Treating preschooler behavioural sleep problems via parent‐mediated telehealth: A randomized controlled trial

Abstract

Objectives

Behavioural sleep problems in the preschool developmental period (ages 3–5 years) are highly prevalent and associated with a myriad of deleterious consequences including anxiety, in the short‐ and long‐term. This study examined a parent‐focused behavioural sleep intervention for children aged from 3 to 5 years, delivered individually via three × fortnightly 90‐min telehealth (synchronous videoconference) sessions, in terms of its ability to improve child sleep, nighttime fears and anxiety.

Methods

Parents of children aged 3 to 5 years (M = 3.57; SD = .56) were randomly allocated to either the Lights Out Videoconference (LOV) or care‐as‐usual (CAU) conditions and completed measures of child sleep problems, anxiety and nighttime fears at pre‐treatment (T1), two weeks post‐treatment (T2) and at 3‐month follow‐up (T3). Parents also completed a measure of treatment satisfaction.

Results

Compared with the CAU condition (n = 16), children whose parents participated in the LOV condition (n = 19) reported a significantly greater reduction in sleep problems, anxiety and nighttime fears from T1 to T2, with treatment effects being maintained at T3. Treatment satisfaction of both the programme, resources and mode of delivery was very high.

Conclusions

A brief, behavioural sleep intervention delivered via videoconferencing for young children is acceptable to parents and represents an efficacious and convenient alternative to face‐to‐face treatment for sleep that has secondary effects on nighttime fears and broader anxiety issues.

Universal Trial Number (UTN): U1111‐1264‐8191.

Australian and New Zealand Clinical Trial Registry (ANZCTR): 12621000466842 retrospective.

The trial was registered retrospectively as the application for registration was submitted after the first participant was registered for the programme. This was a clerical oversite of the authors as to the timing of registration submission. The sleep diaries included in the registration of the trial were not analysed due to significant missing data in the CAU condition. Additionally, some of the secondary outcomes in the trial registry will be published in a separate, paper, which focuses on parents' impressions of the programme and parenting factors.

Practitioner points.

• Sleep problems in preschool aged children can be treated using parent‐mediated interventions delivered via videoconferencing in just three sessions, with high parent satisfaction.

• Treating sleep problems can improve comorbid anxiety and nighttime fears, providing practitioners with structure when treating transdiagnostic behavioural insomnia and anxiety.

• Despite this study being preliminary in nature and difficult to generalize to a more diverse population, it provides promising results for both parents and clinicians alike, with positive outcomes on sleep, anxiety and nighttime fears.

INTRODUCTION

Sleep is essential for physical and emotional health and wellbeing and is integral for both brain and body growth and repair across the lifespan (Anastasiades et al., 2022; Cheng et al., 2020; Mason et al., 2021; Palmer et al., 2024). Young children experience immense physiological growth and neurological development, and therefore sleep is critically important during this phase of life (Anstead, 2000; Carno et al., 2003). However, between 15% and 34% of preschool children (aged 3 to 5 years) have behavioural sleep problems (BSPs; Hiscock et al., 2007; Ophoff et al., 2018) that tend to persist into the school‐age years and beyond (Falch‐Madsen et al., 2020; Williamson et al., 2020). BSPs in the preschool years are also associated with numerous adverse emotional, social and health outcomes (Armstrong et al., 2014; Gregory et al., 2004; Hiscock et al., 2007; Simola et al., 2012).

BSPs are broadly defined as difficulties initiating and/or maintaining sleep, bedtime resistance and difficulty sleeping without parental presence or assistance (Hiscock et al., 2007; Meltzer & Crabtree, 2015). When it comes to emotional and mental health, preschooler BSPs are highly comorbid with anxiety (Chase & Pincus, 2011; Crowe & Spiro‐Levitt, 2024). In fact, longitudinal studies and reviews attempting to disentangle the bidirectional relationship between pediatric sleep problems and anxiety have concluded that sleep problems likely predate and lead to the emergence of anxiety problems (Armstrong et al., 2014; Leahy & Gradisar, 2012; Simola et al., 2012; Whalen et al., 2017). The strong association between BSPs and anxiety is likely due to a complex combination of biological (shared neurobiological pathways, brain structure, dysregulated cortisol, etc) and environmental (including familial) factors (Armstrong et al., 2014; Cox & Olatunji, 2016; Gregory, Caspi, et al., 2005; Gregory, Eley, et al., 2005).

BSPs are the most common form of sleep problems in preschoolers, occurring more frequently than other sleep problems such as parasomnias, sleep breathing disorders and delayed sleep phase disorder (Hiscock et al., 2007; Meltzer & Crabtree, 2015). Fortunately, BSPs can be treated effectively with behavioural interventions (Fehr et al., 2020; Meltzer et al., 2021). However, there are two problems with the BSP intervention literature to date. First, few studies have focused specifically on preschool aged children, with preschool children tending to be included within broader age ranges of younger or older children. Indeed, in a recent scoping review of 47 sleep trials conducted with young (0–5‐year‐old) children, Meltzer et al. (2021) found that only three studies included children over 3 years of age (Cooney et al., 2018; Moore et al., 2007; Seymour et al., 1989), of which only one (Moore et al., 2007) examined preschoolers specifically. Moore et al.'s randomized controlled trial (RCT) evaluated a ‘Bedtime Pass' strategy for children exhibiting bedtime resistance and found that children in the treatment condition left their rooms, called and cried out significantly less frequently than controls. The tendency for researchers to cluster preschool age children with infants and toddlers is inherently problematic, as preschool children exhibit very different and specific sleep problems compared to their younger (and older) counterparts. For example, preschoolers are beginning to develop a capacity and desire for self‐control and independence, which can result in increased bedtime resistance and argumentative behaviours at nighttime (Eisenberg & Sulik, 2011; Griffin et al., 2016). Furthermore, a preschooler's imagination and cognitive skills increase rapidly compared to their logical reasoning during this time, resulting in the development of nighttime fears as they struggle to discern between fantasy and reality (Fehr et al., 2020; Fisak & Barrett, 2019).

