David J. Gavaghan

Assessing the quality of reports of randomized clinical trials: Is blinding necessary?

It has been suggested that the quality of clinical trials should be assessed by blinded raters to limit the risk of introducing bias into meta-analyses and systematic reviews, and into the peer-review process. There is very little evidence in the literature to substantiate this. This study describes the development of an instrument to assess the quality of reports of randomized clinical trials (RCTs) in pain research and its use to determine the effect of rater blinding on the assessments of quality. A multidisciplinary panel of six judges produced an initial version of the instrument. Fourteen raters from three different backgrounds assessed the quality of 36 research reports in pain research, selected from three different samples. Seven were allocated randomly to perform the assessments under blind conditions. The final version of the instrument included three items. These items were scored consistently by all the raters regardless of background and could discriminate between reports from the different samples. Blind assessments produced significantly lower and more consistent scores than open assessments. The implications of this finding for systematic reviews, meta-analytic research and the peer-review process are discussed.

An evaluation of homogeneity tests in meta-analyses in pain using simulations of individual patient data

&NA; In this paper we consider the validity and power of some commonly used statistics for assessing the degree of homogeneity between trials in a meta‐analysis. We show, using simulated individual patient data typical of that occurring in randomized controlled trials in pain, that the most commonly used statistics do not give the expected levels of statistical significance (i.e. the proportion of trials giving a significant result is not equal to the proportion expected due to random chance) when used with truly homogeneous data. In addition, all such statistics are shown to have extremely low power to detect true heterogeneity even when that heterogeneity is very large. Since, in most practical situations, failure to detect heterogeneity does not allow us to say with any helpful degree of certainty that the data is truly homogeneous, we advocate the quantitative combination of results only where the trials contained in a meta‐analysis can be shown to be clinically homogeneous. We propose as a definition of clinical homogeneity that all trials have (i) fixed and clearly defined inclusion criteria and (ii) fixed and clearly defined outcomes or outcome measures. In pain relief, for example, the first of these would be satisfied by all patients having moderate or severe pain, whilst the second would be satisfied by using at least 50% pain relief as the successful outcome measure.

Deriving dichotomous outcome measures from continuous data in randomised controlled trials of analgesics

&NA; Reports of RCTs of analgesics frequently describe results of studies in the form of mean derived indices, rather than using discontinuous events — such as number or proportion of patients with 50% pain relief. Because mean data inadequately describe information with a non‐normal distribution, combining mean data in systematic reviews may compromise the results. Showing that dichotomous data can reliably be derived from mean data, at least in acute pain models, indicates that more meaningful overviews or meta‐analysis may be possible. This study investigated the relationship between continuous and dichotomous analgesic measures in a set of individual patient data, and then used that relationship to derive dichotomous from continuous information in randomised controlled trials (RCTs) of analgesics. Individual patient information from 13 RCTs of parallel‐group and crossover design in acute postoperative pain was used to calculate the percentage of the maximum possible pain relief score (%maxTOTPAR) and the proportion of patients with greater than 50% pain relief (> 50%maxTOTPAR) for the different treatments. The relationship between the measures was investigated in 45 actual treatments and 10 000 treatments simulated using the underlying actual distribution; 1283 patients had 45 separate treatments. Mean %maxTOTPAR correlated with the proportion of patients with >50%maxTOTPAR (r2 = 0.90). The relationship calculated from all the 45 treatments predicted to within three patients the number of patients with more than 50% pain relief in 42 of 45 treatments, and 98.8% of 10 000 simulated treatments. For seven effective treatments, actual numbers‐needed‐to‐treat (NNT) to achieve >50%maxTOTPAR compared with placebo were very similar to those derived from calculated data.

Simulation of multiple ion channel block provides improved early prediction of compounds' clinical torsadogenic risk

