Atlas›Domain II · Behavioural & cognitive›Guide 070
Guide II-070

Paradoxical intention for insomnia

Espie, Ascher1979
A 16 · L 63 · M 2118 interventions in the Trinity explorer
Client cardCompare
[source: …] where a claim comes from[uncertain] not confirmed in a primary source[synthesis] the writer's reconstruction

1. Introduction

Version followed: Espie's paradoxical intention (PI) protocol as set out in his chapter "Paradoxical Intention Therapy" (Espie, 2011, in Perlis, Aloia & Kuhn, eds, Behavioral Treatments for Sleep Disorders). The core in-bed instruction is cross-checked against Ascher & Turner (1979), as reported by Jansson-Fröjmark et al. (2023). The two-session delivery format comes from Ong, Lazar & Broomfield (2025), and sequencing within CBT-I follows Jansson-Fröjmark, Sandlund & Norell-Clarke (2023). Where the earlier Ascher versions differ, this is noted.

PI for insomnia is a single-component cognitive-behavioural technique, not a full therapy. It is brief and individual, used either on its own or added to CBT for insomnia (CBT-I). It descends from Viktor Frankl's logotherapy: Frankl gave the insomnia example of someone who fears sleeplessness and so "hyper-intends" sleep, and proposed the paradoxical intention of trying not to fall asleep [source: Frankl 1959 as summarised in a secondary source; primary text not drawn on — uncertain]. Ascher and colleagues adapted it for sleep-onset insomnia in the late 1970s. Espie later reframed it around sleep effort [source: Espie 2011].

Theory of change. Sleep is an involuntary process. Insomnia is kept going by trying to sleep: effort, intention, monitoring, and anxiety about performing. These raise arousal and block the automatic onset of sleep. When the person stops trying, and even gently intends to stay awake, sleep effort and sleep-performance anxiety fall and sleep can arrive by itself. PI is meant to change the person's outlook as well as their behaviour [source: Espie 2011; Broomfield & Espie 2003]. Other mechanisms have been proposed: the attention–intention–effort pathway, Wegner's ironic-process theory, exposure to bedroom cues with habituation, accumulated sleep drive, and acceptance of being awake. These are hypotheses [source: Espie 2011; Jansson-Fröjmark et al. 2023].

Where it sits. Current reviewers place PI as a second option after CBT-I or as an adjunct to it. They also suggest it may suit people for whom sleep restriction is unsuitable. The 2021 AASM guideline made no recommendation for PI as a single-component treatment [source: Jansson-Fröjmark et al. 2023]. Training is not separately credentialed; it is usually acquired within CBT-I or behavioural sleep medicine training [source: APA Div 12 page; codebook].

2. Framework

No source gives a fixed session-by-session manual. Espie describes the elements in sequence but says they can be reordered or picked per patient. The phases below assemble the published elements; the phase labels and transition criteria are [synthesis].

Format and dose.

  • Individual sessions, in person or by video [source: Ong et al. 2025].
  • Usual practice is to introduce PI at the first session and use the next 2–3 sessions for problem-solving and fine-tuning. Standalone PI has typically run 2–4 sessions, and no optimal dose is established [source: Jansson-Fröjmark et al. 2023].
  • Tested examples:
    • Two 1-hour online sessions two weeks apart, with 14 nights of practice [source: Ong et al. 2025].
    • Four weekly online self-help modules with 15 minutes or less of therapist contact per week [source: Salim et al. 2024].
    • One PI session embedded in a nurse-led group CBT-I programme [source: Jansson-Fröjmark et al. 2023, citing Sandlund et al. 2017].

Phase A: Assessment and selection (1 session plus 1–2 weeks of sleep diary)

  • Goals:
    • Confirm insomnia disorder.
    • Screen for other sleep disorders, such as sleep-related breathing disorders.
    • Get a baseline sleep diary.
    • Rate sleep effort, for example with the Glasgow Sleep Effort Scale (GSES).
    • Decide whether PI fits.
  • [source: Ong et al. 2025; Salim et al. 2024; Jansson-Fröjmark et al. 2023]
  • Move on when:
    • Insomnia is the main sleep problem.
    • Sleep effort or performance anxiety is evident.
    • The client can grasp a non-literal instruction.
    • [synthesis]

Phase B: Rationale and formulation (Session 1, about 60 minutes)

  • Content:
    • Discuss sleep normalcy.
    • Have the client complete the GSES as a good sleeper would, then as themselves.
    • Formulate the problem as a "sleep effort syndrome".
    • Draw parallels and use humour where it fits.
    • Introduce giving up trying, via Method 1 and/or Method 2.
  • [source: Espie 2011; Ong et al. 2025]
  • Move on when: the client can say the rationale back in their own words and has agreed a plan for tonight and for night wakings [synthesis].

