Dhaka

This study evaluates whether a mobile phone-based mHealth Behavioural Change Communication BCC educational intervention can improve the adoption and exclusive use of Liquid Petroleum Gas LPG for cooking among households in semi-rural Bangladesh. Household air pollutants HAP from biomass fuels such as wood, agricultural residue, and cow dung expose almost 3 billion people worldwide, including 89% of people in Bangladesh. In the earlier GEOHealth study, 24 months of LPG use reduced personal PM2.5 exposure by about 58.2% and produced changes in innate immune and inflammatory responses, while chronic cardio-pulmonary markers remained relatively stable. More than 70% of households continued using LPG after the earlier study, although not exclusively. The investigators will conduct a large household-level randomized controlled trial using an mHealth-based educational intervention and will continue following the cohort. The study will examine the long-term effects of HAP reduction on subclinical measures of cardiovascular and pulmonary dysfunction, innate and inflammatory immune function, and antibody response to vaccines. Personal 24-hour and area-wise 5-day exposure to PM2.5 and black carbon will be repeatedly assessed before and after the intervention. Lung function and lung pathology will be assessed using spirometry, Chest X-ray, and high-resolution computed tomography of the chest. Cardiovascular measures will include blood pressure and EKG, while metabolic dysfunction will be assessed using HbA1c and fasting lipid profile. Immune function will be evaluated through immune cell phenotyping, functional cytotoxic killer cells, and oxidative stress of lymphocytes.

Diabetic Macular Edema DME is a complication of diabetes that causes swelling in the central part of the retina, leading to vision loss. This condition results from changes in the small blood vessels of the retina, causing leakage and retinal thickening, often linked to increased levels of vascular endothelial growth factor VEGF. DME is a common cause of blindness in people with diabetes, and it is especially significant in Bangladesh due to the high number of adults living with diabetes. This study compares the effectiveness and safety of two treatments for DME a proposed biosimilar of Ranibizumab and Lucentis, both given as 0.5 mg injections into the eye every four weeks. A total of 70 adults with DME will be randomly assigned to receive one of these treatments in three doses over eight weeks. Safety visits will occur 48 hours after each injection, and the study concludes with an evaluation visit at week 12. Participants will undergo vision tests and imaging of the retina at the start and end of the study to measure changes in visual acuity and retinal thickness. Safety will be monitored throughout with follow-up visits shortly after each injection. Researchers will track how many patients maintain or improve their vision and assess any side effects. The total participation time for each patient is approximately 12 weeks.

Methanol poisoning is a serious health problem, especially in low- and middle-income countries, where outbreaks can cause severe harm to communities. Diagnosing methanol poisoning is difficult because its symptoms resemble other conditions, and traditional tests require costly laboratory equipment. This research aims to evaluate a new bedside test that uses a single drop of blood to detect formate, a substance only present in methanol poisoning, to improve diagnosis speed and accuracy. The research includes two parts first, an observational study comparing the new point-of-care formate test to standard laboratory tests to check its accuracy. If the new test shows good sensitivity, a second study will follow, which is a feasibility trial where hospitals are randomly assigned to different diagnostic approaches. This trial will investigate whether using the bedside test can lead to faster diagnosis and treatment, reduce unnecessary treatments, and evaluate clinical and cost outcomes. Participants suspected of methanol poisoning or unexplained metabolic acidosis at large hospitals in Bangladesh and India will be involved. During the studies, timing from patient arrival to diagnosis and treatment will be measured, along with clinical outcomes such as death rates and treatment needs. The research will also assess how well hospitals can be recruited for this type of trial and aims to raise awareness about methanol poisoning and improve care practices over the study period ending in 2028.

