AUROC = area under the receiver operating characteristic curve; LR + = positive likelihood ratio; LR -= negative likelihood ratio; NPV = negative predictive value; PPV = positive predictive value; ROC = receiver operating characteristic. Critical Care December 2004 Vol 8 No 6 Bewick et al. IntroductionA simple diagnostic test for a particular disease or outcome classifies patients into two groups: those with the outcome and those without. A test is assessed by its ability to diagnose the outcome correctly, whether this is positive or negative. If the actual outcome is not evident then it may be supplied by the 'gold standard' test. The data given in Table 1 provide an example in which the outcome is death or survival. The patients were attending an accident and emergency unit and the venous blood analysis for the metabolic marker lactate was used in the early identification of those patients at risk for death. Patients with lactate levels greater than 1.5 mmol/l were considered to be at risk. In general, the results of a diagnostic test may be presented as shown in Table 2. Sensitivity and specificityThe sensitivity of a diagnostic test is the proportion of patients for whom the outcome is positive that are correctly identified by the test. The specificity is the proportion of patients for whom the outcome is negative that are correctly identified by the test.For the data given in Table 1 the sensitivity of the test using lactate level above 1.5 mmol/l as an indicator of mortality is 81/126 = 0.64, and the specificity is 674/1265 = 0.53. Therefore, 64% of patients in this sample who died and 53% who survived were correctly identified by this test. Because both of these measures are simple proportions, their confidence intervals can be calculated as described in Statistics review 8 [1]. The 95% confidence interval for sensitivity is 56-73% and that for specificity is 51-56%.Generally, both the sensitivity and specificity of a test need to be known in order to assess its usefulness for a diagnosis. A discriminating test would have sensitivity and specificity close to 100%. However, a test with high sensitivity may have low specificity and vice versa. The decision to make use of a diagnostic test will also depend on whether a treatment exists should the result of the test be positive, the cost of such a treatment, and whether the treatment is detrimental in cases in which the result is a false positive. Positive and negative predictive valuesThe positive predictive value (PPV) of a test is the probability that a patient has a positive outcome given that they have a positive test result. This is in contrast to sensitivity, which is the probability that a patient has a positive test result given that they have a positive outcome. Similarly, the negative predictive value (NPV) is the probability that a patient has a negative outcome given that they have a negative test result, in contrast to specificity, which is the probability that a patient has a negative test result given that they have a negative outcome. Review...

We propose an integrated and adaptable approach to improve patient care and clinical outcomes through analgesia and light sedation, initiated early during an episode of critical illness and as a priority of care. This strategy, which may be regarded as an evolution of the Pain, Agitation and Delirium guidelines, is conveyed in the mnemonic eCASH—early Comfort using Analgesia, minimal Sedatives and maximal Humane care. eCASH aims to establish optimal patient comfort with minimal sedation as the default presumption for intensive care unit (ICU) patients in the absence of recognised medical requirements for deeper sedation. Effective pain relief is the first priority for implementation of eCASH: we advocate flexible multimodal analgesia designed to minimise use of opioids. Sedation is secondary to pain relief and where possible should be based on agents that can be titrated to a prespecified target level that is subject to regular review and adjustment; routine use of benzodiazepines should be minimised. From the outset, the objective of sedation strategy is to eliminate the use of sedatives at the earliest medically justifiable opportunity. Effective analgesia and minimal sedation contribute to the larger aims of eCASH by facilitating promotion of sleep, early mobilization strategies and improved communication of patients with staff and relatives, all of which may be expected to assist rehabilitation and avoid isolation, confusion and possible long-term psychological complications of an ICU stay. eCASH represents a new paradigm for patient-centred care in the ICU. Some organizational challenges to the implementation of eCASH are identified.

