When the feeling of sleep and its measurement diverge

# When the feeling of sleep and its measurement diverge OCTOBER 16, 2026 Cannabis is often promoted as a sleep aid, but a long-running study suggests it may be linked to poorer sleep in midlife, and people who use it may not notice. How a night feels and how sleep is recorded can tell different stories. A long observational study of cannabis use and midlife sleep brings that difference into focus. The work was published in JAMA Network Open in 2026. It is an observational study in people, so it can show associations but cannot prove that cannabis caused any sleep problem. The researchers analyzed 1,266 adults from the Coronary Artery Risk Development in Young Adults (CARDIA) study. CARDIA followed participants for 35 years, through an examination conducted in 2020 to 2022. The team measured cannabis use in two ways. Cumulative use was counted in "cannabis-years," where one cannabis-year equals 365 days of use. Recent use meant use in the past 30 days, compared with former use or never use. Sleep was measured in two ways. Objective sleep came from seven days of wrist actigraphy, a wrist-worn monitor that tracks movement. Subjective sleep came from the Pittsburgh Sleep Quality Index, a standard questionnaire. In all, 891 people had objective data and 1,128 had questionnaire data. Among those with objective data, the average age was 60.7, 63.3% were women, 41.0% were Black and 59.0% were white. After statistical adjustment, each additional cannabis-year was linked to higher odds of three problems. These were low sleep percentage, meaning less than 85% of time in bed spent asleep (odds ratio 1.10); high sleep fragmentation, a measure of restlessness during sleep (1.04); and prolonged wakefulness after sleep onset, meaning 30 minutes or more awake during the night (1.21). Recent use was also linked to poorer sleep. Compared with former and never users, recent users had higher odds of low sleep percentage, high fragmentation and prolonged time awake at night. Compared with never users, recent users had 3.23 times the odds of prolonged waking (95% confidence interval 1.44 to 7.27). They also had 4.60 times the odds of long sleep, over eight hours, rather than six to eight hours (95% confidence interval 1.81 to 11.70). Those wide ranges mean the exact size of these effects is uncertain. The questionnaire results differed. Cannabis use was not associated with self-reported sleep quality. A person can feel they sleep fine while a monitor records more broken sleep. The authors say this shows why both kinds of measurement matter. There are limits. The study cannot say whether cannabis disrupts sleep, or whether people with disrupted sleep use more cannabis. The sleep monitoring covered one week. The abstract does not say how cannabis was taken, so these findings cannot be broken down by smoking or edibles. The authors conclude only that the results suggest a potential adverse association between long-term recreational use and midlife sleep health. For adults in midlife who use cannabis to sleep, the practical point is that feeling rested may not show the whole picture. Do not start or stop cannabis, or any sleep aid, without talking to your clinician. Next, watch for studies that track sleep before and after people start or stop using cannabis, which could help separate cause from association. Following changes before and after exposure would help clarify the relationship. Both subjective experience and objective measurement deserve a place in the answer. Sources: JAMA Network Open (2026), Northwestern University via EurekAlert! (JAMA Network Open) JAMA Network Open (2026),: https://doi.org/10.1001/jamanetworkopen.2026.38303 Northwestern University via EurekAlert! (JAMA Network Open): https://www.eurekalert.org/news-releases/1147171 ---