Cartas Al Director
Beware of multiplicity problems in clinical trials
Cuidado con los problemas de multiplicidad en ensayos clínicos
Á. Touma-Fernández
Hospital Universitario Morales Meseguer, Avenida Marqués de los Vélez, s/n, 30008 Murcia, Spain
Read
307
Timeswas read the article
125
Total PDF
182
Total HTML
Share statistics
array:23 [ "pii" => "S0034935615000560" "issn" => "00349356" "doi" => "10.1016/j.redar.2015.02.006" "estado" => "S300" "fechaPublicacion" => "2015-10-01" "aid" => "582" "copyright" => "Sociedad Española de Anestesiología, Reanimación y Terapéutica del Dolor" "copyrightAnyo" => "2015" "documento" => "simple-article" "crossmark" => 1 "subdocumento" => "cor" "cita" => "Rev Esp Anestesiol Reanim. 2015;62:477-8" "abierto" => array:3 [ "ES" => false "ES2" => false "LATM" => false ] "gratuito" => false "lecturas" => array:2 [ "total" => 306 "formatos" => array:3 [ "EPUB" => 5 "HTML" => 180 "PDF" => 121 ] ] "itemSiguiente" => array:18 [ "pii" => "S0034935615000778" "issn" => "00349356" "doi" => "10.1016/j.redar.2015.01.018" "estado" => "S300" "fechaPublicacion" => "2015-10-01" "aid" => "586" "copyright" => "Sociedad Española de Anestesiología, Reanimación y Terapéutica del Dolor" "documento" => "simple-article" "crossmark" => 1 "subdocumento" => "cor" "cita" => "Rev Esp Anestesiol Reanim. 2015;62:478-9" "abierto" => array:3 [ "ES" => false "ES2" => false "LATM" => false ] "gratuito" => false "lecturas" => array:2 [ "total" => 279 "formatos" => array:3 [ "EPUB" => 5 "HTML" => 207 "PDF" => 67 ] ] "es" => array:10 [ "idiomaDefecto" => true "cabecera" => "<span class="elsevierStyleTextfn">Carta al Director</span>" "titulo" => "Comentarios al artículo: «Estudio epidemiológico del dolor en hospitales públicos andaluces»" "tienePdf" => "es" "tieneTextoCompleto" => "es" "paginas" => array:1 [ 0 => array:2 [ "paginaInicial" => "478" "paginaFinal" => "479" ] ] "titulosAlternativos" => array:1 [ "en" => array:1 [ "titulo" => "Comments to the article: «Epidemiologic study about pain in andalusian public hospitals»" ] ] "contieneTextoCompleto" => array:1 [ "es" => true ] "contienePdf" => array:1 [ "es" => true ] "autores" => array:1 [ 0 => array:2 [ "autoresLista" => "J.M. López-Millán" "autores" => array:1 [ 0 => array:2 [ "nombre" => "J.M." "apellidos" => "López-Millán" ] ] ] ] ] "idiomaDefecto" => "es" "EPUB" => "https://multimedia.elsevier.es/PublicationsMultimediaV1/item/epub/S0034935615000778?idApp=UINPBA00004N" "url" => "/00349356/0000006200000008/v1_201509110028/S0034935615000778/v1_201509110028/es/main.assets" ] "itemAnterior" => array:18 [ "pii" => "S0034935614002965" "issn" => "00349356" "doi" => "10.1016/j.redar.2014.10.010" "estado" => "S300" "fechaPublicacion" => "2015-10-01" "aid" => "543" "copyright" => "Sociedad Española de Anestesiología, Reanimación y Terapéutica del Dolor" "documento" => "simple-article" "crossmark" => 1 "subdocumento" => "crp" "cita" => "Rev Esp Anestesiol Reanim. 