Qual é o desempenho dos modelos de aprendizado de máquina na detecção de depressão, ansiedade e estresse entre estudantes de graduação? Uma revisão sistemática

Schaab, Bruno Luis; Calvetti, Prisla Ücker; Hoffmann, Sofia; Diaz, Gabriela Bertoletti; Rech, Maurício; Cazella, Sílvio César; Stein, Airton Tetelbom; Barros, Helena Maria Tannhauser; Silva, Pamela Carvalho da; Reppold, Caroline Tozzi

doi:10.1590/0102-311XEN029323

Tabla de contenido Anterior Actual Siguiente

Tabla de contenido

REVIEW • Cad. Saúde Pública 40 (11) • 2024 • https://doi.org/10.1590/0102-311XEN029323 linkcopiar

Qual é o desempenho dos modelos de aprendizado de máquina na detecção de depressão, ansiedade e estresse entre estudantes de graduação? Uma revisão sistemática

Autoria SCIMAGO INSTITUTIONS RANKINGS

Os alunos de graduação são frequentemente afetados por depressão, ansiedade e estresse. O aprendizado de máquina pode apoiar a avaliação da saúde mental. Com base na seguinte questão de pesquisa “Qual é o desempenho dos modelos de aprendizado de máquina na detecção de depressão, ansiedade e estresse entre estudantes de graduação?”, objetivou-se avaliar o desempenho desses modelos. As pesquisas foram realizadas no PubMed, Embase, PsycINFO e Web of Science. Foram pesquisados estudos que atendessem aos seguintes critérios: publicados em inglês, estudantes universitários de graduação como população alvo, empíricos, publicados em uma revista científica e que previssem resultados de ansiedade, depressão ou estresse via aprendizado de máquina. A qualidade das evidências foi analisada usando o GRADE. Em janeiro de 2024, foram encontrados 2.304 artigos, e 48 estudos atenderam aos critérios de inclusão. Foram identificados diferentes tipos de dados, incluindo dados comportamentais, fisiológicos, de uso da Internet, neurocerebrais, marcadores sanguíneos, dados mistos, demográficos e de mobilidade. Entre os 33 estudos que forneceram dados de precisão, 30 relataram valores superiores a 70%. A acurácia na detecção de estresse variou de 63% a 100%, ansiedade de 53,69% a 97,9% e depressão de 73,5% a 99,1%. Embora a maioria dos modelos apresente desempenho adequado, deve-se notar que 47 deles realizaram apenas validação interna, o que pode superestimar os dados de desempenho. Além disso, a avaliação GRADE indicou que a qualidade da evidência é muito baixa. Os resultados indicam que os algoritmos de aprendizado de máquina são promissores no campo da Saúde Pública; no entanto, é crucial examinar sua aplicabilidade prática. Estudos futuros devem investir principalmente na validação externa dos modelos de aprendizado de máquina.