Researchers are evaluating a new approach to improve care for adults with hypertension and diabetes in rural Bangladesh. The study aims to understand how integrated, decentralized primary care involving training, mobile health mHealth technology, task shifting, and community-based care compares to usual care and mHealth alone. This research will explore how these methods affect treatment, prevention, patient lifestyle changes, and healthcare system costs over 36 months. The study includes three groups one receiving multicomponent decentralized care with mHealth support and community health worker involvement one receiving only mHealth intervention using the Simple App and one receiving usual government primary care. The multicomponent care combines training for providers, community clinic services, and supportive monitoring visits. The mHealth-only group uses the app for patient management without additional community care. Usual care involves standard screening and treatment at subdistrict health facilities. Participants will be monitored through community surveys and facility-based data collection over almost three years. Researchers will assess blood pressure and diabetes control, treatment steps, lifestyle factors, patient burdens like travel and wait times, and barriers to care. Evaluations include quantitative measures, qualitative interviews, and economic analysis to inform broader implementation. The primary outcome is control of hypertension and diabetes after 2 years and 9 months.

Osteoporosis is a widespread bone condition that weakens bones and increases fracture risk, posing a major health and economic challenge globally. This research focuses on developing an artificial intelligence AI model to predict bone mineral density BMD from X-ray images, aiming to improve early osteoporosis detection especially in places like Bangladesh where the standard DEXA scan is scarce and costly. The goal is to create a reliable screening tool that can assist in preventing fractures by enabling earlier diagnosis and treatment. The study collects both retrospective and prospective data from patients undergoing hip and spine X-rays and DEXA scans at a radiology center in Bangladesh. Using convolutional neural networks, the AI model will analyze these X-ray images along with clinical data such as age, gender, menopausal status, and comorbidities to predict BMD. The models accuracy will be evaluated by comparing predictions to actual DEXA results using several statistical measures and cross-validation methods to ensure consistency. Participants will include adults of all genders with varying bone density levels, including normal, low bone mass, and osteoporosis. Data collected will include demographic, clinical history, imaging, and diagnostic results. The study will monitor primary outcomes like BMD measurements of hip and spine, along with secondary outcomes such as WHO osteoporosis classification and fracture risk assessments over about six months. The final AI tool is intended to support clinicians in identifying osteoporosis earlier and prioritizing patients for further testing, potentially improving care in resource-limited settings.

Cervical cancer remains a serious health challenge, especially in low-income and middle-income countries where late diagnosis and limited access to screening contribute to high death rates. This research is focused on creating and testing an artificial intelligence AI model that can analyze images taken during colposcopy, a procedure used to look closely at the cervix for signs of cancer. The goal is to improve the accuracy and efficiency of detecting cervical abnormalities, reducing dependence on doctors experience, which can vary, and overcoming the limitations of invasive biopsies. The study involves both collecting new and reviewing past colposcopic images and related clinical information from women undergoing screening at a medical center in Bangladesh. The AI model will be developed using this diverse set of images to recognize transformation zones and detect cancer stages. Colposcopy involves applying saline, acetic acid, and iodine to the cervix during a single session lasting about 10 to 15 minutes. The study group includes women with suspected cervical abnormalities and those without, to build a comprehensive dataset and test the models generalizability. Participants will undergo colposcopic examinations, and their images and medical data will be collected for AI analysis. The study will assess the models accuracy using measures like the Swede Score and its ability to classify transformation zones. Quality assurance, data validation, and statistical analysis plans are in place to ensure reliable results. The research aims to evaluate how well the AI tool supports clinical decisions and improves early detection, with the potential to make cervical cancer screening more accessible and effective in resource-limited settings.

Breast cancer is a common and potentially deadly disease that highlights the importance of early and accurate diagnosis to improve patient outcomes. Traditional diagnosis relies on manual examination of tissue samples under a microscope, which can be subjective and slow. This research aims to improve breast cancer diagnosis by using advanced computer techniques called deep learning to classify types of breast cancer from tissue images. The types include noninvasive cancers, confined within ducts or lobules, and invasive cancers that spread into surrounding tissue and make up about 70% of cases with generally worse outcomes. The study involves analyzing breast tissue samples collected through biopsy or mastectomy using histopathology, the current gold standard diagnostic method. Researchers will use deep learning models to classify the tissues as normal, benign, noninvasive in situ, or invasive cancer. This AI approach hopes to improve the accuracy and speed of diagnosis compared to manual methods. The model is designed to be adaptable to different healthcare settings, aiming to increase access to reliable breast cancer screening, especially in underserved or low-resource areas. Participants are women who have undergone biopsy due to suspected abnormal cell growth in the breast. Their tissue samples will be analyzed with the AI model within one week of biopsy or surgery. Researchers will assess the accuracy of the AI in classifying breast tissue types. The study focuses on improving cancer detection to support early treatment decisions and better outcomes. Participants may undergo biopsy and mastectomy procedures as part of their care, with histopathological analysis used for diagnosis. The study does not include follow-up screenings and excludes pregnant women and those with severe medical conditions that might affect results or safety.