2015;62:472-6" "abierto" => array:3 [ "ES" => false "ES2" => false "LATM" => false ] "gratuito" => false "lecturas" => array:2 [ "total" => 442 "formatos" => array:3 [ "EPUB" => 6 "HTML" => 302 "PDF" => 134 ] ] "es" => array:13 [ "idiomaDefecto" => true "cabecera" => "<span class="elsevierStyleTextfn">CASO CLÍNICO</span>" "titulo" => "Aturdimiento miocárdico neurogénico en Pediatría. A propósito de un caso" "tienePdf" => "es" "tieneTextoCompleto" => "es" "tieneResumen" => array:2 [ 0 => "es" 1 => "en" ] "paginas" => array:1 [ 0 => array:2 [ "paginaInicial" => "472" "paginaFinal" => "476" ] ] "titulosAlternativos" => array:1 [ "en" => array:1 [ "titulo" => "Neurogenic stunned myocardium in Pediatrics. A case report" ] ] "contieneResumen" => array:2 [ "es" => true "en" => true ] "contieneTextoCompleto" => array:1 [ "es" => true ] "contienePdf" => array:1 [ "es" => true ] "resumenGrafico" => array:2 [ "original" => 0 "multimedia" => array:7 [ "identificador" => "fig0010" "etiqueta" => "Figura 2" "tipo" => "MULTIMEDIAFIGURA" "mostrarFloat" => true "mostrarDisplay" => false "figura" => array:1 [ 0 => array:4 [ "imagen" => "gr2.jpeg" "Alto" => 1866 "Ancho" => 1400 "Tamanyo" => 440801 ] ] "descripcion" => array:1 [ "es" => "<p id="spar0020" class="elsevierStyleSimplePara elsevierViewall">Detalle del electrocardiograma, recogido a voltaje estándar de 10<span class="elsevierStyleHsp" style=""></span>mm/mV, donde se aprecia claramente la anormalidad del segmento ST con ascenso/descenso de hasta 4<span class="elsevierStyleHsp" style=""></span>mm en las derivaciones <span class="elsevierStyleSmallCaps">i</span>, <span class="elsevierStyleSmallCaps">iii</span>, aVL y aVF.</p>" ] ] ] "autores" => array:1 [ 0 => array:2 [ "autoresLista" => "F.J. Alados Arboledas, L. Millán-Miralles, M.P. Millán-Bueno, J.F. Expósito-Montes, C. Santiago-Gutierrez, M.C. Martínez Padilla" "autores" => array:6 [ 0 => array:2 [ "nombre" => "F.J." "apellidos" => "Alados Arboledas" ] 1 => array:2 [ "nombre" => "L." "apellidos" => "Millán-Miralles" ] 2 => array:2 [ "nombre" => "M.P." "apellidos" => "Millán-Bueno" ] 3 => array:2 [ "nombre" => "J.F." "apellidos" => "Expósito-Montes" ] 4 => array:2 [ "nombre" => "C." "apellidos" => "Santiago-Gutierrez" ] 5 => array:2 [ "nombre" => "M.C." "apellidos" => "Martínez Padilla" ] ] ] ] ] "idiomaDefecto" => "es" "EPUB" => "https://multimedia.elsevier.es/PublicationsMultimediaV1/item/epub/S0034935614002965?idApp=UINPBA00004N" "url" => "/00349356/0000006200000008/v1_201509110028/S0034935614002965/v1_201509110028/es/main.assets" ] "en" => array:13 [ "idiomaDefecto" => true "cabecera" => "<span class="elsevierStyleTextfn">Cartas Al Director</span>" "titulo" => "Beware of multiplicity problems in clinical trials" "tieneTextoCompleto" => true "saludo" => "<span class="elsevierStyleItalic">Dear Director</span>:" "paginas" => array:1 [ 0 => array:2 [ "paginaInicial" => "477" "paginaFinal" => "478" ] ] "autores" => array:1 [ 0 => array:3 [ "autoresLista" => "Á. Touma-Fernández" "autores" => array:1 [ 0 => array:3 [ "nombre" => "Á." "apellidos" => "Touma-Fernández" "email" => array:1 [ 0 => "angel.touma.fernandez@gmail.com" ] ] ] "afiliaciones" => array:1 [ 0 => array:2 [ "entidad" => "Hospital Universitario Morales Meseguer, Avenida Marqués de los Vélez, s/n, 30008 Murcia, Spain" "identificador" => "aff0005" ] ] ] ] "titulosAlternativos" => array:1 [ "es" => array:1 [ "titulo" => "Cuidado con los problemas de multiplicidad en ensayos clínicos" ] ] "textoCompleto" => "<span class="elsevierStyleSections"><p id="par0005" class="elsevierStylePara elsevierViewall">Multiplicity issues deal with situations when analyzing data from clinical trials in which the common danger is that the rate of false positive results is much higher than intended. Broadly speaking, multiplicity implies more than one chance to win, i.e. in this context, to produce significant results.