Nuestros investigadores

Jesús Fernando López Fidalgo

Unidad de Estadística
Instituto Cultura y Sociedad (ICS). Universidad de Navarra
Líneas de investigación
Diseño Óptimo de Experimentos
Índice H
12, (WoS, 08/05/2017)

Publicaciones científicas más recientes (desde 2010)

Autores: García-Vicente, A. M.; Delgado-Bolton, R. C.; Amo-Salas, M.; et al.
ISSN 1619-7070  Vol. 44  Nº 9  2017  págs. 1575-1587
The detection of occult cancer in patients suspected of having a paraneoplastic neurological syndrome (PNS) poses a diagnostic challenge. The aim of our study was to perform a systematic review and meta-analysis to assess the diagnostic performance of FDG PET for the detection of occult malignant disease responsible for PNS. METHODS: A systematic review of the literature (MEDLINE, EMBASE, Cochrane, and DARE) was undertaken to identify studies published in any language. The search strategy was structured after addressing clinical questions regarding the validity or usefulness of the test, following the PICO framework. Inclusion criteria were studies involving patients with PNS in whom FDG PET was performed to detect malignancy, and which reported sufficient primary data to allow calculation of diagnostic accuracy parameters. When possible, a meta-analysis was performed to calculate the joint sensitivity, specificity, and detection rate for malignancy (with 95% confidence intervals [CIs]), as well as a subgroup analysis based on patient characteristics (antibodies, syndrome). RESULTS: The comprehensive literature search revealed 700 references. Sixteen studies met the inclusion criteria and were ultimately selected. Most of the studies were retrospective (12/16). For the quality assessment, the QUADAS-2 tool was applied to assess the risk of bias. Across 16 studies (793 patients), the joint sensitivity, specificity, and detection rate for malignancy with FDG PET were 0.87 (95%
Autores: Pena-Pardo, F. J.; García-VIcente, A.M.; Amo-Salas, M.; et al.
ISSN 1699-048X  Vol. 19  Nº 1  2017  págs. 111 - 118
PURPOSE: To assess the diagnostic impact of 18F-FDG-PET/CT in patients suspected of paraneoplastic neurological syndrome (PNS) based on our own pre-test risk classification (PRC). METHODS: A multicenter retrospective longitudinal study was conducted from 2006 to 2014. We designed a seven-point scoring system using the clinical syndrome characteristics [classical (CS) and non-classical syndromes (NCS)] and its location (central, peripheral, in the neuromuscular junction or combined), onconeural antibodies and tumor markers. Patients were classified as low (score 0-2), intermediate (3-4) and high (5-7) pre-test risk of PNS. FDG-PET/CT was classified as negative or positive. Final diagnosis according Graus' criteria (definite, possible or no PNS) was established. Relations between clinical and metabolic variables with the final diagnosis were studied. RESULTS: 73 patients were included, with a follow-up time of 33 months. Eleven (15 %) patients were finally diagnosed with neoplasm (8 invasive cancers). Ultimately, 13 (18 %) and 24 (33 %) subjects were diagnosed as definite or possible PNS. All the patients with final diagnosis of neoplasm had a CS (p = 0.005). PET/CT was helpful to diagnose 6/8 (75 %) invasive cancers. PET/CT findings were associated with the final diagnosis of neoplasm (p = 0.003) and the diagnosis of PNS attending to Graus' criteria (p = 0.019). PRC showed significant association with the final diagnosis of neoplasm and PET/CT results. A majority of patients (10/11) diagnosed of neoplasm had intermediate/high-risk. CONCLUSIONS: Our PRC seems to be a valid tool to select candidates for PET/CT imaging in this setting. PET/CT detected malignancy in a significant proportion of patients with invasive cancer.
Autores: Casero-Alonso, V.; López-Fidalgo, J.; Torsney, B.;
ISSN 0169-2607  Vol. 138  2017  págs. 105 - 115
BACKGROUND AND OBJECTIVE: Binary response models are used in many real applications. For these models the Fisher information matrix (FIM) is proportional to the FIM of a weighted simple linear regression model. The same is also true when the weight function has a finite integral. Thus, optimal designs for one binary model are also optimal for the corresponding weighted linear regression model. The main objective of this paper is to provide a tool for the construction of MV-optimal designs, minimizing the maximum of the variances of the estimates, for a general design space. METHODS: MV-optimality is a potentially difficult criterion because of its nondifferentiability at equal variance designs. A methodology for obtaining MV-optimal designs where the design space is a compact interval [a, b] will be given for several standard weight functions. RESULTS: The methodology will allow us to build a user-friendly computer tool based on Mathematica to compute MV-optimal designs. Some illustrative examples will show a representation of MV-optimal designs in the Euclidean plane, taking a and b as the axes. The applet will be explained using two relevant models. In the first one the case of a weighted linear regression model is considered, where the weight function is directly chosen from a typical family. In the second example a binary response model is assumed, where the probability of the outcome is given by a typical probability distribution. CONCLUSIONS: Practitioners can use the provided applet to identify the solution and to know the exact support points and design weights.