Palavras-chave:
Estudantes; Aprendizado de Máquina; Saúde Mental

STUDY (YEAR)	COUNTRY	SAMPLE SIZE	SAMPLE CHARACTERISTICS	STUDY DESIGN
Amalraj et al. ³⁰ (2023)	India	24	Not reported	Cross-sectional
Jiao et al. ³¹ (2023)	China	65	Not reported	Cross-sectional
Pal et al. ³² (2023)	China	66	Not reported	Experimental
Pourmohammadi & Maleki ³³ (2020)	Iran	34	67.64% female; 20-37 years (M = 25.4; SD: 4.2)	Cross-sectional
Sharma et al. ³⁴ (2022)	India	38	52.6% female; 8-25 years (M = 22.4; SD: 2.42)	Cross-sectional
Silva et al. ³⁵ (2020)	Portugal	83	87.95% female; 17-38 years (M = 22.13; SD: 5.5)	Cross-sectional
Tiwari & Agarwal ³⁶ (2021)	India	34	Not reported	Cross-sectional
Anand et al. ³⁷ (2023)	Saudi Arabia	197	Not reported	Cross-sectional
Balli et al. ³⁸ (2023)	Turkey	79	Not reported	Cross-sectional
Daza et al. ³⁹ (2023)	Peru	284	Not reported	Cross-sectional
Estabragh et al. ⁴⁰ (2013)	Iran	438	50.22% female (M = 21.37; SD: 2.43)	Cross-sectional
Herbert et al. ⁴¹ (2021)	Egypt/Germany	220	50.9% female; 18-33 years (M = 20.45; SD: 1.88)	Cross-sectional
Ge et al. ⁴² (2020)	China	2,009	50.95% female	Cross-sectional
Gil et al. ⁴³ (2022)	South Korea	171	Not reported	Cross-sectional
Maitre et al. ⁴⁴ (2023)	Canada	3,878	Not reported	Longitudinal
Morales-Rodríguez et al. ⁴⁵ (2021)	Spain	337	73% female; 18-67 years (M = 33.11; SD: 12.83)	Longitudinal
Ren et al. ⁴⁶ (2021)	China	478	57.1% female	Cross-sectional
Upadhyay et al. ⁴⁷ (2023)	India	137	18-25 years	Longitudinal
Vergaray et al. ⁴⁸ (2022)	Peru	284	Not reported	Cross-sectional
Wang et al. ⁴⁹ (2020)	China	1,172	56% female; 18-22 years	Cross-sectional
AlShorman et al. ⁵⁰ (2022)	Saudi Arabia	182	100% male; 18-23 years	Cross-sectional
He et al. ⁵¹ (2021)	China	589	Not reported	Longitudinal
Li et al. ⁵² (2015)	China	36	33.33% female	Cross-sectional
Modinos et al. ⁵³ (2013)	Greece	34	58.82% female; 17-27 years (M = 20.5; SD: 2.4)	Cross-sectional
Zhang et al. ⁵⁴ (2019)	China	82	54.87% female; 18-26 years	Cross-sectional
Liu et al. ⁵⁵ (2023)	China	523	100% female (M = 18.99; SD: 1.1)	Longitudinal
Topalovic et al. ⁵⁶ (2021)	Serbia	100	40% female (M = 22.8; SD: 2.2)	Cross-sectional
Ding et al. ⁵⁷ (2020)	China	693	Not reported	Longitudinal
Dehghan-Bonari et al. ⁵⁸ (2023)	Iran	Not reported	Not reported	Cross-sectional
Siraji et al. ⁵⁹ (2023)	Bangladesh	444	34.45%; 17-20 years	Cross-sectional
Zhang et al. ⁶⁰ (2020)	United States	49	34.69% female	Longitudinal
Ware et al. ⁶¹ (2020)	United States	182	76.7% female; 18-25 years	Cross-sectional
Aalbers et al. ⁶² (2023)	Netherlands	224	55.8% female (M = 21.97; SD: 3.04)	Longitudinal
Acikmese & Alptekin ⁶³ (2019)	Turkey	48	Not reported	Cross-sectional
Ahmed & Ahmed ⁶⁴ (2023)	Bangladesh	100	13% female; 19-30 years	Cross-sectional
Chikersal et al. ⁶⁵ (2021)	United States	138	Not reported	Longitudinal
Guerrero et al. ⁶⁶ (2023)	Ecuador	120	18-24 years old	Cross-sectional
Mahalingam et al. ⁶⁷ (2023)	Lebanon	329	Not reported	Cross-sectional
Meda et al. ⁶⁸ (2023)	Italy	1,388	71.46% female; 18-30 years	Longitudinal
Nemesure et al. ⁶⁹ (2021)	France	4,184	57.4% female	Cross-sectional
Bhadra & Kumar ⁷⁰ (2024)	France	4,184	57.4% female	Cross-sectional
Rois et al. ⁷¹ (2021)	Bangladesh	355	57.5% female	Cross-sectional
Sano et al. ⁷² (2018)	United States	201	36% female; 18-25 years	Longitudinal
Ware et al. ⁷³ (2022)	United States	59	Not reported	Cross-sectional
Xu et al. ⁷⁴ (2021)	United States	397	61.46% female	Longitudinal
Yue et al. ⁷⁵ (2021)	United States	79	18-25 years	Cross-sectional
Müller et al. ⁷⁶ (2021)	United States	57	45.61% female; 18-45 years	Cross-sectional
Nayan et al. ⁷⁷ (2022)	Bangladesh	2,121	55% female; 21-25 years	Cross-sectional

