Dhaka

Researchers are investigating the long-term effects of reducing household air pollution (HAP) on heart, lung, and immune system health among women and children in semi-rural Bangladesh. Nearly 3 billion people worldwide, including 89% of Bangladesh's population, are exposed to harmful pollutants from burning biomass fuels like wood and cow dung for cooking. This study evaluates whether a mobile phone–based behavioral change communication (mHealth BCC) intervention can encourage exclusive use of cleaner fuel, specifically Liquid Petroleum Gas (LPG), and how this reduction in air pollution affects subclinical cardiovascular, pulmonary, and immune outcomes over time. The study uses a large household-level randomized controlled trial design where participants receive educational messages and notifications through a mobile phone app to promote LPG use. The frequency of messages varies based on participant responses, and usage is monitored by tracking engagement with educational materials. The study continues to assess personal and area air pollution exposure through detailed 24-hour and 5-day monitoring periods. Assessments include lung function tests, chest imaging, blood pressure, EKG, and blood tests for metabolic and immune markers, conducted before and two years after the intervention. Participants are followed over two years with repeated evaluations of personal and ambient air pollution levels, lung and heart function, and immune responses including vaccine antibody production. The study carefully monitors changes in inflammatory and immune cell function alongside cardiovascular and pulmonary health indicators. These comprehensive measurements aim to understand the health impacts of sustained HAP reduction achieved through behavioral change and exclusive LPG use.

Diabetic macular edema (DME) is a complication of diabetic retinopathy that causes retinal thickening near the center of the macula, leading to vision loss. This condition results from damage to the blood-retinal barrier due to diabetes, causing leakage and swelling in the retina. The study focuses on evaluating the efficacy and safety of ranibizumab, a monoclonal antibody targeting vascular endothelial growth factor (VEGF), which plays a key role in increasing retinal vascular permeability and edema. This randomized, double-blind, parallel study will include 70 patients with DME. Participants will receive three doses of 0.5 mg intravitreal injections of either the test product (Ranibizumab 10mg/ml Injection by Incepta) or the reference product (Lucentis) every four weeks, on Day 0, Week 4, and Week 8. After the first injection, a safety visit will be conducted 48 hours later, with additional safety follow-ups before each injection and 48 hours after the second and third injections. The final visit will occur at Week 12 for efficacy and safety evaluation. During the study, participants will undergo assessments including best corrected visual acuity (BCVA) and central subfield thickness (CST) measurements at baseline and Week 12. Safety will be closely monitored throughout the study with regular visits and ocular examinations. The total participation duration is approximately 12 weeks, during which researchers will evaluate changes in vision and retinal swelling to compare the two ranibizumab treatments.

Methanol poisoning is a serious health problem, especially in low- and middle-income countries, where outbreaks can cause significant harm. Diagnosing this condition is difficult because its symptoms resemble many other illnesses, and traditional lab tests require costly equipment, often leading to missed diagnoses. Methanol itself is not highly toxic, but its breakdown product, formate, can cause brain swelling and death. To improve diagnosis, researchers have developed a new bedside test that measures formate using just a single drop of blood, eliminating the need for laboratory equipment. The research involves two connected studies. The first is an observational study comparing the new point-of-care (POC) formate test with standard laboratory tests to see how well it identifies methanol poisoning. If the new test shows good accuracy, a second feasibility study will follow, using a randomized approach where entire hospitals are assigned to different diagnostic methods. This second study aims to assess whether this trial design can be used in larger research to evaluate how the POC test affects clinical care and costs. Participants suspected of methanol poisoning or unexplained metabolic acidosis at hospitals in Bangladesh and India will take part. During the studies, researchers will collect data on how quickly samples are taken and results reported, time to start appropriate treatment, and the impact on clinical outcomes like deaths and intensive treatments. The studies also aim to raise awareness about methanol poisoning, improve early detection and treatment, and develop better protocols for patient care over a total participation involving initial assessment and follow-up.