Autores: Rhee, J. Y.; Garralda, Eduardo; Torrado, C.; et al.
ISSN 1096-6218  Vol. 20  Nº 12  2017  págs. 1372-77
Background: Palliative care (PC) research in Africa has been proposed as a fifth dimension of the World Health Organization PC Public Health Strategy. We conducted a scoping review of published articles (2005-2016) on palliative care development (PCD) in African countries. Forty-seven articles were found across 26 countries. Objective: To study whether the number of published articles on PCD in countries in Africa can be used as an indicator of PCD. Design: This is a secondary analysis of a completed scoping review. Measurements: Spearman correlations were applied to the number of published articles ("published articles") and the number of published articles with a coauthor from a high-income country (HIC) ("HIC published articles") with level of PCD using Lynch et al's updated world map (PC World Map) as a proxy. A subanalysis was undertaken for Anglophone versus non-Anglophone countries. Results: There were positive Spearman correlations (r) between the PC World Map's levels and published articles (r¿=¿0.73; p¿<¿0.001), and with HIC published articles (r¿=¿0.68; p¿<¿0.001). For Anglophone countries, the r was statistically significant (p¿<¿0.001) at 0.69 and 0.70, versus 0.58 and 0.45 for non-Anglophone countries for published articles and HIC published articles, respectively. Kruskal-Wallis test showed a statistically significant difference between Anglophone and non-Anglophone countries for both published articles and HIC published articles (p¿<¿0.01). Conclusion: Publish
Autores: Rivas-López, M.J.; Yu, R.C.; López-Fidalgo, J.; et al.
ISSN 0167-9473  Vol. 113  2017  págs. 363 - 374
The objective is to improve the fatigue characterisation process based on the concept of optimal experimental design. This is carried out through a probabilistic model, previously developed, which takes into account the experimentally observed loading frequency effect on the fatigue life in plain and fibre-reinforced concrete. The Fisher Information Matrix is first obtained for the simplified fatigue model. The optimal design is found to be located at the minimum values allowed for both the maximum stress and stress ratio, whereas the two loading frequencies are the minimum and maximum values in the defined range. Next, the FIM is derived for the extended fatigue model. The previously carried out experimental plan is 65% efficient compared to the optimum. Even though it has been developed for the specific chosen fatigue model, the current procedure can be applied to any other fatigue model to significantly improve the fatigue characterisation process of any material.
Autores: López-Fidalgo, J.; Rodríguez-Hernández, M. M.;
ISSN 0169-7439  Vol. 154  2016  págs. 150 - 161
In a Pharmacokinetic reaction a ligand may be bound to different types of macro-molecules, each with different number of binding sites. They are frequently involved in certain diseases diagnostics. Adair equation is used very often to model the reactions of biological macro-molecules with a ligand. This equation relates the saturation ratio with the free ligand concentration when the equilibrium is reached and it depends on some association constants of the chemical reaction, which have to be estimated. The main problem considered in this paper is the computation of optimal experimental designs for a mixture of Adair models when different types of macromolecules are mixed in the experiment. The main contribution of this work is obtaining the Fisher Information Matrix for a model with a mixture of probability distributions. Since this is not anymore in the Exponential family the expectation cannot be obtained analytically. Then the computation of optimal designs through the information matrix cannot be done with traditional methods. In data analysis when one has the data this expectation can be computed empirically from the data. But in experimental design data are not available when the experiment is being scheduled. Assuming nominal values of the parameters, as is usually done for nonlinear models, simulations were performed for each point in a suitable discretized design space. The number of simulations and the sample size used in each simulation were empirically tuned for both. A sensitivity analysis was performed for different possible true values of the parameters. Since this meant an important computational burden fractional designs were used to cover a reasonable neighborhood of the nominal values of the parameters. In order to display the results in a friendly way for the practitioner, "safe" neighborhoods of the optimal designs are provided.
Autores: Campos-Barreiro, S.; López-Fidalgo, J.;
ISSN 1547-1063  Vol. 13  Nº 1  2016  págs. 67 - 82
The body mass growth of organisms is usually represented in terms of what is known as ontogenetic growth models, which represent the relation of dependence between the mass of the body and time. The paper is concerned with a problem of finding an optimal experimental design for discriminating between two competing mass growth models applied to a beef farm. T-optimality was first introduced for discrimination between models but in this paper, KL-optimality based on the Kullback-Leibler distance is used to deal with correlated obsevations since, in this case, observations on a particular animal are not independent.