The second problem with the preschool sleep treatment literature to date, is that most trials including preschoolers neglect to assess and report on outcomes other than sleep. The few that have, report some improvements in outcomes such as child prosocial behaviour, child quality of life and parent mental health, but no changes in child behavioural difficulties, academic function, or working memory (e.g., Hiscock et al., 2019; Quach et al., 2011). Furthermore, despite the highly comorbid relationship between paediatric sleep problems and anxiety (Fehr et al., 2020; Lewis et al., 2021), anxiety has been relatively neglected as a treatment focus and outcome variable in preschooler sleep treatment studies. It appears that Donovan et al. (2023) is the only preschooler treatment study to date to measure anxiety as an outcome variable, and to include gold standard treatment strategies addressing nighttime fears and anxiety (i.e., psychoeducation regarding anxiety and the role of avoidance, graded exposure using exposure hierarchies and a reinforcement schedule for hierarchy progression) in their Lights Out’ sleep programme, alongside traditional behavioural sleep intervention components for children (sleep psychoeducation, sleep hygiene, bedtime routines, relaxation and graduated extinction). Donovan et al. tested the efficacy of the parent‐mediated, group‐based behavioural sleep intervention (Child Mage = 4.5 years; SD = .54; N = 128) reporting improvements in child sleep, anxiety, internalizing and externalizing behaviours and behaviour problems. They also demonstrated that improvements in child sleep mediated the improvements on secondary outcomes (Donovan et al., 2023). While Donovan et al. (2023) included the treatment of nighttime fears and anxiety, they only measured general anxiety and not nighttime fears, or nighttime anxiety specifically. This is likely due to the lack of a nighttime fear measure available at the time.

In addition to the problems evident in the BSP intervention literature, there is a larger problem associated with service provision. Indeed, the majority of parents do not receive assistance for their child's sleep problem at all (Newton et al., 2021), with one study reporting that only 11.1% of parents whose child had a sleep problem sought any type of help due to barriers such as lack of access to experienced practitioners, cost, availability of services, family schedules and lack of childcare (Newton et al., 2021; Tapp et al., 2018). Although the Lights Out Programme is efficacious and there are numerous benefits associated with group‐based therapy including efficiency and increased social support among group members, the fragmented and time‐consuming nature of parenting a young child, means that these parents are time‐poor and face difficulties organizing childcare to attend sessions. Furthermore, it has been our experience that organizing treatment groups for parents of young children, finding a time that is convenient for all group parents and ensuring consistent attendance, is extremely difficult. With the rise of the COVID‐19 pandemic in 2020, many psychology clinics ceased in‐person therapy and rapidly moved to telehealth treatment delivery, including telephone and videoconference services (Jonnagaddala et al., 2021). Of the various telehealth modalities, synchronous videoconferencing is most akin to face‐to‐face therapy, allowing patient/client direct, real‐time therapist interaction in both visual and audio modalities. Videoconference also includes additional benefits such as the capacity to share screens, show videos, complete active learning tasks and record sessions for later review either by the psychologist or client (Paruthi, 2020). While the transition back to in‐person therapy has occurred post‐COVID‐19, many providers continue to offer telehealth due to consumer demand. Yet, the evidence for therapies delivered via telehealth lags behind the enormous post‐pandemic clinician uptake of this treatment modality.

Telehealth as a delivery mode for child BSP interventions is a burgeoning area of empirical enquiry. Indeed, a systematic review examining telehealth (including telephone, videoconference and web‐based platform) delivery of sleep interventions for children and adolescents (McLay et al., 2020) reported positive overall treatment effects across the 10 studies reviewed. Of the studies reviewed, only the trial by Witmans et al. (2008) used videoconference delivery, and was also the only study to include children in the preschool age range. However, only 27% of the sample in the Witmans et al. study were preschool aged, the treatment programme targeted insomnia, parasomnias and sleep apnoea, there was no control group and programme effectiveness was only assessed through client satisfaction rather than the sleep outcomes. Nevertheless, 94% of the clients in the Witmans et al. study reported being ‘satisfied with the services’, ‘felt they would use the service again’ and ‘felt telehealth provided easier access to specialised services’. Since publication of the McLay et al. (2020) review, two recent RCTs have since tested the efficacy of parent‐mediated interventions for preschooler sleep problems delivered using group videoconference in Hong Kong (Ip et al., 2024) and the USA (Johnson et al., 2023). Both studies found that parents in the sleep intervention groups reported greater improvement in child sleep, providing an important addition to the literature. However, their interventions were developed for, and included elements specific to, autistic rather than neurotypical children. Furthermore, Johnson et al. noted that parents in their group‐based videoconference programme were observed to be multi‐tasking and became distracted, highlighting one potential problem with group‐based compared to individual videoconference delivery. To date, a sleep intervention programme delivered via individual videoconference for neurotypical children has not been developed or tested.

