A score test for binary data with patient non-compliance

A score test is developed for binary clinical trial data, which incorporates patient non-compliance while respecting randomization. It is assumed in this paper that compliance is all-or-nothing, in the sense that a patient either accepts all of the treatment assigned as specified in the protocol, or none of it. Direct analytic comparisons of the adjusted test statistic for both the score test and the likelihood ratio test are made with the corresponding test statistics that adhere to the intention-to-treat principle. It is shown that no gain in power is possible over the intention-to-treat analysis, by adjusting for patient non-compliance. Sample size formulae are derived and simulation studies are used to demonstrate that the sample size approximation holds. Copyright © 2003 John Wiley & Sons, Ltd.

An adaptive approach to implementing bivariate group sequential clinical trial designs

In clinical trials, situations often arise where more than one response from each patient is of interest; and it is required that any decision to stop the study be based upon some or all of these measures simultaneously. Theory for the design of sequential experiments with simultaneous bivariate responses is described by Jennison and Turnbull (Jennison, C., Turnbull, B. W. (1993). Group sequential tests for bivariate response: interim analyses of clinical trials with both efficacy and safety endpoints. Biometrics 49:741-752) and Cook and Farewell (Cook, R. J., Farewell, V. T. (1994). Guidelines for monitoring efficacy and toxicity responses in clinical trials. Biometrics 50:1146-1152) in the context of one efficacy and one safety response. These expositions are in terms of normally distributed data with known covariance. The methods proposed require specification of the correlation, ρ between test statistics monitored as part of the sequential test. It can be difficult to quantify ρ and previous authors have suggested simply taking the lowest plausible value, as this will guarantee power. This paper begins with an illustration of the effect that inappropriate specification of ρ can have on the preservation of trial error rates. It is shown that both the type I error and the power can be adversely affected. As a possible solution to this problem, formulas are provided for the calculation of correlation from data collected as part of the trial. An adaptive approach is proposed and evaluated that makes use of these formulas and an example is provided to illustrate the method. Attention is restricted to the bivariate case for ease of computation, although the formulas derived are applicable in the general multivariate case.

Reviewing the sustainability/stationarity of current account imbalances with tests for bounded integration

We investigate for 26 OECD economies whether their current account imbalances to GDP are driven by stochastic trends. Regarding bounded stationarity as the more natural counterpart of sustainability, results from Phillips–Perron tests for unit root and bounded unit root processes are contrasted. While the former hint at stationarity of current account imbalances for 12 economies, the latter indicate bounded stationarity for only six economies. Through panel-based test statistics, current account imbalances are diagnosed as bounded non-stationary. Thus, (spurious) rejections of the unit root hypothesis might be due to the existence of bounds reflecting hidden policy controls or financial crises.

Forecast encompassing tests and probability forecasts

We consider tests of forecast encompassing for probability forecasts, for both quadratic and logarithmic scoring rules. We propose test statistics for the null of forecast encompassing, present the limiting distributions of the test statistics, and investigate the impact of estimating the forecasting models' parameters on these distributions. The small-sample performance is investigated, in terms of small numbers of forecasts and model estimation sample sizes. We show the usefulness of the tests for the evaluation of recession probability forecasts from logit models with different leading indicators as explanatory variables, and for evaluating survey-based probability forecasts.

An extension of an over-dispersion test for count data

While over-dispersion in capture–recapture studies is well known to lead to poor estimation of population size, current diagnostic tools to detect the presence of heterogeneity have not been specifically developed for capture–recapture studies. To address this, a simple and efficient method of testing for over-dispersion in zero-truncated count data is developed and evaluated. The proposed method generalizes an over-dispersion test previously suggested for un-truncated count data and may also be used for testing residual over-dispersion in zero-inflation data. Simulations suggest that the asymptotic distribution of the test statistic is standard normal and that this approximation is also reasonable for small sample sizes. The method is also shown to be more efficient than an existing test for over-dispersion adapted for the capture–recapture setting. Studies with zero-truncated and zero-inflated count data are used to illustrate the test procedures.