<a class="elsevierStyleCrossRef" href="#bib0025"><span class="elsevierStyleSup">1</span></a> Virtually, it is present in all clinical trials, and may lead to an unintentional, i.e. due to misunderstanding or lack of knowledge, or, what it is worse, a fraudulent intentional claim that a significant treatment difference exists when actually it does not. A common situation that may lead to multiplicity problems in clinical trials is the presence of multiple endpoints, in which the existence of multiple statistical tests increases the chances of false positive inferences.</p><p id="par0010" class="elsevierStylePara elsevierViewall">The study by Olmos-Rodríguez et al,<a class="elsevierStyleCrossRef" href="#bib0030"><span class="elsevierStyleSup">2</span></a> – with a primary goal of “assessing the effectiveness of … and transfusion requirements in heart surgery” and a secondary one of “assessing its impact on the fluid balance … and hospital length of stay” –, is an example of potential multiplicity problems arising when many outcome measures are used to assess the result of a therapy. In this sense, as reported in Table 2 of the paper, the results of the main goal of the study were tested by using 10 single tests, each one with a control rate of false positive result of 5%. The risk is controlled for each individual single test, and that is what is meant by the significance level of the test or <span class="elsevierStyleItalic">p</span>-value. However, it is not so easy to control the overall risk of declaring at least one false positive somewhere if many separate significance tests are performed. In the quoted results of Table 2, 10 (independent) tests at a separate false positive rate of 5% have been performed, implying a 95% chance of not making a mistake on the first test, a 95<span class="elsevierStyleHsp" style=""></span>×<span class="elsevierStyleHsp" style=""></span>95% (=90.25%) of avoiding a mistake on either of the first two, and so nearly a 10% risk of one or other (or both) of the first two tests resulting in false positive. After 10 independent tests at the 5% level the probability of rejecting the null hypotheses in at least one test when the null is true in all the cases, i.e. the family-wise error rate<a class="elsevierStyleCrossRef" href="#bib0025"><span class="elsevierStyleSup">1</span></a> would be 1<span class="elsevierStyleHsp" style=""></span>−<span class="elsevierStyleHsp" style=""></span>(0.95<span class="elsevierStyleSup">10</span>)<span class="elsevierStyleHsp" style=""></span>=<span class="elsevierStyleHsp" style=""></span>0.4, i.e. a 40% chance of declaring a difference when none exists. Claiming a treatment difference when the rate of false positive by chance is as high as 40% is far from what should be expected. A similar situation arises in the quoted results of Tables 3 and 4, in which many independent tests have been performed.