Can cicadas help us measure the effect of lost sleep?

Real-world events can randomize us to less sleep, providing insights into how sleep deprivation affects our health.

If you’ve been enjoying this Substack, then you’ll definitely enjoy our book, which just came out this past week: Random Acts of Medicine: The Hidden Forces That Sway Doctors, Impact Patients, and Shape Our Health. Available here or wherever you buy your books!

Photo by Gregory Pappas on Unsplash

This week we’re joined by Charlie Bray, who works with us closely as a (phenomenal) research assistant and public health student at Harvard. We spend a lot of time each week brainstorming research ideas with collaborators and students. The idea for this post—one of our favorites—came from Charlie.

Measuring the benefits of sleep

We all know that sleep is an important part of our daily lives, namely because we notice when we haven’t had enough of it. In polling, the average American adult reports that they usually get about 6.8 hours of sleep per night—a few minutes shy of the CDC recommended 7+ hours.

It’s not easy to study sleep. In experimental settings, you’d have to find a group of people willing to subject themselves to unusual sleeping conditions while being surveilled and isolated for days or weeks on end. But experiments of people willing to do this suggest insufficient sleep can reduce the speed at which we think and act, impair our attention and ability to complete tasks, and impair memory.

Measuring the real-world effects of sleep deprivation outside the laboratory setting, however, is challenging. We can’t ethically do a randomized study where we purposefully deprive participants of sleep and have them drive a car, operate heavy machinery, or perform surgery.

As confident as we might be that it’s a bad idea for sleep-deprived people to perform these tasks, it can be difficult to separate correlation from causation to determine to what extent lack of sleep causes problems as opposed to other contributors that might accompany it, like stress or other medical conditions. (Chris & Bapu recently discussed similar challenges with nutrition research in this New York Times essay.)

We all miss out on sleep that we know we should be getting. Sometimes we choose to trade sleep for other activities, like staying out late on the weekends or finally doing that load of laundry. Other times we forego sleep unintentionally, perhaps because we drank that coffee too late in the day, got caught up scrolling through social media, or watching “just one more” streaming episode (e.g., check out “Hijack” on Apple TV).

We can also miss out on sleep we want because of situations that we don’t choose, situations that are beyond our control. The neighbor’s yapping dog makes it impossible to fall asleep. The batteries in the smoke detectors die and need replacing—at 3am. The rumble of a low flying jet wakes us up, and we just can’t fall back asleep. When events like these are timed randomly with respect to the outcomes we care about, they create natural experiments that let us study cause and effect of such sleep disruptions.

The switch to daylight savings time is one of those situations beyond our control, and as far as our bodies are concerned, the timing of the switch is random. In a Swedish study of the natural experiment that occurs when we switch to daylight savings time, researchers found that the rate of heart attacks increases for the week following the switch in the spring, when we lose an hour of sleep. Meanwhile in the fall, when we gain an hour of sleep, there was no difference in heart attacks for the week following the switch. The best explanation for this increase in heart attacks was the hour of sleep lost while “springing forward.” Other studies have found mixed results.

Where do the enormous bugs fit in?

If you’ve had personal experience with cicadas, you can probably guess where we’re going to go with this. But for the uninitiated, some background:

Photo by Ashlee Marie on Unsplash

Cicadas are large insects that have an interesting life cycle. Newly hatched cicadas, called nymphs, burrow underground where they live, unnoticed, for years before emerging above ground to find a mate, mating, and dying shortly thereafter. In North America, there are groups of cicadas that mature and then emerge together in certain regions; these groups are known as “broods.” There are broods that emerge after 13 years underground, and others that emerge after 17 years underground.

When these broods emerge, it’s… not subtle. For starters, there are literally billions of them. They’re big, they’re loud, and they’re everywhere. Early in the morning, males start singing in a “chorus” to attract females—they’re so loud and so numerous that the sound can reach ear-damaging decibel levels similar to a jet engine. After a few weeks, the party’s over, and the cicadas die, leaving billions of tasty treats around for animals (and humans) to eat.

If you’re up for it, you can get a better sense of it all in this video:

The natural experiment

The early morning songs of the male cicadas are so loud that they could easily disrupt sleep, curtailing a human’s night’s rest when they start searching for mates early in the morning. With cicadas turning up en masse only every 13 or 17 years, the timing of the emergences are as good as random when it comes to sleep habits.

We wanted to know if any sleep disturbances caused by the chorus of male cicadas in search of mates was enough to cause health problems. For example, could the noise levels reached by cicadas, which go up to 90 decibels (the sound of a lawnmower or motorcycle), drive people to go to the doctor for insomnia medications or other related problems like headaches or cause transportation accidents? We took a look at patients’ insurance claims from the regions where various broods emerged from 2007 to 2018, and we combined it into a single analysis (called an event study) where we look at the weeks leading up to and following a brood emergence within a given region.

Here’s what we found for transportation accidents:

This event study plot shows the average number of insurance claims (per 100,000 people covered by the insurance policies in the database) across all regional brood emergences from 2007-2018. The beginning of the emergences is noted as week “0,” while week -5 is five weeks prior to the beginning of the emergence, and week 5 is five weeks after the beginning of the emergence.

There was no appreciable difference in the rate of transportation accidents during the emergence of the cicadas. Similarly, we didn’t see any meaningful differences for prescriptions for benzodiazepines (can be prescribed for sleep or anxiety) or headaches. In sum, it didn’t look like the brood emergences were enough to send more people to the doctor for the problems we had anticipated.

Why? There’s a few possible reasons. First, it could be that the cicadas weren’t actually all that disruptive to sleep—perhaps the sounds are so monotonous, albeit loud, that they didn’t wake people up any earlier. Second, the sleep disturbance might have been minor enough that, on average, patients didn’t feel the need to seek care. Another possibility is that we weren’t able to accurately localize where the cicada emergences were most concentrated, since they do like to feed as nymphs on certain kinds of trees that won’t be spread evenly throughout the regions we looked at.

Ultimately, we didn’t learn much about the effect of poor sleep from this analysis—though we did end up learning a lot about cicadas (special thanks to cicada scholars Chris Simon, Gene Kritsky, and John Cooley for guiding us). At the very least, the results suggested that at least there wasn’t a rash of transportation accidents, insomnia, or headaches brought on by cicadas.

But the clever idea behind this natural experiment—a product of Charlie’s creative thinking—was one stone we just couldn’t leave unturned.