Phase C: Nightly practice (about 2 weeks)

  • The client applies the chosen method at bedtime and after any prolonged night waking, and keeps the diary.
  • Warn the client in advance that sleep may not improve at once and may briefly worsen [source: Espie 2011; Ong et al. 2025; Jansson-Fröjmark et al. 2023].

Phase D: Review and fine-tuning (1–3 sessions)

  • Content:
    • Review the client's experience of the instructions.
    • Revisit the thin line between commitment and effort.
    • Add parallels such as the white-bear experiment.
    • Correct drift into active wakefulness or effortful "trying to win".
  • [source: Ong et al. 2025; Espie 2011; Jansson-Fröjmark et al. 2023]
  • Move on when:
    • Sleep-onset latency and wake time on the diary are falling.
    • Sleep effort on the GSES is falling.
    • The client describes less struggle in bed.
    • [synthesis]
  • If there is no change after 2–3 weeks of adequate practice: switch to, or add, stimulus control, sleep restriction or full CBT-I, delivered sequentially [synthesis, informed by Jansson-Fröjmark et al. 2023].

Phase E: Consolidation and relapse prevention (final session or module)

  • Plan for setbacks, and keep giving up effort as a standing attitude rather than a nightly ritual [source: Salim et al. 2024, week 4; Espie 2011 on the outlook change].

3. Guiding principles

  • Sleep is involuntary. You cannot will sleep; you can only stop getting in its way. In practice, never set the client a sleep target. [source: Espie 2011; Ong et al. 2025 Appendix A]
  • The target is effort and performance anxiety, not sleep. Measure success by less struggle and less worry first, and diary sleep second. [source: Broomfield & Espie 2003; Espie 2011]
  • The instruction is a vehicle for an outlook change, not a literal command. If the client is "winning" at staying awake, the method has been misunderstood. [source: Espie 2011]
  • Passive, not active, wakefulness. Lights off, eyes open, lying comfortably. No reading, screens, planning or deliberate self-stimulation. [source: Espie 2011; Ascher & Turner 1979 via Jansson-Fröjmark et al. 2023]
  • Gentleness. Resist heavy eyelids softly, never by force. Coach a soft inner tone. [source: Espie 2011; Ong et al. 2025]
  • Commitment is good; effort is not (the thin line). Praise sticking with the programme, and flag forcing, desperation and "trying to win". [source: Espie 2011]
  • Normalise through the good sleeper. Good sleepers do nothing to sleep, so use them as the reference point. [source: Espie 2011]
  • Humour is a legitimate tool. Exaggerate catastrophes until they become absurd, with the client and never at their expense. [source: Espie 2011, Box 6.1]
  • Acceptance over control. Being awake is allowed, and even an opportunity. This links to mindfulness and acceptance approaches. [source: Espie 2011; Jansson-Fröjmark et al. 2023]
  • Collaborative formulation. Reach the "sleep effort syndrome" understanding together, using the client's own examples. [source: Espie 2011]
  • Select on theory. Offer PI where sleep effort or performance anxiety is prominent, there is enough sleep drive, and the client prefers it. [source: Jansson-Fröjmark et al. 2023]
  • Don't deliver PI at the same time as conflicting components. Stimulus control (leaving the bed) and sleep restriction (less time in bed) clash with PI, so use them in sequence. [source: Jansson-Fröjmark et al. 2023]
  • Pre-empt paradoxical "failure". Tell clients that falling asleep quickly while trying to stay awake is the method working, not a failure. [source: Jansson-Fröjmark et al. 2023]
  • Therapist skill matters. Espie calls standalone PI an advanced form of sleep psychotherapy, so rehearse the rationale before using it. [source: Espie 2011]

4. Intervention inventory

Interventions 4.2–4.9 are Espie's protocol elements, in his order [source: Espie 2011]. Items 4.10–4.17 are the earlier variants and the delivery and integration procedures from other sources, each tagged. All example language is illustrative unless it is quoted and cited.

