Joel Martínez-Salazar
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Joel Martínez-Salazar has a total of 12 co-authors. Here are the top co-authors by number of shared publications:
- Filiberto Toledano-Toledano
Unidad de Investigación en Medicina Basada en Evidencias, Hospital Infantil de México Federico Gómez, Instituto Nacional de Salud, Dr. Márquez 162, Doctores, Cuauhtémoc, Mexico City 06720, Mexico
3 shared publications - David Luna
Unidad de Investigación Multidisciplinaria en Salud, Instituto Nacional de Rehabilitación Luis Guillermo Ibarra Ibarra, Calzada México-Xochimilco 289, Arenal de Guadalupe, Tlalpan, Mexico City 14389, Mexico
2 shared publications - Rosa Paola Figuerola-Escoto
Centro Interdisciplinario de Ciencias de la Salud, Unidad Santo Tomás, Instituto Politécnico Nacional, Av. de Los Maestros s/n, Santo Tomás, Miguel Hidalgo, Mexico City 11340, Mexico
2 shared publications - Ahidée Leyva-López
Center for Population Health Research, National Institute of Public Health, Cuernavaca 62100, Mexico
1 shared publication - Ernesto Roldan-Valadez
Division of Research, Instituto Nacional de Rehabilitacion Luis Guillermo Ibarra Ibarra, 14389, Mexico City, Mexico; I.M. Sechenov First Moscow State Medical University (Sechenov University), Department of Radiology, 119992, Moscow, Russia
1 shared publication - Hugo Sandoval
General Directorate, National Institute of Rehabilitation Luis Guillermo Ibarra Ibarra, Ciudad de México, México
1 shared publication
Latest Publications
- Open Access
- 28 October 2026
- Published by MDPI AG in Medical Sciences
- Vol. 14 (7) , 722
Psychometric Properties of the Abbreviated Version of the Center for Epidemiologic Studies Depression Scale (CES-D-7) in the Mexican Population
Background: Depression is a public health problem that requires early detection. The CES-D-7 is a brief instrument designed to screen for depressive symptoms, which may help identify individuals who could benefit from further clinical evaluation. Objective: To identify the psychometric properties of the CES-D-7 in the Mexican population. Methods: Based on the Encuesta Nacional de Salud y Nutrición Continua conducted from 2021 to 2023, responses to the CES-D-7 were obtained from 32,076 participants. Its structure was analyzed using confirmatory factor analysis. The mean factor loading (Mλ), average variance extracted (AVE), and factorial invariance across sex (men vs. women), age group (20–59 vs. ≥60 years), year (2021 vs. 2022 vs. 2023), and 8 different regions in Mexico were estimated. Reliability was assessed using Cronbach’s α, McDonald’s ω, composite reliability (CR), and the H coefficient. Results: Model fit was excellent according to CFI = 0.982, TLI = 0.973, and SRMR = 0.020, and was at the threshold of acceptability according to RMSEA = 0.080 (90% CI: 0.077–0.084). The mean factor loading (Mλ = 0.751) and average variance extracted (AVE = 0.602) were also acceptable. Reliability was acceptable: Cronbach’s α = 0.824 (95% CI: 0.820–0.828); ordinal McDonald’s ω = 0.847 (95% CI: 0.843–0.850); CR = 0.908; and H = 0.951. Factorial invariance was identified across sex, age group, year, and region. Conclusions: The CES-D-7 demonstrates solid psychometric properties supported by evidence of construct validity, convergence among indicators within the measurement model, reliability, and factorial invariance in a Mexican sample aged 20 years and older.
- 28 October 2026
- Published by MDPI AG in Medical Sciences
- Vol. 14 (7) , 722
Open AccessPsychometric Properties of the Abbreviated Version of the Center for Epidemiologic Studies Depression Scale (CES-D-7) in the Mexican Population
Background: Depression is a public health problem that requires early detection. The CES-D-7 is a brief instrument designed to screen for depressive symptoms, which may help identify individuals who could benefit from further clinical evaluation. Objective: To identify the psychometric properties of the CES-D-7 in the Mexican population. Methods: Based on the Encuesta Nacional de Salud y Nutrición Continua conducted from 2021 to 2023, responses to the CES-D-7 were obtained from 32,076 participants. Its structure was analyzed using confirmatory factor analysis. The mean factor loading (Mλ), average variance extracted (AVE), and factorial invariance across sex (men vs. women), age group (20–59 vs. ≥60 years), year (2021 vs. 2022 vs. 2023), and 8 different regions in Mexico were estimated. Reliability was assessed using Cronbach’s α, McDonald’s ω, composite reliability (CR), and the H coefficient. Results: Model fit was excellent according to CFI = 0.982, TLI = 0.973, and SRMR = 0.020, and was at the threshold of acceptability according to RMSEA = 0.080 (90% CI: 0.077–0.084). The mean factor loading (Mλ = 0.751) and average variance extracted (AVE = 0.602) were also acceptable. Reliability was acceptable: Cronbach’s α = 0.824 (95% CI: 0.820–0.828); ordinal McDonald’s ω = 0.847 (95% CI: 0.843–0.850); CR = 0.908; and H = 0.951. Factorial invariance was identified across sex, age group, year, and region. Conclusions: The CES-D-7 demonstrates solid psychometric properties supported by evidence of construct validity, convergence among indicators within the measurement model, reliability, and factorial invariance in a Mexican sample aged 20 years and older.
