Four short videos beat a salt leaflet in a Chinese kindergarten study. Only 27% of the video group reached the final count, against 66% of the leaflet group

How much can up to eight minutes of video teach a parent about salt? Researchers in Chongqing, China, tried to find out with caregivers of kindergarten children, 84% of them mothers. One group got four clips of one to two minutes each, posted every other day for a week in the WeChat group for their child's class. They covered the risks of too much salt, daily limits, salt hidden in snacks and condiments, and practical tips such as reading labels. A second group got a printed leaflet with similar content, handed out at a parents' meeting or kindergarten event. A third got nothing extra. The results appeared in Nutrients on 26 October, from a study run between April and October 2023.

Knowledge is where the gap shows. Scored as a percentage, all three groups started level, at about 37. Two months after the intervention the video group stood at 67.3, the leaflet group at 51.8 and the group given nothing at 50.5. That last figure deserves a second look: people who received no education gained 13 points anyway. The paper doesn't explain it. Answering the same questions three times may teach something, though that's our guess. Once that drift is taken out, the leaflet's extra gain at two months was 1.60 points (95% CI −1.82 to 5.01), which can't be told apart from nothing. The videos' extra gain was 16.59 (12.85 to 20.34). On the 10-point scale behind those percentages, that's about 1.7 points by our conversion.

Now the flow chart, which is Figure 1 of the paper and the first thing we'd want a reader to see. The study recruited 2,528 caregivers: 1,106 to videos, 696 to the leaflet and 726 to nothing. The final analysis kept 300, 460 and 278. So 27% of the video group was counted, against 66% of the leaflet group and 38% of the control group. Overall 59% of those recruited dropped out of the analysis, where the sample-size plan had allowed for 15%. At the one-week step, 280 people were removed for 'failure to complete the intervention as required'. The chart doesn't split that by group, but the leaflet group lost only 71 people in all at that step, so if the control group had nothing to fail at, at least 209 of them were from the video group. That's our arithmetic. The paper's list of limitations doesn't mention dropout.

Why does that matter? Someone who stays through four videos and two more questionnaires may simply care more about the subject than someone who drifts away, and those people could be expected to learn more from anything. In fairness, the people who were kept had very similar starting scores across the groups, and the regression models adjusted for education, income, occupation and more. But interest isn't on that list, and the paper gives no starting scores for the people who left, so we can't compare them with those who stayed.

Behaviour moved much less than knowledge. On the practice score, covering things like cooking salt, salty snacks and foods with hidden salt, the video group went from 61.9 to 66.6 at two months, the leaflet group from 60.5 to 64.2 and the control group from 61.0 to 64.2. The videos' extra gain over the control group was significant at one week (2.52, 95% CI 0.90 to 4.14) but not at two months (1.48, 95% CI −0.15 to 3.10; P = 0.074). A separate regression does report a significant 2.57 at two months, but its listed adjustments don't include the starting score, and the video group started about a point higher. Either way, all of this is self-report, and the authors warn that it may overstate any improvement, because people can report what they think they should. They took no urine samples, so the study can't say whether any household or any child ate less salt.

The authors are careful about one point that a quick summary would miss. The videos and the leaflet shared core content but not their wording, illustrations or exposure time, so they say the result can't be credited to the video format itself. There's another difference in the design: a leaflet arrived once, while the clips landed on a phone four times over a week, so this may be partly a test of repetition. Allocation wasn't random either. Six private kindergartens were picked from 17 and matched into three groups of two, so whatever is particular to those two kindergartens travels with the result. For a sense of what firmer evidence looks like, a trial in Jiangxi, EduSaltS, randomised 20 primary schools to a year of app-based courses plus offline activities and measured salt in 24-hour urine. Adults in the programme ended up about 1.06 g a day lower than controls (95% CI −1.81 to −0.30), while the children's difference of −0.24 g wasn't significant. It's a far bigger programme with older children, so it doesn't test these videos. Short clips remain a cheap way to explain where hidden salt is. Whether they change what ends up on a child's plate is still an open question.