4.1 Assessment, case selection and baseline monitoring

  • Purpose / when: Before starting. Confirms PI is appropriate and gives a baseline to judge change against.
  • Steps:
    1. Take a standard insomnia history: onset versus maintenance pattern, duration, daytime effects, medications and hypnotics, substances, and psychiatric comorbidity.
    2. Screen for other sleep disorders, such as sleep-disordered breathing, restless legs and circadian disorders. Refer for these where indicated. Trials excluded clients with other sleep disorders, ongoing psychological treatment, recent medication changes or sedating medication, and one trial also excluded psychotic or bipolar disorder, active substance use disorder, and severe depression with suicidal intent [source: Ong et al. 2025; Salim et al. 2024].
    3. Do a standard suicide-risk assessment wherever depression is present, and follow your service's risk pathway before starting brief sleep work [synthesis; standard practice].
    4. Ask the client to keep a daily sleep diary for 1–2 weeks: bedtime, sleep-onset latency, wakings, final wake time, rise time and sleep quality. Collect an Insomnia Severity Index [source: Salim et al. 2024; Ong et al. 2025].
    5. Administer the GSES, a brief 7-item self-report of sleep effort over the past week (Broomfield & Espie). It covers:
      • effort to sleep
      • wish to control sleep
      • avoiding bedtime
      • worry about not sleeping
      • feeling poor at sleeping
      • pre-bed anxiety
      • worry about consequences
      Each item is rated on a 3-point scale. Do not reproduce the items; obtain the scale from its authors or publication [source: Espie 2011]. One trial required a GSES score above 2 to enter, which is a research eligibility rule, not a clinical cut-off [source: Ong et al. 2025].
    6. Look for the features that suggest a good fit:
      • visible sleep effort or performance anxiety, such as the client treating sleep as something to be controlled
      • fear of poor sleep
      • bedroom cues that raise arousal
      • enough sleep drive
      • difficulty accepting unwanted thoughts
      • sleep-onset problems (the evidence on this point is mixed)
      • preference for a brief method
      [source: Jansson-Fröjmark et al. 2023]
    7. Look for the features that argue against PI. These are clients who take instructions very concretely and clients with cognitive impairment [source: Espie 2011].
  • Example language (illustrative): "Before we change anything, I'd like a fortnight of your diary as it is, so we can see what actually shifts."
  • Watch for: Clients who start rating their sleep perfectly every night in the diary, which is a sign of preoccupation. Espie notes that diaries can themselves be paradoxical. Keep diary entries brief and done in the morning [source: Espie 2011, Table 6.2; brevity advice is synthesis].

4.2 Considering sleep "normalcy"

  • Purpose / when: Opens Session 1. Establishes that good sleepers do nothing to sleep, which sets up the idea that effort is the problem.
  • Steps:
    1. Ask the client what a good sleeper does to fall asleep.
    2. Draw out that good sleepers do not think about sleep, plan it or try.
    3. Contrast this with the client's own bedtime routine of strategies and monitoring.
    4. Use light humour to puncture catastrophic predictions.
    [source: Espie 2011]
  • Example language (illustrative): "Think of your partner, or a friend who sleeps like a log. What do they do to fall asleep?" … "So the expert sleeper's technique is to have no technique."
  • Watch for: The client defending their strategies, for example "but my routine helps". Don't argue. Note the strategies so they feed into the formulation in 4.4 [synthesis].

4.3 Measuring sleep normalcy (GSES from the good sleeper's perspective)

  • Purpose / when: Session 1. Makes the absence of effort in normal sleep concrete and gives a number to compare against.
  • Steps:
    1. Give the client the GSES and ask them to answer as an imagined good sleeper would.
    2. Score it. The good sleeper's total should be at or near zero.
    3. Discuss what a zero means: no trying, no control, no worry.
    [source: Espie 2011]
  • Example language (illustrative): "Fill this in as if you were that friend who sleeps like a log … Zero. Interesting. They score nothing on effort at all."
  • Watch for: Clients who give the good sleeper a high score. This shows a belief that sleep needs work. Explore it gently before moving on [synthesis].

4.4 Formulating insomnia as a "sleep effort syndrome"

  • Purpose / when: Session 1, straight after 4.3. Gives the shared explanation that makes giving up trying make sense.
  • Steps:
    1. Ask the client to complete the GSES again, this time for themselves.
    2. Compare their score with the good sleeper's.
    3. Build a simple vicious-cycle diagram with the client, using their own examples: fear of not sleeping, then trying hard and monitoring, then arousal and frustration, then lying awake, then more fear [source: Espie 2011; Ong et al. 2025 Appendix A].
    4. Name the core points: you cannot make yourself fall asleep, and effort keeps you awake. One trial's handout states "you cannot make yourself fall asleep" [source: Ong et al. 2025].
    5. Arrive at the label "sleep effort syndrome" together, rather than announcing it [source: Espie 2011].
  • Example language (illustrative): "Your score is 11 and the good sleeper's is 0. Looking at this circle, what do you notice happens the harder you try?"
  • Watch for: The client nodding along without engaging. Ask them to explain the cycle back to you. Real agreement shows up when they offer their own examples [synthesis].

4.5 The "thin line": commitment versus effort

  • Purpose / when: Session 1 or the first review session. Prevents the client from turning PI into another form of effort.
  • Steps:
    1. Draw a line with two sides.
    2. On the helpful side, list commitment: being motivated, following advice, sticking with the programme, returning after lapses, not giving up.
    3. On the unhelpful side, list effort: trying too hard, becoming more preoccupied, forcing sleep, "trying to win", desperation.
    4. Ask the client where their own behaviours fall.
    5. Make clear that commitment is welcome and effort is the target.
    [source: Espie 2011, Figure 6.2; Ong et al. 2025 session 2]
  • Example language (illustrative): "Doing the homework every night is on this side, which is great. Lying there gritting your teeth to make the method work is over here."
  • Watch for: Clients who frame insomnia as an enemy to defeat. Espie warns that clients who try to beat insomnia lose that battle [source: Espie 2011].

