Regarding the article: Mental disorder prevalence and associated risk factors in three prisons of Spain.

Reading the study by Zabala-Baños et al1 on the prevalence of mental disorders and associated risk factors in three correctional facilities in Spain is of great interest since it addresses one of the first-magnitude healthcare issues in prisons in our country. However, I would like to comment on some aspects of its design, since these may have an impact on the validity of its conclusions. First, the discrepancy between sociodemographic features of the individuals included in this study and those observed in the last 2014 General Report of the Secretariat of Penitentiary Institutions2 on overall data obtained from all Spanish prisons is remarkable. The aforementioned study excluded women (they account for 7.8% of the imprisoned population), preventive detainees (13%), foreigners with language difficulties (28% of the imprisoned population, yet it is unclear what criteria have been followed to assess such difficulties). On the other hand, statistical analysis. Authors say that a p<0.05 level of significance has been assumed. Yet, this consideration can lead to type I errors (false associations) motivated by multiple comparisons: when statistical contrast is carried out within data analysis, the probability of some being randomly statistically significant becomes larger. Nominal significance value, which conventionally is <0.05 must then be adjusted according to the number of executed hypothesis. Therefore, while they deserve congratulations for their effort I believe that no definitive conclusions should be made upon a study with a reduced sample size and some bias regarding the selection of individuals. Javier Velasco Montes Médico funcionario en prácticas Centro Penitenciario Madrid VII (Estremera)