Md Sabbir Hossain
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Md Sabbir Hossain has a total of 31 co-authors. Here are the top co-authors by number of shared publications:
- Jamal Uddin
Department of Statistics Shahjalal University of Science & Technology Sylhet Bangladesh
3 shared publications - Fakrul Islam
Department of Statistics, Shahjalal University of Science and Technology, Sylhet, 3114, Bangladesh
2 shared publications - Mahfuzer Rohman
Shahjalal University of Science and Technology
2 shared publications Mohammed Taj Uddin
Department of Statistics Shahjalal University of Science and Technology Sylhet Bangladesh
2 shared publications- Prosenjit Basak Arka
Shahjalal University of Science and Technology
2 shared publications Fakrul Islam
Shahjalal University of Science and Technology
1 shared publication
Latest Publications
- Open Access
- 11 November 2026
- Published by Wiley in Health Science Reports
- Vol. 9 (11)
Spirometric Lung Function Among Rice Mill Workers in Sylhet District, Bangladesh: A Cross‐Sectional Study
Background: Rice mill workers may encounter respiratory hazards, but a spirometric survey without direct dust measurements or an unexposed comparison group cannot establish an exposure‐response relationship. This study described measured spirometric patterns and examined associations of lung function with employment duration, smoking, age, height, and body mass index among rice mill workers in Sylhet District. Method: This cross‐sectional study analyzed 203 male workers recruited through cluster sampling in 2024. Lung function was assessed with an RMS Helios 401 spirometer. Study records indicated that at least three acceptable maneuvers were obtained and the best value was retained. Descriptive statistics, one‐way analysis of variance, and multivariable linear regression were used. Regression models used measured FVC, measured FEV 1 , and measured FEV 1 /FVC as primary outcomes, modeled employment duration continuously, omitted weight when BMI was included, and used HC3 robust standard errors. Result: Among 203 analyzed workers, 169 (83.25%) were normal by the study's legacy fixed thresholds, 31 (15.27%) had low FVC with a preserved measured FEV 1 /FVC ratio (a possible restrictive pattern), one (0.49%) had an obstructive spirometric pattern, and two (0.99%) had both a low ratio and low FVC. Reduced FEF 25–75 (1 /FVC ( β = −0.211, 95% CI −0.415 to −0.007; p = 0.043), but not with measured FVC or FEV 1 . A sensitivity model showed lower percent‐predicted FVC with longer employment ( β = −0.537 percentage points/year, 95% CI −0.988 to −0.086; p = 0.020). Conclusion: Most workers had spirometry within the study's fixed thresholds. Longer employment duration was associated with a modestly lower measured FEV 1 /FVC and percent‐predicted FVC, but the cross‐sectional design, lack of direct exposure assessment, and absence of a control group preclude causal attribution to rice dust. Low FVC requires lung‐volume testing before restriction can be confirmed.
- 11 November 2026
- Published by Wiley in Health Science Reports
- Vol. 9 (11)
Open AccessSpirometric Lung Function Among Rice Mill Workers in Sylhet District, Bangladesh: A Cross‐Sectional Study
Background: Rice mill workers may encounter respiratory hazards, but a spirometric survey without direct dust measurements or an unexposed comparison group cannot establish an exposure‐response relationship. This study described measured spirometric patterns and examined associations of lung function with employment duration, smoking, age, height, and body mass index among rice mill workers in Sylhet District. Method: This cross‐sectional study analyzed 203 male workers recruited through cluster sampling in 2024. Lung function was assessed with an RMS Helios 401 spirometer. Study records indicated that at least three acceptable maneuvers were obtained and the best value was retained. Descriptive statistics, one‐way analysis of variance, and multivariable linear regression were used. Regression models used measured FVC, measured FEV 1 , and measured FEV 1 /FVC as primary outcomes, modeled employment duration continuously, omitted weight when BMI was included, and used HC3 robust standard errors. Result: Among 203 analyzed workers, 169 (83.25%) were normal by the study's legacy fixed thresholds, 31 (15.27%) had low FVC with a preserved measured FEV 1 /FVC ratio (a possible restrictive pattern), one (0.49%) had an obstructive spirometric pattern, and two (0.99%) had both a low ratio and low FVC. Reduced FEF 25–75 (1 /FVC ( β = −0.211, 95% CI −0.415 to −0.007; p = 0.043), but not with measured FVC or FEV 1 . A sensitivity model showed lower percent‐predicted FVC with longer employment ( β = −0.537 percentage points/year, 95% CI −0.988 to −0.086; p = 0.020). Conclusion: Most workers had spirometry within the study's fixed thresholds. Longer employment duration was associated with a modestly lower measured FEV 1 /FVC and percent‐predicted FVC, but the cross‐sectional design, lack of direct exposure assessment, and absence of a control group preclude causal attribution to rice dust. Low FVC requires lung‐volume testing before restriction can be confirmed.