Autores: Amo-Salas, M.; Delgado-Marquez, E.; López-Fidalgo, J.;
ISSN 0040-1706  Vol. 58  Nº 2  2016  págs. 269 - 276
This article focuses on analyzing the process of jam formation during the discharge by gravity of granular material stored in a two-dimensional silo. The aim of the article is two-fold. First, optimal experimental designs are computed, in which four approaches are considered: D-optimality, a combination of D-optimality and a cost/gain function, Bayesian D-optimality, and sequential designing. These results reveal that the efficiency of the design used by the experimenters can be improved dramatically. A sensitivity analysis with respect to the most important parameter is also performed. Second, estimation of the unknown parameters is done using least squares, that is, assuming normality, and also via maximum likelihood assuming the exponential distribution. Simulations for the designs considered in this article show that the variance, the mean squared error, and the bias of the estimators using maximum likelihood are in most cases lower than those using least squares.
Autores: Amos-Salas, M.; Delgado-Márquez, E.; Filová, L.; et al.
ISSN 0932-5026  Vol. 57  Nº 4  2016  págs. 875 - 891
During the discharge of a two-dimensional silo, the flow of grains through an opening is arrested if the size of the outlet is not large enough. In the outpouring of grains, jamming occurs due to the formation of an arch at the outlet. After breaking the arch, the grains fall until a new arch is formed. Several models have been proposed to explain this process. The goal of this work is twofold. First, we developed a comparative study of the models proposed by Janda et al. (Europhys Lett 84(4):44002-1¿44002-6, 2008) and To (Phys Rev E 71(6):060301-1¿060301-4, 2005). We have computed D-optimal and c-optimal designs for the most important parameter of the models. Secondly, using the criterion of KL-optimality developed by López-Fidalgo et al. (J R Stat Soc Ser B 69(2):231¿242, 2007), optimal designs for discriminating between these models have been computed.
Autores: Rappold, C.; López-Fidalgo, J.;
ISSN 2469-9985  Vol. 94  Nº 4  2016  págs. 044616
After the demonstration of the feasibility of hypernuclear spectroscopy with heavy-ion beams, the HypHI Collaboration will next focus on the study of proton- and neutron-rich hypernuclei. The use of a fragment separator for the production and separation of rare-isotope beams is a crucial aspect to producing hypernuclei far from the stability line. Precise spectroscopy of exotic hypernuclei is planned to be carried out at the GSI and later at the FAIR facility with the FRS and Super-FRS fragment separators. A systematic study and an optimization analysis were performed to determine optimal experimental conditions for producing hypernuclei with high isospin. The optimal conditions are obtained based on theoretical models for the heavy-ion induced reaction and hypernuclei production. Experimental efficiencies for the production of exotic secondary beams were also taken into account via Monte Carlo simulations of the fragment separator. The developed methodology is presented to deduce the expected yields of Be-8(Lambda) and subsequently other proton-rich and neutron-rich hypernuclei.
Autores: Jiménez-Londoño, G. A.; García-Vicente, A. M.; Poblete-García, V. M.; et al.
ISSN 2253-654X  Vol. 35  Nº 5  2016  págs. 298 - 305
AIM: To analyze the relationship of clinical variables related to prognosis and tumor burden, with metabolic variables obtained in the staging (18)F-FDG PET/CT, and their value in the prognosis in follicular lymphoma (FL). METHODS: 82 patients with FL, a (18)F-FDG PET/CT at diagnosis and a follow-up for a minimum of 12 months, were retrospectively enrolled in the present study. Clinical variables (Tumor grade, Follicular Lymphoma International Prognostic Index (FLIPI) and Tumor burden) were evaluated. Metabolic variables such as SUVmax in the highest hypermetabolic lesion, extralymphatic locations, number of involved lymph node locations, bone marrow (BM) involvement, PET stage and diameter of the biggest hypermetabolic lesion, were analyzed in order to establish a PET score and classify the studies in low, intermediate and high metabolic risk. Clinical and metabolic variables (included metabolic risk) were compared. The relation among all variables and disease-free survival (DFS) was studied. RESULTS: The 28% of patients had a high-grade tumor. The 30.5% had FLIPI risk low, 29.3% intermediate y 40.2% high. The 42.7% presented a high tumor burden. The PET/CT was positive in 94% of patients. The tumor grade did not show significant relation with metabolic variable. FLIPI risk and tumor burden showed statistical relations with the SUV max and the PET score (p<0.008 and p=0.003 respectively). With respect to DFS, significant differences were detected for the PET stage and FLIPI risk (p=0.015 and p=0.047 respectively). FLIPI risk was the only significant predictor in Cox regression analysis, with a Hazard Ratio of 5.13 between high risk and low risk. CONCLUSION: The present research highlights the significant relation between metabolic variables obtained with FDG PET/CT and clinical variables although their goal as an independent factor of prognosis was not demonstrated in the present work.
