Exact Tests for Contemporaneous Correlation of Disturbances in Seemingly Unrelated Regressions.

This paper proposes finite-sample procedures for testing the SURE specification in multi-equation regression models, i.e. whether the disturbances in different equations are contemporaneously uncorrelated or not. We apply the technique of Monte Carlo (MC) tests [Dwass (1957), Barnard (1963)] to obtain exact tests based on standard LR and LM zero correlation tests. We also suggest a MC quasi-LR (QLR) test based on feasible generalized least squares (FGLS). We show that the latter statistics are pivotal under the null, which provides the justification for applying MC tests. Furthermore, we extend the exact independence test proposed by Harvey and Phillips (1982) to the multi-equation framework. Specifically, we introduce several induced tests based on a set of simultaneous Harvey/Phillips-type tests and suggest a simulation-based solution to the associated combination problem. The properties of the proposed tests are studied in a Monte Carlo experiment which shows that standard asymptotic tests exhibit important size distortions, while MC tests achieve complete size control and display good power. Moreover, MC-QLR tests performed best in terms of power, a result of interest from the point of view of simulation-based tests. The power of the MC induced tests improves appreciably in comparison to standard Bonferroni tests and, in certain cases, outperforms the likelihood-based MC tests. The tests are applied to data used by Fischer (1993) to analyze the macroeconomic determinants of growth.

Quelle régulation pour l’arrêt d’un protocole de recherche clinique de thérapie génique somatique ? État des lieux auprès des cliniciens-chercheurs européens

Depuis 2002, le débat sur les risques associés à la thérapie génique est initié suite à l’annonce que deux enfants inclus dans un essai thérapeutique impliquant une thérapie génique ont développé des effets indésirables important. En Janvier 2005, le débat sur les risques reprit suite à l’interruption du protocole sur les enfants bulle du Pr Fischer à l’hôpital Necker de Paris. Nous avons donc étudié le processus impliqué ainsi que la réflexion éthique associée aux décisions d’arrêt de protocole de recherche. Notre travail a été mené par une équipe pluridisciplinaire combinant chercheurs en santé, généticiens et éthiciens. Nous avons étudié la participation des chercheurs, des patients, des institutions officielles, des comités d’éthique ainsi que des associations de patients dans le processus de décision d’interruption d’un protocole de recherche.Nous avons également analysé les critères jugés les plus pertinents dans l’arrêt d’un protocole de recherche. Enfin nous avons analysé le point de vue des personnes directement impliquées dans la thérapie génique au moyen d’un questionnaire. Toutes les personnes contactées ont présenté un poster de recherche au congrès de la Société Européenne de Thérapie Génique. 62 personnes d’autant d’équipes de recherche différentes, de 17 pays, sur les 350 contactés ont répondu. Selon eux, la décision d’arrêt d’un protocole de recherche doit être prise suite à une consultation des chercheurs, des patients, du ministère de tutelle, d’une agence nationale de régulation ou d’un comité d’éthique ; la légitimité étant accordée à des décisions prises en commun par les chercheurs, les patients et les comités d’éthique. Les incidents sérieux et de façon plus surprenante, les incidents moins graves sont jugés comme étant des critères suffisants pour interrompre un essai. Nous avons fini par analyser les conséquences éthiques, telles que balance bénéfice/risque, processus de régulation ou responsabilité, de ces critères sur l’arrêt d’un protocole de recherche.

The costs of alcohol, illegal drugs, and tobacco in Canada, 2002.

Objective: The aim of this study was to estimate costs attributable to substance use and misuse in Canada in 2002. Method: Based on information about prevalence of exposure and risk relations for more than 80 disease categories, deaths, years of life lost, and hospitalizations attributable to substance use and misuse were estimated. In addition, substance-attributable fractions for criminal justice expenditures were derived. Indirect costs were estimated using a modified human capital approach. Results: Costs of substance use and misuse totaled almost Can. $40 billion in 2002. The total cost per capita for substance use and misuse was about Can. $1,267: Can. $463 for alcohol, Can. $262 for illegal drugs, and Can. $541 for tobacco. Legal substances accounted for the vast majority of these costs (tobacco: almost 43% of total costs; alcohol: 37%). Indirect costs or productivity losses were the largest cost category (61%), followed by health care (22%) and law enforcement costs (14%). More than 40,000 people died in Canada in 2002 because of substance use and misuse: 37,209 deaths were attributable to tobacco, 4,258 were attributable to alcohol, and 1,695 were attributable to illegal drugs. A total of about 3.8 million hospital days were attributable to substance use and misuse, again mainly to tobacco. Conclusions: Substance use and misuse imposes a considerable economic toll on Canadian society and requires more preventive efforts.

Comorbid Depression Among Untreated Illicit Opiate Users : Results From a Multisite Canadian Study

This study aimed to describe patterns of major depression (MDD) in a cohort of untreated illicit opiate users recruited from 5 Canadian urban centres, identify sociodemographic characteristics of opiate users that predict MDD, and determine whether opiate users suffering from depression exhibit different drug use patterns than do participants without depression. Baseline data were collected from 679 untreated opiate users in Vancouver, Edmonton, Toronto, Montreal, and Quebec City. Using the Composite International Diagnostic Interview Short Form for Major Depression, we assessed sociodemographics, drug use, health status, health service use, and depression. We examined depression rates across study sites; logistic regression analyses predicted MDD from demographic information and city. Chi-square analyses were used to compare injection drug use and cocaine or crack use among participants with and without depression. Almost one-half (49.3%) of the sample met the cut-off score for MDD. Being female, white, and living outside Vancouver independently predicted MDD. Opiate users suffering from depression were more likely than users without depression to share injection equipment and paraphernalia and were also more likely to use cocaine (Ps < 0.05). Comorbid depression is common among untreated opiate users across Canada; targeted interventions are needed for this population.