Adherence to digitally delivered group acceptance and commitment therapy for insomnia

The World Health Organization1 defines adherence as the extent to which an individual's behavior aligns with health professional recommendations. In the context of insomnia, high adherence to psychological treatments is generally associated with significant improvements in sleep complaints.2-4 However, dropout remains a challenge, affecting approximately 14%–40% of patients.5 In the case of cognitive behavior therapy for insomnia (CBT-I), discontinuations are often linked to the demanding nature of the intervention, such as the use of sleep restriction techniques. A consistent limitation in the literature is the lack of standardized definitions and measurement approaches for adherence, which complicates comparisons across studies.6, 7 Despite this, current evidence suggests that greater psychosocial support, increased self-efficacy, and fewer dysfunctional beliefs about sleep are associated with better adherence to CBT-I. Research on newer approaches, such as acceptance and commitment therapy for insomnia (ACT-I), remains scarce. Initial findings suggest that ACT-I may facilitate adherence to behavioral components of CBT-I, such as stimulus control and sleep restriction.8 Within the scenario of ACT-I, what is understood as adherence differs from the metrics of CBT-I and whose parameters are better defined (e.g., adherence to the sleep window can be assessed using total sleep time, and time to go to bed). Although there is no consensus in the literature, adherence to ACT-I probably correlates with the construct of psychological flexibility, which can be understood as the capacity to be fully aware and in contact with the present moment using the circumstances offered by the situation to adjust or persist in behaviors that are aligned with personal values and goals.9 Within this context, adherence appears to be directly linked to the ability to deal with discomfort without avoiding difficult experiences, accepting it, in addition to contact with the present moment. Since studies investigating the adherence to psychological treatments for insomnia are scarce and new therapeutic modalities are promising, the aim of the present study is to evaluate the factors associated with adherence to ACT-I. In view of the lack of data on whether psychological flexibility influences adherence to ACT-I, our hypothesis is that higher levels of psychological inflexibility and higher levels of anxiety predict nonadherence to treatment. Data were collected as part of a study that evaluated the efficacy of ACT-I versus CBT-I. Details of the procedures have been published previously.10 In summary, participants were recruited, screened, and randomly assigned to ACT-I or CBT-I or a waiting list after baseline assessments. For the current aim of assessing the factors associated with adherence to treatment, only insomnia patients randomized to ACT-I were included in this study (n = 76). Therefore, these represent secondary analyses. The study was approved by the Research Ethics Committee of the university where it was conducted (CAAE: 65743917.2.0000.0068 and Approval number: 4.582.587). All participants provided informed consent electronically. The study design was preregistered at ClinicalTrials.gov (Identifier: NCT04866914) (https://classic.clinicaltrials.gov/ct2/show/study/NCT04866914). Although the clinical trial was preregistered, the present research question on factors associated with adherence was not preregistered. Participants were eligible if they were between 18 and 59 years of age and met the criteria for chronic insomnia: (i) difficulty initiating or maintaining sleep, defined as a sleep onset latency and/or time awake after sleep onset ≥30 min, and/or early morning awakenings with a corresponding total sleep time ≤6.5 h per night; (ii) insomnia complaints occurring more than three nights per week and persisting for more than 3 months; and (iii) sleep disturbance (or associated daytime fatigue) causing significant distress or impairment in social, occupational, or other important areas of functioning. These definitions represented a combination of criteria from the American Academy of Sleep Medicine, the International Classification of Sleep Disorders, and the Diagnostic and Statistical Manual of Mental Disorders, along with quantitative cutoffs typically used in insomnia research.11-13 Exclusion criteria comprised the presence of progressive or unstable medical conditions (e.g., cancer and acute pain) or neurodegenerative disease, unstable psychiatric comorbidities (e.g., lifetime diagnosis of psychotic or bipolar disorder, more than two lifetime episodes of major depressive disorder, or untreated current major depressive disorder), other untreated sleep disorders (moderate to severe sleep apnea, restless legs syndrome, and circadian rhythm sleep–wake disorders), alcohol or substance misuse (including marijuana use more than once per week), illiteracy, and night-shift work incompatible with treatment participation. Participants were also excluded if they were unable to attend weekly group sessions or did not have access to a computer with internet connection to complete online sessions. Importantly, participants taking sleep medication were not excluded, provided that their dosage had been stable for at least 3 months prior to enrollment. Previous exposure to cognitive or psychological treatments for insomnia was not an exclusion criterion; however, none of the participants were engaged in such therapies at the time of recruitment. Psychiatric diagnoses at baseline were assessed using the Mini International Neuropsychiatric Interview (MINI14), which was used exclusively for eligibility and descriptive characterization of the sample. Detailed data on the therapeutic guidelines have been reported previously.10 ACT-I consisted of six group sessions via a platform for digital meetings, each lasting 120 min. All sessions were conducted by two clinical psychologists certified in sleep science, who were trained in the ACT-I protocol (the first and the last author of this paper). The ACT-I protocol was based on the ACT-I manual of Hayes et al.15 and the protocol proposed by El Rafihi-Ferreira et