Risk Scoring in Perioperative and Surgical Intensive Care Patients: A Review

Ability to predict surgical outcomes by surgical Apgar score: a systematic review

BACKGROUND

The Surgical Apgar score (SAS) is a straightforward and unbiased measure to assess the probability of experiencing complications after surgery. It is calculated upon completion of the surgical procedure and provides valuable predictive information. The SAS evaluates three specific factors during surgery: the estimated amount of blood loss (EBL), the lowest recorded mean arterial pressure (MAP), and the lowest heart rate (LHR) observed. Considering these factors, the SAS offers insights into the probability of encountering postoperative complications.

METHODS

Three authors independently searched the Medline, PubMed, Web of Science, Scopus, and Embase databases until June 2022. This search was conducted without any language or timeframe restrictions, and it aimed to cover relevant literature on the subject. The inclusion criteria were the correlation between SAS and any modified/adjusted SAS (m SAS, (Modified SAS). eSAS, M eSAS, and SASA), and complications before, during, and after surgeries. Nevertheless, the study excluded letters to the editor, reviews, and case reports. Additionally, the researchers employed Begg and Egger's regression model to evaluate publication bias.

RESULTS

In this systematic study, a total of 78 studies \were examined. The findings exposed that SAS was effective in anticipating short-term complications and served as factor for a long-term prognostic following multiple surgeries. While the SAS has been validated across various surgical subspecialties, based on the available evidence, the algorithm's modifications may be necessary to enhance its predictive accuracy within each specific subspecialty.

CONCLUSIONS

The SAS enables surgeons and anesthesiologists to recognize patients at a higher risk for certain complications or adverse events. By either modifying the SAS (Modified SAS) or combining it with ASA criteria, healthcare professionals can enhance their ability to identify patients who require continuous observation and follow-up as they go through the postoperative period. This approach would improve the accuracy of identifying individuals at risk and ensure appropriate measures to provide necessary care and support.

Detecting Major Complications and Death After Emergency Abdominal Surgery Using the Surgical Apgar Score: A Retrospective Analysis in a Caribbean Setting

Objective

The Surgical Apgar Score (SAS) is a simple 10-point scoring system that has been shown to be predictive of major postoperative complications and death after surgery. We evaluated the predictive ability of this score in a cohort of patients undergoing emergency abdominal surgery in a Caribbean tertiary hospital.

Methods

The SAS was calculated retrospectively from the anaesthesia records of all patients undergoing emergency abdominal surgery during a 12-month period. The postoperative surgical records of these patients were then examined for the presence of major complications and death. The association between the SAS and outcomes was tested using binary logistic regression, and the SAS discriminatory ability was determined from the receiver-operating curve (ROC) analysis.

Results

Of the 220 patients studied, 72 (33%) suffered an in-hospital major complication or death. The highest complication rate occurred in the low-scoring groups, with 68% of those scoring <4 being affected. Low-scoring patients (<4) had four times the risk of major complications when compared to higher-scoring groups (relative risk [RR], 4.21; 95% confidence interval [CI], 2.5-7.3; p<0.001). The odds ratio (OR) for major complications or death per unit increase in the SAS was 0.58 (95% CI, 0.47-0.72; p<0.001). The c-statistic of the SAS for predicting major complications or death was 0.71 (95% CI, 0.68-0.73; p<0.0001).

Conclusion

The SAS is a simple 10-point score that can be used in patients undergoing emergency surgery in a Caribbean setting to help identify those that are at a higher risk of postoperative complications. Due to its ease in calculation, it can be added to other commonly used criteria to help triage the postoperative patient.

"Neo-PIRO": Introducing a Novel Grading System for Surgical Infections of Neonates

Introduction:

Quantification of surgical sepsis was never done beyond superficial, subfascial, and deep surgical site infection (SSI). Invasive surgical sepsis with systemic manifestation has not been tried to be quantified in general and pediatric surgery in particular. Hence, this attempts to develop a novel grading system to quantify neonatal surgical infections.