Autores: Tommasi, C.; Martin-Martin, R.; López-Fidalgo, J.;
ISSN 0960-3174  Vol. 26  Nº 6  2016  págs. 1163 - 1172
In the literature, different optimality criteria have been considered for model identification. Most of the proposals assume the normal distribution for the response variable and thus they provide optimality criteria for discriminating between regression models. In this paper, a max-min approach is followed to discriminate among competing statistical models (i.e., probability distribution families). More specifically, k different statistical models (plausible for the data) are embedded in a more general model, which includes them as particular cases. The proposed optimal design maximizes the minimum KL-efficiency to discriminate between each rival model and the extended one. An equivalence theorem is proved and an algorithm is derived from it, which is useful to compute max-min KL-efficiency designs. Finally, the algorithm is run on two illustrative examples.
Autores: García-Vicente, A.; Soriano-Castrejón, A.; Amo-Salas, M.; et al.
ISSN 2253-654X  Vol. 35  Nº 3  2016  págs. 152 - 158
Aim To explore the relationship between basal 18F-FDG uptake in breast tumors and survival in patients with breast cancer (BC) using a molecular phenotype approach. Material and Methods This prospective and multicentre study included 193 women diagnosed with BC. All patients underwent an 18F-FDG PET/CT prior to treatment. Maximum standardized uptake value (SUVmax) in tumor (T), lymph nodes (N), and the N/T index was obtained in all the cases. Metabolic stage was established. As regards biological prognostic parameters, tumors were classified into molecular sub-types and risk categories. Overall survival (OS) and disease free survival (DFS) were obtained. An analysis was performed on the relationship between semi-quantitative metabolic parameters with molecular phenotypes and risk categories. The effect of molecular sub-type and risk categories in prognosis was analyzed using Kaplan¿Meier and univariate and multivariate tests. Results Statistical differences were found in both SUVT and SUVN, according to the molecular sub-types and risk classifications, with higher semi-quantitative values in more biologically aggressive tumors. No statistical differences were observed with respect to the N/T index. Kaplan¿Meier analysis revealed that risk categories were significantly related to DFS and OS. In the multivariate analysis, metabolic stage and risk phenotype showed a significant association with DFS. Conclusion High-risk phenotype category showed a worst prognosis with respect to the other categories with higher SUVmax in primary tumor and lymph nodes.
Autores: Stehlik, M.; López-Fidalgo, J.; Casero-Alonso, V.; et al.
ISSN 1436-3240  Vol. 29  Nº 2  2015  págs. 379 - 395
Optimal design is a crucial issue in Environmental measurement with typical time-space correlated observations. A modified Arrhenius model with a particular correlation structure will be applied to the methane removal in the atmosphere, a very important environmental issue at this moment. We introduce a class of integrated compound criteria for obtaining robust designs. In particular, the paper provides an insight into the relationship of a compound D-optimality criterion for both the trend and covariance parameters, and the Integrated Mean Squared Prediction Error (IMSPE) criterion. In general, if there are two or more approaches of a given problem, e.g. two rival models or two different parts of a model, an integral relationship may be constructed with the aim of finding a suitable compromise between them. The Fisher information matrix (FIM) will be used in both cases. Then the integral compound criterion with respect to a density from a given parametric family of distributions is optimized. We also discuss some general conditions around the behavior of the introduced approach for comparing the FIMs and provide computing methods.
Autores: García-Vicente, A.; López-Fidalgo, J.; Soriano-Castrejón, Á.; et al.
ISSN 1619-7070  Vol. 42  Nº 12  2015  págs. 1804 - 1813
Aim: To explore the relationship between basal (18)¿F-FDG PET/CT information in breast tumours and survival in locally advanced breast cancer (LABC). Methods: This prospective, multicentre study included 198 women diagnosed with LABC. All patients underwent (18)¿F-FDG PET/CT prior to treatment. The maximum standardized uptake value (SUVmax) in tumor (T), lymph nodes (N) and the N/T ratio was obtained in all cases. Stage according to PET/CT imaging (metabolic stage) and conventional imaging techniques (clinical stage) was established. During follow-up, patient status was established (disease free status or not). The relationship between all the variables and overall survival (OS) and disease-free survival (DFS) was analysed using the Kaplan-Meier and Cox regression methods. A ROC analysis was performed to obtain a cut-off value of SUVmax that was useful in the prediction of outcome. Results: The mean SUVmax¿±¿SD values in the primary tumour, lymph nodes and the SUVmax N/T index were 7.40¿±¿5.57, 4.17¿±¿4.74 and 0.73¿±¿1.20, respectively. Higher semiquantitative metabolic values were found in more advanced metabolic and clinical stages. During follow-up, 78.4 % of patients were free of disease. Significant relationships were observed between SUVT and SUVN and patient status. With respect to OS and DFS, significant differences were detected for the metabolic stage. Kaplan-Meier analysis revealed that using the cut-off values, a primary-tumour SUVmax¿¿¿6.05 or a nodal SUVmax ¿2.25 were significantly correlated with DFS and OS. Conclusion: PET imaging with (18)¿F-FDG offers prognostic information for LABC that can be obtained preoperatively and noninvasively.