TYPE OF DATA	STUDY (YEAR)	OUTCOME MEASURE	GOAL	ALGORITHM	PERFORMANCE	VALIDATION
Physiological data	Amalraj et al. ³⁰ (2023)	Not reported	To classify stress among university students	ANN-GA	Accuracy = 99%	Internal validation
Physiological data	Jiao et al. ³¹ (2023)	Not reported	To classify students with depression and without depression and with stress and without stress	Not reported	Accuracy depression = 95.26% Accuracy stress = 98.46% Accuracy depression vs. stress = 100%	Internal validation
Physiological data	Pal et al. ³² (2023)	SAS	To classify anxiety among university students	RF	Accuracy = 80% AUC = 82% PPV = 80% Sensitivity = 80% Specificity = 73%	Internal validation with “leave-one-out” cross-validation
Physiological data	Pourmohammadi & Maleki ³³ (2020)	STAI	To classify stress among university students	SVM	Accuracy = 100% two levels Accuracy = 97.6% three levels Accuracy = 92.2% four levels	Internal validation with nested 10-fold cross-validation
Physiological data	Sharma et al. ³⁴ (2022)	BDI-II	To classify the level of depression among university students	AEN	Accuracy = 95.2% five levels	Internal validation with 10-fold cross-validation
Physiological data	Silva et al. ³⁵ (2020)	PSS	To classify stress among university students	NN	Sensitivity = 78.1% Specificity = 74.2%	Internal validation with 10-fold cross validation
				NB	Sensitivity = 62.7% Specificity = 74.2%
				SVM	Sensitivity = 47.5% Specificity = 82.1%
				RF	Sensitivity = 74.8% Specificity = 71.2%
				KNN	Sensitivity = 69% Specificity = 75.6%
Physiological data	Tiwari & Agarwal ³⁶ (2021)	PSS	To classify students’ mental state into four categories: relaxed, stressed, partially stressed, and happy	LR	Accuracy = 83.3%	Internal validation with 10-fold cross-validation
				SVM	Accuracy = 88.3%
				KNN	Accuracy = 82.4%
				BAG	Accuracy = 98.4%
				RF	Accuracy = 97.9%
				GB	Accuracy = 98.2%
				ANN	Accuracy = 99.4%
Behavioral data	Anand et al. ³⁷ (2023)	QF	Classify the stress of university students into three categories: highly stressed, manageable stress, and no stress	DT + RF + AdaBoost	Accuracy = 93.48% PPV = 92.99% F1 = 93.14% Sensitivity = 93.30%	Internal validation with 5-fold cross-validation
Behavioral data	Balli et al. ³⁸ (2023)	BDI	To detect students with depression and without depression	XGBoost	Accuracy = 89.6%	Not reported
Behavioral data	Daza et al. ³⁹ (2023)	GAD-7	To predict anxiety level of university students	KNN	Accuracy = 97.83% Sensitivity = 98.44% Specificity = 99.32% F1 = 97.88%	Internal validation with 10-fold cross-validation
Behavioral data	Estabragh et al. ⁴⁰ (2013)	SPI	To diagnose college students with social anxiety	BN	AUC = 89.8%	Not reported
Behavioral data	Herbert et al. ⁴¹ (2021)	STAI	To predict trait anxiety among university students	SVR, GBR	RMSE = 0.90 % of RMSE in range = 15.04%	Internal validation with test set
Behavioral data	Ge et al. ⁴² (2020)	GAD-7	To predict university students with anxiety	XGBoost	Accuracy = 97.3% Sensitivity = 97.3% Specificity = 96.3%	Internal validation with 5-fold cross-validation
Behavioral data	Gil et al. ⁴³ (2022)	CES-D	To predict the risk of depression in college students	RF	Accuracy = 86.27% PPV = 80.59% Sensitivity = 85.00% Specificity = 87.10% F1 = 82.74% AUC = 86.05%	Internal validation
Behavioral data	Maitre et al. ⁴⁴ (2023)	GAD-7	To investigate the anxiety level of university students	LR	R² = 0.5300	Internal validation with 10-fold cross-validation
				LASSO	R² = 0.5294
				RF	R² = 0.5383
				XGBoost	R² = 0.5630
				CatBoost	R² = 0.5656
Behavioral data	Morales-Rodríguez et al. ⁴⁵ (2021)	PSS	To predict the stress level of college students	ANN	AUC = 74.8%	Internal validation
Behavioral data	Ren et al. ⁴⁶ (2021)	SAS, PHQ-9	To assess depression and anxiety in university students	LR	Accuracy anxiety = 81.42% AUC anxiety = 88.50% Sensitivity anxiety = 83.21% Specificity anxiety = 80.38% Accuracy depression = 73.5% AUC depression = 80.60% Sensitivity depression = 75.3% Specificity depression = 71.80%	Internal validation with 5-fold cross-validation