Materials and Methods:

Predisposing factors, infection, response, and organ failure (PIRO) is being used in critical care institutions for medical sepsis; it was modified with neonate-specific surgical parameters. Authors have developed a grading of these parameters into Grade I, II, and III.

Results:

A blinded statistical test was performed and results were put to test. Extended Mantel–Haenszel Chi-square test validated linear relationship with grade and outcome, hospital stay, deep SSI, and organ dysfunction. Analysis of variance also showed the significant relationship of changing trends in grade and outcome. (1) Higher the grade indicated the probability of death. (2) Grade I patients had less duration of hospital stay compared to Grade II and III (P = 0.04). (3) The requirement of organ support and SSI were also more in Grade III. (4) Grade I patients had less increase in trends compared to Grade II and III (F = 4.86). Authors therefore feel Neo-PIRO seems to be the first scoring system that shows a linear relationship between scores and grade.

Conclusion:

Neo-PIRO is a novel grading system with surgical neonate-specific parameters. Future versions to include molecular parameters, as well as parameters selected by regression analysis.

Incidence and characteristic analysis of in-hospital falls after anesthesia

Background

In-hospital falls may result in serious clinical adverse consequences, but the effects of anesthesia in the occurrence of postoperative falls are still undetermined. Anesthesia may theoretically cause postoperative falls due to the residual pharmacologic and neuromuscular blocking effects of anesthetics. We retrospectively reviewed events of in-hospital falls occurred after anesthesia management to identify the incidence and risk factors of postanesthesia falls.

Methods

We reviewed the postanesthesia visit of patients received anesthesia in the Hualien Buddhist Tzu Chi General Hospital from January 2009 to December 2013. Falls happened within 24 h after anesthesia were recorded. The Poisson regression model was used for simultaneous analysis of the association between incidence proportion of postanesthesia falls and the potential risk factors.

Results

A total of 60,796 inpatients received anesthesia management over the past 5 years, and ten patients fell within 24 h after anesthesia. All cases happened in the general wards. Falls occurred more often at the bedside, presence of caregivers, and during the daytime. Patients underwent regional anesthesia, and old age significantly increased the risk of postanesthesia falls, while differences in gender and ASA physical status did not affect the occurrence of postanesthesia falls.

Conclusions

The overall incidence proportion of postanesthesia falls is 1.6 cases per 10,000 patients (95 % CI 0.006 to 0.026 %) over a 24-h observation period. Falls are more commonly happened during the less expected periods after operation and are increased in the elderly and patients received regional anesthesia. This study highlights that more comprehensive clinical practice guidelines for postoperative care should be exercised to prevent the in-hospital falls.

Do Serially Recorded Prognostic Scores Predict Outcome Better Than One-Time Recorded Score on Admission? A Prospective Study in Adult Intensive Care Patients

OBJECTIVES

The prognosticating ability of one-time recorded Acute Physiology and Chronic Health Evaluation (APACHE) IV score was compared with serially recorded Mortality Prediction Model (MPM) II scores.

DESIGN AND METHODS

A prospective observational study was conducted for a period of 6 months. Acute Physiology and Chronic Health Evaluation IV score was recorded during the first day on intensive care unit (ICU) admission. Mortality Prediction Model II was recorded on admission, 24, 48, and 72 hours. Predicted mortality was compared with observed mortality. The systems were calibrated and tested for discriminant functions.

RESULTS

One hundred and fifty patients were studied. The observed mortality was 21.3%. The mean predicted hospital mortality by APACHE IV was 20.6%. The mean predicted hospital mortality rate by serial MPM II measurements was 27.7%, 24.3%, 25.5%, and 25.8%. The area under the receiver-operating characteristic curve was 0.87 for APACHE IV and 0.82, 0.84, 0.85, and 0.89 for MPM II series. Both systems calibrated well with similar degree of goodness of fit.

CONCLUSION

Acute Physiology and Chronic Health Evaluation IV on admission predicted hospital mortality better than serially recorded MPM, which overestimated mortality. Also, APACHE IV had a slightly better discrimination compared to MPM II on admission. One-time recording of APACHE IV on admission may be sufficient for prognostication of ICU patients rather than serial MPM scores.

