Efficacy of a Paradoxical Intervention in the Treatment of Sleep-Onset Insomnia in Women During and After the Climacteric Stage: A Pilot Study

Abstract

Paradoxical interventions involve asking patients to maintain or exaggerate the symptoms they wish to eliminate, with the aim of reducing them. These techniques received empirical support, particularly in the treatment of sleep-onset insomnia, during the 1980s. However, changes in sleep habits over recent decades warrant further research on their current applicability. This article reports a pilot study involving 26 women (aged 40–70) with significant difficulty initiating sleep, who were randomly assigned to either a paradoxical or a non-paradoxical treatment; a prior waiting period was also used as a waitlist control. Both treatments produced statistically significant improvements in insomnia scores between pre- and post-treatment, with no significant differences between groups. These findings suggest that paradoxical interventions may be effective for the current treatment of sleep-onset insomnia and support the need for randomized clinical trials with larger samples.

1. Introduction

Paradoxical interventions consist of instructing the patient to intentionally intensify the behavior or thoughts that they seek to overcome in order to diminish it. Although they may seem absurd to those unfamiliar with such techniques, paradoxical interventions follow certain logics. First, asking a patient to increase the symptoms they wish to eliminate can help them view the problem from a new perspective, as well as achieve a rapid behavioral change [1]. On the other hand, as O’Connell [2] points out, paradoxical interventions transfer control of the symptom directly to the patient. Typically, the symptoms experienced by the client are beyond their control: they do not want to have anxiety attacks, yet they do; they want to sleep, yet they cannot. Asking the patient to exaggerate these symptoms gives them control over them.

Paradoxical interventions have been examined by numerous authors throughout history. Among the figures who that contributed the most to their development and dissemination was Viktor Frankl, who introduced the technique of paradoxical intention as a central component of his widely recognized logotherapy [3]. Frankl applied paradoxical techniques to overcome phobias and anticipatory anxiety by encouraging patients to do, or wish to happen, the things that are feared.

Other key figures in the history of paradoxical interventions were Milton Erickson, who skillfully integrated the symptom into the treatment process [4], and the research group from the Mental Research Institute in Palo Alto, California, that includes Watzlawick, Jackson, and Weakland, as well as others. They argued that advice such as “Don’t get into trouble” or “Be kind to your siblings,” although defining the desired change, cannot be regarded as therapeutic. Such advice assumes that the patient can choose whether or not to be kind, or whether or not to get into trouble. However, in clinical reality, symptoms lie beyond the patient’s control, making it impossible for them to simply stop experiencing them voluntarily [5].

However, it should be noted that paradoxical interventions have often been criticized as manipulative or unethical [6] because they ask patients to voluntarily increase their symptoms. However, many authors argue that such critiques stem from misunderstandings since these techniques have shown effectiveness in treating psychological symptoms. Hunsley [7] highlights the importance of distinguishing between compliance-based interventions, where improvement occurs when patients follow the therapist’s instructions as given, and defiance-based ones, where improvement occurs when patients disobey the therapist’s directives. Ethical concerns mainly arise in the latter. In compliance-based approaches—our focus here—patients may, for example, be asked to exaggerate their anxious symptoms, which paradoxically leads to a reduction in anxiety. As noted before, these interventions return a sense of control to the patient: by deliberately performing symptoms that normally arise involuntarily, they learn to manage and eventually extinguish them. This experience of gaining control is rewarding.

Moreover, previous research has provided some evidence for the therapeutic effectiveness of paradoxical interventions [8,9,10,11]. In their meta-analysis, Shoham-Salomon and Rosenthal [12] reported that these interventions were as effective as standard treatment approaches. Hampton [13] further concluded that paradoxical interventions outperformed all comparison groups—including non-paradoxical treatments, placebo, and waiting list conditions—in addressing certain psychopathologies. These results agree with those previously obtained by Hill [14] who observed that patients receiving paradoxical interventions showed significantly greater improvements than those treated with non-paradoxical approaches. Furthermore, paradoxical interventions appear to be particularly effective in the treatment of sleep difficulties, especially sleep-onset insomnia [1,15]. Several studies performed in the 1980s reported encouraging results regarding the effectiveness of paradoxical interventions in reducing sleep onset latency [16,17,18,19].