Researchers are evaluating integrated, decentralized primary care approaches for managing hypertension and diabetes in rural Bangladesh. This study aims to compare the effectiveness and cost-effectiveness of a multicomponent care package, including healthcare provider training, mHealth tools, task shifting, and community-based care, against usual care and mHealth intervention alone. The study also explores factors influencing treatment and prevention, lifestyle changes, and barriers or facilitators to care from multiple perspectives including patients, providers, and public health officials. The study involves three groups: one receiving the multicomponent decentralized care which combines government care models with the Simple App for coordinated management, community health worker involvement, and supportive supervision; a second group receiving only the mHealth intervention with training and use of the Simple App; and a third group receiving usual government primary care without added interventions. Components include training for clinicians and community health workers, use of the Simple App for patient management, screening, counseling, medication dispensing, and follow-up support. Supervision visits and refresher trainings are also part of the intervention. Participants aged 40 and above with hypertension and diabetes who live in selected rural areas are involved for up to 2 years and 9 months. Researchers will collect data on blood pressure and diabetes control, lifestyle behaviors, patient burden, and healthcare system factors through repeated surveys, cohort data, and qualitative interviews. Health economic evaluations will assess the cost-effectiveness of the interventions. The study aims to provide evidence to help improve scaling and sustainability of care for these chronic conditions in resource-limited settings.

Osteoporosis is a common bone condition that reduces bone strength and increases the risk of fractures, posing a major public health concern worldwide. Despite effective treatments, osteoporosis often goes undiagnosed until a fracture occurs. In places like Bangladesh, where the condition is prevalent, access to the standard diagnostic tool, DEXA scans, is limited, especially in rural and low-resource areas. This research aims to develop an artificial intelligence (AI) model to predict bone mineral density (BMD) from widely available X-ray images as a simpler, more accessible screening method. The study uses deep learning techniques, specifically a convolutional neural network (CNN), to analyze X-ray images of the hip and spine to estimate BMD. It includes both new data collection from patients undergoing X-rays and DEXA scans and retrospective data from clinical records. The AI model will be trained and tested on this diverse dataset, which also includes clinical information like age, gender, menopausal status, and other health factors. The goal is to create a tool that can support early osteoporosis detection and reduce reliance on DEXA scans, which are costly and scarce in some settings. Participants will provide X-ray images and DEXA scan results along with clinical data such as medical history and demographics. Researchers will evaluate the AI model's accuracy by comparing its BMD predictions against DEXA measurements using statistical metrics. The study will monitor data quality and privacy throughout. If successful, this AI tool could offer a fast, non-invasive, and cost-effective way to screen for osteoporosis, helping healthcare providers prioritize patients for further testing and improve osteoporosis management, especially in regions with limited resources.

Cervical cancer is a major health concern especially in low-income countries where late diagnosis and limited access to screening contribute to high death rates. This research aims to create and test an artificial intelligence (AI) model that analyzes colposcopic images to detect cervical cancer more accurately and efficiently. Colposcopy, a procedure to examine the cervix for cancer signs, depends heavily on doctors' skills, which can lead to inconsistent results. The study hopes the AI model will improve early detection by providing a reliable and scalable tool, particularly for underserved areas with fewer trained specialists. The study involves collecting colposcopic images and related clinical data from patients at Ibn Sina Medical College Hospital in Dhaka, Bangladesh, using both prospective and retrospective case-control methods. The intervention is colposcopy, where saline, acetic acid, and iodine solutions are applied to the cervix to help visualize abnormalities, followed by image analysis using the AI model. The procedure typically lasts 10 to 15 minutes during a single session. The AI model will be developed to classify cervical transformation zones and detect abnormalities, aiming to support clinical decision-making and improve screening workflows. Participants are women aged 18 or older who agree to cervical cancer screening and have no history of hysterectomy or cervical cancer. The study will assess the AI model's accuracy, usability, and generalizability in real clinical settings. Researchers will monitor screening results, image quality, and patient outcomes to measure the model's performance and impact on early cancer detection. Quality assurance measures will ensure reliable data collection and analysis. The study also evaluates safety and plans for handling missing or inconsistent data to support robust conclusions about the AI tool's effectiveness.