Implementation and preliminary validation of a new score that predicts post-operative complications

BACKGROUND

An accurate pre-operative risk assessment could reduce morbidity and mortality for high-risk surgical patients. The aim of the study was to implement and preliminary validate a new score that could predict the occurrence of post-operative complications (PoCs): the Anesthesiological and Surgical Postoperative Risk Assessment (ASPRA) score.

METHODS

The ASPRA score was created through a literature's review; a score of 1-3 was given to each identified risk factor, according to its statistical correlation with PoC. ASPRA was retrospectively applied to a derivation set of 176 surgical patients. A receiver operating characteristic (ROC) analysis evaluated the discriminating ability of the score and cutoff value in predicting the occurrence of PoCs, according to the Clavien-Dindo classification of surgical complications. The statistical validation of the score and related cutoff values was prospectively ran within a validation set of 1928 surgical patients.

RESULTS

Through ROC analysis, an ASPRA score of 7 was chosen as the cutoff value in the derivation set. In the validation set, 65.3% of patients presented a PoC (Clavien ≥ 1). In this group, ROC analysis showed an area under the curve (AUC) of 0.72, and although potentially related to the high rate of complications a high positive predictive value of 87.0% has been observed. No significant differences were found in ROC-AUC, sensitivity, specificity, or positive or negative predictive value between the derivation and validation sets (P > 0.05).

CONCLUSION

The new ASPRA score has a high positive predictive value to predict the occurrence of PoCs. Further prospective studies are required to confirm these results.

Application of pediatric index of mortality version 2: score in pediatric intensive care unit in an African developing country

Introduction

Outcome of patients admitted to PICU can be evaluated by many illness severity scoring systems. This prospective observational study evaluated the outcome of patients admitted to PICU in Fayoum University hospital of a developing country using the pediatric index of mortality version 2 scoring system.

Methods

All patients included in this study were subjected to data collection including demographics, diagnoses at admission, duration of ICU stay (DOS), pediatric index of mortality version 2 (PIM2) score and hospital outcome. The ratio of observed to predicted mortality (standardized mortality ratio (SMR)) was calculated for the set of patients.

Results

The study included 205 patients. The main causes of admission were respiratory, cardiovascular and neurological illnesses. Patients stay in ICU ranged from 1 - 45 days with a median 6 (interquartile range (IQ): 3-9) days. Discriminatory function of PIM2 scoring system was acceptable with the area under the ROC curve 0.76 (95%CI: 0.60-0.91). PIM2 calibrated well using Hosmer Lemeshow analysis (H-L X2= 1.410, df= 8, p=0.9). The mean predicted mortality was 5.6 (95% CI: 3.43 - 7.91) and the observed mortality was 8.8% giving a SMR 1.55.

Conclusion

PIM2 scoring system show adequate discriminatory function and well calibrated for the case mix of patients in PICU of Fayoum, Egypt. It can be used as beneficial tool for evaluation of risk adjusted mortality. Further larger scale studies in cooperation with other Egyptian universities and neighboring countries can improve the performance of our PICUs and critical care services.

Surgical systems biology and personalized longitudinal phenotyping in critical care

Systems-wide molecular analysis of the metabolic, inflammatory and immune response to surgical trauma has yet to be translated into the operating room. Surgical patients are exposed to a large number of heterogeneous environmental insults that cannot only be quantified by genome-orientated 'omics platforms. Furthermore, surgery demands rapid or near real-time analysis. Systems-level metabolic phenotyping provides a novel 'global' perspective of an organism's metabolic response to surgical injury and, therefore, serves as an ideal platform for the development of personalized therapies in surgery. This article reviews current personalized approaches to healthcare in surgery and explores future directions for personalized surgical biomarker discovery and therapeutics. In particular, this article discusses our vision of 'personalized metabolic phenotyping' in surgery, and outlines next-generation technologies that will make this approach a reality.

Incidence of cardiac arrests and unexpected deaths in surgical patients before and after implementation of a rapid response system

BACKGROUND

Rapid response systems (RRSs) are considered an important tool for improving patient safety. We studied the effect of an RRS on the incidence of cardiac arrests and unexpected deaths.