Aims The level of inhibition of the human Ether-à-go-go-related gene (hERG) channel is one of the earliest preclinical markers used to predict the risk of a compound causing Torsade-de-Pointes (TdP) arrhythmias. While avoiding the use of drugs with maximum therapeutic concentrations within 30-fold of their hERG inhibitory concentration 50% (IC50) values has been suggested, there are drugs that are exceptions to this rule: hERG inhibitors that do not cause TdP, and drugs that can cause TdP but are not strong hERG inhibitors. In this study, we investigate whether a simulated evaluation of multi-channel effects could be used to improve this early prediction of TdP risk. Methods and results We collected multiple ion channel data (hERG, Na, l-type Ca) on 31 drugs associated with varied risks of TdP. To integrate the information on multi-channel block, we have performed simulations with a variety of mathematical models of cardiac cells (for rabbit, dog, and human ventricular myocyte models). Drug action is modelled using IC50 values, and therapeutic drug concentrations to calculate the proportion of blocked channels and the channel conductances are modified accordingly. Various pacing protocols are simulated, and classification analysis is performed to evaluate the predictive power of the models for TdP risk. We find that simulation of action potential duration prolongation, at therapeutic concentrations, provides improved prediction of the TdP risk associated with a compound, above that provided by existing markers. Conclusion The suggested calculations improve the reliability of early cardiac safety assessments, beyond those based solely on a hERG block effect.

Chaste: An Open Source C++ Library for Computational Physiology and Biology

Chaste — Cancer, Heart And Soft Tissue Environment — is an open source C++ library for the computational simulation of mathematical models developed for physiology and biology. Code development has been driven by two initial applications: cardiac electrophysiology and cancer development. A large number of cardiac electrophysiology studies have been enabled and performed, including high-performance computational investigations of defibrillation on realistic human cardiac geometries. New models for the initiation and growth of tumours have been developed. In particular, cell-based simulations have provided novel insight into the role of stem cells in the colorectal crypt. Chaste is constantly evolving and is now being applied to a far wider range of problems. The code provides modules for handling common scientific computing components, such as meshes and solvers for ordinary and partial differential equations (ODEs/PDEs). Re-use of these components avoids the need for researchers to ‘re-invent the wheel’ with each new project, accelerating the rate of progress in new applications. Chaste is developed using industrially-derived techniques, in particular test-driven development, to ensure code quality, re-use and reliability. In this article we provide examples that illustrate the types of problems Chaste can be used to solve, which can be run on a desktop computer. We highlight some scientific studies that have used or are using Chaste, and the insights they have provided. The source code, both for specific releases and the development version, is available to download under an open source Berkeley Software Distribution (BSD) licence at http://www.cs.ox.ac.uk/chaste, together with details of a mailing list and links to documentation and tutorials.

Deriving dichotomous outcome measures from continuous data in randomised controlled trials of analgesics: use of pain intensity and visual analogue scales

Abstract The aim of this study was to examine whether mean data from categorical pain intensity and visual analogue scales for both pain intensity and relief could be used reliably to derive dichotomous outcome measures for meta‐analysis. Individual patient data from randomised controlled trials of single‐dose analgesics in acute postoperative pain were used. The methods used were as follows: data from 132 treatments with over 4700 patients were used to calculate mean %maxSPID (categorical pain intensity), %maxVAS‐SPID (visual analogue pain intensity) and %maxVAS‐TOTPAR (visual analogue pain relief); these were used to derive relationships with the number of patients who achieved at least 50% pain relief (%maxTOTPAR). Good agreement was obtained between the actual number of patients with >50%maxTOTPAR and the number calculated for all three measures. For SPID, verification included independent data sets. For calculations involving each measure, summing the positive and negative differences between actual and calculated numbers of patients with >50%maxTOTPAR gave an average difference of less than 0.25 patients per treatment arm. Reports of randomised trials of analgesics frequently describe results of studies in the form of mean derived indices, rather than using discontinuous events, such as number or proportion of patients obtaining at least 50% pain relief. Because mean data inadequately describe information with a non‐normal distribution, combining such mean data in systematic reviews may compromise the results. Showing that dichotomous data can reliably be derived from mean SPID, VAS‐SPID and VAS‐TOTPAR as well as TOTPAR data in previously published acute pain studies makes much more information accessible for meta‐analysis.

Oral morphine in cancer pain: influences on morphine and metabolite concentration

One hundred fifty‐one patients with chronic cancer pain were studied during chronic treatment with oral morphine. Plasma concentrations of morphine and metabolites (M3G and M6G) were measured. The ratio of plasma morphine to metabolites was not affected by dose. Generalized linear interactive modeling analysis using morphine dose, age, sex, renal and hepatic dysfunction, and concomitant medication as explanatory variables accounted for 70% of the variance in plasma concentrations of morphine, morphine‐3‐glucuronide (M3G) and morphine‐6‐glucuronide (M6G). Increasing morphine dose was a significant factor for increased plasma concentrations of morphine, M3G, and M6G. Other significant factors were: age greater than 70 years (increased M3G and M6G plasma concentrations), plasma creatinine >150 µmol/L (increased M3G and M6G plasma concentrations), male sex (decreased morphine and M6G plasma concentrations), raised creatinine plus coadministration of tricyclic antidepressants (increased M3G plasma concentrations), ranitidine (increased morphine plasma concentrations), and raised creatinine plus coadministration of ranitidine (increased M6G plasma concentrations).