- Open Access
- 27 October 2026
- Published by MDPI AG in Medical Sciences
- Vol. 14 (7) , 713
Mortality Prediction in Imbalanced Pediatric Oncology Data: A Cost-Sensitive Hybrid Resampling Framework Using Recurrent Deep Learning
Background: Severe class imbalance hampers mortality prediction in pediatric oncology, where clinical-improvement discharges greatly outnumber deaths and bias standard models toward the majority class. Objectives: We aimed to develop and evaluate a Cost-Sensitive Hybrid Resampling Framework (CSHRF) across recurrent deep-learning architectures to mitigate this bias and improve discharge-outcome classification. Methods: A retrospective cross-sectional study was conducted using electronic health record data from a National Institute of Health in Mexico City during 2023. We analyzed 1607 hospital-discharge records, comprising 1554 clinical improvements and 53 deaths. The CSHRF combines cost-sensitive learning, which amplified the mortality class by a factor of 2.5, with a fixed-seed hybrid resampling protocol embedded in a regularized funnel-shaped architecture (N → N/2 → N/4) trained with gradient clipping (clipnorm = 1.0). Simple Recurrent Neural Network (SimpleRNN), Long Short-Term Memory (LSTM), and Gated Recurrent Unit (GRU) models were compared with a naïve recurrent baseline and standard tabular classifiers (Weighted Logistic Regression, Balanced Random Forest, and XGBoost) using stratified cross-validation. Performance was evaluated using the area under the receiver operating characteristic curve (AUC), G-mean, sensitivity, specificity, precision, F1-score, mean absolute error (MAE), and root mean squared error (RMSE). Results: The naïve baseline showed limited mortality detection (sensitivity 0.150 ± 0.147). The proposed framework substantially improved minority-class detection, with CSHRF-GRU achieving sensitivity of 0.740 ± 0.145, specificity of 0.700 ± 0.076, G-mean of 0.714 ± 0.045, and AUC of 0.761 ± 0.075. Ablation analysis demonstrated that cost-sensitive weighting contributed most substantially to mortality detection, whereas hybrid resampling alone produced more modest gains. Among tabular baselines, Weighted Logistic Regression achieved competitive performance, with sensitivity of 0.684 ± 0.130 and G-mean of 0.767 ± 0.073. Precision remained low (0.079 ± 0.015) and probabilistic calibration errors were substantial (MAE 0.448; RMSE 0.488), reflecting the trade-off required to prioritize mortality detection under severe imbalance. Conclusions: The CSHRF improved discharge-outcome classification in a highly imbalanced pediatric oncology dataset and provides a methodological framework for data-constrained hospital environments. Given its high false-positive rate and probability overestimation, the approach should be interpreted as a discharge-related outcome classification model rather than a standalone clinical decision-support system, and future work should incorporate richer clinical and disease-specific variables to improve predictive performance and calibration.
