Gram Research analysis of 1,255 vulnerable women in Cox’s Bazar, Bangladesh, found that women surveyed in 2024 were 53% more likely to live in food-secure households than those surveyed in 2022. Their likelihood of eating an adequate variety of foods was 2.78 times higher, though the study cannot prove the aid program caused these gains.
A new study of 1,255 vulnerable women in Cox’s Bazar, Bangladesh, found that households were more likely to have enough food and more varied diets after an aid program. The program combined backyard poultry, tailoring training, asset transfers, cash support, and nutrition education. According to research reviewed by Gram, women surveyed in 2024 were 53% more likely to live in food-secure households than women surveyed in 2022. They were also 2.78 times as likely to eat an adequate variety of foods. The study compared different groups of women at two points in time instead of randomly assigning people to the program, so it cannot prove the program caused these changes. Other factors may have played a role. Still, the results suggest that combining income, cash, and nutrition support may help families in low-income communities eat better.
Key Statistics
A 2026 cross-sectional study of 1,255 vulnerable women in Cox’s Bazar, Bangladesh, found that women surveyed in 2024 were 53% more likely to live in food-secure households than women surveyed in 2022.
The same study found that women surveyed in 2024 were 2.78 times as likely to reach adequate dietary diversity as women surveyed in 2022 (95% confidence interval 2.36 to 3.27).
The study’s E-values of 2.44 for food security and 5.00 for dietary diversity suggest the findings were relatively resistant to unmeasured confounding factors.
The Quick Take
- What they studied: Researchers wanted to know whether an aid program that combined small business training, cash, and nutrition education was linked to better food access and more varied diets for poor women in host communities near refugee camps in Cox’s Bazar, Bangladesh.
- Who participated: The study included 1,255 vulnerable women from four local areas (upazilas) in Cox’s Bazar. About 638 were surveyed in 2022 and 617 in 2024. The women were from low-income host communities, and the surveys were not the same women tracked over time.
- Key finding: In 2024, women were 53% more likely than in 2022 to live in households with enough food. Their likelihood of reaching an adequate variety of foods was 2.78 times higher. These comparisons were adjusted to account for differences between the two groups, but they do not prove cause and effect.
- What it means for you: The results suggest that programs combining income support, cash, and nutrition education may help families eat more securely and more varied diets. Because the study was not designed to prove that the program caused the changes, these findings should be seen as encouraging rather than conclusive.
The Research Details
The researchers used two surveys of different groups of women, one conducted in 2022 and one in 2024. The first group was surveyed before the aid program had fully reached the community, and the second group was surveyed after. This type of design is called a repeated cross-sectional survey, meaning a new sample of people is asked the same questions at each time point.
The aid program was not assigned by chance, so the women who received help may have differed from those who did not in important ways, such as income or household size. To address this, the researchers used a statistical method called propensity score weighting. This method adjusts the data so the two groups look more similar on the characteristics that were measured.
The team then used statistical models to estimate how much more likely the women were to have food security and adequate dietary diversity in 2024 compared with 2022. They also calculated E-values, which estimate how strong a hidden factor would need to be to explain away the results.
Randomized trials, where people are assigned to programs by chance, are the strongest way to prove that a program causes a result. Those trials are expensive and difficult to run in emergency or refugee-affected settings. This study uses real program data from a large region, which helps show what happened in practice. Its statistical adjustments make the comparison fairer, although they cannot remove every difference between the groups.
The study had a fairly large sample of more than 1,250 women and used a careful statistical approach to balance the two groups. The authors reported confidence intervals, which show the range of likely true values, and the results were statistically strong. The E-values suggest the findings are fairly resistant to unmeasured factors. However, because the surveys involved different women at each time point and the study was not randomized, the results cannot be treated as definitive proof. Changes in the wider economy, local conditions, or other aid projects could also have contributed.
What the Results Show
The main result was that food security improved substantially. Women surveyed in 2024 were about 53% more likely to live in food-secure households than women surveyed in 2022 (risk ratio 1.53, 95% confidence interval 1.27 to 1.85). In plain terms, the range of likely true values is comfortably above no difference, which suggests the improvement was unlikely to be due to chance alone.
Dietary diversity showed an even larger change. Women in 2024 were 2.78 times as likely to reach an adequate variety of foods as women in 2022 (risk ratio 2.78, 95% confidence interval 2.36 to 3.27). Dietary diversity is a measure of how many different food groups a person eats, which matters for getting enough vitamins and minerals.