Generation of histo−anatomically representative models of the individual heart: tools and application

This paper presents methods to build histo-anatomically detailed individualized cardiac models. The models are based on high-resolution three-dimensional anatomical and/or diffusion tensor magnetic resonance images, combined with serial histological sectioning data, and are used to investigate individualized cardiac function. The current state of the art is reviewed, and its limitations are discussed. We assess the challenges associated with the generation of histo-anatomically representative individualized in silico models of the heart. The entire processing pipeline including image acquisition, image processing, mesh generation, model set-up and execution of computer simulations, and the underlying methods are described. The multifaceted challenges associated with these goals are highlighted, suitable solutions are proposed, and an important application of developed high-resolution structure–function models in elucidating the effect of individual structural heterogeneity upon wavefront dynamics is demonstrated.

Development of an anatomically detailed MRI-derived rabbit ventricular model and assessment of its impact on simulations of electrophysiological function

Recent advances in magnetic resonance (MR) imaging technology have unveiled a wealth of information regarding cardiac histoanatomical complexity. However, methods to faithfully translate this level of fine-scale structural detail into computational whole ventricular models are still in their infancy, and, thus, the relevance of this additional complexity for simulations of cardiac function has yet to be elucidated. Here, we describe the development of a highly detailed finite-element computational model (resolution: ∼125 μm) of rabbit ventricles constructed from high-resolution MR data (raw data resolution: 43 × 43 × 36 μm), including the processes of segmentation (using a combination of level-set approaches), identification of relevant anatomical features, mesh generation, and myocyte orientation representation (using a rule-based approach). Full access is provided to the completed model and MR data. Simulation results were compared with those from a simplified model built from the same images but excluding finer anatomical features (vessels/endocardial structures). Initial simulations showed that the presence of trabeculations can provide shortcut paths for excitation, causing regional differences in activation after pacing between models. Endocardial structures gave rise to small-scale virtual electrodes upon the application of external field stimulation, which appeared to protect parts of the endocardium in the complex model from strong polarizations, whereas intramural virtual electrodes caused by blood vessels and extracellular cleft spaces appeared to reduce polarization of the epicardium. Postshock, these differences resulted in the genesis of new excitation wavefronts that were not observed in more simplified models. Furthermore, global differences in the stimulus recovery rates of apex/base regions were observed, causing differences in the ensuing arrhythmogenic episodes. In conclusion, structurally simplified models are well suited for a large range of cardiac modeling applications. However, important differences are seen when behavior at microscales is relevant, particularly when examining the effects of external electrical stimulation on tissue electrophysiology and arrhythmia induction. This highlights the utility of histoanatomically detailed models for investigations of cardiac function, in particular for future patient-specific modeling.

Deriving dichotomous outcome measures from continuous data in randomised controlled trials of analgesics: verification from independent data.

Abstract A previously established relationship for deriving dichotomous from continuous information in randomised controlled trials (RCTs) of analgesics has been tested using an independent data set. Individual patient information from 18 RCTs of parallel‐group design in acute postoperative pain (after abdominal, gynaecological and oral surgery) was used to calculate the percentage of the maximum possible pain relief score (%maxTOTPAR) and the proportion of patients with >50%maxTOTPAR for the different treatments. The relationship between the measures was investigated in 85 treatments with over 3400 patients. In 80 of 85 treatments (94%) agreement between calculated and actual number of patients with >50%maxTOTPAR was within four patients per treatment and in 72 (85%) was within three (average of 40 patients per treatment, range 21–58 patients). Summing the positive and negative differences between actual and calculated numbers of patients with >50%maxTOTPAR gave an average difference of 0.30 patients per treatment arm. Reports of RCTs of analgesics frequently describe results of studies in the form of mean derived indices, rather than using discontinuous events, such as number or proportion of patients with 50% pain relief. Because mean data inadequately describe information with a non‐normal distribution, combining mean data in systematic reviews may compromise the results. Showing that dichotomous data can reliably be derived from mean data in acute pain studies enables data published as means to be used for quantitative systematic reviews which require data in dichotomous form.