Breast cancer is a common and potentially deadly disease that requires early and accurate detection to improve patient outcomes. Traditional diagnosis depends on manual histopathological examination, but this method has drawbacks like subjectivity and limited efficiency. This research uses deep learning technology to classify breast cancer into invasive and noninvasive types from histopathological images, aiming to improve accuracy and efficiency in diagnosis. Noninvasive cancers such as ductal carcinoma in situ (DCIS) and lobular carcinoma in situ (LCIS) are confined within milk ducts or lobules, while invasive cancers spread to surrounding breast tissue and make up about 70% of cases, often with worse prognosis. The study focuses on developing an AI model that automates the analysis of breast tissue images obtained through biopsy or mastectomy, classifying them as normal benign, in situ, or invasive cancer. This model aims to surpass traditional manual methods in accuracy and speed and to provide a scalable diagnostic tool suitable for diverse healthcare settings, including underserved and low-resource areas. The research also emphasizes adapting the model to work reliably across different facilities with varying equipment and patient populations. Participants provide breast tissue samples through biopsy or mastectomy, which are then examined microscopically using histopathology. Researchers measure how accurately the AI model classifies breast tissue types compared to standard diagnostic results within one week of sample collection. The study includes collecting medical records and ensuring participant eligibility, while excluding pregnant women and those with severe medical conditions. The goal is to make breast cancer screening more accessible, efficient, and reliable worldwide, ultimately supporting early detection and better treatment outcomes.

This research evaluates whether combining poverty alleviation with depression treatment improves outcomes for low-income rural Bangladeshi women with depression. It aims to compare this integrated intervention to depression treatment alone in terms of reducing depressive symptoms at 6 months, lowering relapse rates at 18 months, and improving treatment uptake and retention. The study also explores effects on economic vulnerability, anxiety, culture-specific symptoms, quality of life, and function, alongside a mixed methods evaluation of how the interventions are delivered and received. Participants in the control group receive a 10-session, 6-month manualized group psychotherapy based on the Problem Management Plus program, which includes problem solving, social support, behavioral activation, and relaxation techniques led by trained peers. Those in the experimental group receive the same psychotherapy plus a poverty alleviation program adapted from the Graduation Program. This includes financial literacy classes, savings accounts, food support, transfer of goats with feed and veterinary care, gardening supplies, and agricultural skill training over up to 12 months. During the study, participants complete research interviews at 6, 12, and 18 months to assess depressive symptoms and other outcomes. The study measures include changes in depression scores, relapse rates, treatment participation, economic status, and psychosocial factors. Qualitative interviews with participants and staff also assess how the interventions are adopted, retained, and delivered. Total participation spans at least 18 months with ongoing monitoring of symptoms and implementation outcomes.

Researchers are evaluating the effects of treating asymptomatic bacteriuria (bacterial infection without symptoms) in pregnant individuals to potentially reduce the occurrence of small vulnerable newborns (SVN) and stillbirths (SB). This phase 3 randomized controlled trial is conducted in low- and middle-income countries as part of a global effort to improve newborn and child health, aligning with the World Health Organization's goal to end preventable deaths of newborns and children under five years old. The study involves about 1,134 pregnant participants across seven international sites working with partners from the United States. Participants are randomly assigned to receive either a 7-day course of oral nitrofurantoin monohydrate/macrocrystals (100 mg taken twice daily for 14 doses) or a matching placebo with the same schedule. This treatment phase occurs during pregnancy between 12 and 20 weeks of gestation. The study compares these two groups to assess the potential benefits of antibiotic treatment for asymptomatic bacteriuria in pregnancy. During the study, participants are closely monitored from pregnancy until 42 days after giving birth. Researchers collect data on the number of small vulnerable newborns and stillbirths, along with other health outcomes. The study also tracks adherence to the medication and assesses safety throughout the follow-up period. The total participation time covers screening, treatment, and postpartum monitoring to evaluate the impact of the interventions on both mothers and newborns.

This research evaluates whether combining cabergoline and letrozole improves ovulation rates compared to letrozole alone in women with polycystic ovary syndrome (PCOS) who seek fertility treatment. The study is a randomized controlled trial involving infertile women aged 18 to 35 diagnosed with PCOS based on Rotterdam criteria. The goal is to see if the combination therapy results in better ovulation induction. Participants will be randomly assigned to either the experimental group receiving cabergoline 0.5 mg tablets on days 2 and 9 of their menstrual cycle plus letrozole 5 mg daily for 5 consecutive days starting on day 2, or the comparator group receiving letrozole alone with the same dosing schedule. Treatment will be given for up to three cycles, with monitoring of ovarian response through transvaginal ultrasound and ovulation confirmation through urinary LH testing and serum progesterone measurement. During the study, participants will have multiple visits each cycle for baseline assessments, monitoring follicle growth, checking ovulation, and confirming pregnancy if applicable. Researchers will collect data through interviews, clinical exams, laboratory tests, and ultrasound scans. The main outcome measured is the ovulation rate within 21 days after each induction cycle, assessed over 12 weeks. Safety and drug tolerance will also be monitored throughout the study.