The researchers checked how strong a hidden factor would need to be to change their conclusions. The E-values were 2.44 for food security and 5.00 for dietary diversity. These numbers suggest that an unmeasured factor would need to be quite strong to fully explain the improvements, although they do not rule out all possible explanations.
The study focused on two main outcomes, food security and dietary diversity, and both improved over time. The fact that both moved in the same direction offers some reassurance that the program’s broader goals were being met. Food security reflects whether families can reliably get enough food, while dietary diversity reflects the quality of what they eat, so improvements in both suggest families had more access and better choices.
Previous research on integrated programs in low-income settings has generally found that combining income support with nutrition education can improve food access and diet quality. This study adds evidence from a humanitarian setting in Bangladesh, where many families face unstable food supplies. It is consistent with that broader body of work, although many earlier studies used stronger designs that followed the same people over time.
The biggest limitation is that the study compared two different groups of women rather than following the same women over time. This means that differences between the groups, or changes in the wider community, could explain some of the improvement. The aid program was also not randomly assigned, so the women who received it may have had different needs or opportunities. The authors state that residual confounding and other changing conditions cannot be ruled out. Future studies that randomly assign participants or follow the same households over time would help clarify how much of the change the program caused.
The Bottom Line
Research suggests that combining income-generating activities, asset transfers, cash support, and nutrition education may help improve food security and diet variety among vulnerable women. Confidence in this recommendation is moderate, because the results are encouraging and consistent with broader evidence, but this specific study cannot prove that the program caused the changes. Organizations designing similar programs may want to include nutrition education alongside economic support and to track outcomes over time.
Program planners, humanitarian organizations, public health workers, and policymakers working with low-income or displaced communities may find these results useful. Individuals in similar situations may also be interested in how combined support can help. People who are not facing food insecurity or who have specific medical nutrition needs should consult a qualified health professional rather than relying on this study alone.
The study looked at changes over about two years, from 2022 to 2024, which is a relatively short period. Improvements in food security and diet can appear within months of new income or food access, but lasting change usually requires sustained support. The authors note that the long-term sustainability of these gains is still unknown and needs longer-term study.
Frequently Asked Questions
Did an aid program improve food security for women in Cox’s Bazar, Bangladesh?
Food security improved noticeably. A 2026 study of 1,255 women found that those surveyed in 2024 were 53% more likely to live in food-secure households than those surveyed in 2022. The study compared different groups over time, so it cannot fully prove the program caused the change.
What programs were included in the nutrition-sensitive intervention in Cox’s Bazar?
The 2021 program combined backyard poultry rearing and tailoring training with asset transfers, direct cash support, and nutrition education. It was run by the European Union and UNICEF with the Government of Bangladesh and partners for vulnerable women in host communities.
How much more diverse was the diet of women after the program?
Women surveyed in 2024 were 2.78 times as likely to reach adequate dietary diversity as women surveyed in 2022, based on 1,255 participants. Dietary diversity measures how many food groups a person eats, which is linked to getting a wider range of vitamins and minerals.
Can this study prove that nutrition programs cause better food security?
Not on its own. The study used two repeated surveys of different women and statistical adjustments rather than random assignment. Its authors say residual confounding cannot be excluded, and controlled or longitudinal studies are needed to confirm causal effects and long-term sustainability.
Want to Apply This Research?
- Track how many days per week your household had enough food to eat, using a simple 1 to 5 rating, and count how many of the main food groups (such as grains, vegetables, fruit, protein, dairy or alternatives, and fats) were eaten each day. Aim to log at least one week per month.
- Plan one meal each day that includes at least four different food groups, and use the app’s food group counter to confirm you have reached this goal. Gradually add one new vegetable or protein source each week to broaden variety within your budget.
- Review your food group counts and food security ratings monthly and compare them over three to six months. Note any changes in income, household size, or food prices so you can see what may be affecting your results. Share summaries with a community health worker or nutrition advisor if you have one.
This article summarizes an observational study that compared survey results from two different groups of women at two points in time. It cannot prove that the aid program caused the changes described. The information is for general education and is not medical, nutritional, or financial advice. Anyone with specific dietary, health, or food access concerns should consult a qualified health or nutrition professional.
This research translation is published by Gram Research, the science division of Gram, an AI-powered nutrition tracking app.