4.6 Drawing helpful parallels

  • Purpose / when: Session 1 and reviews. Makes the paradox credible through everyday experience.
  • Steps:
    1. Offer one or two examples from Espie's list:
      • sleeping better in a sleep laboratory than at home
      • finally dozing off once relieved that morning is near
      • choking in sport
      • not scratching an itch
      • trying to get rid of a tune in your head
      • an audition going worse than expected
      • overthinking a golf swing
      • threading a needle
    2. Optionally run the white-bear thought-suppression demonstration: ask the client to try not to think of a white bear for a minute, then debrief on how often it came back.
    3. Ask the client for examples from their own life.
    [source: Espie 2011; Ong et al. 2025 session 2]
  • Example language (illustrative): "For the next minute, whatever you do, don't think of a white bear … How did that go?"
  • Watch for: The analogy that lands is usually the client's own. Reuse it in later sessions as shorthand [synthesis].

4.7 Humour to deflate catastrophising

  • Purpose / when: Any session where the client catastrophises about the consequences of poor sleep. Espie's Box 6.1 models it.
  • Steps:
    1. Take the client's feared consequence and extend it playfully through further catastrophes, for example poor work, then the boss noticing, then losing the job, then family hardship.
    2. Continue until the client pushes back that this is overblown.
    3. Share the laugh.
    4. Link it to a wider pattern of trying too hard, if one exists, and invite the client to give themselves a break.
    [source: Espie 2011, Box 6.1, paraphrased]
  • Example language (illustrative): "So one bad night, then the report is late, then the boss fires you, then the house goes, then the dog has to be rehomed…?" (pause) "Bit much?"
  • Watch for: Use this only with a working alliance and a client who can take a joke. If the client looks hurt or unheard, stop, apologise and validate (see section 5, Ruptures). This is not cognitive restructuring: you are not testing probabilities [source: Espie 2011 on the difference from standard cognitive restructuring; rupture guidance is synthesis].

4.8 Method 1: "Turning the tables" (giving up trying via acceptance)

  • Purpose / when: The lighter, acceptance-based route to giving up trying. Espie derives it from Fogle and Dyal's "paradoxical giving up". Suits clients who find an explicit stay-awake instruction odd, or who are strongly drawn to acceptance [source: Espie 2011; client-fit note is synthesis].
  • Steps:
    1. Invite the client to take a carefree attitude to the insomnia wherever possible, and to accept that they have it.
    2. Encourage them to welcome chances to get out of bed rather than clinging to it.
    3. In bed or before bed, have them deliberately imagine as many disastrous consequences of being awake as they can, until these look exaggerated and absurd. The aim is to go with it, not to dispute it.
    4. Reframe wakefulness as an opportunity, and plan something useful or enjoyable to do with waking time.
    [source: Espie 2011, Table 6.1, paraphrased]
  • Example language (illustrative): "Tonight, if you're awake, see how ridiculous you can make the disaster story. Then notice you're simply a person lying in a dark room."
  • Watch for: Getting out of bed in this method overlaps with stimulus control. Agree in advance whether the client gets up or stays, so the two methods are not mixed in one night (see 4.15) [synthesis informed by Jansson-Fröjmark et al. 2023]. Progress shows as lower distress about being awake, even before sleep changes [source: Espie 2011 on mechanism; inference].