Autores: Casero-Alonso, V.; López-Fidalgo, J.;
Revista: TEST
ISSN 1133-0686  Vol. 24  Nº 4  2015  págs. 701 - 713
A procedure based on a multiplicative algorithm for computing optimal experimental designs subject to cost constraints in simultaneous equations models is presented. A convex criterion function based on a usual criterion function and an appropriate cost function is considered. A specific L-optimal design problem and a numerical example are taken from Conlisk (J Econ 11:63¿76, 1979) to compare the procedure. The problem would need integer nonlinear programming to obtain exact designs. To avoid this, he solves a continuous nonlinear programming problem and then he rounds off the number of replicates of each experiment. The procedure provided in this paper reduces dramatically the computational efforts in computing optimal approximate designs. It is based on a specific formulation of the asymptotic covariance matrix of the full-information maximum likelihood estimators, which simplifies the calculations. The design obtained for estimating the structural parameters of the numerical example by this procedure is not only easier to compute, but also more efficient than the design provided by Conlisk.
Autores: Amo-Salas, M.; López-Fidalgo, J.; Pedregal, D. J.;
ISSN 0951-8320  Vol. 133  2015  págs. 87 - 94
Some time series applications require data which are either expensive or technically difficult to obtain. In such cases scheduling the points in time at which the information should be collected is of paramount importance in order to optimize the resources available. In this paper time series models are studied from a new perspective, consisting in the use of Optimal Experimental Design setup to obtain the best times to take measurements, with the principal aim of saving costs or discarding useless information. The model and the covariance function are expressed in an explicit form to apply the usual techniques of Optimal Experimental Design. Optimal designs for various approaches are computed and their efficiencies are compared. The methods working in an application of industrial maintenance of a critical piece of equipment at a petrochemical plant are shown. This simple model allows explicit calculations in order to show openly the procedure to find the correlation structure, needed for computing the optimal experimental design. In this sense the techniques used in this paper to compute optimal designs may be transferred to other situations following the ideas of the paper, but taking into account the increasing difficulty of the procedure for more complex models.
Autores: Campos-Barreiro, S.; López-Fidalgo, J.;
ISSN 1618-2510  Vol. 24  Nº 3  2015  págs. 491 - 505
The body mass growth of organisms is usually represented in terms of what is known as ontogenetic growth models, which represent the relation of dependence between the mass of the body and time. This paper discusses design issues of West's ontogenetic growth model applied to a Holstein-Friesian dairy farm in the northwest of Spain. D-optimal experimental designs were computed to obtain an optimal fitting of the model. A correlation structure has been included in the statistical model due to the fact that observations on a particular animal are not independent. The choice of a robust correlation structure is an important contribution of this paper; it provides a methodology that can be used for any correlation structure. The experimental designs undertaken provide a tool to control the proper weight of heifers, which will help improve their productivity and, by extension, the competitiveness of the dairy farm.
Autores: Casero-Alonso, V. M.; López-Fidalgo, J.;
ISSN 0932-5026  Vol. 56  Nº 2  2015  págs. 273 - 290
Optimal experimental designs are considered for models with simultaneous equations. In particular, a model with two equations is assumed where one of the explanatory variables (exogenous) of the first equation is then the response variable (endogenous) of the second equation. In both equations there is a control variable, which is being designed through the celebrated D-optimality criterion. This work is based on a more restricted approach using just the first equation and assuming the distribution of the exogenous/endogenous variable completely known. Then a conditionally restricted optimal design was computed afterwards. In this paper the conditional model is assumed partially known but it has to be fitted as well. Although both approaches identify different prior knowledge a comparison of the optimal designs for both approaches is made. Since the model is not linear in the usual sense the optimal designs will depend on the parameters and a sensitivity analysis against its choice is performed.
Autores: López-Fidalgo, J.; Rodríguez-Hernández, M. M.;
ISSN 0169-7439  Vol. 138  2014  págs. 133 - 141
The Adair equation is used to model biological macro-molecule reactions. This equation relates the saturation rate to the free ligand concentration. But, the latter is not a variable completely under the control of the experimenter. The ligand is a random variable depending on an initial ligand added by the experimenter, which can be designed, but the dependence of the saturation rate on the initial ligand has not been considered in the literature. In this paper a transformed model based on the Adair model of first order (monomer) is derived in order to obtain a proper model that depends on the initial ligand. This model will allow proper fitting and optimal designs using the initial ligand. It will be called the transformed Adair model (TAM). Optimal designs as well as seven-point quasi-optimal designs forced to follow a harmonic, geometric or uniform progression, are computed. The parameters are estimated and compared for simulated data from these designs. A sensitivity analysis against the choice of nominal values of the parameters is also performed for the TAM. The analytic version of the transformed model is only possible for the first class model. But the good efficiencies of the optimal designs obtained directly from the monomer model for fitting the TAM justify doing something similar for the second order model (dimer). Designs were computed numerically in this case.