This trial focuses on low income rural Bangladeshi women with depression, studying whether combining poverty alleviation with depression treatment leads to better outcomes compared to depression treatment alone. The main questions include improvements in depression symptoms at six months, relapse rates at 18 months, and treatment uptake and retention. Additional outcomes studied include economic vulnerability, anxiety, culturally specific symptoms, quality of life, and functioning, with a mixed methods implementation evaluation involving interviews with participants and staff. Participants are assigned to one of two groups. The control group receives six sessions of a manualized group psychotherapy based on the WHO Problem Management Plus program, including problem solving, social support, behavioral activation, and relaxation techniques, along with four support group meetings. The experimental group receives the same psychotherapy plus a poverty alleviation program consisting of financial education sessions, savings accounts, consumption support for six months, a productive asset transfer of goats, animal feed and veterinary care for a year, gardening supplies, and agricultural skill building. Participants will take part in research interviews at 6, 12, and 18 months to assess depressive symptoms, relapse, economic status, anxiety, function, quality of life, and tension. The study monitors treatment adoption, retention, and fidelity through quantitative and qualitative methods. The total participation spans up to 18 months, allowing evaluation of both clinical and implementation outcomes.

Researchers are evaluating the effects of antibiotic treatment on asymptomatic bacteriuria AB during pregnancy in low- and middle-income countries. This phase 3 randomized controlled trial aims to reduce the incidence of small vulnerable newborns SVN and stillbirths SB, contributing to global health goals to prevent early childhood deaths. The study involves pregnant individuals and newborns across seven international sites collaborating with the NICHD Global Network for Womens and Childrens Health Research. Participants with AB are randomly assigned to one of two groups one group receives a 7-day course of oral nitrofurantoin monohydratemacrocrystals at 100 mg twice daily 14 doses total, while the other group receives a matching placebo on the same schedule. The study compares the impact of this antibiotic treatment versus placebo during pregnancy. During the study, participants will complete screening, randomization, treatment, and follow-up visits, with monitoring continuing until 42 days after birth. Researchers will assess the number of small vulnerable newborns or stillbirths as the main outcome. Participants must remain in the study area for the full postpartum follow-up period, during which health and safety are monitored carefully.

Researchers are studying women with polycystic ovary syndrome PCOS who are seeking to become pregnant. The study compares whether combining two drugs, cabergoline and letrozole, helps increase the rate of ovulation better than using letrozole alone. This randomized controlled trial involves participants diagnosed with PCOS according to the Rotterdam criteria and focuses on ovulation induction to improve fertility outcomes. Participants are randomly assigned to one of two groups. The experimental group takes cabergoline tablets twice on days 2 and 9 of their menstrual cycle plus letrozole tablets daily for five days starting on day 2. The comparison group takes only letrozole tablets daily for five days starting on day 2. Treatments continue for up to three menstrual cycles. Ovarian response is monitored using transvaginal ultrasound to check follicle growth on day 12 of each cycle. Participants will attend multiple visits each cycle for initial assessments, medication administration, ultrasound monitoring, and blood tests to confirm ovulation by measuring progesterone levels. Pregnancy testing and ultrasound for fetal heartbeat are done if menstruation is missed. The main outcome measured is ovulation rate within 21 days after each cycle, with additional monitoring of follicle size, ovulation confirmation, and pregnancy occurrence over 12 weeks.