- Open Access
- 9 October 2026
- Published by Wiley in Public Health Challenges
- Vol. 5 (4)
Comparison of Machine Learning Classifiers for Predicting Reported Antibiotic Treatment After Recent Fever/Cough Among Under‐Five Children: A Pooled Analysis of DHS Data From 64 LMICs
Background: Machine learning (ML) methods are increasingly used for health‐related prediction. This study compared logistic regression with five ML algorithms for predicting reported antibiotic treatment following recent fever/cough among children under five using pooled Demographic and Health Survey (DHS) data. Methods: Data from 64 low‐ and middle‐income countries (LMICs) were analyzed. The final analytic sample included 183,425 children with recent fever/cough, of whom 56,769 (30.95%) were reported to have received antibiotics. Logistic regression, K ‐nearest neighbors ( K NN), decision tree, random forest (RF), gradient boosting, and XGBoost were evaluated. The data were divided into model‐development and common heldout test samples, with 55,028 observations in the test set. Hyperparameters were selected using the same three stratified cross‐validation folds for all models, and Synthetic Minority Over‐sampling Technique for Nominal and Continuous (SMOTENC) was applied only within training folds. Model performance was assessed using ROC‐AUC, area under the precision–recall curve (AUPRC), balanced accuracy, sensitivity, specificity, precision, F 1 score, Matthews correlation coefficient (MCC), Brier score, and calibration measures. Primary classification results used a prespecified probability threshold of 0.50. Results: Overall predictive performance was modest. RF achieved the highest ROC‐AUC (0.568, 95% CI: 0.552–0.591), whereas K NN achieved the highest AUPRC (0.352, 95% CI: 0.308–0.401) and balanced accuracy (0.531, 95% CI: 0.526–0.536). At the 0.50 threshold, sensitivity was low across all models, ranging from 0.105 to 0.252, whereas specificity ranged from 0.810 to 0.910. Calibration was also limited, with calibration slopes ranging from 0.021 to 0.328. No model consistently performed best across all evaluation measures. Conclusions: ML models did not demonstrate clear overall superiority over logistic regression. Although RF and K NN performed better on selected measures, discrimination and calibration remained limited. ML may be useful as a complementary predictive approach, but stronger validation, additional predictors, and more complete consideration of the DHS complex survey design are needed before practical public‐health application.
- 9 October 2026
- Published by Wiley in Public Health Challenges
- Vol. 5 (4)
Open AccessComparison of Machine Learning Classifiers for Predicting Reported Antibiotic Treatment After Recent Fever/Cough Among Under‐Five Children: A Pooled Analysis of DHS Data From 64 LMICs
Background: Machine learning (ML) methods are increasingly used for health‐related prediction. This study compared logistic regression with five ML algorithms for predicting reported antibiotic treatment following recent fever/cough among children under five using pooled Demographic and Health Survey (DHS) data. Methods: Data from 64 low‐ and middle‐income countries (LMICs) were analyzed. The final analytic sample included 183,425 children with recent fever/cough, of whom 56,769 (30.95%) were reported to have received antibiotics. Logistic regression, K ‐nearest neighbors ( K NN), decision tree, random forest (RF), gradient boosting, and XGBoost were evaluated. The data were divided into model‐development and common heldout test samples, with 55,028 observations in the test set. Hyperparameters were selected using the same three stratified cross‐validation folds for all models, and Synthetic Minority Over‐sampling Technique for Nominal and Continuous (SMOTENC) was applied only within training folds. Model performance was assessed using ROC‐AUC, area under the precision–recall curve (AUPRC), balanced accuracy, sensitivity, specificity, precision, F 1 score, Matthews correlation coefficient (MCC), Brier score, and calibration measures. Primary classification results used a prespecified probability threshold of 0.50. Results: Overall predictive performance was modest. RF achieved the highest ROC‐AUC (0.568, 95% CI: 0.552–0.591), whereas K NN achieved the highest AUPRC (0.352, 95% CI: 0.308–0.401) and balanced accuracy (0.531, 95% CI: 0.526–0.536). At the 0.50 threshold, sensitivity was low across all models, ranging from 0.105 to 0.252, whereas specificity ranged from 0.810 to 0.910. Calibration was also limited, with calibration slopes ranging from 0.021 to 0.328. No model consistently performed best across all evaluation measures. Conclusions: ML models did not demonstrate clear overall superiority over logistic regression. Although RF and K NN performed better on selected measures, discrimination and calibration remained limited. ML may be useful as a complementary predictive approach, but stronger validation, additional predictors, and more complete consideration of the DHS complex survey design are needed before practical public‐health application.