4.9 Method 2: Explicit paradoxical intention (Espie's version)

  • Purpose / when: This is the core technique and the default for clients with marked sleep effort. It is introduced at the end of Session 1 and practised nightly for about two weeks [source: Espie 2011; Ong et al. 2025].
  • Consent and safety first:
    • Explain that this is a counter-intuitive, partly exposure-like method, and that early sleep may worsen briefly, sometimes with more anxiety [source: Jansson-Fröjmark et al. 2023].
    • Get explicit agreement to try it for the agreed period.
    • Agree that if the client becomes markedly sleepy in the daytime, they will avoid driving or risky work and contact you [synthesis].
    • Do not use PI when the client takes instructions very concretely or has cognitive impairment [source: Espie 2011].
    • Take extra care where sleep loss is itself a risk, such as a history of bipolar disorder or seizure disorder. PI asks for passive time awake in bed, not deliberate sleep deprivation, but agree the plan with the treating team. The sources do not establish safety in these groups [synthesis; uncertain].
  • Steps (client instructions, paraphrased from Espie 2011 Table 6.1 and Ong et al. 2025 Appendix A):
    1. Go to bed at your usual time, when sleepy.
    2. Turn the lights off and lie comfortably.
    3. Keep your eyes open.
    4. Give up any effort to fall asleep, and any concern about still being awake.
    5. When your eyelids feel heavy, gently tell yourself to stay awake a few minutes more, and that sleep will come naturally when it is ready.
    6. If you feel worried or irritable, repeat a short reassurance. One trial's handout used "Staying awake will help me get to sleep quicker" [source: Ong et al. 2025].
    7. Do not force yourself awake. Don't stimulate yourself with thoughts, movement, reading or screens. This is "a passive approach rather than an active one" [source: Espie 2011].
    8. Apply the same instructions if you wake in the night and don't drop back off quickly [source: Espie 2011].
  • Therapist steps:
    • Rehearse the instructions in session by having the client describe tonight step by step.
    • Give a one-page written summary in your own words.
    • Pre-empt the "failure" of falling asleep fast (see section 5).
    • Agree the review date.
    [rehearsal and written summary: Ong et al. 2025 used an instruction sheet; rehearsal is synthesis]
  • Example language (illustrative): "Tonight, lights out, lie comfortably, eyes open. Your only job is to stop trying. If your eyelids get heavy, just say to yourself, kindly: 'a couple more minutes awake, sleep can come when it's ready.' No forcing either way."
  • Watch for:
    • Clients actively fighting sleep, for example pinching themselves, sitting up or using the phone. That is effort, so return to passivity.
    • Clients reporting that they fell asleep "too quickly to do it". That is the method working.
    • Signs it is working: lower GSES and lower bedtime anxiety, then shorter diary sleep-onset latency and less wake after sleep onset. In the mechanism study, effort and anxiety fell even when actigraphic onset did not change significantly, so expect subjective change first [source: Broomfield & Espie 2003].

4.10 Classic paradoxical intention (Ascher & Turner "Type A")

  • Purpose / when: The original, simpler behavioural version used in the 1979 experiment. Use it with clients who want a plain instruction without the effort formulation, or as a minimal version in groups [source: Ascher & Turner 1979 via Jansson-Fröjmark et al. 2023; client-fit note is synthesis].
  • Steps:
    1. Give a full rationale: performance anxiety about sleep interferes with sleep, so removing the demand to sleep removes the anxiety. Ascher & Turner's participants received this explanation [source: Ascher & Turner 1979 abstract].
    2. Instruct the client to lie comfortably in bed in a dark room with eyes open and try to remain awake as long as possible.
    3. Tell them not to engage in activities that are incompatible with sleep, such as reading, TV or eating, while doing so.
    [source: Jansson-Fröjmark et al. 2023 summarising Ascher & Turner 1979]
  • Example language (illustrative): "In bed, lights off, eyes open: see how long you can stay awake, without doing anything to keep yourself awake."
  • Watch for: This wording ("as long as possible") invites more effort than Espie's gentler version. If the client turns it into a contest, switch to the 4.9 wording [synthesis].

4.11 Paradoxical intention added to relaxation (Ascher & Efran)

  • Purpose / when: For clients whose relaxation training has not helped, often because they use relaxation as a tool to force sleep. Ascher & Efran (1978) tested PI only as an add-on to relaxation that had previously failed [source: Jansson-Fröjmark et al. 2023].
  • Steps:
    1. Keep the relaxation practice.
    2. Instruct the client to continue relaxing longer than usual, while resisting the urge to fall asleep instead of hoping relaxation will produce sleep.
    3. Reframe relaxation as an end in itself.
    [source: Jansson-Fröjmark et al. 2023 summarising Ascher & Efran 1978; primary not drawn on]
  • Example language (illustrative): "Keep doing the relaxation, but now it isn't a sleep button. Relax, and gently stay awake while you do it."
  • Watch for: Ascher & Efran reported that some clients had trouble carrying out PI because they fell asleep too quickly. Treat this as success [source: Jansson-Fröjmark et al. 2023]. Relaxation and mindfulness generally combine more easily with PI when practised outside bed [source: ibid.].

4.12 Observational ("reframing", Type B) variant

  • Purpose / when: The same in-bed instruction, plus asking the client to notice and later report their thoughts while awake. It has mainly been used as an active control in research. Clinically it can gather material on dysfunctional sleep beliefs and repetitive thinking for later cognitive work [source: Jansson-Fröjmark et al. 2023]. Its standalone clinical value is not established [uncertain].
  • Steps:
    1. Give the 4.10 or 4.9 instruction.
    2. Add a brief morning note of the main thoughts noticed while awake. Do not take notes in bed [morning-note timing is synthesis].
    3. Bring the notes to the next session for the formulation or cognitive work.
  • Example language (illustrative): "While you're lying there awake, just notice what your mind serves up. In the morning, jot two or three lines."
  • Watch for: Observation turning into monitoring. If the notes grow long or anxious, drop this variant [synthesis].

