Liz Cheek

Statistics review 13: Receiver operating characteristic curves

This review introduces some commonly used methods for assessing the performance of a diagnostic test. The sensitivity, specificity and likelihood ratio of a test are discussed. The uses of the receiver operating characteristic curve and the area under the curve are explained.

Statistics review 12: Survival analysis

This review introduces methods of analyzing data arising from studies where the response variable is the length of time taken to reach a certain end-point, often death. The Kaplan–Meier methods, log rank test and Coxs proportional hazards model are described.

Statistics review 9: One-way analysis of variance

This review introduces one-way analysis of variance, which is a method of testing differences between more than two groups or treatments. Multiple comparison procedures and orthogonal contrasts are described as methods for identifying specific differences between pairs of treatments.

Statistics review 10: Further nonparametric methods

This review introduces nonparametric methods for testing differences between more than two groups or treatments. Three of the more common tests are described in detail, together with multiple comparison procedures for identifying specific differences between pairs of groups.

A meta-analysis on the clinical effectiveness of transversus abdominis plane block

STUDY OBJECTIVE To study the efficacy of the transversus abdominal plane (TAP) block. DESIGN Meta-analysis. SETTING District general hospital. PATIENTS 86 patients in the TAP block group and 88 in the non-TAP block group. MEASUREMENTS Statistical analyses were performed using Microsoft Excel 2007 for Windows XP. Hedges g statistic was used for the calculation of standardized mean differences (SMD). Binary data (nausea) were summarized as risk ratios (RR). MAIN RESULTS Patients with TAP block required less morphine after 24 hours than those who did not have the block (random effects model: SMD -4.81, 95% CI [-7.45, -2.17], z = -3.57, P < 0.001). There was less time to first request of morphine in the non-TAP block group (random effects model: SMD 4.80, 95% CI [2.16, 7.43], z = 3.57, P < 0.001). Patients in the TAP block group had less pain up to 24 hours postoperatively. No statistical differences were found with respect to nausea. CONCLUSIONS TAP block reduces the need for postoperative opioid use, it increases the time first request for further analgesia, it provides more effective pain relief, and it reduces opioid-associated side effects.

Predictors of Death after Clostridium difficile Infection: A Report on 128 Strain-Typed Cases from a Teaching Hospital in the United Kingdom

We assessed the relationship between strain type, clinical factors, and outcome in 128 patients with Clostridium difficile infection. Strain type was not associated with any outcome measure. On multivariate analysis, ischemic heart disease and hypoalbuminemia predicted death. Metronidazole treatment in severe disease was associated with a higher rate of treatment failure and death.

Statistics review 8: Qualitative data – tests of association

This review introduces methods for investigating relationships between two qualitative (categorical) variables. The χ2 test of association is described, together with the modifications needed for small samples. The test for trend, in which at least one of the variables is ordinal, is also outlined. Risk measurement is discussed. The calculation of confidence intervals for proportions and differences between proportions are described. Situations in which samples are matched are considered.

Developing environment-specific water quality guidelines for suspended particulate matter

It is generally well recognised that suspended particulate matter (SPM), from nano-scale particles to sand-sized sediments, can cause serious degradation of aquatic ecosystems. However, at present there is a poor understanding of the SPM conditions that water quality managers should aim to achieve in contrasting environments in order to support good ecological status. In this article, we analyse long-term SPM data collected from a wide range of reference-condition temperate environments in the UK (638 stream/river sites comprising 42 different ecosystem-types). One-way analysis of variance reveals that there is a statistically significant difference (p < 0.001) between the background SPM concentrations observed in contrasting ecosystems that are in reference condition (minimal anthropogenic disturbance). One of the 42 ecosystems studied had mean background concentrations of SPM in excess of the current European Union (EU) water quality guideline, despite being in reference condition. The implications of this finding are that the EUs current blanket water quality guideline (25 mg L(-1) for all environments) is inappropriate for this specific ecosystem-type which will be non-compliant with the guideline regardless of the intensity of land-use. The other 41 ecosystems studied had mean concentrations below the current EU water quality guideline. However, this does not necessarily mean that the guideline is appropriate for these ecosystems, as previous research has demonstrated that detrimental impacts can be experienced by some freshwater organisms, of all trophic levels, when exposed to concentrations below 25 mg L(-1). Therefore, it is suggested here that it is likely that some ecosystems, particularly those with mean concentrations in the 0.00-5.99 mg L(-1) range, require much lower guideline values in order to be effectively protected. We propose a model for predicting environment-specific water quality guidelines for SPM. In order to develop this model, the 638 reference condition sites were first classified into one of five mean background SPM ranges (0.00-5.99, 6.00-11.99, 12.00-17.99, 18.00-23.99 and >24.00 mg L(-1)). Stepwise Multiple Discriminant Analysis (MDA) of these ranges showed that a sites SPM range can be predicted as a function of: mean annual air temperature, mean annual precipitation, mean altitude of upstream catchment, distance from source, slope to source, channel width and depth, the percentage of catchment area comprised of clay, chalk, and hard rock solid geology, and the percentage of the catchment area comprised of blown sand as the surface (drift) material. The MDA technique, with cross-validation (Wilks-Lambda 0.358, p 0.000), can predict the correct or the next closest SPM range of a site in 90% of cases. This technique can also predict SPM range membership in a probabilistic manner, allowing for an estimate of uncertainty to be made in the allocation of a site to an environment-specific SPM range.

Statistics review 11: assessing risk.

Relative risk and odds ratio have been introduced in earlier reviews (see Statistics reviews 3, 6 and 8). This review describes the calculation and interpretation of their confidence intervals. The different circumstances in which the use of either the relative risk or odds ratio is appropriate and their relative merits are discussed. A method of measuring the impact of exposure to a risk factor is introduced. Measures of the success of a treatment using data from clinical trials are also considered.

The Impact of Shape on the Perception of Euler Diagrams

Euler diagrams are often used for visualizing data collected into sets. However, there is a significant lack of guidance regarding graphical choices for Euler diagram layout. To address this deficiency, this paper asks the question ‘does the shape of a closed curve affect a user’s comprehension of an Euler diagram?’ By empirical study, we establish that curve shape does indeed impact on understandability. Our analysis of performance data indicates that circles perform best, followed by squares, with ellipses and rectangles jointly performing worst. We conclude that, where possible, circles should be used to draw effective Euler diagrams. Further, the ability to discriminate curves from zones and the symmetry of the curve shapes is argued to be important. We utilize perceptual theory to explain these results. As a consequence of this research, improved diagram layout decisions can be made for Euler diagrams whether they are manually or automatically drawn.