Can Demographic and Health Surveys (2007-2024) Capture Alcohol Use Trends in Zambia?
This study demonstrates that Zambia's Demographic and Health Surveys (2007–2024) cannot reliably track national alcohol use trends due to significant inconsistencies in survey instruments, reference periods, fieldwork seasons, and proxy reporting, thereby preventing the country from effectively monitoring its alcohol reduction commitments.
Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer
Alcohol is a powerful force in public health, capable of causing significant harm to individuals and communities long before a person ever develops a disease. Because of this, governments and health organizations track how much people drink to see if their policies are working. In Zambia, officials have set a goal to reduce the harmful use of alcohol by ten percent. To know if they are succeeding, they need a reliable way to measure drinking habits over time. The most common tool for this is a massive household survey called the Demographic and Health Survey, which asks thousands of people about their lives, including their alcohol consumption. For years, researchers and policymakers have treated the results from these surveys, conducted every few years, as a single, continuous story of how drinking is changing across the nation.
However, a new analysis suggests that this story might be an illusion. By closely examining the specific questions asked in four different rounds of the survey between 2007 and 2024, researchers found that the tools used to measure drinking changed so much that the results cannot be compared directly. It is as if a thermometer was replaced with a ruler, then with a clock, and then with a different kind of thermometer, and someone tried to claim they were tracking the same temperature over time. The study concludes that the apparent drop in alcohol use that Zambia has been celebrating is likely not a real change in behavior, but rather a result of these shifting questions, the time of year the surveys were conducted, and the way people report their habits.
The researchers, working with data from the University of Zambia and the National HIV/AIDS/STI/TB Council, began by looking at the raw numbers that had been published. On the surface, the data told a clear story of improvement. In 2007, about forty-two percent of men reported drinking alcohol. By 2024, that number had fallen to twenty-eight percent. For women, the numbers also showed a decline, dropping from eleven percent to less than nine percent. If one simply accepted these numbers at face value, it would look like Zambia was successfully meeting its national health goals. But the team suspected that the questions themselves were the problem. They discovered that the surveys did not ask the same thing every time. In 2007, the question asked men only about beer, ignoring other types of alcohol that are very common in Zambia. In later years, the questions changed to ask about all alcohol, but then shifted again to ask about drinking in the last thirty days, whereas earlier surveys asked about an undated current status or just the last week. Furthermore, one entire round of the survey in 2018 did not ask anyone about their own drinking at all.
To test whether the decline was real, the researchers tried to reconstruct the data using different methods to see if the trend held up. They first looked at the time of year the surveys were conducted. Drinking habits often change with the seasons, and the surveys were not run during the same months every time. When the researchers restricted their comparison to only the months that the 2013–2014 survey and the 2024 survey had in common, the clear drop in male drinking disappeared. The numbers became so uncertain that the decline could not be confirmed; it was just as likely that drinking had stayed the same or even increased slightly. This suggested that the earlier appearance of a drop was largely due to comparing different seasons rather than different years.
The team also looked at a different type of question that remained the same in every survey: asking a woman if her husband or partner drank alcohol. This question provided a consistent way to track trends because the wording never changed. The results here showed a steep decline in reported partner drinking from 2007 to 2018, but then the numbers flattened out and stopped falling between 2018 and 2024. This pattern contradicted the idea of a steady, continuous national decline. Moreover, the researchers checked how accurate these reports were by comparing what wives said about their husbands with what the husbands said about themselves. They found that while wives were very good at identifying husbands who did not drink, they became less and less likely to report that their husbands were drinking over time. This suggests that the decline seen in the partner reports might be due to people becoming less willing to talk about drinking, rather than an actual drop in consumption.
The study also examined the 2024 survey on its own, without trying to compare it to previous years, to understand the current situation in Zambia. The data showed that about twenty-eight percent of men and less than nine percent of women were current drinkers. The analysis revealed interesting differences in who was drinking. Among men, those with more money were actually less likely to drink than those with less money. Among women, the opposite was true: those with more money were more likely to drink. Drinking was also more common in cities than in rural areas for both sexes. These findings provide a clear picture of the current landscape, but they cannot be used to say whether things are getting better or worse compared to the past.
The authors conclude that the current national surveys in Zambia are not capable of tracking the trend in alcohol use. The changes in the questions, the time of year the surveys were run, and the shifting willingness of people to report drinking have created a fog that hides the true trend. Because of this, the country cannot yet tell if it is meeting its goal to reduce harmful alcohol use. The researchers recommend that future surveys keep the exact same questions about alcohol to ensure consistency. They also suggest adding a specific question about heavy drinking on a single occasion, which is a key part of the global health target, and reporting exactly which months the survey took place. Until these changes are made, any claim that alcohol use in Zambia is falling is based on data that cannot support it. The real story is not that drinking has stopped, but that we have not yet built the right tools to measure it accurately.
Drowning in papers in your field?
Get daily digests of the most novel papers matching your research keywords — with technical summaries, in your language.