Behavioral data	Upadhyay et al. ⁴⁷ (2023)	HDRS and CDRS along with clinician diagnostic	To assess persistent depression disorder among university students	Stacked SVM	Accuracy = 89.4% Sensitivity = 89.92% Specificity = 89.96% PPV = 89.82% F1 = 89.96%	Internal validation
Behavioral data	Vergaray et al. ⁴⁸ (2022)	PHQ-9	To predict depression among college students	SVM	Accuracy = 94.69% Sensitivity = 94.22% PPV = 94.09% F1 = 94.12%	Internal validation with 10-fold cross-validation
Behavioral data	Wang et al. ⁴⁹ (2020)	SAS	To predict the level of stress (normal, mild, moderate, severe) at the beginning of the academic semester and one month after the beginning of the academic semester	XGBoost	Model 1 Accuracy anxiety level = 83.81% Accuracy anxiety change = 79.26% Model 2 Accuracy anxiety level = 82.10% Accuracy anxiety change = 84.38%	Internal validation with test set
Neurocerebral data	AlShorman et al. ⁵⁰ (2022)	DASS-21	To detect mental stress among university students	SVM with RBF kernel	Accuracy = 81.40% AUC = 86.10% F1 = 81.40% PPV = 81.50% Sensitivity = 84.40% Specificity = 81.50%	Internal validation
Neurocerebral data	He et al. ⁵¹ (2021)	STAI	To classify anxiety in university students in comparison to healthy controls, individuals with depression and individuals with schizophrenia	BLR	Accuracy control vs. anxiety = 68.72% AUC control vs. anxiety = 72% Sensitivity control vs. anxiety = 71.40% Specificity control vs. anxiety = 65% Accuracy major depression vs. anxiety = 53.68% AUC major depression vs. anxiety = 53% Sensitivity major depression vs. anxiety = 72.20% Specificity major depression vs. anxiety = 33.13% Accuracy schizophrenia vs. anxiety = 59.1% AUC schizophrenia vs. anxiety = 59% Sensitivity schizophrenia vs. anxiety = 32.88% Specificity schizophrenia vs. anxiety = 73.91%	Internal validation with 10-fold cross-validation and external validation
Neurocerebral data	Li et al. ⁵² (2015)	BDI-II	To classify students with depression and without depression	KNN	Accuracy = 99.1% AUC = 99.9%	Internal validation with 10-fold cross-validation
Neurocerebral data	Modinos et al. ⁵³ (2013)	BDI-II	To classify depression among university student	SVM	Accuracy = 77% Sensitivity = 71% Specificity = 82%	Internal validation with “leave-one-out” cross-validation
Neurocerebral data	Zhang et al. ⁵⁴ (2019)	TAS along with clinician diagnostic	To classify students with high anxiety and low anxiety	CNN	Accuracy = 86.5% PPV = 84% Sensitivity= 100% F1 = 91.1%	Internal validation with 5-fold cross-validation
Blood marker data	Liu et al. ⁵⁵ (2023)	CES-D	To predict depression among university women over a 1-year period	SVM	R = 0.81; p < 0.001	Internal validation with test set
Blood marker data	Topalovic et al. ⁵⁶ (2021)	DASS-21	To predict the increase in stress levels among university students	BLR	Accuracy = 70% Nagelkerke R² = 0.38 Snell R² = 0.28	Not reported
Internet dada	Ding et al. ⁵⁷ (2020)	QF	To classify depression among university students	RBF-NN	Accuracy = 82%	Internal validation with test set
				SVM	Accuracy = 80%
				KNN	Accuracy = 79%
				DISVM	Accuracy = 86%
Internet data	Dehghan-Bonari et al. ⁵⁸ (2023)	Not reported	To diagnose students with and without depression	RF	Accuracy = 94%	Internal validation
Internet data	Siraji et al. ⁵⁹ (2023)	DASS-21	To detect college students with and without depression	SVM	Accuracy = 85.14% F1 = 84.92% AUC = 98.41%	Internal validation with 5-fold cross-validation
Internet data	Zhang et al. ⁶⁰ (2020)	PHQ-9, GAD-7	To predict deterioration of depression and anxiety among university students	OLS	Depression (MSE = 2.37, R² = 0.84) Anxiety (MSE = 2.48, R² = 0.81)	Internal validation with “leave-one-out” cross-validation
Internet data	Ware et al. ⁶¹ (2020)	PHQ-9/QIDS along with clinician diagnostic	To predict various symptoms of depression in university students	SVM with RBF kernel	Model 1 F1 = 67% PPV = 71% Sensitivity = 64% Specificity = 73% Model 2 F1 = 72% PPV = 65% Sensitivity l = 81% Specificity = 51%	Internal validation with “leave-one-out” cross-validation