The current study

This study involved a small, pilot, randomized controlled trial (RCT) of a modified, synchronous, videoconference version of the ‘Lights Out’ programme (Donovan et al., 2023), delivered to individual parents of neurotypical children in the preschool developmental period (3 to 5 years) relative to a care‐as‐usual (CAU) condition. Assessments were conducted at baseline (T1), 2‐weeks post‐treatment (T2; primary endpoint) and 3‐months following treatment (T3; secondary endpoint). The study had two primary aims. The first was to evaluate the efficacy of the Lights Out Videoconference (LOV) programme in terms of reducing parent‐reported child sleep problems (as measured by the Child Sleep Habits Questionnaire (CSHQ; Owens et al., 2000) and a parental rating of ‘mild’, ‘moderate’ or ‘severe’) relative to a CAU condition. The second was to evaluate parent satisfaction with the programme. Secondary aims were to examine the relative reduction of nighttime fears and anxiety for the LOV compared to CAU condition. It was hypothesized that:

1. Children in the LOV condition would demonstrate significantly greater reductions in sleep problems from T1 to T2 relative to the CAU condition.

2. Improvements in child sleep would be maintained at T3 for the LOV condition.

3. Parents who completed the intervention would be satisfied with the programme.

4. Children in the LOV condition would demonstrate significantly greater reductions in anxiety and nighttime fears from T1 to T2 relative to the CAU condition.

5. improvements in anxiety and nighttime fears would be maintained at T3 for the LOV condition.

METHOD

Design

Parents were randomly assigned to either the 5‐week Lights Out Videoconference (LOV) or Care‐as‐Usual (CAU) conditions in a 1:1 ratio, using blocks of six created using a computer‐generated random number sequence. Parents were required to complete assessments at pre‐treatment (T1, week 0), 2 weeks post‐treatment (T2, week 8, primary endpoint) and 3‐month follow‐up (T3, week 21, secondary endpoint). Those in either condition could access other services during the study if they so desired, and those in the CAU condition were provided with LOV following the wait period when they ceased to be part of the trial. Primary outcomes included child sleep problems and treatment satisfaction. Secondary outcomes included nighttime fears and anxiety.

Power estimation

Power was estimated using G*Power 3.1.9.2 (Faul et al., 2007) on the CSHQ primary outcome measure. A medium to large effect size (f = .25–.40) was expected based on systematic reviews and analyses of behavioural interventions for paediatric sleep problems, including telehealth (McLay et al., 2020; Meltzer & Mindell, 2014). We chose to power the study on a medium effect size to ensure maximum, and therefore sufficient, sample size. Results indicated that the required sample size to achieve 80% power at a significance criterion of α = .05, with a r = .5 to detect a medium effect size (f = .25) was N = 36 for a 2 (treatment, waitlist) × 2 (pre to post, then post to follow‐up) repeated measures ANOVA.

Participants

Participants were 37 parents aged 21–47 years (M = 37.24, SD = 5.04) who reported being either the mother (86.5%) or father (13.5%) of a child aged between 3 and 5 years (M = 3.57, SD = .56) with a sleep problem, and who were randomized to either the LOV (n = 19) or CAU (n = 18) conditions. Participants were recruited through social media advertisements and media publicity as well as through referrals from childcare centres, paediatricians, general practitioners, mental health professionals and parenting support organizations. Participants were initially screened via an online survey to ascertain broad inclusion and exclusion criteria. Participants were included in the study if (a) the child scored ≥41 on the Children's Sleep Habits Questionnaire (Owens et al., 2000) and (b) the parent rated the child as having a ‘moderate’ or ‘severe’ sleep problem (see Measure section). Participants were excluded if their child had an intellectual or developmental disorder, as the treatment programme was designed for neurotypical children. Participants reporting children with medical sleep problems (sleep apnoea, parasomnias) were excluded and referred to their general practitioner for assistance. Those who met criteria for inclusion were emailed an information and consent sheet with a link to the pre‐treatment assessment. Only those who provided digital consent went on to complete the online survey. Data was collected online using Lime Survey hosted by Griffith University.

Detailed demographic information of the participants is outlined in Table 1, and Figure 1 illustrates the flow of participants through the study. As is evident from Table 1, the majority of adult respondents were Caucasian (89.2%) and married (75.7%). In terms of household income, 45.9% had a household income between AUD$100,00 and AUD$200,000 and 40.5% had completed a bachelor's degree. Just over half of the children were male (51.4%) and the majority lived with both parents (89.2%). As is evident from Figure 1, 100% of participants in the LOV group (n = 19) and 84.2% of those in the CAU group (n = 16) completed post‐treatment and follow‐up assessments. Of the 19 LOV treatment participants, three (15.79%) had their partner attend the session. In these three cases, the partner attended all three sessions. Only the parent who completed the baseline survey completed T2 and T3 assessments and was included the descriptives.

TABLE 1.

Demographic characteristics and descriptive statistics for participants at pre‐treatment.

