Wrist trackers picked up poorer sleep among cannabis users that their questionnaires did not
Each additional year of cannabis use is linked to 21% higher odds of prolonged waking during sleep in a JAMA Network Open analysis of CARDIA data. Participants' own sleep ratings did not fall, and the observational design cannot show that cannabis caused the difference.
The Scientist · Science desk
Three figures: each additional year of cannabis use was linked with 21 percent higher odds of prolonged waking (at least 30 minutes awake in total), 10 percent higher odds of a low sleep percentage and 4 percent higher odds of high sleep fragmentation.
| Figure | Value | Note | Claim |
|---|---|---|---|
| Prolonged waking, per added year of use | 21% higher odds | 8 | |
| Low sleep percentage | 10% higher odds | 8 | |
| High fragmentation | 4% higher odds | 8 |
What happened
- CARDIA began recruiting in 1985 at four US sites and enrolled over 5,000 people; this analysis draws on its year-35 follow-up in 2022.
- Wrist trackers recorded seven days of sleep for 891 participants, and 753 of them also gave self-reported sleep assessments.
- The same one-year increment in cannabis use carried 10% higher odds of a low sleep percentage and 4% higher odds of high sleep fragmentation.
- People who had used cannabis in the past 30 days slept 20 minutes longer per night than never-users, yet scored worse on the other three objective measures.
Why it matters
- Who's exposed exposure About 22 million Americans already use cannabis for sleep, and in this cohort long-term users reported no worse sleep, so how rested they feel would not flag the pattern the tracker found.
- What doesn't add up contradiction Extra sleep time and worse sleep efficiency point in opposite directions in recent users, so a study that counted only duration would have scored cannabis as helping.
- What it limits constraint CARDIA is observational, so it cannot show whether cannabis disrupts sleep or whether poor sleepers use more cannabis; advice drawn from it stays a caution.
- Who has to decide decision Anyone testing cannabis for sleep has to choose the outcome: a questionnaire showed no decline here, while the tracker did.
The counts fit together. Add the 1,128 participants with self-reported sleep [5] to the 891 with wrist-tracker data, subtract the 753 who gave both [6], and the total is the 1,266 in the analysis [12]. That is about a quarter or less of the more than 5,000 people CARDIA recruited from 1985 [13]. ScienceAlert's account does not say how that subset was chosen.
The 10, 4 and 21 percent figures are changes in the odds of crossing a threshold: less than 85 percent of the sleep period spent asleep, high fragmentation, or at least 30 minutes awake in total [8]. They do not convert directly into minutes of lost sleep. The tracking window was seven days [6].
The researchers' explanation for the gap between tracker and questionnaire starts from the extra 20 minutes in recent users. "Individuals may not perceive declines in sleep quality or may interpret longer sleep duration as improved sleep despite objectively poorer sleep," they write in their paper [15]. That is a hypothesis. ScienceAlert says the reason for the discrepancy is not entirely certain [16].
Thanh-Huyen T. Vu, an epidemiologist at Northwestern University, put the cohort's finding this way: "The wearable device detected more nighttime disruption among participants with long-term cannabis exposure, but the participants themselves generally did not report worse sleep," [10]. Of the paper's broader reach, she said: "This study challenges the common belief that cannabis helps people sleep better," [14].
In our view, this cohort undercuts the case for cannabis as a sleep aid on objective measures, though it cannot test the claim directly. The thing this doesn't tell you is direction. The researchers adjusted for age, race, sex, education, smoking, alcohol use and other factors [7], but an observational design cannot settle whether cannabis disrupts sleep or whether people with disrupted sleep use more of it [2]. The cohort does have what the researchers say earlier work often lacked: long follow-up and objective sleep data [11].
What to watch
- Whether the full JAMA Network Open paper reports confidence intervals around the 10, 4 and 21 percent odds, and absolute minutes awake.
- Whether other long-follow-up cohorts with wearable data find the same tracker-versus-questionnaire gap, and which way the link runs.
- Whether trial designs that assign cannabis and track sleep objectively can separate cause from association.
Clarity's read
What the record supports and how the coverage leans. The claims behind it follow.
Reality
- Evidence50
- Adoption
- Not scored
- Hype gap+20
- Incentives
- Not scored
- Confidence45
Claim ledger
Ranked by verification strength, evidence, and original report placement.
- [1]
An estimated 22 million Americans currently use cannabis as a form of sleep aid, and almost one in four says they are likely to try the drug to help them sleep.
- [2]
A JAMA Network Open analysis of 35 years of observational data from the CARDIA study found that cannabis use was associated with multiple objective measures of poorer sleep.
- [3]
Neither lifetime nor recent cannabis use was associated with participants reporting lower subjective sleep quality on questionnaires.
- [4]
The CARDIA study began recruiting participants in 1985 from four sites across the US, ultimately involving over 5,000 people, and its ninth follow-up assessment took place in 2022 (year 35).
- [5]
The analysis used data from 1,266 CARDIA participants, including 1,128 people who self-reported sleep assessments at the end of the study period.
- [6]
Objective sleep data came from 891 participants who wore a wrist tracker for seven days, and 753 people contributed to both the subjective and objective analyses.
- [7]
After adjusting for age, race, sex, education level, smoking and alcohol use, and other factors, greater lifetime cannabis use was associated with a lower sleep percentage, higher sleep fragmentation, and waking up for longer periods during sleep.
- [8]
Each additional year of cannabis use was linked with 10 percent higher odds of a low sleep percentage (less than 85 percent of the sleep period asleep), 4 percent higher odds of high sleep fragmentation, and 21 percent higher odds of prolonged waking up (at least 30 minutes awake in total during sleep periods).
- [9]
Among recent cannabis users (used within the last 30 days), cannabis use was linked with longer sleep duration (20 minutes extra per night) but also with lower sleep percentage, higher sleep fragmentation and longer periods of waking up, compared with people who had never used the drug.
- [10]
"The wearable device detected more nighttime disruption among participants with long-term cannabis exposure, but the participants themselves generally did not report worse sleep,"
- [11]
The researchers note that previous studies of cannabis and sleep have shown mixed results and were often limited by small samples, short follow-up periods or cross-sectional designs, and that few incorporated objective sleep data.
- [12]
The 1,128 self-report participants plus the 891 wrist-tracker participants, minus the 753 in both, gives 1,266, which matches the reported analysis sample.
- [13]
The 1,266 analyzed participants are about 25 percent or less of the more than 5,000 people CARDIA recruited.
- [14]
"This study challenges the common belief that cannabis helps people sleep better,"
- [15]
"Individuals may not perceive declines in sleep quality or may interpret longer sleep duration as improved sleep despite objectively poorer sleep," the researchers write in their paper.
- [16]
ScienceAlert reports that the reason for the discrepancy between objective and subjective sleep in cannabis users is not entirely certain.
Get this ledger inside ChatGPT. Ask about any of our stories and the answer brings its figure, each claim’s verdict and the sources. Free, no account.
Sources
1 independent publisher whose own reporting we read for this story.
- sciencealert.comMillions of Americans Use Cannabis to Help Them Sleep. It Could Be Doing The Opposite.
1 article · October 15, 2026
Topics and entities
Follow any of these and your For You feed starts watching them — no settings page required.
Topics
- Cannabis and sleepFollow
- Sleep and healthFollow
- Observational epidemiologyFollow
Entities
- ScienceAlertFollow
- JAMA Network OpenFollow
- Kristen L. KnutsonFollow
- Thanh-Huyen T. VuFollow
- CARDIA studyFollow