4.13 Applying PI to night-time awakenings

  • Purpose / when: For sleep-maintenance problems. Espie reports no evidence that PI works differently for onset versus maintenance insomnia, and the instruction applies after waking [source: Espie 2011]. Some reviewers speculate onset may respond better because sleep drive is higher then [source: Jansson-Fröjmark et al. 2023; uncertain].
  • Steps:
    1. On waking and not quickly returning to sleep, lights stay off.
    2. Lie comfortably with eyes open and give up trying.
    3. Gently allow a few more minutes awake.
    4. Do not clock-watch; turn the clock away [clock advice is synthesis].
  • Example language (illustrative): "If you wake at three, you don't need to rescue the night. Eyes open, lie easy, let yourself have a few more minutes awake."
  • Watch for: Clients using PI as an emergency technique only on bad nights, which keeps sleep framed as a performance. Encourage the attitude every night [synthesis].

4.14 PI as a one-night behavioural experiment

  • Purpose / when: A brief way to introduce PI within group or multicomponent CBT-I. In one Swedish nurse-led group programme it was introduced as a single-night experiment before stimulus control and sleep restriction [source: Jansson-Fröjmark et al. 2023 describing Sandlund et al. 2017; trial paper not drawn on; uncertain on details].
  • Steps:
    1. Set up the prediction ("If I try to stay awake, I'll be awake all night").
    2. Run the 4.9 or 4.10 instruction for one night.
    3. Record what happened.
    4. Debrief at the next session against the prediction.
    [prediction–test–debrief structure is synthesis using standard behavioural-experiment practice]
  • Example language (illustrative): "What's your prediction if you try to stay awake tonight? Let's write it down and test it."
  • Watch for: About 30% of group participants rated PI useful immediately after treatment, so many clients will not take to it from one exposure. Offer the fuller protocol (4.2–4.9) to those who do [source: Jansson-Fröjmark et al. 2023; offer advice is synthesis].

4.15 Sequencing PI with stimulus control, sleep restriction and other CBT-I components

  • Purpose / when: Whenever PI is used alongside CBT-I.
  • Steps:
    1. Do not deliver PI at the same time as stimulus control or sleep restriction. Leaving the bed when awake contradicts lying in bed awake, and restricting time in bed leaves little room for PI. Run them sequentially, in either order [source: Jansson-Fröjmark et al. 2023].
    2. Elements that do fit with PI include going to bed only when sleepy and not using the bed for other activities [source: ibid.].
    3. Cognitive components combine more easily with PI than behavioural ones [source: ibid.].
    4. Point out the paradox already present in other components: staying up later under sleep restriction, getting up under stimulus control, tensing in order to relax, testing the worst case in cognitive therapy [source: Espie 2011, Table 6.2].
    5. Choose the order on clinical grounds [synthesis]:
      • PI first when effort and anxiety dominate and the client doubts sleep restriction.
      • Sleep restriction first when time in bed greatly exceeds sleep. Sleep restriction builds sleep drive, which may later help PI [source: Jansson-Fröjmark et al. 2023 on sleep drive].
  • Example language (illustrative): "For the next two weeks we're doing one thing only, the staying-awake approach. After that we'll review and decide whether to add the time-in-bed schedule."
  • Watch for: Clients mixing rules on the same night, for example getting up after 15 minutes and also trying to stay awake in bed. Write down a single rule per phase [synthesis].

4.16 Review and fine-tuning sessions

  • Purpose / when: Sessions 2–4. Consolidate the outlook change and correct misapplication [source: Jansson-Fröjmark et al. 2023; Ong et al. 2025].
  • Steps:
    1. Review the diary and, optionally, repeat the GSES.
    2. Ask for a night-by-night account of applying the instructions, covering what they actually did and thought.
    3. Revisit the thin line (4.5) and add parallels (4.6), for example the white-bear demonstration in Session 2 [source: Ong et al. 2025].
    4. Identify drift into effort or active wakefulness and correct it.
    5. Encourage nightly use and persistence [source: Salim et al. 2024, weeks 2–3].
    6. Decide whether to continue, switch method (Method 1 or Method 2), or move to other components [synthesis].
  • Example language (illustrative): "Talk me through Tuesday night, from the moment the light went off."
  • Watch for: Improvement in the client's attitude and distress often comes before diary change. Some studies told clients not to expect sleep gains for a few weeks [source: Jansson-Fröjmark et al. 2023].