Demographic characteristics	LOV N = 19		CAU N = 18		Entire sample N = 37
	M (SD)	%	M (SD)	%	M (SD)	%
Parent age (years)	37.37 (5.97)		37.11 (4.00)		37.24 (5.04)
Mother (%)	78.9		94.4			86.5
Father (%)	21.1		5.6			13.5
Child age (years)	3.58 (.61)		3.56 (.51)		3.57 (.56)
Female (%)	52.6		44.4			48.6
Male (%)	47.4		55.6			51.4
Household income (AUD)
<$60,000		5.3		22.2		13.5
$60,001–$80,000		0		0		0
$80,001–$100,000		10.5		5.6		8.1
$100,001–$200,000		57.9		33.3		45.9
More than $200,000		21.1		38.9		29.7
No response		5.3		0		2.7
Highest level of education
Below grade 12		5.3		5.6		5.4
Grade 12		10.5		0		5.4
Associate or technical degree (TAFE)		26.3		5.6		16.2
Bachelor degree		36.8		44.4		40.5
Postgraduate degree		21.1		44.4		32.4
Parent ethnicity
Caucasian		89.5		94.4		89.2
Asian		0		5.6		5.4
Aboriginal or Torres Strait Islander		5.3		0		2.7
Black/African		5.3		0		2.7
Child ethnicity
Caucasian		84.2		88.9		86.5
Asian		5.3		11.1		8.1
Aboriginal or Torres Strait Islander		5.3		0		2.7
Pacific Islander		5.3		0		2.7
Employment status
Student		0		5.6		2.7
Employed full‐time		57.9		44.4		51.4
Employed part‐time or casually		31.6		27.8		29.7
Unemployed		10.5		22.2		16.2
Marital status
Single		5.3		11.1		8.1
Defacto		15.8		11.1		13.5
Married		78.9		72.2		75.7
Divorced/separated		0		5.6		2.7
Living arrangement
Living with both parents		94.7		83.3		89.2
Living with mother only		0		11.1		5.4
Living with mother and another parent figure		5.3		5.6		5.4
Number of children living in the household
1		21.1		22.2		21.6
2		47.4		50.0		18.6
3		10.5		22.2		16.2
4		21.1		5.6		13.5

Abbreviations: CAU; Care‐As‐Usual group; De facto, unmarried couple living together; LOV, Lights Out Videoconference behavioural intervention group.

FIGURE 1.

CONSORT Flow of participants through the study.

Measures

The primary outcomes of parent‐rated child sleep and the secondary outcomes of parent‐rated child anxiety and nighttime fears were assessed at T1, T2 and T3. Demographic information was collected at T1 only. The primary outcome of treatment satisfaction was collected at T2 only.

Demographic information

Parents were required to report their own age, gender, ethnicity, annual household income, highest level of education and marital status, as well as the age, gender, ethnicity and living arrangements of their child.

Primary outcomes

Sleep problems 1

In line with studies conducted by Donovan et al. (2023) and Hiscock et al. (2019), parents were asked to rate their child's sleep problem severity as ‘none’, ‘mild’, ‘moderate’, or ‘severe’. Also, in line with these studies, ‘none’ and ‘mild’ were coded as absence of a sleep problem and ‘moderate’ and ‘severe’ were coded as presence of a sleep problem.

Sleep problems 2

The 33‐item Children's Sleep Habits Questionnaire (CSHQ; Owens et al., 2000) was used to assess sleep behaviours in children. Items are rated on a three‐point Likert scale (1 = rarely [0–1 night per week] to 2 = usually [5–7 nights per week]), and summed to produce a total score that may range from 33 to 99, with higher scores indicating more problematic child sleep behaviours and total scores over 41 being indicative of a clinical‐level paediatric sleep problem (Owens et al., 2000). The CSHQ has been used with parents of children from early childhood to early adolescence (Sen & Spruyt, 2020) and has shown acceptable internal consistency (α = .77) in a RCT targeting children aged 3 to 5 years with sleep problems (Donovan et al., 2023). It has also demonstrated adequate test–retest reliability in previous studies (.62 to .79; Owens et al., 2000). The internal consistency for the total CSHQ score in the present study was acceptable (α = .78).

Treatment satisfaction

Treatment satisfaction was assessed with an author‐developed questionnaire. Parents were required to rate on a 5‐point Likert scale ranging from 1 (not at all) to 5 (very much) their agreement with a range of items assessing (a) the likelihood of recommending the programme to others, (b) how effective they believe the programme was in helping to improve their child's current, (c) how effective they believe the programme would be in helping with future sleep problems and (d) their overall satisfaction with the programme. Parents were asked to consider the sessions, mode of delivery and resources when responding. Each item ranged from 1 to 5, with higher scores representing higher programme satisfaction.

Secondary outcomes

Anxiety symptoms

The 28‐item Revised Preschool Anxiety Scale (PAS‐R; Edwards et al., 2010) was used to measure child anxiety. Items are rated on a 5‐point scale from 0 (not at all true) to 4 (very often true) and summed to produce a total score that may range from 0 to 112, with higher scores indicating greater anxiety. The PAS‐R has demonstrated strong internal consistency (α = .72–.92), maternal/paternal agreement (r = .60–.75) and 12‐month stability (r = .60–.75; Edwards et al., 2010) in previous studies. The internal consistency for the total PAS‐R score in the present study was very good (α = .93).

Nighttime fears

The 17‐item Fears and Worries at Nighttime—Young Children (FAWN‐YC; Shiels et al., 2024) was used to measure nighttime fear. Items are rated on a 6‐point scale from 0 (never true) to 5 (always true), and items are summed to produce a total score that may range from 0 to 85, with higher scores indicating greater nighttime fears. The FAWN‐YC has demonstrated strong internal consistency (α = .92), convergent and divergent validity (r = .41–.82) and excellent 2‐week stability (ICC = .95; Shiels et al., 2024) in previous studies. The internal consistency of the FAWN‐YC score in the present study was very good (α = .91).

Access to sleep services

At T3, participants indicated whether, at any point during the study, they had (yes/no): (a) accessed additional support for sleep or not, (b) accessed sleep assistance via the internet, (c) consulted with a general practitioner or paediatrician regarding sleep problems, (d) visited a psychologist or counsellor regarding sleep problems, (e) sought advice from their child's teachers or educators regarding sleep problems and/or (f) accessed other additional support for sleep problems (required to specify). Optional text boxes were also included for the parent to provide more information.

