bedside sign stress of major surgery and critical illness.
GCS and of impending transtentorial herniation.
A fixed and Consciousness: The Glasgow Coma Scale provides a dilated pupil .
is the classic sign of vital functional assessment of the cerebral cortex and compression of the third cranial nerve .
culomotor A low GCS score in ASDH reflects not only nerv.
as the uncinate gyrus of the temporal lobe is Brain Pupillary Herniation:
The forced captured by the preoperative RASH score alone.
supratentorial mass.
The parasympathetic fibers that Time to Surgery and the Irreversible Cascade: The control pupillary constriction run along the exterior of finding that a longer time to surgery independently predicts mortality is of paramount clinical importance.
Every hour of delay allows the secondary injury advanced brainstem compression and is often a pre- cascade to progress unchecked.
Sustained high ICP terminal sign, reflecting irreversible damage to the leads to worsening ischemia, excitotoxicity, and Our multivariable model confirmed its Timely surgical decompression is the status as a powerful independent predictor.
Midline single most effective intervention to halt this vicious Shift and Anatomical Disruption: Radiographically, cycle by reducing intracranial volume and restoring midline shift is a direct, quantifiable measure of the anatomical severity of the mass effect.
An MLS > 5 mm neurosurgical axiom that for ASDH, "time is brain," signifies a substantial intracranial pressure gradient, and delays in intervention can negate the potential leading to a cascade of mechanical failures.
This shift benefits of an otherwise successful operation.
Bilateral non-reactivity Our not only compresses the ipsilateral hemisphere but The true clinical power of the RASH score in our also distorts the brainstem, obstructs CSF pathways study is its exceptionally high negative predictive value leading to hydrocephalus, and can cause kinking of (NPV) of 93.
In a chaotic emergency setting, a low perforating arteries, resulting in secondary ischemic score (<.
provides a valuable, data-driven measure of strokes in the brainstem and thalamus.
In our cohort, reassurance for both the clinical team and the this was the single most powerful predictor in the patient's family.
18 It can help frame counseling by bivariate analysis, with 100% mortality among those suggesting a high probability (>90%) of survival, which exhibiting this finding.
While potentially a statistical is invaluable for managing expectations and building artifact of our small sample, this "quasi-complete a therapeutic alliance.
19 Conversely, the modest positive predictive value (PPV) of 47.
8% requires nuanced interpretation.
It is critical to understand represent a point of near-irreversible injury in this that a high score is not a definitive death sentence.
Type of Surgery and Intracranial Rather, it identifies a patient at extremely high risk Hypertension: Our analysis introduced the type of who requires the most aggressive monitoring and A PPV of 47.
8% means that even among decompressive craniectomy (DC) is an independent these high-risk patients, more than half survive.
predictor of mortality is not surprising.
A DC is not a Therefore, a high RASH score should be interpreted as cause of poor outcome, but rather a marker of it.
16 The a call to action, not a justification for futility or decision to perform a DC instead of a simple withdrawal of care.
It could be used to triage patients craniotomy is made when the surgeon identifies to higher-level ICU beds, justify the placement of intractable brain swelling and anticipates dangerously invasive ICP monitors, or trigger more advanced high postoperative ICP.
It is a salvage maneuver for the neuromonitoring protocols.
This score should always most severe end of the injury spectrum, where the be used as an adjunct to, and never a replacement for, secondary injury cascade has already produced holistic clinical judgment and a thorough discussion profound cerebral edema.
Thus, the need for DC of goals of care with the family.
The reflects an underlying injury severity that is not fully Figure 2.
Pathophysiological correlates of the rash score and surgical factors.
This study must be interpreted within the context sick for surger.
were excluded.
This means our of its significant limitations.
The most profound of findings are only applicable to the intermediate group these is the selection bias inherent in its retrospective, of patients for whom the decision to operate was made.
surgical-only design.
Our cohort of 67 patients was This selection process likely explains our relatively low selected from a larger pool of 95 ASDH patients.
Those mortality rate of 20.
9% compared to the 30-70% cited managed non-operatively .
ue to being too well or too in broader literature.
our cohort was pre-selected to exclude the most futile cases.
This bias may also stratification, facilitate more objective prognostication, artificially inflate the score's performance, as it was and improve communication with families during a tested on the very group with the most prognostic critical phase of care.
However, its use must be Secondly, our modest sample size (N=.
tempered by a clear understanding of the significant and low number of mortality events .
limit our selection bias inherent in a surgical cohort.
It shows statistical power.
This is evident in the wide confidence promise as a valuable adjunct to clinical judgment in this region, pending further validation in larger, multi- substantial uncertainty.
The low events-per-variable center prospective studies.
ratio in our regression model means these results should be considered exploratory.
The non-significant findings for some variables in the bivariate analysis.
References