Leon Bax

A systematic comparison of software dedicated to meta-analysis of causal studies.

BackgroundOur objective was to systematically assess the differences in features, results, and usability of currently available meta-analysis programs.MethodsSystematic review of software. We did an extensive search on the internet (Google, Yahoo, Altavista, and MSN) for specialized meta-analysis software. We included six programs in our review: Comprehensive Meta-analysis (CMA), MetAnalysis, MetaWin, MIX, RevMan, and WEasyMA. Two investigators compared the features of the software and their results. Thirty independent researchers evaluated the programs on their usability while analyzing one data set.ResultsThe programs differed substantially in features, ease-of-use, and price. Although most results from the programs were identical, we did find some minor numerical inconsistencies. CMA and MIX scored highest on usability and these programs also have the most complete set of analytical features.ConclusionIn consideration of differences in numerical results, we believe the user community would benefit from openly available and systematically updated information about the procedures and results of each programs validation. The most suitable program for a meta-analysis will depend on the users needs and preferences and this report provides an overview that should be helpful in making a substantiated choice.

ST Analysis of the Fetal Electrocardiogram in Intrapartum Fetal Monitoring: A Meta-Analysis.

OBJECTIVE: To compare the effects of ST-waveform analysis in combination with cardiotocography with conventional cardiotocography for intrapartum fetal monitoring. DATA SOURCES: We searched MEDLINE, Embase, and PubMed for randomized controlled trials (RCTs) evaluating ST-waveform analysis for intrapartum fetal monitoring. METHODS OF STUDY SELECTION: We identified RCTs that compared ST-waveform analysis and conventional cardiotocography for intrapartum fetal monitoring of singleton pregnancies in cephalic presentation beyond 34 weeks of gestation and evaluating at least one of the following: metabolic acidosis, umbilical cord pH less than 7.15, umbilical cord pH less than 7.10, umbilical cord pH less than 7.05, umbilical cord pH less than 7.00, Apgar scores less than 7 at 5 minutes, admittance to the neonatal intensive care unit, need for intubation, presence of hypoxic ischemic encephalopathy, perinatal death, operative delivery, and number of fetal blood samplings. TABULATION, INTEGRATION, AND RESULTS: Five RCTs, which included 15,352 patients, met the selection criteria. Random-effects models were used to estimate the combined relative risks (RRs) of ST analysis compared with conventional cardiotocography. Compared with conventional cardiotocography, ST analysis showed a nonsignificant reduction in metabolic acidosis (RR 0.72, 95% confidence interval 0.43–1.19, number needed to treat [NNT] 357). ST analysis significantly reduced the incidence of additional fetal blood sampling (RR 0.59, 95% confidence interval 0.44–0.79, NNT 11), operative vaginal deliveries (RR 0.88, 95% confidence interval 0.80–0.97, NNT 64), and total operative deliveries (RR 0.94, 95% confidence interval 0.89–0.99, NNT 64). For other outcomes, no differences in effect were seen between ST analysis and conventional cardiotocography, or data were not suitable for meta-analysis. CONCLUSION: The additional use of ST analysis for intrapartum monitoring reduced the incidence of operative vaginal deliveries and the need for fetal blood sampling but did not reduce the incidence of metabolic acidosis at birth.

Detection of Lymph Node Metastases by Gadolinium-Enhanced Magnetic Resonance Imaging: Systematic Review and Meta-analysis

BACKGROUND Gadolinium-based contrast agents are used with magnetic resonance imaging (MRI) to highlight tumor vascularity in organs. They are also widely used for primary tumor visualization. We conducted a systematic review and meta-analysis of the existing evidence of the accuracy of gadolinium-enhanced MRI for staging lymph node metastases. METHODS We systematically searched the MEDLINE, Cochrane, CANCERLIT, and EMBASE databases for studies published in English or German from January 1, 1988, to January 1, 2008, that assessed the diagnostic accuracy of gadolinium-enhanced MRI in the evaluation of lymphatic metastases compared with histopathologic examination as the reference test. Based on a priori-defined clinical considerations, we studied three subgroups of studies: those that used a single malignancy criterion and those that used multiple malignancy criteria with or without contrast highlighting. Summaries of MRI sensitivity and specificity for detecting lymph node metastases were calculated using a bivariate regression model. All statistical tests were two-sided. RESULTS The literature search yielded 43 full-text papers that were considered for inclusion in the meta-analysis. We performed quantitative pooled analyses on the 32 studies that provided data on patient-level diagnosis. The weighted estimates of sensitivity and specificity for all studies combined were 0.72 (95% confidence interval [CI] = 0.66 to 0.79) and 0.87 (95% CI = 0.82 to 0.91). Estimates of sensitivity and specificity were essentially unchanged for studies that used a single malignancy criterion (0.71 [95% CI = 0.61 to 0.79] and 0.88 [95% CI = 0.80 to 0.93], respectively; n = 11 studies) or multiple malignancy criteria without contrast enhancement (0.70 [95% CI = 0.58 to 0.79] and 0.86 [95% CI = 0.68 to 0.94], respectively; n = 6 studies). The sensitivity increased to 0.84 (95% CI = 0.70 to 0.92), with a specificity of 0.82 (95% CI = 0.72 to 0.89) for the nine studies that incorporated contrast enhancement in their multiple malignancy criteria. Six studies did not define the malignancy criteria they used. CONCLUSIONS The overall accuracy of gadolinium-enhanced magnetic resonance imaging for the detection of nodal metastases is moderate. Incorporating contrast enhancement in the malignancy criteria substantially improves the accuracy of this diagnostic test.

Polychotomization of continuous variables in regression models based on the overall C index

BackgroundWhen developing multivariable regression models for diagnosis or prognosis, continuous independent variables can be categorized to make a prediction table instead of a prediction formula. Although many methods have been proposed to dichotomize prognostic variables, to date there has been no integrated method for polychotomization. The latter is necessary when dichotomization results in too much loss of information or when central values refer to normal states and more dispersed values refer to less preferable states, a situation that is not unusual in medical settings (e.g. body temperature, blood pressure). The goal of our study was to develop a theoretical and practical method for polychotomization.MethodsWe used the overall discrimination index C, introduced by Harrel, as a measure of the predictive ability of an independent regressor variable and derived a method for polychotomization mathematically. Since the naïve application of our method, like some existing methods, gives rise to positive bias, we developed a parametric method that minimizes this bias and assessed its performance by the use of Monte Carlo simulation.ResultsThe overall C is closely related to the area under the ROC curve and the produced di(poly)chotomized variables predictive performance is comparable to the original continuous variable. The simulation shows that the parametric method is essentially unbiased for both the estimates of performance and the cutoff points. Application of our method to the predictor variables of a previous study on rhabdomyolysis shows that it can be used to make probability profile tables that are applicable to the diagnosis or prognosis of individual patient status.ConclusionWe propose a polychotomization (including dichotomization) method for independent continuous variables in regression models based on the overall discrimination index C and clarified its meaning mathematically. To avoid positive bias in application, we have proposed and evaluated a parametric method. The proposed method for polychotomizing continuous regressor variables performed well and can be used to create probability profile tables.