</p><p id="par0015" class="elsevierStylePara elsevierViewall">There is no consensus on the ideal way multiplicity issues should be approached, and even if they should be corrected at all,<a class="elsevierStyleCrossRef" href="#bib0035"><span class="elsevierStyleSup">3</span></a> but it is clear that, if author makes multiple comparisons readers should expect some cautious interpretation of results taking into account potential dangers associated with multiplicity,<a class="elsevierStyleCrossRef" href="#bib0040"><span class="elsevierStyleSup">4</span></a> mainly if the trial has a confirmatory aim as it was the case. Many strategies are available depending on the multiplicity problem, being the most important, if possible, concentrate on one primary endpoint. Other widely known approaches are the conservatives Bonferroni and Sidak corrections, control of the false discovery rates, the use of an appropriate multivariate analysis and the increasingly used gate keeping procedures. All these procedures are not easy to implement by non professional medical statisticians so that contact with them in an early design phase is advisable.</p><p id="par0020" class="elsevierStylePara elsevierViewall">In summary, multiplicity problems are virtually present in all trials and should always be taken into account when designing and interpreting trials results. A strict adherence to CONSORT guidelines<a class="elsevierStyleCrossRef" href="#bib0040"><span class="elsevierStyleSup">4</span></a> may help with this aim.</p><span id="sec0005" class="elsevierStyleSection elsevierViewall"><span class="elsevierStyleSectionTitle" id="sect0005">Conflict of interest</span><p id="par0025" class="elsevierStylePara elsevierViewall">I hereby declare I have taken into account the ethical responsibilities included in REDAR standards and meet the requirements for authorship. Also I declare no conflict of interest.</p></span></span>" "textoCompletoSecciones" => array:1 [ "secciones" => array:2 [ 0 => array:2 [ "identificador" => "sec0005" "titulo" => "Conflict of interest" ] 1 => array:1 [ "titulo" => "References" ] ] ] "pdfFichero" => "main.pdf" "tienePdf" => true "bibliografia" => array:2 [ "titulo" => "References" "seccion" => array:1 [ 0 => array:2 [ "identificador" => "bibs0005" "bibliografiaReferencia" => array:4 [ 0 => array:3 [ "identificador" => "bib0025" "etiqueta" => "1" "referencia" => array:1 [ 0 => array:2 [ "contribucion" => array:1 [ 0 => array:2 [ "titulo" => "Adjusting for multiple testing – when and how" "autores" => array:1 [ 0 => array:2 [ "etal" => false "autores" => array:2 [ 0 => "R. Bender" 1 => "S. Lange" ] ] ] ] ] "host" => array:1 [ 0 => array:1 [ "Revista" => array:6 [ "tituloSerie" => "J Clin Epidemiol" "fecha" => "2001" "volumen" => "54" "paginaInicial" => "343" "paginaFinal" => "349" "link" => array:1 [ 0 => array:2 [ "url" => "https://www.ncbi.nlm.nih.gov/pubmed/11297884" "web" => "Medline" ] ] ] ] ] ] ] ] 1 => array:3 [ "identificador" => "bib0030" "etiqueta" => "2" "referencia" => array:1 [ 0 => array:2 [ "contribucion" => array:1 [ 0 => array:2 [ "titulo" => "Repercusión del cebado con coloides y la ultrafiltración sobre el sangrado posoperatorio y la transfusión sanguínea