Do wearable-measured and self-reported sleep duration give different estimates of the link between sleep and inflammatory markers?

Why a wrist tracker and a sleep diary can tell different stories

A wearable estimates sleep from movement and heart-rate patterns, while a diary records what a person remembers about their sleep. These are not the same measurement, so they can disagree even when both are trying to capture 'sleep duration.' In a study of 130 people wearing three generations of Garmin smartwatches, the average gap between smartwatch-recorded and self-reported sleep duration ranged from about 7 to 21 minutes depending on the watch generation, and only the oldest watch differed significantly from self-report [2]. That suggests the size of the gap can depend on the device, not just on the person.

The disagreement is not limited to duration. In cancer survivors, Fitbit and a standard sleep diary produced similar total sleep time on average, but the tracker recorded substantially more wake time after sleep onset, more awakenings, and higher sleep efficiency [3]. So even when two methods agree on 'how long did you sleep,' they may disagree on how broken that sleep was — and that distinction matters when the question is inflammation, because different inflammatory markers may track different aspects of sleep.

Where the two measures point in different directions

The clearest direct comparison comes from a placebo-controlled vaccine trial in young adults. There, wearable-derived short sleep duration was significantly correlated with higher CRP, an inflammatory marker [4]. The same study also collected self-reported sleep quality using the Pittsburgh Sleep Quality Index (PSQI), but the abstract reports that wearable-derived sleep duration was correlated with higher PSQI scores — that is, with worse self-reported sleep quality — rather than reporting a direct self-report-to-CRP link [4]. In plain terms, the wearable measure carried the inflammation signal in that trial, while the self-report measure was tied to the wearable measure rather than independently to CRP.

A larger study of 377 nurses complicates any simple 'wearables are better' conclusion. That study used both seven days of actigraphy (a wearable movement monitor) and seven days of sleep diaries, then measured four inflammatory markers: CRP, interleukin-6 (IL-6), interleukin-1 beta (IL-1β), and tumor necrosis factor-alpha (TNF-α) [5]. The sleep–inflammatory associations varied by racial/ethnic group, and within groups the two methods sometimes highlighted different sleep features. Among White nurses, lower mean and greater variability in actigraphy-measured total sleep time, plus greater variability in diary-measured total sleep time, were associated with higher IL-6 [5]. Among Black nurses, lower mean diary-measured sleep efficiency was associated with higher IL-6 and IL-1β [5]. Among Hispanic nurses, greater diary-measured mean total sleep time was associated with higher CRP [5]. Among Asian nurses, greater variability in actigraphy-measured sleep efficiency was associated with lower CRP [5].

When the choice of measure is likely to change the answer

The nurse study suggests that the measure matters most when the question is about a specific inflammatory marker in a specific group. The same sleep feature — say, variability in total sleep time — was linked to IL-6 when measured by actigraphy among White nurses, but the diary version of that feature also appeared for the same group, while other groups showed different patterns entirely [5]. That is not evidence that one method is uniformly more accurate; it is evidence that the two methods can capture different things, and those different things can map onto different inflammatory markers.

Device generation and population also shape the comparison. The smartwatch validation study found that the oldest watch differed significantly from self-report while newer watches did not, and it noted that agreement between smartwatch sleep scores and self-reported sleep quality broke down at the very low and very high ends of the quality scale [2]. In cancer survivors, the tracker and diary agreed on total sleep time but diverged on wake time and sleep efficiency, and only the diary — not the Fitbit — distinguished people with clinical insomnia from those with no or mild insomnia [3]. Those findings are about measurement agreement, not inflammation, but they show why a wearable and a diary can produce different estimates when either is used as the sleep variable in an inflammation analysis.

What to keep in mind when you see a sleep–inflammation result

The most defensible reading of the current evidence is that wearable and self-reported sleep duration are not interchangeable when the outcome is inflammation. In the vaccine trial, the wearable measure was the one tied to CRP [4]; in the nurse study, the two methods each contributed distinct associations depending on racial/ethnic group and inflammatory marker [5]. Neither study shows that one method is the 'true' measure of sleep for inflammation. Instead, they suggest that the choice of measure can change which sleep–inflammatory relationship you detect, and that the answer may depend on the marker, the population, and the specific sleep feature being examined.

One practical implication is that a study reporting a link between 'sleep duration' and an inflammatory marker should be read with attention to how sleep was measured. A wearable-derived duration and a diary-derived duration can differ by minutes on average [2] and can disagree on sleep continuity [3], and those differences appear large enough to matter when the outcome is a marker like CRP or IL-6 [4][5]. The evidence does not yet establish which method is preferable for which marker; it does establish that the two methods can give different estimates of the sleep–inflammation link.

Sources used in this answer

Related research questions