New meta-analysis – handwashing promotion reduces respiratory infections by 17%

Our study was just published in the Lancet, a great collaboration between a team from LSHTM*, WHO, UBC/IHME, and Emory. It’s a meta-analysis of interventions promoting handwashing with soap (HWWS) and their effectiveness in preventing acute respiratory infections (ARI) in low-income and middle-income countries (L&MICs).

Our key result is that HWWS can reduce cases of ARI in L&MICs by 17% on average (95% confidence interval: 10-24%). This is important because the ARI disease burden is big. By the end of 2021, the COVID-19 pandemic had caused between 15 and 18 million excess deaths. However, endemic lower respiratory infections (LRIs) such as pneumonia kill 2.5 million people each and every year (83% in L&MICs), giving the same total as COVID after about 6-7 years. Making a 17% dent in such a large burden would be valuable. Upper respiratory infections (URIs) such as the common cold do not kill many people but nonetheless contribute to a disease burden similar to that of measles. LRIs and URIs together (=ARIs) cause 104 million DALYs every year, and promoting handwashing could prevent a big chunk of that disease burden.

We came to our 17% estimate using meta-analysis, a statistical method for combining the results of multiple studies based on their estimated effect sizes and their confidence intervals around those estimates. While there have been meta-analyses on this topic before, our study is the first for 15 years to do so for a broad definition of ARIs, and the first ever to provide separate estimates for effects on LRI versus URI. It’s important to understand that the studies included in our meta-analysis are ones which evaluated interventions promoting handwashing. Some similar reviews to ours often include studies with designs which look at a sample of people with and without a disease, and ask them about their HWWS behaviour, then infer effectiveness from that. We do not include such study designs, which were abundant early in the COVID pandemic and carry a high risk of bias.

Our pooled estimates are not of the effect of practising HWWS at an individual level, they’re of the effect of promoting HWWS to a given population. The 26 studies we include used a wide variety of modes of promotion (e.g. household visits, group activities) and some provided materials (e.g. handwashing facility, soap) as well. Look at our Table 1 to see the variety there.

Governments rush to emphasise HWWS during respiratory epidemics, whether flu or COVID-19, but they less often pay attention to HWWS in “normal times”. Our review suggests under-prioritising routine promotion of HWWS could be a missed opportunity to reduce the large endemic burden of respiratory disease. However, successful promotion is not plain sailing. Look at our Figure 3 which covers interventions in which HWWS was the majority of intervention content, and you’ll see plenty of interventions which had no significant effect on ARI (confidence interval line overlaps 1). Plenty of these interventions were also at small scale. Promoting handwashing has to be done carefully to be effective.

In conclusion, while it is not straightforward to achieve uptake of handwashing behaviours and subsequent adherence to them, the dividends for health can be huge given the size of the endemic respiratory burden. The below “Research in Context” box summarises what our review added to the existing knowledge about the relationship between HWWS and ARI. For more detail on our methods, please read our paper!

* Sarah Bick deserves special mention for being the second reviewer. That role involved independently reviewing all the abstracts and papers I reviewed, as well as cranking the stats to get each study’s effect size and standard error in a comparable shape for meta-analysis. Thanks Sarah!