The double triangular test in practice

The double triangular test was introduced twenty years ago, and the purpose of this paper is to review applications that have been made since then. In fact, take-up of the method was rather slow until the late 1990s, but in recent years several clinical trial reports have been published describing its use in a wide range of therapeutic areas. The core of this paper is a detailed account of five trials that have been published since 2000 in which the method was applied to studies of pancreatic cancer, breast cancer, myocardial infarction, epilepsy and bedsores. Before those accounts are given, the method is described and the history behind its evolution is presented. The future potential of the method for sequential case-control and equivalence trials is also discussed. Copyright © 2004 John Wiley & Sons, Ltd.

The sequential analysis of repeated binary responses: a score test for the case of three time points

In this paper a robust method is developed for the analysis of data consisting of repeated binary observations taken at up to three fixed time points on each subject. The primary objective is to compare outcomes at the last time point, using earlier observations to predict this for subjects with incomplete records. A score test is derived. The method is developed for application to sequential clinical trials, as at interim analyses there will be many incomplete records occurring in non-informative patterns. Motivation for the methodology comes from experience with clinical trials in stroke and head injury, and data from one such trial is used to illustrate the approach. Extensions to more than three time points and to allow for stratification are discussed. Copyright © 2005 John Wiley & Sons, Ltd.

The F test for comparing two normal variances: correct and incorrect calculation of the two-sided p-value?

This article illustrates that not all statistical software packages are correctly calculating a p-value for the classical F test comparison of two independent Normal variances. This is illustrated with a simple example, and the reasons why are discussed. Eight different software packages are considered.

Evaluation of a sequential global test of improved recovery following stroke as applied to the ICTUS trial of citicoline

The International Citicoline Trial in acUte Stroke is a sequential phase III study of the use of the drug citicoline in the treatment of acute ischaemic stroke, which was initiated in 2006 in 56 treatment centres. The primary objective of the trial is to demonstrate improved recovery of patients randomized to citicoline relative to those randomized to placebo after 12 weeks of follow-up. The primary analysis will take the form of a global test combining the dichotomized results of assessments on three well-established scales: the Barthel Index, the modified Rankin scale and the National Institutes of Health Stroke Scale. This approach was previously used in the analysis of the influential National Institute of Neurological Disorders and Stroke trial of recombinant tissue plasminogen activator in stroke. The purpose of this paper is to describe how this trial was designed, and in particular how the simultaneous objectives of taking into account three assessment scales, performing a series of interim analyses and conducting treatment allocation and adjusting the analyses to account for prognostic factors, including more than 50 treatment centres, were addressed. Copyright (C) 2008 John Wiley & Sons, Ltd.

A combined score test for binary and ordinal endpoints from clinical trials

There is growing interest, especially for trials in stroke, in combining multiple endpoints in a single clinical evaluation of an experimental treatment. The endpoints might be repeated evaluations of the same characteristic or alternative measures of progress on different scales. Often they will be binary or ordinal, and those are the cases studied here. In this paper we take a direct approach to combining the univariate score statistics for comparing treatments with respect to each endpoint. The correlations between the score statistics are derived and used to allow a valid combined score test to be applied. A sample size formula is deduced and application in sequential designs is discussed. The method is compared with an alternative approach based on generalized estimating equations in an illustrative analysis and replicated simulations, and the advantages and disadvantages of the two approaches are discussed.

Can portmanteau nonlinearity tests serve as general mis-specification tests?

A number of recent papers have employed the BDS test as a general test for mis-specification for linear and nonlinear models. We show that for a particular class of conditionally heteroscedastic models, the BDS test is unable to detect a common mis-specification. Our results also demonstrate that specific rather than portmanteau diagnostics are required to detect neglected asymmetry in volatility. However for both classes of tests reasonable power is only obtained using very large sample sizes.

Portmanteau model diagnostics and tests for nonlinearity: a comparative Monte Carlo study of two alternative methods

This paper employs an extensive Monte Carlo study to test the size and power of the BDS and close return methods of testing for departures from independent and identical distribution. It is found that the finite sample properties of the BDS test are far superior and that the close return method cannot be recommended as a model diagnostic. Neither test can be reliably used for very small samples, while the close return test has low power even at large sample sizes