- Open Access
- 18 August 2026
- Published by Springer Nature in BMC Public Health
Prevalence and factors associated with HIV among women aged 15–49 years in Zambia: a multilevel mixed-effects analysis
- 18 August 2026
- Published by Springer Nature in BMC Public Health
Open AccessPrevalence and factors associated with HIV among women aged 15–49 years in Zambia: a multilevel mixed-effects analysis
- Open Access
- 21 May 2026
- Published by Wiley in Health Science Reports
- Vol. 9 (5) , e72548
Frailty Among the Older Adults in Bangladesh: A Comparative Analysis of Rural and Urban Populations
Background and Aims: The increasing incidence of frailty among older adults in Bangladesh is a significant public health concern. This study aims to compare the level and determinants of frailty among older adults living in rural and urban areas of Bangladesh. Methods: A cross‐sectional survey of 944 older adults aged 55 years and above, conducted in the Sylhet District of Bangladesh between 2022 and 2023, selected through a two‐stage random sampling method in urban areas and a multistage random approach in rural areas. The health status of participants was assessed using a 33‐indicator Frailty Index (FI 33 ), which categorized individuals into four groups: good health (0 < = FI 33 < = 12), mild frailty (0.12 < FI 33 < = 0.24), moderate frailty (0.24 < FI 33 < = 0.36), and severe frailty (FI 33 > 0.36). Data analysis involved descriptive statistics, Pearson χ 2 tests, and multinomial logistic regression. Results: The prevalence of moderate to severe frailty was higher in rural areas (57.6%) compared to urban areas (49%). This difference was statistically significant ( χ ² = 37.09, p < 0.001). Rural older adults also had a higher likelihood of severe frailty than urban older adults (AOR: 1.80, 95% CI: 0.92–3.52). Age was a significant predictor of frailty, with adults aged 65 years and above having a much higher risk (AOR: 9.06, 95% CI: 4.33–18.94). Women were also more likely to experience moderate frailty than men (AOR: 4.43, 95% CI: 2.27–8.62). Conclusion: These findings highlight the need for targeted interventions in rural areas and among older women to address frailty in Bangladesh's aging population.
- 21 May 2026
- Published by Wiley in Health Science Reports
- Vol. 9 (5) , e72548
Open AccessFrailty Among the Older Adults in Bangladesh: A Comparative Analysis of Rural and Urban Populations
Background and Aims: The increasing incidence of frailty among older adults in Bangladesh is a significant public health concern. This study aims to compare the level and determinants of frailty among older adults living in rural and urban areas of Bangladesh. Methods: A cross‐sectional survey of 944 older adults aged 55 years and above, conducted in the Sylhet District of Bangladesh between 2022 and 2023, selected through a two‐stage random sampling method in urban areas and a multistage random approach in rural areas. The health status of participants was assessed using a 33‐indicator Frailty Index (FI 33 ), which categorized individuals into four groups: good health (0 < = FI 33 < = 12), mild frailty (0.12 < FI 33 < = 0.24), moderate frailty (0.24 < FI 33 < = 0.36), and severe frailty (FI 33 > 0.36). Data analysis involved descriptive statistics, Pearson χ 2 tests, and multinomial logistic regression. Results: The prevalence of moderate to severe frailty was higher in rural areas (57.6%) compared to urban areas (49%). This difference was statistically significant ( χ ² = 37.09, p < 0.001). Rural older adults also had a higher likelihood of severe frailty than urban older adults (AOR: 1.80, 95% CI: 0.92–3.52). Age was a significant predictor of frailty, with adults aged 65 years and above having a much higher risk (AOR: 9.06, 95% CI: 4.33–18.94). Women were also more likely to experience moderate frailty than men (AOR: 4.43, 95% CI: 2.27–8.62). Conclusion: These findings highlight the need for targeted interventions in rural areas and among older women to address frailty in Bangladesh's aging population.