4.17 Guided online / self-help delivery

  • Purpose / when: When face-to-face sessions are impractical.
  • Steps:
    • Either deliver the 4.2–4.9 content by video with screen-share and a whiteboard for the formulation [source: Ong et al. 2025], or use four weekly self-help modules:
      1. Week 1: overview, insomnia information, rationale, and starting PI.
      2. Week 2: common challenges.
      3. Week 3: troubleshooting and persistence.
      4. Week 4: relapse prevention.
    • Add a brief weekly phone or video check-in of 15 minutes or less, plus daily diary entries [source: Salim et al. 2024].
  • Example language (illustrative): "In our 10-minute call, I mainly want to hear how the eyes-open part is going."
  • Watch for: Monitor for adverse effects at the end of treatment. Salim et al. used the Negative Effects Questionnaire at post-treatment [source: Salim et al. 2024].

4.18 Relapse prevention

  • Purpose / when: The final session or module.
  • Steps:
    1. Have the client summarise their own sleep-effort cycle and the attitude that breaks it.
    2. Identify early warning signs of renewed effort, such as new sleep rituals, checking the clock, or anxiety in the evening.
    3. Plan to return to Method 1 or Method 2 for a set period after any setback.
    4. Normalise the occasional bad night.
    [source: Salim et al. 2024 week 4 for the topic; content is synthesis]
  • Example language (illustrative): "When a run of bad nights comes — and it will — what's the first sign you're starting to try again?"
  • Watch for: The client treating PI as a nightly ritual that must be performed. The aim is an attitude that needs no technique [source: Espie 2011 on the outlook change; framing is synthesis].

5. Handling problems

The primary sources contain little explicit troubleshooting. Each item is tagged with its source; anything tagged [synthesis] is my own clinical guidance.

"I fell asleep straight away, so I couldn't do it."

  • The source reports this for one group CBT-I participant, who concluded PI had failed, and links it to perfectionism or unclear instructions [source: Jansson-Fröjmark et al. 2023].
  • Response: Celebrate it. Falling asleep while trying to stay awake is the method working. Re-state that the goal was never to stay awake, it was to stop trying to sleep.
  • Prevention: Say this before the first night [synthesis].

Effort transferred to staying awake ("trying to win")

  • Espie and Lindsay raised the concern that PI may move anxiety onto the effort of staying awake [source: Espie & Lindsay 1985 as summarised on Wikipedia; primary not drawn on — uncertain].
  • Response:
    • Return to the thin line (4.5).
    • Re-emphasise that the approach is passive and gentle, not forced [source: Espie 2011].
    • If the client is competitive by nature, switch to Method 1 (4.8), which carries no stay-awake instruction [synthesis].

Early worsening or increased anxiety

  • In one study, about a third of clients had longer sleep-onset latency in week 1, which then improved by the end of treatment. In the 1985 six-case series, onset worsened in 3 of 6 clients. Raised early anxiety is the main side effect noted [source: Jansson-Fröjmark et al. 2023].
  • Response:
    • Forewarn the client.
    • Use the diary to show the trend over weeks rather than nights.
    • Hold steady for the agreed 2 weeks unless distress is severe.
    • If onset is clearly worse at review, or the client describes it as "too easy" to stay awake, consider that sleep drive is low. Move to sleep restriction or stimulus control, delivered sequentially [synthesis informed by Jansson-Fröjmark et al. 2023 on sleep drive and the case-series explanation].

"This is odd" or "it's a trick": scepticism and reluctance

  • APA Division 12 notes that clients are often reluctant and that adherence can be a problem [source: APA Div 12].
  • Response:
    • Lead with the formulation (4.2–4.4) and the client's own parallels before giving any instruction.
    • Be open that the method is counter-intuitive, and invite a time-limited experiment (4.14).
    • Don't oversell it.
    • If the client still declines, offer sleep restriction, which Div 12 notes works on a similar principle and is often more readily accepted, or offer standard CBT-I [source: APA Div 12; framing is synthesis].

Concrete or literal clients and cognitive impairment

  • Espie says PI is unsuitable for these clients [source: Espie 2011].
  • Response: Use other CBT-I components. If a literal client has already started PI, simplify it to "lights off, eyes open, no trying" and drop the stay-awake framing [synthesis].

Non-adherence

  • Response:
    • Find out what happened the night they didn't practise.
    • Common reasons [synthesis]:
      • fear of losing a whole night
      • the bed partner's routine
      • the habit of using a phone in bed
      • forgetting at 3 a.m.
    • Problem-solve each one: a single-sentence cue card by the bed, and a partner briefing.
    • Reduce the ask to bedtime only, before adding night wakings.
    • Keep the instruction short [synthesis].
  • Ong et al. reported good attendance, but no source describes how nightly PI practice was checked [source: Ong et al. 2025].

Using screens or activities "to stay awake"

  • Response: Clarify that wakefulness should be passive, with no self-stimulation [source: Espie 2011]. Agree that phones go out of the bedroom during the PI phase [synthesis].