al.10 Adherence was assessed through a self-report questionnaire administered posttreatment. Participants classified their engagement as follows: (a) having completed the proposed tasks (“total adherence”) or (b) not having completed the tasks (“non-adherence”). For statistical analyses, adherence was dichotomized into adherent (≥75% of tasks completed, corresponding to “all” or “most” tasks) versus nonadherent (<75%). This operational cutoff was chosen to ensure that adherence reflected consistent engagement across the six-session program, while acknowledging occasional lapses. Tasks primarily included mindfulness exercises and behavioral recommendations discussed in each session. Data were collected in the posttreatment phase using the adherence and satisfaction questionnaire. Examples of items include: “After each session, did you feel engaged to carry out the proposed tasks?” and “Did you follow the recommendations and complete the tasks requested in the intervention program?”. Sociodemographic data (age and sex), symptoms of insomnia, anxiety, depression, and psychological inflexibility were evaluated as potential predictors of treatment adherence. Insomnia severity was assessed using the Insomnia Severity Index (ISI),16 in which higher scores indicate greater severity of insomnia symptoms. Symptoms of anxiety and depression were measured using the Hospital Anxiety and Depression Scale (HADS),17 composed of two subscales: HADS-A (anxiety) and HADS-D (depression), with higher scores reflecting greater symptom intensity. For the regression analyses, HADS scores were treated as continuous variables. Psychological inflexibility was assessed using the Acceptance and Action Questionnaire-II (AAQ-II),18 where higher scores indicate greater psychological inflexibility. Data were analyzed using Jamovi (version 2.3), adopting a significance level of 0.05 for all comparisons. Descriptive statistics are presented as proportions, means, and standard deviations. Multivariable logistic regression analysis was performed to assess psychological inflexibility as a factor associated with adherence to ACT-I. The multivariable model included age, sex, symptoms of insomnia, anxiety and depression as covariates, and psychological inflexibility as a factor associated with adherence (following recommendations) to ACT-I. Odds ratios (ORs) with 95% confidence intervals were reported as the primary effect measures. Additionally, participants were dichotomized into adherent (≥75% of tasks completed) and nonadherent (<75%) groups. Between-group comparisons were conducted using independent samples t-tests for continuous variables and chi-squared tests (or Fisher's exact tests, when applicable) for categorical variables. A total of 76 participants were enrolled in the study and completed baseline assessments (59 women; mean age = 40.5 years and SD = 10.6). Baseline demographic and clinical characteristics of the full sample are presented in Table 1. Of the 76 participants enrolled at baseline, 66 completed the Adherence and Satisfaction to Treatment questionnaire and were therefore included in the adherence and posttreatment analyses. Among these participants, 25 (37.9%) were classified as adherent (≥75% of tasks completed) and 41 (62.1%) as nonadherent. Participants with available adherence data (n = 66) were dichotomized into adherent (n = 25) and nonadherent (n = 41) groups. Table 1 presents baseline sociodemographic characteristics and mean scores (±SD) for HADS-A, HADS-D, ISI, and AAQ-II for the full sample as well as comparisons between adherent and nonadherent participants. No statistically significant differences were observed between adherent and nonadherent participants regarding sex, insomnia severity, or depressive symptoms. Adherent participants were significantly older and reported lower psychological inflexibility, with AAQ-II scores on average 6.7 points lower than those of nonadherent participants. The results of the multivariable logistic regression analysis examining baseline predictors of treatment adherence are presented in Table 2. All analyses were conducted among participants with available adherence data (n = 66). Multivariable analysis showed that each additional year of age was associated with a 12% increase in the odds of adherence to the intervention (OR = 1.12 and 95% CI [1.04–1.21]). Higher anxiety symptoms were also significantly associated with greater odds of adherence, such that each additional point on the HADS-A increased the odds of adherence by 32% (OR = 1.32 and 95% CI [1.02–1.70]). In addition, lower psychological inflexibility was associated with greater adherence, with each one-point decrease on the AAQ-II increasing the odds of adherence to therapy (OR = 0.84 and 95% CI [0.74–0.95]). The current paper aimed to describe potential predictors of adherence to ACT-I. The results of the logistic regression model demonstrated that age and AAQ-II increased the odds of adherence. In addition to age and psychological inflexibility, anxiety symptoms (HADS-A) were also statistically significantly associated with adherence. Participants reporting higher anxiety scores were more likely to engage consistently with treatment tasks. One possible explanation is that individuals with elevated anxiety may perceive greater urgency to manage their sleep difficulties and therefore show stronger motivation to apply therapeutic strategies. Another possibility is that higher levels of anxiety is associated with higher conscientiousness, which in itself is a general predictor of finishing tasks that one has taken on. However, this finding should be interpreted with caution as the effect was modest and the sample size limited. Future studies should further examine the role of anxiety symptoms in predicting adherence to ACT-I, ideally using larger samples and multimethod assessments of both anxiety and adherence. The finding suggest that age was positively associated with treatment adherence and offers an important perspective on the mechanisms underlying engagement in psychological interventions for insomnia. Older adults may present greater routine stability, a clearer perception of the consequences of poor sleep, and consequently higher motivation to follow therapeutic