Procedure

Ethics was obtained from the Griffith University Human Research Ethics Committee, and the trial was registered with the Australian and New Zealand Clinical Trials Registry (ANZCTR: 12621000466842 retrospective). The CONSORT (Consolidated Standards of Reporting Trials) statement for randomized trials of nonpharmacological treatment update was followed in the reporting of this trial (Boutron et al., 2017). Referrals were screened using an online survey to ascertain inclusion and exclusion criteria before enrolment and randomisation (see above). The randomisation schedule was held in an encrypted file with a post‐doctoral research fellow responsible for informing participants over email of their allocation after parents had completed their online pre‐treatment survey. In the email, LOV parents were informed that their therapist would phone them to schedule sessions, and CAU parents were reminded that they would be emailed within 2 months (aligning with T2) and then within 5 months (aligning with T3) to complete measures via an online survey before being provided with the videoconference sessions. The therapist was then provided with the email address and phone number of the participant to schedule their sessions.

One therapist (the first author) treated all parents in the LOV condition. The therapist was a provisionally registered psychologist, was enrolled in the clinical psychology training programme at Griffith University and therefore had minimal previous experience in delivering therapeutic interventions. The therapist had, however, been heavily involved in the first trial of Lights Out, was a co‐author of the programme, and had co‐facilitated several of the face‐to‐face groups in the RCT of the original programme. The therapist completed 6 hours of structured training on the LOV treatment protocol, which focused on both treatment content and telehealth delivery skills and received supervision fortnightly by the second author, a registered Clinical Psychologist and developer of the original Lights Out programme.

To ensure fidelity with the comprehensive, 51‐page digital therapist manual, an independent research assistant used a checklist to rate treatment adherence for a random sample of 12 recorded sessions (21.05% of the total sessions conducted). Fidelity also included structured facilitator training, facilitator supervision sessions, as well as retention rates (Feely et al., 2018; Sanetti et al., 2021). Email reminders for participants to complete T2 and T3 assessments were sent at the scheduled times by a research assistant who was blind to condition. In two‐parent families, the parent who completed the baseline measures also completed the T2 and T3 measures. During the trial, parents in the CAU condition were not offered any assistance or resources as part of the trial. However, unlike traditional waitlist controls, they were not asked to refrain from seeking assistance elsewhere during the trial period. All CAU participants were offered the videoconference programme (by intern psychologists who received the same training and supervision as the LOV condition therapist) after completion of their final T3 assessment when they ceased to be part of the study.

Intervention

Based on parent feedback that the original programme included content that was not relevant to their particular child, the Lights Out Videoconference (LOV) programme was adapted to be more personalized and therefore included fewer sessions (see Table S1 for more details of programme differences and Table 2 for details of session content). A subset of the developers of the original programme designed the LOV manualised programme which consisted of 3 × 90‐min, individual videoconference sessions that were scheduled fortnightly and that were tailored to the behavioural sleep problems each child presented with. At the completion of each session, parents were emailed comprehensive, personalized session materials, including a 54‐page digital manual at the end of Session 3. Session 1 focussed on psychoeducation, training in sleep hygiene, bedtime routines, praise and rewards and goal setting. Session 2 covered relaxation, self‐soothing and token rewards. Parents of children with fear‐based sleep problems (i.e., fear of the dark, bedtime or nighttime separation fears, fear of nighttime sounds and shadows, imagination‐based fears, etc) also received training in fear/anxiety avoidance and exposure, while parents of children with oppositional behaviours‐based sleep problems received training in behaviour management (see Table 2). If a parent reported that both fear and oppositional behaviours were problematic, Session 2 focused on the most problematic area. Session 3 involved future planning and training in parental self‐care. Parents of children with fear‐based sleep problems were also briefly instructed in the use of behaviour management skills, while parents of children with oppositional behaviour‐based sleep problems were briefly instructed in the use of exposure hierarchy skills. Parents were also instructed on where in the parent manual to find more information on psychoeducation and techniques they had not covered in session. By the end of the programme, all parents received the same information, albeit delivered at different times (Session 2 or Session 3) and different modes (discussed and/or written information provided).

TABLE 2.

Concepts covered in the Lights Out Videoconference programme.

Session	Concepts covered
1	Psychoeducation on sleep problems, nighttime fears and anxiety Goal setting—nighttime fears or difficult/oppositional bedtime behaviours Sleep hygiene (both adult and child) Bedtime routine reward charts (constructed in session) How to praise and reward young children
2	Adding nighttime relaxation strategies to their bedtime routine How to prompt their child to self‐soothe Introducing a reward token system
	Fear‐based	Oppositional behaviours
	Avoidance and exposure psychoeducation How to acknowledge yet ignore fearful behaviours How to construct and implement a personalized exposure hierarchy for their child	Dealing with oppositional behaviours—behaviour flowchart (e.g., remaining calm, using statements vs. questions when giving instructions, providing limited options for behaviours, how to use time out in the bedtime routine, reasonable consequences, importance of consistency and persistence) Additional brief strategies to compliment new skills
3	Trouble shooting How to respond when children argue, tantrum, call out and refuse to go to bed (i.e., graduated extinction—‘Night Ticket’ and ‘Back Soon’) Strategies for self‐care Maintenance and relapse prevention
	Fear‐based	Oppositional behaviours
	Additional brief strategies to compliment new skills Briefly cover how to manage difficult bedtime behaviours (and demonstrate where to find this information in the emailed booklet)	Briefly cover how to manage nighttime fear behaviours using exposure (and demonstrate where to find this information in the emailed booklet)

Data analytic plan

Preliminary analysis

The data were analyzed using IBM SPSS Statistics (Version 29) analytic software.

Assumptions for ANOVA (normality, equal variance, independence and sphericity) and Pearson's X ² square (independence, mutually exclusive cells, cell values >5) were checked.