Overwhelm in bed (rumination, panic, intrusive worry)

  • Response:
    • Validate the experience.
    • Do structured worry or cognitive work outside bed in the evening. Cognitive components combine well with PI [source: Jansson-Fröjmark et al. 2023].
    • Consider Method 1's acceptance stance [source: Espie 2011].
    • If arousal escalates to panic, have the client get up and settle before returning. This means temporarily dropping PI for that night, and it is acceptable [synthesis].
    • Treat any comorbid anxiety disorder in its own right [synthesis].

Ruptures (client feels mocked, dismissed or told "it's all in your head")

  • These are a risk with the humour and the counter-intuitive message [synthesis].
  • Response:
    • Name what happened and apologise for the joke if it landed badly.
    • Re-validate how real and costly the insomnia is.
    • Explain that effort is a normal, understandable reaction, not a personal fault. Espie's framing of trying too hard as the problem, not the person, helps here [source: Espie 2011 Box 6.1 tone; repair steps are synthesis].
    • Resume only when the client agrees.

Stuck: no change after 2–3 weeks of adequate practice

  • Response:
    • Check the diagnosis: undetected sleep apnoea, circadian delay, restless legs, substances or medication.
    • Check whether the client has enough sleep drive, for example from napping or long time in bed.
    • Check whether PI was really practised passively.
    • Then move to stimulus control, sleep restriction or full CBT-I, in sequence [synthesis; second-line positioning from Jansson-Fröjmark et al. 2023].

Conflict with concurrent CBT-I components

  • Response: Pause one of them. Don't run PI alongside stimulus control or sleep restriction [source: Jansson-Fröjmark et al. 2023].

Gains fade when sleep improves

  • Response: The aim is a lasting attitude. Use relapse planning (4.18) and give the client permission to re-run the PI phase after a setback [synthesis].

Hypnotic users and medication changes

  • Trials excluded recent medication changes and as-needed hypnotic use [source: Salim et al. 2024; Ong et al. 2025].
  • Response: Coordinate any tapering with the prescriber and do not combine PI with a taper on the same timetable. Generalising to this group is uncertain [synthesis; uncertain].

Daytime sleepiness and safety

  • PI doesn't restrict time in bed, but early worsening can occur [source: Jansson-Fröjmark et al. 2023].
  • Response: Ask about sleepiness at each review. Advise against driving or safety-critical tasks when the client is drowsy [synthesis].
  • Reviewers suggest PI may suit people for whom sleep restriction is risky, such as some drivers or people with seizure disorders, but these groups have not been studied [source: Jansson-Fröjmark et al. 2023; uncertain].

Depression and suicidality emerging during treatment

  • Response: Follow standard risk assessment and safety planning. Brief sleep work does not replace treatment for depression. One trial excluded severe depression with suicidal intent [source: Salim et al. 2024; risk practice is synthesis].

Sources

Sources

Further reading (not drawn on unless listed above)

  • Espie, C. A. (2011). Paradoxical intention therapy. In M. Perlis, M. Aloia & B. Kuhn (Eds.), Behavioral Treatments for Sleep Disorders (pp. 61–70). Academic Press/Elsevier.
  • Ascher, L. M., & Efran, J. S. (1978). Use of paradoxical intention in a behavioral program for sleep onset insomnia. J Consult Clin Psychol, 46, 547–550.
  • Ascher, L. M., & Turner, R. M. (1979). Paradoxical intention and insomnia: an experimental investigation. Behav Res Ther, 17, 408–411.
  • Fogle, D. O., & Dyal, J. A. (1983). Paradoxical giving up and the reduction of sleep performance anxiety in chronic insomniacs. Psychotherapy: Theory, Research & Practice, 20, 21–30. [uncertain]
  • Espie, C. A., & Lindsay, W. R. (1985). Paradoxical intention in the treatment of chronic insomnia: six case studies. Behav Res Ther, 23, 703–709. [uncertain]
  • Broomfield, N. M., & Espie, C. A. (2003). Behav Cogn Psychother, 31, 313–324.
  • Broomfield, N. M., & Espie, C. A. (2005). Towards a valid, reliable measure of sleep effort. J Sleep Res, 14, 401–407. [uncertain]
  • Morin, C. M., & Espie, C. A. (2003). Insomnia: A Clinical Guide to Assessment and Treatment. Kluwer. [uncertain]
  • Edinger, J. D., et al. (2021). Behavioral and psychological treatments for chronic insomnia disorder in adults: AASM clinical practice guideline. J Clin Sleep Med, 17, 255–262. [full text not drawn on; status reported via Jansson-Fröjmark et al. 2023]
  • Frankl, V. E. (1959/2006). Man's Search for Meaning. Beacon Press. [uncertain]
← Previous · II-069CBT for insomniaNext · II-071 →Illness management and recovery