Autores: Dragalin, V.; López-Fidalgo, J.;
ISSN 0323-3847  Vol. 56  Nº 5  2014  págs. 730 - 731
Autores: López-Fidalgo, J.; Rivas-López, M. J.;
ISSN 0167-9473  Vol. 71  Nº SI  2014  págs. 859 - 867
In a follow-up study the time-to-event may be censored either because of dropouts or the end of study is earlier. This situation is frequently modeled through a Cox-proportional hazards model including covariates, some of which are under the control of the experimenter. When the model is to be fitted to n observed times these are known and for each of them it is also known whether that time is censored or not. When the experiment is to be designed neither the observed times nor the information about whether a particular unit will be censored are known. For censoring some additional prior probability distribution has to be assumed. Thus, the design problem faces two sources of imprecision when the experiment is to be scheduled. On the one hand, the censored times are not known. On the other hand, there is uncertainty about occurrence of censoring. A prior probability distribution is needed for this. Moreover, the Cox partial likelihood instead of the full likelihood is usually considered for these models. A partial information matrix is built in this case and optimal designs are computed and compared with the optimal designs for the full likelihood information. The usual tools for computing optimal designs with full likelihood are no longer valid for partial information. Some general results are provided in order to deal with this new approach. An application to a simple case with two possible treatments is used to illustrate it. The partial information matrix depends on the parameters and therefore a sensitivity analysis is conducted in order to check the robustness of the designs for the choice of the nominal values of the parameters.
Autores: Rivas-López, M. M.; López-Fidalgo, J.; Del-Campo, R.;
ISSN 0323-3847  Vol. 56  Nº 5  2014  págs. 819 - 837
Proportional Hazards models have been widely used to analyze survival data. In many cases survival data do not verify the assumption of proportional hazards. An alternative to the PH models with more relaxed conditions are Accelerated Failure Time models. These models are fairly commonly used in the field of manufacturing, but they are more and more frequent for modeling clinical trial data. They focus on the direct effect of the explanatory variables on the survival function allowing an easier interpretation of the effect of the corresponding covariates on the survival time. Optimal experimental designs are computed in this framework for Type I and random arrival. The results are applied to clinical models used to prevent tuberculosis in Ugandan adults infected with HIV.
Autores: Amo-Salas, M.; Arroyo-Jiménez, M. M.; Bustos-Escribano, D.; et al.
ISSN 1687-952X  Vol. 2014  2014  págs. 240263
Multiple choice questions (MCQs) are one of the most popular tools to evaluate learning and knowledge in higher education. Nowadays, there are a few indices to measure reliability and validity of these questions, for instance, to check the difficulty of a particular question (item) or the ability to discriminate from less to more knowledge. In this work two new indices have been constructed: (i) the no answer index measures the relationship between the number of errors and the number of no answers; (ii) the homogeneity index measures homogeneity of the wrong responses (distractors). The indices are based on the lack-of-fit statistic, whose distribution is approximated by a chi-square distribution for a large number of errors. An algorithm combining several traditional and new indices has been developed to refine continuously a database of MCQs. The final objective of this work is the classification of MCQs from a large database of items in order to produce an automated-supervised system of generating tests with specific characteristics, such as more or less difficulty or capacity of discriminating knowledge of the topic.
Autores: Biswas, A.; López-Fidalgo, J.;
ISSN 1539-1604  Vol. 12  Nº 2  2013  págs. 92 - 101
Compound optimal designs are considered where one component of the design criterion is a traditional optimality criterion such as the D-optimality criterion, and the other component accounts for higher efficacy with low toxicity. With reference to the dose-finding problem, we suggest the technique to choose weights for the two components that makes the optimization problem simpler than the traditional penalized design. We allow general bivariate responses for efficacy and toxicity. We then extend the procedure in the presence of nondesignable covariates such as age, sex, or other health conditions. A new breast cancer treatment is considered to illustrate the procedures.
Autores: Amo-Salas, M.; López-Fidalgo, J.; Porcu, E.;
Revista: TEST
ISSN 1133-0686  Vol. 22  Nº 1  2013  págs. 159 - 187
This paper considers optimal experimental designs for models with correlated observations through a covariance function depending on the magnitude of the responses. This suggests the use of stochastic processes whose covariance structure is a function of the mean. Covariance functions must be positive definite. This fact is nontrivial in this context and constitutes one of the challenges of the present paper. We show that there exists a huge class of functions that, composed with the mean of the process in some way, preserves positive definiteness and can be used for the purposes of modeling and computing optimal designs in more realistic situations. We offer some examples for an easy construction of such covariances and then study the problem of locally D-optimal designs through an illustrative example as well as a real radiation retention model in the human body.