METHODS

Retrospective before- after study in a university medical centre. We included 1376 surgical patients before (period 1) and 2410 patients after introduction of the RRS (period 2). Outcome measures were corrected for the baseline covariates age, gender and ASA.

RESULTS

The number of patients who experienced a cardiac arrest and/or who died unexpectedly decreased non significantly from 0.50% (7/1376) in period 1 to 0.25% (6/2410) in period 2 (odds ratio (OR) 0.43, CI 0.14-1.30). The individual number of cardiac arrests decreased non-significantly from 0.29% (4/1367) to 0.12% (3/2410) (OR 0.38, CI 0.09-1.73) and the number of unexpected deaths decreased non-significantly from 0.36% (5/1376) to 0.17% (4/2410) (OR 0.42, CI 0.11-1.59). In contrast, the number of unplanned ICU admissions increased from 2.47% (34/1376) in period 1 to 4.15% (100/2400) in period 2 (OR 1.66, CI 1.07-2.55). Median APACHE ll score at unplanned ICU admissions was 16 in period 1 versus 16 in period 2 (NS). Adherence to RRS procedures. Observed abnormal early warning scores ≤72 h preceding a cardiac arrest, unexpected death or an unplanned ICU admission increased from 65% (24/37 events) in period 1 to 91% (91/101 events) in period 2 (p < 0.001). Related ward physician interventions increased from 38% (9/24 events) to 89% (81/91 events) (p < 0.001). In period 2, ward physicians activated the medical emergency team in 65% of the events (59/91), although in 16% (15/91 events) activation was delayed for one or two days. The overall medical emergency team dose was 56/1000 admissions.

CONCLUSIONS

Introduction of an RRS resulted in a 50% reduction in cardiac arrest rates and/or unexpected death. However, this decrease was not statistically significant partly due to the low base-line incidence. Moreover, delayed activation due to the two-tiered medical emergency team activation procedure and suboptimal adherence of the ward staff to the RRS procedures may have further abated the positive results.

Risk factors for major peri-operative complications in adult spinal deformity surgery: a multi-center review of 953 consecutive patients

PURPOSE

Major peri-operative complications for adult spinal deformity (ASD) surgery remain common. However, risk factors have not been clearly defined. Our objective was to identify patient and surgical parameters that correlate with the development of major peri-operative complications with ASD surgery.

METHODS

This is a multi-center, retrospective, consecutive, case-control series of surgically treated ASD patients. All patients undergoing surgical treatment for ASD at eight centers were retrospectively reviewed. Each center identified 10 patients with major peri-operative complications. Randomization tables were used to select a comparably sized control group of patients operated during the same time period that they did not suffer major complications. The two groups were analyzed for differences in clinical and surgical factors. Analysis was restricted to non-instrumentation related complications.

RESULTS

At least one major complication occurred in 80 of 953 patients (8.4 %), including 72 patients with non-instrumentation related complications. There were no significant differences between the complications and control groups based on the demographics, ASA grade, co-morbidities, body mass index, prior surgeries, pre-operative anemia, smoking, operative time or ICU stay (p > 0.05). Hospital stay was significantly longer for the complications group (14.4 vs. 7.9 days, p = 0.001). The complications group had higher percentages of staged procedures (46 vs. 37 %, p = 0.011) and combined anterior-posterior approaches (56 vs. 32 %, p = 0.011) compared with the control group.

CONCLUSION

The major peri-operative complication rate was 8.4 % for 953 surgically treated ASD patients. Significantly higher rates of complications were associated with staged and combined anterior-posterior surgeries. None of the patient factors assessed were significantly associated with the occurrence of major peri-operative complications. Improved understanding of risk profiles and procedure-related parameters may be useful for patient counseling and efforts to reduce complication rates.