Pre‐treatment comparisons analytic plan

Between groups (LOV vs. CAU) ANOVAs were used to test for pre‐treatment differences on demographic (child and parent age) and outcome (CSHQ, PAS‐R and FAWN‐YC) variables. Pearson's X ² tests were used to assess for pre‐treatment group differences on gender (child). To test for differences between those retained versus not retained at T2 and T3, between groups (retained vs. not retained) ANOVAs were conducted on age (child and parent), CSHQ, PAS‐R and FAWN‐YC. Pearson's X ² tests were used to assess for differences between those retained versus not retained on gender (child).

Intervention effects analytic plan

In line with the hypotheses, analyses were conducted from T1 to T2 (pre‐post effects—hypotheses 1 and 3), and from T2 to T3 (hypotheses 2 and 4). Pearson's X2 tests were used to analyse the between groups (LOV vs. CAU) difference between T1‐T2 (hypothesis 1) and T2‐T3 (hypothesis 2). To address hypotheses 1 and 3 (i.e., assessment of T1‐T2 effects) for the categorical variables of the CSHQ, PAS‐R and FAWN‐YC, a series of between groups (LOV vs. CAU), repeated measures (T1 vs. T2) ANOVAs were conducted. To address hypotheses 2 and 4 (i.e., assessment of T2‐T3 maintenance effects) for the categorical variables of CSHQ, PAS‐R and FAWN‐YC, a series of between groups (LOV vs. CAU), repeated measures (T2 vs. T3) ANOVAs were conducted. Effect size was reported using partial eta squared (${\eta }_p^2$) where .01 indicates a small effect, .06 indicates a medium effect and .14 or higher indicates a large effect. T‐tests were then conducted to assess differences between groups at each time point, and differences between time‐points for each group, with effect size reported using Cohen's d (where .2 indicates a small effect, .5 indicates a medium effect and .8 or higher indicates a large effect).

RESULTS

Assumptions were met for all analyses, with the exception that the FAWN‐YC did not meet the assumption of sphericity, and therefore Greenhouse–Geisser and Huynh Feldt corrections were used for this measure. Data was missing from 2 participants in the CAU condition who dropped out between T1 (pre‐treatment) and T2 (post‐treatment), resulting in 3% missing data overall. Using Jakobsen et al.'s (2017) flowchart for dealing with missing data in RCTs, multiple imputation was not employed and only observed data were analysed.

Pre‐treatment comparisons

Therapist adherence was >95%, completion of sessions was 100% and participant retention was 100%. Treatment session duration in the LOV group was 90.21 min (SD = 13.34) for session one, 80.53 min (SD = 10.18) for session two and 84.00 min (SD = 11.57) for session three. Of the treatment group, the sleep problems of 52.63% and 47.63% of children were determined to be fear‐based and oppositional behaviour‐based, respectively. The means and standard deviations for all outcomes measures for the LOV and CAU conditions at each assessment point are reported in Table 3. There were no significant differences between groups at T1 on parent age F (1,36) = .02, p = .879, child age F (1,36) = .02, p = .90, or child gender (X ² = .25, p = .618). There were no significant differences between groups at T1 on the CSHQ, F (1,35) = .788, p = .381, or FAWN‐YC F (1,35) =1.52, p = .226. However, the LOV group was found to have significantly higher scores than the CAU group on the PAS‐R F (1,35) = 6.89, p = .013.

TABLE 3.

Means and standard deviations of LOV and CAU groups at each assessment point.

	LOV						CAU
	Pre‐treatment (n = 19)		Post‐treatment (n = 19)		3‐month follow‐up (n = 19)		Pre‐treatment (n = 18)		Post‐treatment (n = 16)		3‐month follow‐up (n = 16)
	Mean	SD	Mean	SD	Mean	SD	Mean	SD	Mean	SD	Mean	SD
CSHQ	59.21	9.29	44.79	7.09	42.32	7.63	56.89	6.23	53.13	6.99	49.87	7.24
PAS‐R	58.21	18.76	41.11	18.45	34.26	17.00	42.06	18.67	40.00	19.74	34.94	17.83
FAWN‐YC	39.89	14.46	23.74	16.13	21.79	13.33	33.28	18.10	24.94	13.88	22.63	14.91

Abbreviations: CSHQ, Children's Sleep Habits Questionnaire; FAWN‐YC, Fears and Worries at Night in Young Children; PAS‐R, Preschool Anxiety Scale Revised.

There were no significant T1 differences between those retained and those not retained at T2 and T3 on parent age, F (1,36) = .05, p = .834, child age, F (1,36) = .03, p = .86, or child gender (X ² = .002, p = .97). There were no significant T1 differences between those retained and those not retained on the CSHQ, F (1,35) = 2.48, p = .124, PAS‐R, F (1,35) = .04, p = .841, or FAWN‐YC, F (1,35) = 2.91, p = .133.

Intervention effects

Primary outcomes

Sleep problems

At T1, 100% of parents reported that their child had a moderate or severe sleep problem (presence of a sleep problem) as it was an inclusion criterion for participation in the study. By T2, 93.7% of the CAU compared to 26.3% of the LOV condition children were rated by their parents as having a sleep problem, and this difference was significant (X ² = 16.13, p = <.001). By T3, 62.5% of the CAU compared to 21.1% of the LOV condition children were rated by their parents as having a sleep problem, and this difference was significant (X ² = 6.22, p = .01).