Autores: Campos-Barreiro, S.; López-Fidalgo, J.;
ISSN 1742-4682  Vol. 10  2013  págs. 21
Background: The pathology of the Benign Paroxysmal Positional Vertigo (BPPV) is detected by a clinician through maneuvers consisting of a series of consecutive head turns that trigger the symptoms of vertigo in patient. A statistical model based on a new maneuver has been developed in order to calculate the volume of endolymph displaced after the maneuver. Methods: A simplification of the Navier-Stokes problem from the fluids theory has been used to construct the model. In addition, the same cubic splines that are commonly used in kinematic control of robots were used to obtain an appropriate description of the different maneuvers. Then experimental designs were computed to obtain an optimal estimate of the model. Results: D-optimal and c-optimal designs of experiments have been calculated. These experiments consist of a series of specific head turns of duration Delta t and angle alpha that should be performed by the clinician on the patient. The experimental designs obtained indicate the duration and angle of the maneuver to be performed as well as the corresponding proportion of replicates. Thus, in the D-optimal design for 100 experiments, the maneuver consisting of a positive 30 degrees pitch from the upright position, followed by a positive 30 degrees roll, both with a duration of one and a half seconds is repeated 47 times. Then the maneuver with 60 degrees/60 degrees pitch/roll during half a second is repeated 16 times and themaneuver 90 degrees/90 degrees pitch/roll during half a second is repeated 37 times. Other designs with significant differences are computed and compared. Conclusions: A biomechanical model was derived to provide a quantitative basis for the detection of BPPV. The robustness study for the D-optimal design, with respect to the choice of the nominal values of the parameters, shows high efficiencies for small variations and provides a guide to the researcher. Furthermore, c-optimal designs give valuable assistance to check how efficient the D-optimal design is for the estimation of each of the parameters. The experimental designs provided in this paper allow the physician to validate the model. The authors of the paper have held consultations with an ENT consultant in order to align the outline more closely to practical scenarios.
Autores: Amo-Salas, M.; López-Fidalgo, J.; López-Rios, V.;
ISSN 0361-0918  Vol. 41  Nº 7  2012  págs. 944 - 963
Two nested pharmacokinetic models are considered in this article. Several observations are taken on the same subject so they are correlated. The covariance function assumed is an exponential covariance function. Optimal exact designs are computed with different criteria, both for discriminating between models and estimating parameters. Compound criteria to estimate the parameters and nonlinear functions of the parameters are used. An iterative algorithm based on T-optimality and an algorithm from Brimkulov et al. (1986) are combined in order to compute T-optimal designs with correlated observations. Finally, compound designs to discriminate between the models and estimate the nonlinear functions are considered. A test power study is performed to adjust the compound parameter.
Autores: Feo-Brito, F.; Mur-Gimeno, P.; Carnés, J.; et al.
ISSN 1081-1206  Vol. 106  Nº 2  2011  págs. 146 - 152
Background: Allergic symptoms are commonly related to atmospheric pollen counts in sensitized allergic individuals. However, concordance between symptoms, pollen counts, and aeroallergen concentrations is not always good. Objectives: To determine the correlation between olive pollen counts, aeroallergen levels, and clinical symptoms in patients with allergic asthma or rhinitis in Ciudad Real (Spain). Methods: Two types of samplers were used to determine pollen exposure: a Burkard spore trap to collect pollen grains and a high-volume air sampler to collect airborne particles. A total of 366 air filters were collected. After extraction, they were analyzed by specific immunoglobulin E enzyme-linked immunosorbent assay inhibition using a serum pool containing high titers of olive-specific immunoglobulin E. Twenty olive-pollen monosensitized patients were asked to record their daily symptoms before, during, and after the olive pollen season. Results: Olive pollen was detected between April 21 and June 30, 2004. Symptoms showed positive and significant correlations with pollen counts (r = 0.700, P < .001) and aeroallergen levels (r = 0.803, P < .001). Using a Poisson regression model, relative changes in aeroallergen concentrations and pollen counts were found to be similar and significant. Threshold levels for the induction of symptoms were 162 olive pollen grains/m(3) and 22.7 ng of olive pollen allergen/m(3) (equivalent to 0.9 ng/m(3) of Ole e 1). Conclusions: Olive aeroallergen concentrations and pollen counts are positively associated with symptoms of rhinitis and asthma in olive-allergic patients. Both data may be used in the clinical follow-up of these patients.