Mixed data	Aalbers et al. ⁶² (2023)	SESS	To assess the stress of university students	LASSO	MAE = 0.84	Internal validation with 10-fold cross-validation
				SVM	MAE = 0.84
				RF	MAE = 0.84
Mixed data	Acikmese & Alptekin ⁶³ (2019)	QF	To classify university students into stressed and non-stressed	LSTM	Accuracy = 63% PPV = 63% Sensitivity = 63% F1 = 63%	Internal validation with test set
Mixed data	Ahmed & Ahmed ⁶⁴ (2023)	PHQ-9	To identify depressed and non-depressed students	BFS	Accuracy = 78% PPV = 77.4% AUC = 78% Specificity = 75.50% Sensitivity = 80.4% F1 = 78.80%	Internal validation with 10-fold cross-validation
Mixed data	Chikersal et al. ⁶⁵ (2021)	BDI-II	To classify depression among university students at the end of the semester, as well as depression worsening	AdaBoost	Accuracy = 85.7% depression end of semester F1 = 82% depression end of semester Accuracy = 88.1% depression worsening F1 = 81% depression worsening	Internal validation with “leave-one-out” cross-validation
Mixed data	Guerrero et al. ⁶⁶ (2023)	AMAS-C	To identify college students with and without anxiety	Not reported	Model 1 (facial expression) PPV = 86.84% Model 2 (emotions recognition) PPV = 84.21%	Internal validation
Mixed data	Mahalingam et al. ⁶⁷ (2023)	BAI	To classify university students with and without anxiety	MLP	AUC = 80.70% Accuracy = 67.5%	Not reported
				LR	AUC = 77.25% Accuracy = 67.67%
				SVM	AUC = 76.01% Accuracy = 69.70%
				RF	AUC = 74.75% Accuracy = 67.68%
				XGBoost	AUC = 72.58% Accuracy = 63.64%
Mixed data	Meda et al. ⁶⁸ (2023)	BDI-II	To assess the change of depression symptoms in university students after six months	RF	Overall PPV = 77% PPV in depression worsening = 49%	Internal validation
Mixed data	Nemesure et al. ⁶⁹ (2021)	Clinician diagnostic	To classify university students with generalized anxiety disorder and major depressive disorder	XGBoost classifier	Generalized anxiety disorder AUC = 73% Generalized anxiety disorder sensitivity = 70% Generalized anxiety disorder specificity = 66% Major depressive disorder AUC = 67% Major depressive disorder sensitivity = 55% Major depressive disorder specificity = 70%	Internal validation with 5-fold cross-validation
Mixed data	Bhadra & Kumar ⁷⁰ (2024)	Clinician diagnostic	To detect students with and without depression	ANN	Accuracy = 88.46%	Internal validation
				SVM	Accuracy = 88%
				RF	Accuracy = 88.46%
				XGBoost	Accuracy= 84.18%
Mixed data	Rois et al. ⁷¹ (2021)	QF	To classify stress among university students	AdaBoost	Accuracy = 89%	Internal validation with 10-fold cross-validation
Mixed data	Sano et al. ⁷² (2018)	PSS	To classify the stress of university students into high stress and low stress	SVM with RBF	Accuracy = 81.5% F1 = 83%	Internal validation with “leave-one-out” cross-validation
				SVM linear	Accuracy = 70.3% F1 = 72%
				LASSO	Accuracy = 67.6% F1 = 74%
Mixed data	Ware et al. ⁷³ (2022)	PHQ-9/QIDS along with clinician diagnostic	To predict depression among college students	SVM	F1 = 82% PPV = 78% Sensitivity = 86% Specificity = 74%	Internal validation
Mixed data	Xu et al. ⁷⁴ (2021)	BDI-II	To detect students with and without depression	Not reported	Accuracy = 79.10% PPV = 81.40% Sensitivity = 85.40% F1 = 83.30%	Internal validation with leave-one-out cross-validation
Mixed data	Yue et al. ⁷⁵ (2021)	PHQ-9 along with clinician diagnostic	To predict depression among college students	SVM with RBF kernel	F1 = 79% PPV = 77% Sensitivity = 79% Specificity = 72%	Internal validation with “leave-one-out” cross-validation
Mobility data	Müller et al. ⁷⁶ (2021)	QF based on ICD-10	To predict depression among college students	RF	AUC = 82%	Internal validation
Demographic data	Nayan et al. ⁷⁷ (2022)	PHQ-9, GAD-7	To predict college students with depression and anxiety	KNN	Accuracy depression = 88.28% Sensitivity depression = 66.67% Specificity depression = 96.13%	Internal validation with 10-fold cross-validation
Demographic data	Nayan et al. ⁷⁷ (2022)	PHQ-9, GAD-7	To predict college students with depression and anxiety	RF	Accuracy anxiety = 91.49% Sensitivity anxiety = 67.77% Specificity anxiety = 98.53%	Internal validation with 10-fold cross-validation

