How do diagnostic enrollment criteria affect treatment effect estimates in circadian rhythm sleep disorder trials?

Why the same treatment can look stronger or weaker depending on who is enrolled

Circadian rhythm sleep-wake disorders are defined by a mismatch between a person's internal body clock and the external day-night schedule, or by an alteration in the clock system itself [1]. But the symptoms—insomnia, early waking, excessive daytime sleepiness—overlap heavily with other sleep problems. If a trial enrolls people based mainly on symptoms like poor sleep, it may include many participants whose underlying problem is not circadian at all. Those participants are unlikely to respond to a treatment that targets the clock, so their inclusion can make the treatment's average effect look smaller than it really is for the intended group [1].

The reverse is also true. When enrollment requires objective evidence of a circadian misalignment—such as sleep diaries, actigraphy, or phase markers like dim light melatonin onset—the trial selects a more uniform group [1]. In that group, a treatment that works on circadian timing has a clearer signal to show. This is not because the treatment is fundamentally different, but because the comparison is being made among people who actually have the condition the treatment is designed to address.

Heterogeneous patient groups make treatment comparisons harder to interpret

A review of circadian rhythm sleep disorders noted that comparing treatment effects is limited by the heterogeneous nature of the patient populations studied [4]. That observation points directly to enrollment criteria: if different trials use different diagnostic thresholds or include different subtypes, their results are not measuring the same thing. One trial might enroll only people with delayed sleep-wake phase disorder confirmed by phase markers, while another might include anyone reporting difficulty falling asleep at a desired time. The treatment effect estimates from these two trials would not be directly comparable, and pooling them could obscure whether a treatment truly works.

Clinical guidelines for intrinsic circadian rhythm sleep-wake disorders emphasize diagnosis, study design, and patient population as key elements [3]. That emphasis reflects a practical concern: the enrolled population is part of the experiment, not an afterthought. A guideline that specifies which diagnostic features define the target population is implicitly acknowledging that treatment effect estimates depend on getting that population right.

What this means when you read a trial result

When you see a treatment effect estimate from a circadian rhythm sleep disorder trial, the first question to ask is who was enrolled and how the diagnosis was confirmed. If the trial relied only on symptom reports, the estimate may understate the treatment's benefit for people with a confirmed circadian disorder, because the enrolled group likely includes non-circadian sleep problems [1]. If the trial required objective phase markers, the estimate is more specific to circadian misalignment but may not generalize to people who would be diagnosed in a typical clinic without those markers.

This is not a reason to distrust any particular trial. It is a reason to match the enrollment criteria to the question you are asking. A treatment effect estimate is always an estimate for a particular population, and in circadian rhythm sleep disorders, the diagnostic criteria are the main lever that defines that population [1][3].

Sources used in this answer

Related research questions