Safety and caregiver satisfaction with gastrostomy in patients with Ataxia Telangiectasia

Background

Ataxia Telangiectasia (A-T) is a rare monogenetic neurodegenerative disease with pulmonary, nutritional, and dysphagic complications. Gastrostomy tube (GT) feedings are commonly recommended to manage these co-morbidities. In general, outcomes of GT placement in patients with progressive diseases that develop during childhood are not well characterized. The primary purposes of this study were to determine whether GT placement in patients with A-T would be tolerated and associated with caregiver satisfaction.

Methods

We completed a retrospective review of 175 patients who visited the A-T Children's Center at Johns Hopkins Hospital from 2001 through 2008, and identified 28 patients with A-T (19 males, 9 females) who underwent GT placement for non-palliative reasons. Information was obtained from medical records, interviews with primary health care providers, and 24 (83%) caregivers of patients with GT's who responded to survey requests.

Results

Twenty-five (89%) patients tolerated GT placement and were a median of 5.0 (0.4-12.6) years post GT placement at the time of this investigation. Three (11%) patients died within one month of GT placement. In comparison to patients who tolerated GT placement, patients with early mortality were older when GT's were placed (median 24.9 vs. 12.3 years, p = 0.006) and had developed a combination of dysphagia, nutritional, and respiratory problems. Caregivers of patients tolerating GT placement reported significant improvements in mealtime satisfaction and participation in daily activities.

Conclusions

GT placement can be well tolerated and associated with easier mealtimes in patients with A-T when feeding tubes are placed at young ages. Patients with childhood onset of disorders with predictable progression of the disease process and impaired swallowing may benefit from early versus late placement of feeding tubes.

Validation of Pediatric Index of Mortality-2 scoring system in a pediatric intensive care unit, Barbados

This study evaluated the outcome of patients in a pediatric intensive care unit (ICU) of a developing country applying Pediatric Index of Mortality (PIM) version-2 scoring system. A total of 163 consecutive patients were prospectively studied. Data included demographics, diagnoses at admission, PIM-2 score, the duration of ICU stay and hospital outcome. Predicted mortality and standardized mortality ratio (SMR) were calculated. Respiratory and neurological illnesses were the main admission diagnoses. The mean length of stay was 5.4 [95% Confidence Intervals (CI): 4-6.9] days. The mean predicted mortality was 6.2% (95% CI: 4.3-8.1); the observed mortality rate was 5.5%, the SMR being 0.89. Hosmer-Lemeshow analysis calibrated PIM-2 for the case mix [χ(2) = 5.64 (df = 7), p = 0.58]. The area under the ROC curve was 0.82 (95% CI: 0.72-0.92) showing a good discriminant function. Performance of the pediatric ICU in Barbados is comparable to that of developed world by risk-adjusted outcome evaluation.

[Scoring system for evaluation of clinical outcome of severe injuries in patients]

A trauma-scoring system converts the severity of injury into a number, so helping clinicians to define patient's condition. Aim of our investigation was assessment of scoring systems in clinical outcome of patients with severe traumatic injury, as well as ISS, AIS, APACHE II and SOFA score were counted. Mean age of traumatized patients was 35 yrs, predominantly males. Based on results of our investigation we concluded that ISS, APACHE II and SOFA score adequately can be used for predicting clinical outcome of severe traumatized patients.

Influence of routine computed tomography on predicted survival from blunt thoracoabdominal trauma

Introduction

Many scoring systems have been proposed to predict the survival of trauma patients. This study was performed to evaluate the influence of routine thoracoabdominal computed tomography (CT) on the predicted survival according to the trauma injury severity score (TRISS).

Patients and methods

1,047 patients who had sustained a high-energy blunt trauma over a 3-year period were prospectively included in the study. All patients underwent physical examination, conventional radiography of the chest, thoracolumbar spine and pelvis, abdominal sonography, and routine thoracoabdominal CT. From this group with routine CT, we prospectively defined a selective CT (sub)group for cases with abnormal physical examination and/or conventional radiography and/or sonography. Type and extent of injuries were recorded for both the selective and the routine CT groups. Based on the injuries found by the two different CT algorithms, we calculated the injury severity scores (ISS) and predicted survivals according to the TRISS methodology for the routine and the selective CT algorithms.