recommendations. In addition, factors, such as increased personal responsibility for health and fewer competing time demands (e.g., work or family caregiving), may facilitate the regularity required for mindfulness and behavioral exercises proposed in ACT-I. Conversely, younger individuals—often exposed to more variable schedules and with less experience in emotional self-regulation—may face greater challenges in sustaining acceptance- and values-based practices. Thus, age may reflect not only contextual and availability differences but also distinct levels of psychological maturity and flexibility, which are central to the ACT model. These results highlight the need for tailored engagement strategies for younger participants, such as personalizing clinical examples and incorporating more interactive digital tools, to gradually foster psychological flexibility throughout treatment. As mentioned earlier, the concept of adherence to ACT is not yet consensual. However, since theoretical models of third-generation therapies are broader and more contextual and involve lifestyle changes,19 the concept of adherence may follow the same premise. Within this context, the measurement of adherence may depend on understanding what maintains each patient's complaint, thereby defining which behaviors would be appropriate as indicators of adherence. However, in the present study, adherence was defined idiosyncratically by the participants, who provided a general self-assessment of whether they had been following the treatment. This methodological aspect highlights the challenge of operationalizing adherence within contextual behavioral frameworks and reinforces the need for future studies to incorporate behaviorally anchored measures of adherence. Given the above, quantitative measures of psychological flexibility can help compose the range of components for the assessment of adherence since some instruments are already available and validated. This is the first study reporting the findings of evaluation of role of psychological flexibility in adherence to interventions of ACT-I that did not involve behavioral elements of CBT-I. Corroborating our hypothesis, the greater the psychological flexibility at baseline, the greater the treatment adherence. This means that the lower the experiential avoidance and cognitive fusion and the greater the capacity to be in contact with the present moment, acting toward one's own values, with a contextual self, the higher the probability of adhering to the proposed treatment.9 From a clinical perspective, identifying patients with higher levels of psychological flexibility may also help predict treatment adherence, enabling the design of more effective and personalized interventions. Additionally, the predictive potential of psychological flexibility for treatment adherence appears to go beyond insomnia; for example, studies that investigated other clinical population suggested that psychological inflexibility can negatively influence behaviors related to treatment adherence, such as consistent attendance at sessions and persistence, in following therapeutic recommendations despite discomfort.20 Understanding the mechanism of psychological flexibility as a predictor of adherence can provide insights into the therapeutic process. Future empirical research should examine whether interventions grounded in values clarification, mindfulness, and acceptance could enhance engagement among younger participants or those demonstrating higher levels of experiential avoidance. Exploring these mechanisms may help clarify how psychological flexibility contributes to sustained treatment adherence across developmental stages. Despite the promising results, one of the main limitations of the present study, as well as of the current literature on adherence to psychological treatment for insomnia, is related to the inconsistencies in the measurement of adherence. For example, Harvey et al.21 compared adherence between patients undergoing CBT-I and patients who were submitted only to the behavior components of the intervention and cognitive therapy. Despite the proposal, the authors did not evaluate adherence of the cognitive therapy group, probably because of the lack of adherence variables that were measurable from a cognitive point of view. A further limitation relates to the group-based digital format (via a platform for digital meetings), which may have influenced adherence—potentially supporting engagement through peer accountability and reducing logistical barriers, but also introducing challenges such as technology fatigue. In addition, the present analysis did not include dysfunctional beliefs and attitudes about sleep, although this instrument was administered in the primary trial. Given the limited sample size, we prioritized a parsimonious model to reduce the risk of overfitting. Future studies with larger samples should systematically evaluate dysfunctional sleep-related beliefs as potential predictors of adherence in ACT-based interventions. In conclusion, the current study provides evidence supporting the understanding of adherence to psychological treatment for insomnia. Adherence to certain treatment recommendations may be more predictive of treatment outcome than others. Future studies should further examine subjective and objective measures of adherence, including in interventions that use ACT-I. Ila Marques Porto Linares: Investigation; writing—original draft; formal analysis; project administration. Susanna Jernelöv: Writing—review and editing; data curation; formal analysis. Renatha El Rafihi-Ferreira: Conceptualization; supervision; resources; project administration; funding acquisition; writing—review and editing; software. The authors would like to thank all participants; M. Carmo, H. Lira, and A. Souza for statistical advice; and the ASONO-HCFMUSP group for their contribution. Renatha El Rafihi-Ferreira was the recipient of a grant from the São Paulo Research Foundation (FAPESP) (grant number 2018/19506-5). The authors declare no conflicts of interest. The study was approved by the appropriate Institutional Review Board/Ethics Committee, and all participants provided written informed consent prior to participation. Research data are not shared.