From T1 to T2, there was a significant main effect for Time (F[1, 33] = 39.25, p < .001) and a significant Group × Time interaction (F[1, 33] = 19.33, p < .001) on the CSHQ, indicating that although there was an overall improvement in sleep from T1 to T2, the LOV group reported a significantly greater reduction compared to the CAU group, with a large effect size (${\eta }_p^2$ = .37; see Figure 2a). There was a significant decrease in CSHQ scores from T1 to T2 for the LOV (t = 6.89, p < .001, d = 1.58), but not the CAU group (p = .13). Sleep problems were significantly lower in the LOV compared to the CAU group at T2 (t = 3.49, p = .001, d = 1.19).

FIGURE 2.

Outcomes at T1; pre‐treatment, T2; 2‐week post‐treatment and T3; 3‐month follow‐up.

From T2 to T3, there was a significant main effect for Time (F[1, 33] = 7.50, p = .01) but not Group × Time (F[1, 33] = .13, p = .71; see Figure 2b), indicating that there was an overall reduction in sleep problems from T2 to T3. A significant decrease in CSHQ scores was found from T2 to T3 for the CAU group (t = 2.59, p = .02, d = .64) but not the LOV group (p = .14), and sleep problems were significantly lower in the LOV compared to the CAU group at T3 (t = 2.99, p = .005, d = 1.01).

Treatment satisfaction

Parental treatment satisfaction was very high, with parents reporting a very high likelihood of recommending the programme to others (M = 4.84 SD = .38), belief that the programme was very effective in helping to improve their child's current sleep problem (M = 4.26 SD = .73), and belief that the programme addressed their child's future sleep difficulties (M = 4.58 SD = .69). Overall, parental satisfaction with sessions, resources and mode of delivery was also very high (M = 4.63 SD = .50).

Secondary outcomes

Anxiety

From T1 to T2, there was a significant main effect for Time (F[1, 33] = 17.10, p < .001) and a significant Group × Time interaction (F[1, 33] = 12.39, p = .001), on the PAS‐R, indicating that although there was a general reduction in anxiety from T1 to T2, the LOV group demonstrated a significantly greater improvement, with a large effect size (${\eta }_p^2$ = .27; see Figure 2b). There was a significant decrease in PAS‐R scores from T1 to T2 for the LOV (t = 4.68, p < .001, d = 1.07), but not the CAU group (p = .54). However, there was no significant difference between conditions in PAS‐R scores at T2 (p = .87).

From T2 to T3, there was a significant main effect for Time (F[1, 33] = 6.40, p = .02) but not Group × Time (F[1, 33] = .14, p = .71; see Figure 2b), indicating that there was a general decrease in PAS‐R scores from T2 to T3. There was a significant decrease in PAS‐R scores from T2 to T3 for the LOV group (t = 2.23, p = .03, d = .51), but not the CAU group (p = .13). Finally, there was no significant difference between conditions on the PAS‐R at T3 (p = .91).

Nighttime fears

From T1 to T2 there was a significant main effect for Time (F[1, 33] = 15.38, p < .001) and a significant Group × Time interaction (F[1, 33] = 4.35, p = .04) on the FAWN‐YC, indicating that although there was a general reduction in nighttime fears from T1 to T2, the LOV group demonstrated a significantly greater improvement, with a large effect size (${\eta }_p^2$ = .15; see Figure 2c). There was a significant decrease in FAWN‐YC scores from T1 to T2 for the LOV (t = 3.70, p < .001, d = .85), but not the CAU group (p = .09), and there was no significant difference between the groups in nighttime fears at T2 (p = .81).

There was no significant time (F[1, 33] = 2.18, p = .15) or Group × Time interaction (F[1, 33] = .02, p = .90; see Figure 2c) evident on the FAWN‐YC from T2 to T3. There was no significant difference in FAWN‐YC scores from T2 to T3 for either the LOV group (p = .33) or the CAU group (p = .15). Finally, there were no significant differences between conditions on nighttime fears at T3 (p = .86).

Access to sleep services during wait period

Of the 16 participants from the CAU group who completed the follow‐up survey, 6 (37.5%) did not seek any additional sleep services during the trial, 7 (43.75%) read materials regarding sleep online, 1 (6.25%) consulted with a general practitioner, 1 (6.25%) visited a psychologist and 1 (6.25%) both consulted with a general practitioner and read materials online. Of the 19 participants from the LOV group, none of the parents sought any additional sleep services during the trial.

DISCUSSION

This study aimed to evaluate the efficacy of a parent‐focused, videoconference, behavioural sleep intervention for children aged 3 to 5 years. It was predicted that relative to the care‐as‐usual (CAU) control condition, children in the Lights Out Videoconference (LOV) treatment condition would demonstrate significantly greater improvements in sleep, anxiety and nighttime fears from pre‐ to post‐assessment that would be maintained at a 3‐month follow‐up. It was also predicted that parents would be satisfied with the programme.

As predicted, the intervention was successful in reducing sleep problems. At post‐assessment, 93.7% of parents in the CAU group reported that their child had a moderate or severe sleep problems compared with only 26.3% of parents in the LOV group. By follow‐up, these rates had dropped to 21.1% for the LOV group and 62.5% for the CAU group. The LOV group also demonstrated a significantly greater reduction in CSHQ scores compared to the CAU group from T1 to T2, with both groups significantly and equally improving in sleep from T2 to T3. The positive treatment findings of this study with respect to sleep are consistent with research examining face‐to‐face behavioural sleep interventions that have included (albeit not always exclusively) children in the preschool developmental period (Donovan et al., 2023; Fehr et al., 2020; Meltzer et al., 2021), thus adding to the evidence that brief, parent‐focused behavioural interventions are effective in treating sleep problems in young children.