112 AUROC = area under the receiver operating characteristic curve; C.I. = confidence interval; ln = natural logarithm; logit = natural logarithm of the odds; MLE = maximum likelihood estimate; OR = odds ratio; ROC = receiver operating characteristic curve. Critical CareFebruary 2005 Vol 9 No 1 Bewick et al. IntroductionLogistic regression provides a method for modelling a binary response variable, which takes values 1 and 0. For example, we may wish to investigate how death (1) or survival (0) of patients can be predicted by the level of one or more metabolic markers. As an illustrative example, consider a sample of 2000 patients whose levels of a metabolic marker have been measured. Table 1 shows the data grouped into categories according to metabolic marker level, and the proportion of deaths in each category is given. The proportions of deaths are estimates of the probabilities of death in each category. Figure 1 shows a plot of these proportions. It suggests that the probability of death increases with the metabolic marker level. However, it can be seen that the relationship is nonlinear and that the probability of death changes very little at the high or low extremes of marker level. This pattern is typical because proportions cannot lie outside the range from 0 to 1. The relationship can be described as following an 'S'-shaped curve. Logistic regression with a single quantitative explanatory variableThe logistic or logit function is used to transform an 'S'-shaped curve into an approximately straight line and to change the range of the proportion from 0-1 to -∞ to +∞. Although this model looks similar to a simple linear regression model, the underlying distribution is binomial and the parameters a and b cannot be estimated in exactly the same way as for simple linear regression. Instead, the parameters are usually estimated using the method of maximum likelihood, which is discussed below. Binomial distributionWhen the response variable is binary (e.g. death or survival), then the probability distribution of the number of deaths in a sample of a particular size, for given values of the explanatory AbstractThis review introduces logistic regression, which is a method for modelling the dependence of a binary response variable on one or more explanatory variables. Continuous and categorical explanatory variables are considered.

Patients with limited cardiac reserve are less likely to survive and develop more complications following major surgery. By augmenting oxygen delivery index (DO2I) with a combination of intravenous fluids and inotropes (goal directed therapy (GDT)), postoperative mortality and morbidity of high-risk patients may be reduced. However, although most studies suggest that GDT may improve outcome in high-risk surgical patients, it is still not widely practiced. We set out to test the hypothesis that GDT results in greatest benefit in terms of mortality and morbidity in patients with the highest risk of mortality and have undertaken a systematic review of the current literature to see if this is correct. We performed a systematic search of Medline, Embase and CENTRAL databases for randomized controlled trials (RCTs) and reviews of GDT in surgical patients. To minimize heterogeneity we excluded studies involving cardiac, trauma, and paediatric surgery. Extremely high risk, high risk and intermediate risks of mortality were defined as >20%, 5 to 20% and <5% mortality rates in the control arms of the trials, respectively. Meta analyses were performed and Forest plots drawn using RevMan software. Data are presented as odd ratios (OR; 95% confidence intervals (CI), and P-values). A total of 32 RCTs including 2,808 patients were reviewed. All studies reported mortality. Five studies (including 300 patients) were excluded from assessment of complication rates as the number of patients with complications was not reported. The mortality benefit of GDT was confined to the extremely high-risk group (OR = 0.20, 95% CI 0.09 to 0.41; P < 0.0001). Complication rates were reduced in all subgroups (OR = 0.45, 95% CI 0.34 to 0.60; P < 0.00001). The morbidity benefit was greatest amongst patients in the extremely high-risk subgroup (OR = 0.27, 95% CI 0.15 to 0.51; P < 0.0001), followed by the intermediate risk subgroup (OR = 0.43, 95% CI 0.27 to 0.67; P = 0.0002), and the high-risk subgroup (OR 0.56, 95% CI 0.36 to 0.89; P = 0.01). Despite heterogeneity in trial quality and design, we found GDT to be beneficial in all high-risk patients undergoing major surgery. The mortality benefit of GDT was confined to the subgroup of patients at extremely high risk of death. The reduction of complication rates was seen across all subgroups of GDT patients.

The present review introduces the notion of statistical power and the hazard of under-powered studies. The problem of how to calculate an ideal sample size is also discussed within the context of factors that affect power, and specific methods for the calculation of sample size are presented for two common scenarios, along with extensions to the simplest case.