OUTCOME	STUDIES (PARTICIPANTS)	RISK OF BIAS	INDIRECTNESS	INCONSISTENCY	IMPRECISION	PUBLICATION BIAS	CERTAINTY OF EVIDENCE
Accuracy	33 (15,105)	Serious	Serious	Not serious	Serious	None	⨁◯◯◯ Very low
Sensitivity/Sensibility	5 (4,535)	Serious	Serious	Not serious	Serious	None	⨁◯◯◯ Very low
PPV	2 (1,508)	Serious	Serious	Not serious	Serious	None	⨁◯◯◯ Very low
AUC	3 (832)	Serious	Serious	Not serious	Serious	None	⨁◯◯◯ Very low
Other outcomes	5 (4,894)	Serious	Serious	Not serious	Serious	None	⨁◯◯◯ Very low

STUDY	SAMPLE REPRESENTATIVENESS	CONTROL CONFOUNDING VARIABLES	ASSESSMENT OF THE OUTCOME	MACHINE LEARNING ALGORITHM	PERFORMANCE METRICS	MISSING DATA	TEST UNSEEN	CLASS IMBALANCE	FEATURE SELECTION + HYPERPARAMETER
Amalraj et al. ³⁰ (2023)	No	No	No	Yes	Yes	Yes	Yes	No	No
Jiao et al. ³¹ (2023)	No	No	Yes	Yes	Yes	No	Yes	No	No
Pal et al. ³² (2023)	No	No	Yes	Yes	Yes	Yes	Yes	No	No
Pourmohammadi & Maleki ³³ (2020)	No	No	Yes	Yes	Yes	No	Yes	No	Yes
Sharma et al. ³⁴ (2022)	No	No	Yes	Yes	Yes	No	Yes	No	Yes
Silva et al. ³⁵ (2020)	No	No	Yes	Yes	Yes	No	Yes	No	No
Tiwari & Agarwal ³⁶ (2021)	No	No	Yes	Yes	Yes	No	Yes	Yes	Yes
Anand et al. ³⁷ (2023)	Yes	No	Yes	Yes	Yes	No	Yes	Yes	No
Balli et al. ³⁸ (2023)	No	No	Yes	Yes	Yes	No	No	No	No
Daza et al. ³⁹ (2023)	Yes	No	Yes	Yes	Yes	Yes	Yes	Yes	No
Estabragh et al. ⁴⁰ (2013)	Yes	No	Yes	Yes	Yes	No	No	No	No
Herbert et al. ⁴¹ (2021)	Yes	No	Yes	Yes	Yes	Yes	Yes	No	Yes
Ge et al. ⁴² (2020)	Yes	No	Yes	Yes	Yes	No	Yes	No	No
Gil et al. ⁴³ (2022)	Yes	No	Yes	Yes	Yes	Yes	Yes	No	Yes
Maitre et al. ⁴⁴ (2023)	Yes	No	Yes	Yes	Yes	No	Yes	No	No
Morales-Rodríguez et al. ⁴⁵ (2021)	Yes	No	Yes	Yes	Yes	Yes	Yes	No	Yes
Ren et al. ⁴⁶ (2021)	Yes	No	Yes	Yes	Yes	No	Yes	Yes	No
Upadhyay et al. ⁴⁷ (2023)	Yes	No	Yes	Yes	Yes	No	Yes	No	No
Vergaray et al. ⁴⁸ (2022)	Yes	No	Yes	Yes	Yes	Yes	Yes	Yes	No
Wang et al. ⁴⁹ (2020)	Yes	No	Yes	Yes	Yes	No	Yes	No	No
AlShorman et al. ⁵⁰ (2022)	Yes	No	Yes	Yes	Yes	No	Yes	No	No