Results

Based on injuries detected by the selective CT algorithm, the mean ISS was 14.6, resulting in a predicted mortality of 12.5%. Because additional injuries were found by the routine CT algorithm, the mean ISS increased to 16.9, resulting in a predicted mortality of 13.7%. The actual observed mortality was 5.4%.

Conclusion

Routine thoracoabdominal CT in high-energy blunt trauma patients reveals more injuries than a selective CT algorithm, resulting in a higher ISS. According to the TRISS, this results in higher predicted mortalities. Observed mortality, however, was significantly lower than predicted. The predicted survival according to MTOS seems to underestimate the actual survival when routine CT is used.

A review of risk scoring systems utilised in patients undergoing gastrointestinal surgery

INTRODUCTION

Adequate stratification and scoring of risk is essential to optimise clinical practice; the ability to predict operative mortality and morbidity is important. This review aims to outline the essential elements of available risk scoring systems in patients undergoing gastrointestinal surgery and their differences in order to enable effective utilisation.

METHODS

The English literature was searched over the last 50 years to provide an overview of systems pertaining to the adult surgical patient.

DISCUSSION

Scoring systems can provide objectivity and mortality prediction enabling communication and understanding of severity of illness. Incorporating subjective factors within scoring systems can allow clinicians to apply their experience and understanding of the situation to an individual but are not reproducible. Limitations relating to obtaining variables, calculating predicted mortality and applicability were present in most systems. Over time scoring systems have become out-dated which may reflect continuing improvement in care. APACHE II shows the importance of reproducibility and comparability particularly when assessing critically ill patients. Both NSQIP in the USA and P-POSSUM in the UK seem to have many benefits which derive from their comprehensive dataset. The "Surgical Apgar" score offers relatively objective criteria which contrasts against the subjective nature of the ASA score.

CONCLUSION

P-POSSUM and NSQIP are comprehensive but are difficult to calculate. In the search for a simple and easy to calculate score, the "Surgical Apgar" score may be a potential answer. However, more studies need to be performed before it becomes as widely taken up as APACHE II, NSQIP and P-POSSUM.

Model for end-stage liver disease predicts mortality for tricuspid valve surgery

BACKGROUND

Patients undergoing tricuspid valve surgery have a mortality of 9.8%, which is higher than expected given the complexity of the procedure. Despite liver dysfunction seen in many patients with tricuspid disease, no existing risk model accounts for this. The Model for End-Stage Liver Disease (MELD) score accurately predicts mortality for abdominal surgery. The objective of this study was to determine if MELD could accurately predict mortality after tricuspid valve surgery and compare it to existing risk models.

METHODS

From 1994 to 2008, 168 patients (mean age, 61 +/- 14 years; male = 72, female = 96) underwent tricuspid repair (n = 156) or replacement (n = 12). Concomitant operations were performed in 87% (146 of 168). Patients with history of cirrhosis or MELD score 15 or greater (MELD = 3.8*LN [total bilirubin] + 11.2*log normal [international normalized ratio] + 9.6*log normal [creatinine] + 6.4) were compared with patients without liver disease or MELD score less than 15. Preoperative risk, intraoperative findings, and complications including operative mortality were evaluated. Statistical analyses were performed using chi(2), Fisher's exact test, and area under the curve (AUC) analyses.

RESULTS

Patients with a history of liver disease or MELD score of 15 or greater had significantly higher mortality (18.9% [7 of 37] versus 6.1% [8 of 131], p = 0.024). To further characterize the effect of MELD, patients were stratified by MELD alone. No major differences in demographics or operation were identified between groups. Mortality increased as MELD score increased, especially when MELD score of 15 or greater (p = 0.0015). A MELD score less than 10, 10 to 14.9, 15 to 19.9, and more than 20 was associated with operative mortality of 1.9%, 6.8%, 27.3%, and 30.8%, respectively. By multivariate analysis, MELD score of 15 or greater remained strongly associated with mortality (p = 0.0021). The MELD score predicted mortality (AUC = 0.78) as well as the European System for Cardiac Operative Risk Evaluation logistic risk calculator (AUC = 0.78, p = 0.96).