- 27 October 2026
- Published by MDPI AG in Medical Sciences
- Vol. 14 (7) , 713
Open AccessMortality Prediction in Imbalanced Pediatric Oncology Data: A Cost-Sensitive Hybrid Resampling Framework Using Recurrent Deep Learning
Background: Severe class imbalance hampers mortality prediction in pediatric oncology, where clinical-improvement discharges greatly outnumber deaths and bias standard models toward the majority class. Objectives: We aimed to develop and evaluate a Cost-Sensitive Hybrid Resampling Framework (CSHRF) across recurrent deep-learning architectures to mitigate this bias and improve discharge-outcome classification. Methods: A retrospective cross-sectional study was conducted using electronic health record data from a National Institute of Health in Mexico City during 2023. We analyzed 1607 hospital-discharge records, comprising 1554 clinical improvements and 53 deaths. The CSHRF combines cost-sensitive learning, which amplified the mortality class by a factor of 2.5, with a fixed-seed hybrid resampling protocol embedded in a regularized funnel-shaped architecture (N → N/2 → N/4) trained with gradient clipping (clipnorm = 1.0). Simple Recurrent Neural Network (SimpleRNN), Long Short-Term Memory (LSTM), and Gated Recurrent Unit (GRU) models were compared with a naïve recurrent baseline and standard tabular classifiers (Weighted Logistic Regression, Balanced Random Forest, and XGBoost) using stratified cross-validation. Performance was evaluated using the area under the receiver operating characteristic curve (AUC), G-mean, sensitivity, specificity, precision, F1-score, mean absolute error (MAE), and root mean squared error (RMSE). Results: The naïve baseline showed limited mortality detection (sensitivity 0.150 ± 0.147). The proposed framework substantially improved minority-class detection, with CSHRF-GRU achieving sensitivity of 0.740 ± 0.145, specificity of 0.700 ± 0.076, G-mean of 0.714 ± 0.045, and AUC of 0.761 ± 0.075. Ablation analysis demonstrated that cost-sensitive weighting contributed most substantially to mortality detection, whereas hybrid resampling alone produced more modest gains. Among tabular baselines, Weighted Logistic Regression achieved competitive performance, with sensitivity of 0.684 ± 0.130 and G-mean of 0.767 ± 0.073. Precision remained low (0.079 ± 0.015) and probabilistic calibration errors were substantial (MAE 0.448; RMSE 0.488), reflecting the trade-off required to prioritize mortality detection under severe imbalance. Conclusions: The CSHRF improved discharge-outcome classification in a highly imbalanced pediatric oncology dataset and provides a methodological framework for data-constrained hospital environments. Given its high false-positive rate and probability overestimation, the approach should be interpreted as a discharge-related outcome classification model rather than a standalone clinical decision-support system, and future work should incorporate richer clinical and disease-specific variables to improve predictive performance and calibration.
- Open Access
Comparative Analysis of Three Predictive Models of Performance Indicators with Results-Based Management: Cancer Data Statistics in a National Institute of Health
Predictive models play a crucial role in RBMs to analyze performance indicator results to manage unexpected events and make timely decisions to resolve them. Their use in Mexico is deficient, and monitoring and evaluation are among the weakest pillars of the model. In response to these needs, the aim of this study was to perform a comparative analysis of three predictive models to analyze 10 medical performance indicators and cancer data related to children with cancer. To accomplish these purposes, a comparative and retrospective study with nonprobabilistic convenience sampling was conducted. The predictive models were exponential smoothing, autoregressive integrated moving average, and linear regression. The lowest mean absolute error was used to identify the best model. Linear regression performed best regarding nine of the ten indicators, with seven showing p < 0.05. Three of their assumptions were checked using the Shapiro–Wilk, Cook’s distance, and Breusch–Pagan tests. Predictive models with RBM are a valid and relevant instrument for monitoring and evaluating performance indicator results to support forecasting and decision-making based on evidence and must be promoted for use with cancer data statistics. The place numbers obtained by cancer disease inside the main causes of death, morbidity and hospital outpatients in a National Institute of Health were presented as evidence of the importance of implementing performance indicators associated with children with cancer.
Open AccessComparative Analysis of Three Predictive Models of Performance Indicators with Results-Based Management: Cancer Data Statistics in a National Institute of Health
Predictive models play a crucial role in RBMs to analyze performance indicator results to manage unexpected events and make timely decisions to resolve them. Their use in Mexico is deficient, and monitoring and evaluation are among the weakest pillars of the model. In response to these needs, the aim of this study was to perform a comparative analysis of three predictive models to analyze 10 medical performance indicators and cancer data related to children with cancer. To accomplish these purposes, a comparative and retrospective study with nonprobabilistic convenience sampling was conducted. The predictive models were exponential smoothing, autoregressive integrated moving average, and linear regression. The lowest mean absolute error was used to identify the best model. Linear regression performed best regarding nine of the ten indicators, with seven showing p < 0.05. Three of their assumptions were checked using the Shapiro–Wilk, Cook’s distance, and Breusch–Pagan tests. Predictive models with RBM are a valid and relevant instrument for monitoring and evaluating performance indicator results to support forecasting and decision-making based on evidence and must be promoted for use with cancer data statistics. The place numbers obtained by cancer disease inside the main causes of death, morbidity and hospital outpatients in a National Institute of Health were presented as evidence of the importance of implementing performance indicators associated with children with cancer.