It is interesting to note that Donovan et al.'s (2023) five‐session, in‐person, group Lights Out programme and this three‐session, individual, videoconference version produced very similar results in terms of sleep outcomes. Both versions found similar reductions in parent‐reported sleep problems as measured by the CSHQ, with similar large effect sizes. At post‐treatment, the treatment group was reported as having significantly lower CSHQ scores than the control group in both the original (d = 1.00) and videoconference studies (d = 1.19). Additionally, at post‐treatment, 28.8% of the face‐to‐face treatment group and 26.3% of the videoconference treatment group retained a moderate to severe sleep problem. It would therefore appear that compared to the original face‐to‐face version of Lights Out, this shorter, more accessible programme is similarly effective in reducing sleep problems in preschoolers.

Interestingly, both the Donovan et al. (2023) study and the current study found the control group also demonstrated a reduction in sleep problems over time, highlighting the natural remission in sleep problems that occurs for some young children, and perhaps the fact that 62.5% of the control group accessed some type of assistance for their sleep over the course of the study. However, the greater reduction in sleep problems for the treatment versus control groups in both studies, and the greater percentage of control versus treatment children who retained a sleep problem at T2 and T3 in both studies, speak to the ability of the programme to assist those children for whom natural remission does not occur.

The findings that a videoconference version of the Lights Out programme has effects on anxiety and nighttime fears are promising and in line with Donovan et al.'s (2023) findings. Compared to the control group, those receiving the Lights Out Videoconference programme demonstrated a greater reduction in both anxiety and nighttime fears over time, therefore supporting hypothesis 3. However, it should be noted that there were no significant differences between groups on anxiety and nighttime fears at either T2 or T3. Although this may be due to pre‐treatment group differences on anxiety and nighttime fears (which were significant in the case of anxiety) coupled with the small sample size, the results must be interpreted with caution and the study must be replicated with a larger sample size before the effects of the videoconference programme on anxiety and nighttime fears can be determined. Nevertheless, given the very high comorbidity rates of behavioral sleep problems and anxiety in childhood (Chase & Pincus, 2011; Crowe & Spiro‐Levitt, 2024) and the fact that anxiety in young children is often overlooked, a videoconference programme that has the potential to reduce both problems in only three sessions represents a convenient and cost‐effective method for reducing the two most common health and mental health complaints of this developmental period.

Also, in support of hypotheses, parental treatment satisfaction ratings of the programme were very high, perhaps contributing to the 100% retention rates in the treatment group. Caregiver treatment satisfaction is known to be associated with both treatment engagement and outcomes (Acri et al., 2016), and given that the programme is parent‐focused, it is important that parents are both engaged and confident in the strategies taught so that they are able to implement the programme at home. The 100% attendance and retention rates are significant strengths of the study and may be attributed to the programme being more personalized to the particular sleep problems of each child, a strong working alliance and parents having shared decision making and autonomy throughout the programme. Future research would benefit from collecting both qualitative data (through interviews with parents) regarding the drivers of high attendance and retention rates, as well as administering a standardized measure of the therapeutic alliance to assess its importance.

Limitations and future directions

Despite its strengths, this study was not without limitations. First, while the study was adequately powered and there was no attrition in the treatment condition, the sample size was very small and may have been responsible for the statistically significant difference between groups at pre‐treatment on anxiety. Stratification of the sample by anxiety when randomizing families to condition may be one way that future research could ensure greater initial equivalence between groups on anxiety. Furthermore, the small sample was largely homogenous, consisting predominantly of Caucasian, highly educated, wealthy families with married, biological parents living together. The results may therefore not be generalizable to other socioeconomic groups, ethnicities and family structures and future studies would benefit from a larger sample size with greater diversity. Unlike the original Lights Out study and the inclusion of strategies targeting behaviour problems, the current study did not include a measure of oppositional behaviour. Future research would benefit from including a validated measure of oppositional behaviour. An additional limitation of the current study was that only one research assistant completed fidelity checks (specifically therapist adherence). Future research may benefit from multiple research assistants completing this task and including an inter‐rater reliability analysis.

In addition to the suggestions for future research described above, there are a number of other areas for future empirical enquiry. First, the use of gold standard clinical diagnostic interviews to assess sleep, anxiety and fear status would not only strengthen future studies but may also assist in assessing potential differential effectiveness of the programme on children with insomnia and/or specific anxiety disorders (i.e., separation anxiety, specific phobia of the dark). In addition, given that the ‘Lights Out’ programme is a comprehensive, multi‐component intervention, dismantling studies may be conducted in the future to investigate the most effective strategies for sleep and anxiety outcomes. Finally, given the cross‐disciplinary nature of sleep problems, and that the programme is manualised, the intervention could be delivered by other allied health professionals in a future, larger study in an attempt to expand access for families.

CONCLUSIONS

This study is promising for children, parents and clinicians. It demonstrates that the Lights Out programme is amenable to an individualized, videoconference format, which may be a more efficient, scalable modality of care for addressing behavioural sleep problems in the broader population. Delivering programmes via telehealth alleviates many barriers to parental treatment seeking. Given high treatment satisfaction combined with only a small number of sessions, the videoconference version of Lights Out provides clinicians with a user‐friendly, efficient, evidence‐based programme to extend evidence‐based treatment reach to families and reduce difficulties associated with time and financial burdens that frequently afflict parents of young children. The importance of identifying ways of optimizing access for families to evidence‐based treatments cannot be overstated. It is hoped that this study goes some way towards achieving this important goal.

AUTHOR CONTRIBUTIONS

Amy Shiels: Investigation; writing – original draft; conceptualization; methodology; formal analysis; project administration; data curation; resources; writing – review and editing. Lara J. Farrell: Writing – review and editing. Caroline L. Donovan: Supervision; writing – review and editing; conceptualization; methodology.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflict of interest.

Supporting information

Table S1.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.