Autores: López-Fidalgo, J.; Ortiz-Rodríguez, I. M.; Wong, W. K.;
ISSN 0266-4763  Vol. 38  Nº 3  2011  págs. 501 - 512
We briefly review and discuss design issues for population growth and decline models. We then use a flexible growth and decline model as an illustrative example and apply optimal design theory to find optimal sampling times for estimating model parameters, specific parameters and interesting functions of the model parameters for the model with two real applications. Robustness properties of the optimal designs are investigated when nominal values or the model is mis-specified, and also under a different optimality criterion. To facilitate use of optimal design ideas in practice, we also introduce a website for generating a variety of optimal designs for popular models from different disciplines.
Autores: Arévalo-Villena, M.; Fernández-Guerrero, M. M.; López-Fidalgo, J.; et al.
ISSN 2156-8456  Vol. 2  Nº 2  2011  págs. 89 - 96
Pectinases are used in Enology for some different utilities. Enzymatic preparations from moulds are a mixed of different enzymes with strong and unspe-cific activities. Some Saccharomyces cerevisiae pro-duce pectinases which can be used instead of com-mercial preparations. The objectives of this work were to study the enzyme secretion by one Saccharo-myces cerevisiae (CECT 11783) for growing on grape skin (industry oenological by-product) as carbon source. Preliminary experiments showed that the strain produced pectinases for growing on grape skin without any other carbon source. Statistical treat-ment (factorial design 25) was applied to evaluate the influences of related factors (agitation, temperature, presence of peptone and detergent in the medium and time of growth) Variables with the most significant interactions for pectinase production were agitation and nitrogen source concentration. Response surface methodology showed that a first order model was not adequate for results. Nevertheless, the built of a sec-ond order model offered a polynomial equation which surface predicted a maximum of activity (52.68 enzymatic units) for specific values of the studied variables (147.8 rpm of agitation and 15.9 g of pep-tone/ L culture medium).
Autores: López-Fidalgo, J.; Garcet-Rodríguez, S.;
ISSN 0233-1888  Vol. 45  Nº 2  2011  págs. 143 - 154
This paper considers the problem of constructing optimal approximate designs when an independent variable might be censored. The problem is which design should be applied in practice to obtain the best approximate design when a censoring distribution is assumed known in advance. The approach for finite or continuous design spaces deserves different attention. In both cases, equivalent theorems and algorithms are provided in order to calculate optimal designs. Some examples illustrate this approach for D-optimality.
Autores: Amo-Salas, M.; López-Fidalgo, J.; Rodríguez-Díaz, J. M.;
ISSN 1539-1604  Vol. 9  Nº 1  2010  págs. 55 - 66
The model that describes the retention in lungs of radioisotope particles is studied in this paper, considering the situation of an accident in facilities that handle radioactive materials. Optimal times to make the bioassays are computed for D- and c-optimality, and efficiencies for the computed designs are provided and compared. Moreover, the test power is checked by means of simulations and replications. After that the inverse of the Fisher information matrix is compared to an estimation of the covariance matrix of the parameters. Finally, a study taking into consideration the randomness of the designs space is performed.
Autores: Tommasi, C.; López-Fidalgo, J.;
ISSN 0167-9473  Vol. 54  Nº 1  2010  págs. 143 - 150
The Bayesian KL-optimality criterion is useful for discriminating between any two statistical models in the presence of prior information. If the rival models are not nested then, depending on which model is true, two different Kullback-Leibler distances may be defined. The Bayesian KL-optimality criterion is a convex combination of the expected values of these two possible Kullback-Leibler distances between the competing models. These expectations are taken over the prior distributions of the parameters and the weights of the convex combination are given by the prior probabilities of the models. Concavity of the Bayesian KL-optimality criterion is proved, thus classical results of Optimal Design Theory can be applied. A standardized version of the proposed criterion is also given in order to take into account possible different magnitudes of the two Kullback-Leibler distances. Some illustrative examples are provided.
Autores: Amo-Salas, M.; Jiménez-Alcazar, A.; López-Fidalgo, J.;
Libro:  MODA11-Advances in model oriented design and analysis
2016  págs. 11 - 18
In this paper the tools provided by the theory of the optimal design of experiments are applied to a model where the function is given in implicit form. This work is motivated by a dosimetry problem, where the dose, the controllable variable, is expressed as a function of the observed value from the experiment. The best doses will be computed in order to obtain precise estimators of the parameters of the model. For that, the inverse function theorem will be used to obtain the Fisher information matrix. Properly the D-optimal design must be obtained directly on the dose using the inverse function theorem. Alternatively a fictitious D-optimal design on the observed values can be obtained in the usual way. Then this design can be transformed through the model into a design on the doses. Both designs will be computed and compared for a real example. Moreover, different optimal sequences and their D-effiencies will be computed as well. Finally, c-optimal designs for the parameters of the model will be provided.