- Open Access
- 23 January 2026
- Published by Springer Nature in Research Square
Evaluating Imputation Methods for Handling Missing Data in Complex Survey Designs: Evidence from the India DHS 2017–18
Missing data are a pervasive challenge in large-scale population-based surveys such as the DHS, and inadequate handling of missingness can lead to biased estimates and reduced statistical power. This study examined patterns of missing data and compared the performance of multiple imputation approaches using data from the IDHS 2017–18. The prevalence and structure of missingness across key socio-demographic and health-related variables were assessed, and Little’s test was applied to evaluate whether data were missing completely at random. Five analytical strategies were compared: complete case analysis, MICE, machine learning–based Decision Tree imputation, KNN imputation, and latent class–based imputation. Survey-weighted logistic regression models accounting for stratification, clustering, and sampling weights were fitted to each dataset, with antibiotic use for childhood fever selected as the outcome for model comparison. Several variables exhibited substantial missingness, particularly partner/husband’s occupation, respondent’s occupation, and vaccination status. Little’s test indicated that the missing data were not completely at random. Compared with complete case analysis, all imputation approaches improved estimate precision and revealed meaningful associations. Among the evaluated methods, the Decision Tree approach produced the most stable and consistent results, identifying significant predictors that were not detected using traditional methods. Although MICE and latent class–based imputation yielded improved estimates, their performance declined under extreme levels of missingness. These findings highlight the importance of appropriate missing data handling in complex survey analyses and demonstrate that machine learning–based Decision Tree imputation offers a flexible and robust alternative for addressing extensive missingness in large public health datasets.
- 23 January 2026
- Published by Springer Nature in Research Square
Open AccessEvaluating Imputation Methods for Handling Missing Data in Complex Survey Designs: Evidence from the India DHS 2017–18
Missing data are a pervasive challenge in large-scale population-based surveys such as the DHS, and inadequate handling of missingness can lead to biased estimates and reduced statistical power. This study examined patterns of missing data and compared the performance of multiple imputation approaches using data from the IDHS 2017–18. The prevalence and structure of missingness across key socio-demographic and health-related variables were assessed, and Little’s test was applied to evaluate whether data were missing completely at random. Five analytical strategies were compared: complete case analysis, MICE, machine learning–based Decision Tree imputation, KNN imputation, and latent class–based imputation. Survey-weighted logistic regression models accounting for stratification, clustering, and sampling weights were fitted to each dataset, with antibiotic use for childhood fever selected as the outcome for model comparison. Several variables exhibited substantial missingness, particularly partner/husband’s occupation, respondent’s occupation, and vaccination status. Little’s test indicated that the missing data were not completely at random. Compared with complete case analysis, all imputation approaches improved estimate precision and revealed meaningful associations. Among the evaluated methods, the Decision Tree approach produced the most stable and consistent results, identifying significant predictors that were not detected using traditional methods. Although MICE and latent class–based imputation yielded improved estimates, their performance declined under extreme levels of missingness. These findings highlight the importance of appropriate missing data handling in complex survey analyses and demonstrate that machine learning–based Decision Tree imputation offers a flexible and robust alternative for addressing extensive missingness in large public health datasets.
- Open Access
- 12 August 2025
- Published by Springer Nature in BMC Psychiatry
- Vol. 25 (1) , 783
The association of cyberbullying with major depressive disorders among Bangladeshi female adolescents: findings from the Bangladesh adolescent health and wellbeing survey 2019-20
Cyberbullying refers to the act of using digital technology to engage in bullying. It involves the intentional use of the internet to demean or denigrate individuals, and it has been linked to substantial psychological distress globally. Despite its increasing prevalence, there remains a gap in nationwide research on its association with the mental well-being of female adolescents in Bangladesh. The main objective of this study was to understand the association of cyberbullying with major depressive disorder (MDD) among female adolescents in Bangladesh. We used secondary data from the 2019–20 Bangladesh Adolescent Health and Wellbeing Survey (BAHWS), a nationally representative survey. A multiple logistic regression model was used to identify the possible association between cyberbullying and MDD. Among the 4,984 female adolescents surveyed, 8% (95% CI: 7–9) reported experiencing cyberbullying within the past 12 months, and 12% (95% CI: 11–14) were found to have MDD. Among those who had experienced cyberbullying, the prevalence of MDD was 31%. Additionally, a dose‒response relationship was observed, where the burden of MDD increases with the frequency of cyberbullying experiences. Compared with those who did not experience cyberbullying, adolescents who faced any form of cyberbullying in the past 12 months were almost four times more likely (AOR: 3.97, 95% CI: 3.12–5.05) to have MDD. Cyberbullying is a significant issue among female adolescents in Bangladesh, with a higher risk of MDD linked to increased exposure. Accessible reporting mechanisms and timely counselling through teachers or healthcare providers can help reduce the long-term mental health impact of cyberbullying. Not applicable. The online version contains supplementary material available at 10.1186/s12888-025-07234-z.
