RESUMO
Background: Household food insecurity (HFI) increased in Latin America by 9% between 2019 and 2020. Scant evidence shows who was unable to recover from the COVID-19 pandemic. Our aim was to use a Machine Learning (ML) approach to identify consistent and influential predictors of persistent moderate or severe HFI over 2 years. Methods: We use a three-wave longitudinal telephone survey with a probabilistic sample representative of the Mexican population. With a response rate of 51.3 and 60.8% for the second and third waves, the final sample size consisted of 1,074 individuals. The primary outcome was persistent HFI, i.e., respondents who reported moderate or severe HFI in 2021 and 2022. Twelve income-related predictors were measured in 2020, including baseline HFI. We employed 6 supervised ML algorithms to cross-validate findings in models, examined its precision with 4 standard performance indicators to assess precision, and used SHAP values (Shapley Additive exPlanations) to identify influential predictors in each model. Results: Prevalence of persistent moderate/severe HFI in 2021 and 2022 was 8.8%. Models with only a HFI 2020 baseline measure were used as a reference for comparisons; they had an accuracy of 0.79, a Cohen's Kappa of 0.57, a sensitivity of 0.68, and a specificity of 0.88. When HFI was substituted by the suite of socioeconomic indicators, accuracy ranged from 0.70 to 0.84, Cohen's Kappa from 0.40 to 0.67, sensitivity from 0.86 to 0.90, and specificity from 0.75 to 0.82. The best performing models included baseline HFI and socioeconomic indicators; they had an accuracy between 0.81 and 0.92, a Cohen's Kappa between 0.61 and 0.85, a sensitivity from 0.74 to 0.95, and a specificity from 0.85 to 0.92. Influential and consistent predictors across the algorithms were baseline HFI, socioeconomic status (SES), adoption of financial coping strategies, and receiving government support. Discussion: Persistent HFI can be a relevant indicator to identify households that are less responsive to food security policies. These households should be prioritized for innovative government support and monitored to assess changes. Forecasting systems of HFI can be improved with longitudinal designs including baseline measures of HFI and socioeconomic predictors.
Assuntos
COVID-19 , Insegurança Alimentar , Humanos , COVID-19/epidemiologia , México/epidemiologia , Estudos Longitudinais , Feminino , Masculino , Adulto , Pessoa de Meia-Idade , Aprendizado de Máquina , Características da Família , Inquéritos e Questionários , Fatores Socioeconômicos , Adulto Jovem , SARS-CoV-2 , Adolescente , Pandemias , Abastecimento de Alimentos/estatística & dados numéricosRESUMO
Introduction: The COVID-19 pandemic has exacerbated mental health challenges, particularly depression among college students. Detecting at-risk students early is crucial but remains challenging, particularly in developing countries. Utilizing data-driven predictive models presents a viable solution to address this pressing need. Aims: 1) To develop and compare machine learning (ML) models for predicting depression in Argentinean students during the pandemic. 2) To assess the performance of classification and regression models using appropriate metrics. 3) To identify key features driving depression prediction. Methods: A longitudinal dataset (N = 1492 college students) captured T1 and T2 measurements during the Argentinean COVID-19 quarantine. ML models, including linear logistic regression classifiers/ridge regression (LogReg/RR), random forest classifiers/regressors, and support vector machines/regressors (SVM/SVR), are employed. Assessed features encompass depression and anxiety scores (at T1), mental disorder/suicidal behavior history, quarantine sub-period information, sex, and age. For classification, models' performance on test data is evaluated using Area Under the Precision-Recall Curve (AUPRC), Area Under the Receiver Operating Characteristic curve, Balanced Accuracy, F1 score, and Brier loss. For regression, R-squared (R2), Mean Absolute Error, and Mean Squared Error are assessed. Univariate analyses are conducted to assess the predictive strength of each individual feature with respect to the target variable. The performance of multi- vs univariate models is compared using the mean AUPRC score for classifiers and the R2 score for regressors. Results: The highest performance is achieved by SVM and LogReg (e.g., AUPRC: 0.76, 95% CI: 0.69, 0.81) and SVR and RR models (e.g., R2 for SVR and RR: 0.56, 95% CI: 0.45, 0.64 and 0.45, 0.63, respectively). Univariate models, particularly LogReg and SVM using depression (AUPRC: 0.72, 95% CI: 0.64, 0.79) or anxiety scores (AUPRC: 0.71, 95% CI: 0.64, 0.78) and RR using depression scores (R2: 0.48, 95% CI: 0.39, 0.57) exhibit performance levels close to those of the multivariate models, which include all features. Discussion: These findings highlight the relevance of pre-existing depression and anxiety conditions in predicting depression during quarantine, underscoring their comorbidity. ML models, particularly SVM/SVR and LogReg/RR, demonstrate potential in the timely detection of at-risk students. However, further studies are needed before clinical implementation.
RESUMO
Background: Despite numerous efforts to assess the impact of the COVID-19 pandemic on mental health, there are few longitudinal studies that examine the change in the burden of psychological distress over time and its associated factors, especially in developing countries. Objective: The primary aim of this study was to assess the levels of psychological distress at two time points during the COVID-19 outbreak based on a representative community sample in Chile. The secondary aim was to identify groups that are more vulnerable to psychological distress during the pandemic. Methods: A nationally representative, longitudinal telephone survey of Chilean adults was conducted. This study analyses panel data from two waves in 2020: May 30 to June 10 and September 15 to October 9. A total of 823 people participated in both surveys. Changes in mental health outcomes (anxiety and depressive symptoms) were assessed, estimating the effect of demographic characteristics, psychosocial and economic factors, household conditions, and health status. Results: There was a significant increase in psychological distress (PHQ-4 ≥ 6) between Waves 1 (22.6%) and 2 (27.0%), especially among younger participants. Overall, the results of this study show that being female, living in or near the capital, living in overcrowded households and having a perceived lack of space in the home, loneliness or perceived social isolation, and having received mental health treatment within the last year are significantly associated with psychological distress over time (p < 0.05). Conclusion: This study highlights the need to implement psychosocial programs to protect people's psychological well-being, as well as social policies to improve household living conditions and levels of social connectedness during the COVID-19 outbreak.
RESUMO
BACKGROUND: Spastic paraplegia, optic atrophy and neuropathy (Spoan syndrome) is an autosomal recessive disease with approximately 70 cases recorded in Brazil and Egypt. METHODS: This is a prospective longitudinal study performed with 47 patients affected with Spoan syndrome of seven communities of Rio Grande do Norte (Brazil) to investigate changes in motor function based on comparative data obtained from a 10-year follow-up. RESULTS: The mean age of the participants was 47.21 ± 12.42 years old, and the mean age at loss of ambulation and hand function were 10.78 ± 5.55 and 33.58 ± 17.47 years old, respectively. Spearman's correlation analysis between the score on the Modified Barthel Index and the investigated variables evidenced statistical significance for age (p < 0.001) and right- and left-hand grip strength (p = 0.042 and p = 0.021, respectively). Statistical significance was not evidenced for the remainder of the variables, including age at onset of symptoms (p = 0.634), age at loss of ambulation (p = 0.664) and age at loss of hand function (p = 0.118). CONCLUSIONS: Our analysis allows asserting that the participants exhibited slight dependence until age 35. The greatest losses occurred from ages 35 to 41, and starting at 50, practically all patients become completely dependent. These findings are relevant for determining the prognosis as well as suitable treatment, rehabilitation and assistive technology for these individuals.