He et al. ⁵¹ (2021)	Yes	Yes	Yes	Yes	Yes	Yes	Yes	No	Yes
Li et al. ⁵² (2015)	No	No	Yes	Yes	Yes	No	Yes	Yes	No
Modinos et al. ⁵³ (2013)	No	Yes	Yes	Yes	Yes	No	Yes	No	No
Zhang et al. ⁵⁴ (2019)	Yes	No	Yes	Yes	Yes	No	Yes	No
Liu et al. ⁵⁵ (2023)	Yes	Yes	Yes	Yes	Yes	No	Yes	No	No
Topalovic et al. ⁵⁶ (2021)	No	No	Yes	Yes	Yes	No	No	No	No
Ding et al. ⁵⁷ (2020)	Yes	No	Yes	Yes	Yes	No	Yes	No	No
Dehghan-Bonari et al. ⁵⁸ (2023)	No	No	No	Yes	Yes	Yes	Yes	No	No
Siraji et al. ⁵⁹ (2023)	Yes	No	No	Yes	Yes	No	Yes	No	Yes
Zhang et al. ⁶⁰ (2020)	No	Yes	Yes	Yes	Yes	No	Yes	No	Yes
Ware et al. ⁶¹ (2020)	Yes	No	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Aalbers et al. ⁶² (2023)	Yes	No	Yes	Yes	Yes	No	Yes	No	Yes
Acikmese & Alptekin ⁶³ (2019)	No	No	Yes	Yes	Yes	No	Yes	Yes	No
Ahmed & Ahmed ⁶⁴ (2023)	No	No	Yes	Yes	Yes	No	Yes	No	Yes
Chikersal et al. ⁶⁵ (2021)	No	No	Yes	Yes	Yes	Yes	Yes	No	Yes
Guerrero et al. ⁶⁶ (2023)	Yes	No	Yes	Yes	Yes	No	Yes	No	No
Mahalingam et al. ⁶⁷ (2023)	Yes	No	Yes	Yes	Yes	Yes	No	No	Yes
Meda et al. ⁶⁸ (2023)	Yes	No	Yes	Yes	Yes	No	Yes	No	Yes
Nemesure et al. ⁶⁹ (2021)	Yes	No	Yes	Yes	Yes	Yes	Yes	No	Yes
Bhadra & Kumar ⁷⁰ (2024)	Yes	No	Yes	Yes	Yes	Yes	Yes	No	Yes
Rois et al. ⁷¹ (2021)	Yes	No	Yes	Yes	Yes	No	Yes	No	Yes
Sano et al. ⁷² (2018)	Yes	No	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Ware et al. ⁷³ (2022)	No	No	Yes	Yes	Yes	No	Yes	No	No
Xu et al. ⁷⁴ (2021)	Yes	No	Yes	Yes	Yes	Yes	Yes	No	No
Yue et al. ⁷⁵ (2021)	No	No	Yes	Yes	Yes	Yes	Yes	No	Yes
Müller et al. ⁷⁶ (2021)	No	No	Yes	Yes	Yes	No	Yes	No	Yes
Nayan et al. ⁷⁷ (2022)	Yes	No	Yes	Yes	Yes	Yes	Yes	No	No

location_on

Escola Nacional de Saúde Pública Sergio Arouca, Fundação Oswaldo Cruz Rua Leopoldo Bulhões, 1480 , 21041-210 Rio de Janeiro RJ Brazil, Tel.:+55 21 2598-2511, Fax: +55 21 2598-2737 / +55 21 2598-2514 - Rio de Janeiro - RJ - Brazil
E-mail: cadernos@ensp.fiocruz.br

rss_feed Acompañe los números de esta revista en su lector de RSS

[TFN1] M: mean; SD: standard deviation.