- 12 August 2025
- Published by Springer Nature in BMC Psychiatry
- Vol. 25 (1) , 783
Open AccessThe association of cyberbullying with major depressive disorders among Bangladeshi female adolescents: findings from the Bangladesh adolescent health and wellbeing survey 2019-20
Cyberbullying refers to the act of using digital technology to engage in bullying. It involves the intentional use of the internet to demean or denigrate individuals, and it has been linked to substantial psychological distress globally. Despite its increasing prevalence, there remains a gap in nationwide research on its association with the mental well-being of female adolescents in Bangladesh. The main objective of this study was to understand the association of cyberbullying with major depressive disorder (MDD) among female adolescents in Bangladesh. We used secondary data from the 2019–20 Bangladesh Adolescent Health and Wellbeing Survey (BAHWS), a nationally representative survey. A multiple logistic regression model was used to identify the possible association between cyberbullying and MDD. Among the 4,984 female adolescents surveyed, 8% (95% CI: 7–9) reported experiencing cyberbullying within the past 12 months, and 12% (95% CI: 11–14) were found to have MDD. Among those who had experienced cyberbullying, the prevalence of MDD was 31%. Additionally, a dose‒response relationship was observed, where the burden of MDD increases with the frequency of cyberbullying experiences. Compared with those who did not experience cyberbullying, adolescents who faced any form of cyberbullying in the past 12 months were almost four times more likely (AOR: 3.97, 95% CI: 3.12–5.05) to have MDD. Cyberbullying is a significant issue among female adolescents in Bangladesh, with a higher risk of MDD linked to increased exposure. Accessible reporting mechanisms and timely counselling through teachers or healthcare providers can help reduce the long-term mental health impact of cyberbullying. Not applicable. The online version contains supplementary material available at 10.1186/s12888-025-07234-z.
- Open Access
- 30 January 2024
- Published by Wiley in Health Science Reports
- Vol. 7 (2) , e1849
Understanding the impact of socioeconomic and health factors on geriatric depression: A comparative study in rural and urban Bangladesh
Background and Aims: The prevalence of depression among the elderly is a growing concern, and this study examines the differences between urban and rural areas in terms of geriatric depression. Methods: Using a two‐stage random sampling approach in urban areas and a multistage random sampling approach in rural areas, the study surveyed 944 elderly individuals of both sexes. Results: The results indicate that the prevalence of depression was high, with 52.5% of the elderly population experiencing mild to severe depression. The study found that increasing age, female gender, nuclear family structure, and involvement of housewives or others were significant factors affecting depression in urban areas, while increasing age and elderly people without spouses were significant factors in rural areas. Additionally, the study identified hearing impairment, asthma, and arthritis as risk factors for depression in rural areas, and bronchitis, heart disease, and thyroid illness as significant factors in urban areas. Conclusion: These findings highlight the need for policymakers to focus on addressing the mental health needs of older people, particularly women and those without spouses.
- 30 January 2024
- Published by Wiley in Health Science Reports
- Vol. 7 (2) , e1849
Open AccessUnderstanding the impact of socioeconomic and health factors on geriatric depression: A comparative study in rural and urban Bangladesh
Background and Aims: The prevalence of depression among the elderly is a growing concern, and this study examines the differences between urban and rural areas in terms of geriatric depression. Methods: Using a two‐stage random sampling approach in urban areas and a multistage random sampling approach in rural areas, the study surveyed 944 elderly individuals of both sexes. Results: The results indicate that the prevalence of depression was high, with 52.5% of the elderly population experiencing mild to severe depression. The study found that increasing age, female gender, nuclear family structure, and involvement of housewives or others were significant factors affecting depression in urban areas, while increasing age and elderly people without spouses were significant factors in rural areas. Additionally, the study identified hearing impairment, asthma, and arthritis as risk factors for depression in rural areas, and bronchitis, heart disease, and thyroid illness as significant factors in urban areas. Conclusion: These findings highlight the need for policymakers to focus on addressing the mental health needs of older people, particularly women and those without spouses.