en cirugía cardiaca. Ensayo clínico aleatorizado" "autores" => array:1 [ 0 => array:2 [ "etal" => false "autores" => array:6 [ 0 => "M. Olmos-Rodríguez" 1 => "J.A. Ballester-Hernández" 2 => "M.T. Arteta-Bárcenas" 3 => "A. Rodríguez-Cerezo" 4 => "M.A. Vidarte-Ortiz de Artiñano" 5 => "C. Veiga-Alameda" ] ] ] ] ] "host" => array:1 [ 0 => array:2 [ "doi" => "10.1016/j.redar.2014.04.008" "Revista" => array:6 [ "tituloSerie" => "Rev Esp Anestesiol Reanim" "fecha" => "2015" "volumen" => "62" "paginaInicial" => "81" "paginaFinal" => "89" "link" => array:1 [ 0 => array:2 [ "url" => "https://www.ncbi.nlm.nih.gov/pubmed/25048999" "web" => "Medline" ] ] ] ] ] ] ] ] 2 => array:3 [ "identificador" => "bib0035" "etiqueta" => "3" "referencia" => array:1 [ 0 => array:2 [ "contribucion" => array:1 [ 0 => array:2 [ "titulo" => "Clinical trials must cope better with multiplicity" "autores" => array:1 [ 0 => array:2 [ "etal" => false "autores" => array:1 [ 0 => "J. Wittes" ] ] ] ] ] "host" => array:1 [ 0 => array:2 [ "doi" => "10.1038/nm1112-1607" "Revista" => array:5 [ "tituloSerie" => "Nat Med" "fecha" => "2012" "volumen" => "18" "paginaInicial" => "1607" "link" => array:1 [ 0 => array:2 [ "url" => "https://www.ncbi.nlm.nih.gov/pubmed/23135505" "web" => "Medline" ] ] ] ] ] ] ] ] 3 => array:3 [ "identificador" => "bib0040" "etiqueta" => "4" "referencia" => array:1 [ 0 => array:2 [ "contribucion" => array:1 [ 0 => array:2 [ "titulo" => "CONSORT 2010 statement: updated guidelines for reporting parallel group randomized trials" "autores" => array:1 [ 0 => array:3 [ "colaboracion" => "for the CONSORT Group" "etal" => false "autores" => array:3 [ 0 => "K.F. Schultz" 1 => "D.G. Altman" 2 => "D. Moher" ] ] ] ] ] "host" => array:1 [ 0 => array:1 [ "Revista" => array:4 [ "tituloSerie" => "Br Med J" "fecha" => "2010" "volumen" => "340" "paginaInicial" => "c332" ] ] ] ] ] ] ] ] ] ] ] "idiomaDefecto" => "en" "url" => "/00349356/0000006200000008/v1_201509110028/S0034935615000560/v1_201509110028/en/main.assets" "Apartado" => array:4 [ "identificador" => "7552" "tipo" => "SECCION" "en" => array:2 [ "titulo" => "Cartas al Director" "idiomaDefecto" => true ] "idiomaDefecto" => "en" ] "PDF" => "https://static.elsevier.es/multimedia/00349356/0000006200000008/v1_201509110028/S0034935615000560/v1_201509110028/en/main.pdf?idApp=UINPBA00004N&text.app=https://www.elsevier.es/" "EPUB" => "https://multimedia.elsevier.es/PublicationsMultimediaV1/item/epub/S0034935615000560?idApp=UINPBA00004N" ]
| Year/Month | Html | Total | |
|---|---|---|---|
| 2023 March | 2 | 4 | 6 |
| 2017 July | 3 | 3 | 6 |
| 2017 June | 13 | 8 | 21 |
| 2017 May | 13 | 2 | 15 |
| 2017 April | 18 | 1 | 19 |
| 2017 March | 25 | 47 | 72 |
| 2017 February | 15 | 2 | 17 |
| 2017 January | 10 | 2 | 12 |
| 2016 December | 15 | 7 | 22 |
| 2016 November | 9 | 3 | 12 |
| 2016 October | 35 | 12 | 47 |
| 2016 September | 0 | 2 | 2 |
| 2016 August | 0 | 2 | 2 |
| 2016 July | 0 | 1 | 1 |
| 2016 June | 0 | 2 | 2 |
| 2016 May | 0 | 2 | 2 |
| 2016 April | 0 | 3 | 3 |
| 2016 March | 0 | 3 | 3 |
| 2016 February | 0 | 2 | 2 |
| 2016 January | 2 | 4 | 6 |
| 2015 December | 1 | 2 | 3 |
| 2015 November | 2 | 2 | 4 |
| 2015 October | 7 | 6 | 13 |
| 2015 September | 12 | 3 | 15 |
Show all