CONCLUSIONS

The MELD score predicts mortality in patients undergoing tricuspid valve surgery and offers a simple and effective method of risk stratification in these patients.

Perioperative Mortality Risk Score using Pre- and Post-operative Risk Factors in Older Patients

We developed a risk score for 30-day postoperative mortality: the Perioperative Mortality risk score. We used a derivation cohort from a previous study of surgical patients aged 70 years or more at three large metropolitan teaching hospitals, using the significant risk factors for 30-day mortality from multivariate analysis. We summed the risk score for each of six factors creating an overall Perioperative Mortality score. We included 1012 patients and the 30-day mortality was 6%.

The three preoperative factors and risk scores were (“three A's”): 1) age, years: 70 to 79=1, 80 to 89=3, 90+=6; 2) ASA physical status: ASA I or II=0, ASA III=3, ASA IV=6, ASA V=15; and 3) preoperative albumin <30 g/l=2.5.

The three postoperative factors and risk scores were (“three I's”) 1) unplanned intensive care unit admission =4.0; 2) systemic inflammation =3; and 3) acute renal impairment=2.5. Scores and mortality were: <5=1%, 5 to 9.5=7% and ≥10=26%.

We also used a preliminary validation cohort of 256 patients from a regional hospital. The area under the receiver operating characteristic curve (C-statistic) for the derivation cohort was 0.80 (95% CI 0.74 to 0.86) similar to the validation C-statistic: 0.79 (95% CI 0.70 to 0.88), P=0.88. The Hosmer-Lemeshow test (P=0.35) indicated good calibration in the validation cohort. The Perioperative Mortality score is straightforward and may assist progressive risk assessment and management during the perioperative period. Risk associated with surgical complexity and urgency could be added to this baseline patient factor Perioperative Mortality score.

Is the European System for Cardiac Operative Risk Evaluation model valid for estimating the operative risk of patients considered for percutaneous aortic valve replacement?

OBJECTIVE

The European System for Cardiac Operative Risk Evaluation has been used to define a particularly high-risk group of patients for aortic valve replacement in whom alternative procedures, such as stent-mounted percutaneous valve procedures, may be appropriate. Our objective was to assess the validity of this risk assessment at a large-volume, tertiary cardiac surgical center.

METHODS

From January 1, 2000, to December 30, 2006, a total of 1177 patients underwent isolated aortic valve replacement at the Mayo Clinic. Patient and operative demographics were recorded in a prospective database. Early mortality (< or = 30 days) was obtained. Additive and logistic European System for Cardiac Operative Risk Evaluations were calculated for each patient.

RESULTS

The mean patient age was 68.0 years (+/-14.7 years) at the time of surgery, and 36.8% were female. Variables used in the calculation of the European System for Cardiac Operative Risk Evaluation included chronic lung disease (15% of our cohort), extracardiac arteriopathy (13.8%), neurologic dysfunction (0.2%), previous cardiac surgery (23.2%), renal failure (6.5%), active endocarditis (3.1%), recent myocardial infarction (1.1%), unstable angina (0.1%), and severe pulmonary hypertension (6.5%). The ejection fraction was severely reduced (< or = 30%) in 4.9% of patients and moderately reduced (< or = 50%) in 12.7% of patients. One percent of patients were in a critical state, and operation was performed urgently in 3.4% of patients. Although mean mortality estimates were 6.9% +/- 3.4% (additive European System for Cardiac Operative Risk Evaluation) and 10.9% +/- 12.7% (logistic European System for Cardiac Operative Risk Evaluation), actual overall operative mortality in our patients was 2.5%. Additive and logistic European System for Cardiac Operative Risk Evaluations overestimated operative mortality in low, intermediate, and high-risk subgroups by up to 17.8%.

CONCLUSIONS

The European System for Cardiac Operative Risk Evaluation should not be used to determine the operability of patients for isolated aortic valve replacement. Elevated European System for Cardiac Operative Risk Evaluations alone do not appropriately define a population for use of a percutaneous aortic valve.

Predicting ICU survival: a meta-level approach

Background

The performance of separate Intensive Care Unit (ICU) status scoring systems vis-à-vis prediction of outcome is not satisfactory. Computer-based predictive modeling techniques may yield good results but their performance has seldom been extensively compared to that of other mature or emerging predictive models. The objective of the present study was twofold: to propose a prototype meta-level predicting approach concerning Intensive Care Unit (ICU) survival and to evaluate the effectiveness of typical mining models in this context.

Methods

Data on 158 men and 46 women, were used retrospectively (75% of the patients survived). We used Glasgow Coma Scale (GCS), Acute Physiology And Chronic Health Evaluation II (APACHE II), Sequential Organ Failure Assessment (SOFA) and Injury Severity Score (ISS) values to structure a decision tree (DTM), a neural network (NNM) and a logistic regression (LRM) model and we evaluated the assessment indicators implementing Receiver Operating Characteristics (ROC) plot analysis.

Results

Our findings indicate that regarding the assessment of indicators' capacity there are specific discrete limits that should be taken into account. The Az score ± SE was 0.8773± 0.0376 for the DTM, 0.8061± 0.0427 for the NNM and 0.8204± 0.0376 for the LRM, suggesting that the proposed DTM achieved a near optimal Az score.

Conclusion

The predicting processes of ICU survival may go "one step forward", by using classic composite assessment indicators as variables.

Evaluation of trauma care applying TRISS methodology in a Caribbean developing country

There have been conflicting reports regarding the applicability of Trauma Injury Severity Score (TRISS) methodology to evaluate trauma care in a developing country setting. The objective of this study was to apply TRISS methodology to evaluate trauma care in the public hospitals of a Caribbean developing country. A prospective, observational study was conducted in the three major general hospitals in Trinidad. Major trauma patients were included. Demographic data, waiting time in the Emergency Department, and nature of injury (blunt or penetrating) were noted. Revised Trauma Score, Injury Severity Score, and Glasgow Coma Scale were applied to all patients on admission. Hospital outcomes were noted. Predicted outcomes were calculated for adult patients using TRISS methodology. M, Z statistics and receiver operating characteristic (ROC) curve analysis were done. There were 326 trauma patients studied, of whom 279 adults were evaluated by the TRISS methodology. Men were more frequently involved in trauma than women; there was more blunt trauma than penetrating trauma. The M statistic was 0.98 and the overall Z statistic was 5.81. The ROC curve analysis showed TRISS to be a fair discriminator in the study case-mix with an area under the curve of 0.82 (95% confidence interval 0.69-0.96). There is a considerable disparity between predicted and observed outcomes when trauma patients are evaluated by the TRISS methodology in a developing country setting.

Surgical management of oncogeriatric patients

Major changes are taking place at a great pace in modern medicine, and surgical oncologists are at the forefront when new skills are to be tested and implemented. Perhaps the most significant change we are facing relates to the aging of our population, with most solid tumors presenting at age 70 years (+/- 5 years). The demographics and epidemiological details are covered in the appropriate sections of this special issue, but it is important to realize how such a shift influences our day-to-day practice. These principally are occurring in improved anesthetic care, minimally invasive surgery, nonoperative therapies, risk assessment, and quality-of-life estimates.

Effects of SVM parameter optimization on discrimination and calibration for post-procedural PCI mortality

Support vector machines (SVM) have become popular among machine learning researchers, but their applications in biomedicine have been somewhat limited. A number of methods, such as grid search and evolutionary algorithms, have been utilized to optimize model parameters of SVMs. The sensitivity of the results to changes in optimization methods has not been investigated in the context of medical applications. In this study, radial-basis kernel SVM and polynomial kernel SVM mortality prediction models for percutaneous coronary interventions were optimized using (a) mean-squared error, (b) mean cross-entropy error, (c) the area under the receiver operating characteristic, and (d) the Hosmer-Lemeshow goodness-of-fit test (HL chi(2)). A threefold cross-validation inner and outer loop method was used to select the best models using the training data, and evaluations were based on previously unseen test data. The results were compared to those produced by logistic regression models optimized using the same indices. The choice of optimization parameters had a significant impact on performance in both SVM kernel types.