However, despite these results—where significant differences were found between pre- and post-treatment measures in all cases, with special emphasis on the study conducted by Gould [18] in which the paradoxical treatment proved to be more effective than two non-paradoxical treatments and a waiting list control—the study of the effectiveness of paradoxical interventions for the treatment of insomnia has not been revisited. In the study of insomnia, it is particularly important to rely on recent research as sleep habits have changed considerably in recent decades. One of the reasons for these changes is the widespread use of screens. According to a study conducted by Hjetland and colleagues in 2025 [20], which included 45,202 participants, each additional hour of screen use before bedtime was associated with a reduction of 24 min of sleep and a 59% increase in the risk of developing insomnia. Therefore, to properly assess the effectiveness of paradoxical interventions in the treatment of sleep-onset insomnia, current experimental research is needed to address today’s challenges.

On the other hand, another aspect that has not been studied as thoroughly as it deserves is insomnia during the climacteric period. According to Salari and colleagues [21], the prevalence of sleep disorders after menopause is 51.6%. However, other review studies, such as the one performed by Gyun-Ho [22], acknowledge that there is still a shortage of standardized protocols addressing the treatment of insomnia in women during and after the climacteric stage. It is worth noting that sleep-onset insomnia, defined by the DSM-5 as difficulty initiating sleep [23], may entail negative consequences for overall functioning in daily life for those who suffer from it.

For all these reasons, the aim of the present study was to evaluate the efficacy of a paradoxical intervention in the treatment of sleep-onset insomnia in women aged 40 to 70 years. To this end, the paradoxical treatment described by Gould in 1988 [18] was employed since it was the one that achieved the best results at that time [15]. In that study, as previously mentioned, paradoxical therapy was compared with two non-paradoxical treatments: coping imagery and sleep information. Among these, the one that proved to be the most effective in reducing sleep onset latency was coping imagery; therefore, this was the non-paradoxical treatment used as the comparison condition for the paradoxical intervention.

2. Materials and Methods

2.1. Participants

This study involved 26 women between 40 and 70 years of age with difficulty falling asleep. The mean age was 52.8 years (SD = 6.60). Regarding sample size, Billingham et al. [24] conducted an audit in 2013 of pilot and feasibility trials registered in the United Kingdom Clinical Research Network database. They reported a median sample size of n = 33 for publicly funded pilot studies and n = 25 for those with private funding. Therefore, the sample included in the present study falls within the range typically observed for this type of design.

All participants reported a sleep onset latency greater than 30 min. For the sleep onset latency data, participants were asked to report their average sleep onset latency, defined as the time elapsed from attempting to fall asleep until sleep onset. The participants’ mean latency was 75.8 min (SD = 26.2).

Regarding the duration of insomnia, all participants had experienced difficulties initiating sleep for at least one year. The mean duration of insomnia was 17.2 years (SD = 18.2).

2.2. Procedure

Participant recruitment was carried out through an advertisement on radio and social media. It offered free psychological treatment to women aged 40 to 70 years with significant difficulty initiating sleep. No compensation was provided for participating in the study.

A total of 88 women expressed interest in participating in the study, of whom, 34 met the eligibility criteria. The inclusion criteria were being between 40 and 70 years of age, experiencing difficulty falling asleep, and having full proficiency in Spanish, in both written and oral comprehension. The exclusion criteria were currently undergoing another psychological treatment, taking medication that induces drowsiness—whether prescribed specifically for sleep problems or for other purposes (e.g., anxiolytics), including melatonin supplements or other sedative dietary supplements—and having a physical condition that prevented or hindered sleep.

Finally, of the 34 participants who met the eligibility criteria, only 26 began the study as 8 declined participation before signing the informed consent form. Therefore, 26 participants took part in the study.

2.3. Design

This study employed a randomized controlled pilot design with two experimental groups (paradoxical vs. non-paradoxical intervention) and three measurement points (PRE1, PRE2, and POST).

Prior to participation, all subjects received verbal and written information regarding the general objectives of the study, the procedures and expected duration, as well as the potential benefits and risks. Confidentiality of the data was guaranteed, and the voluntary nature of participation, including the right to withdraw at any time without consequences, was emphasized. Written informed consent was then obtained from all participants. Subsequently, an initial assessment session was carried out in which the Insomnia Severity Index (ISI) and the Menopause Rating Scale (MRS) were administered. This constituted the first measurement point, hereafter referred to as PRE1. Following this, all participants underwent a four-week waiting list period. The purpose of this period was to examine how the passage of time without treatment affected the severity of their insomnia.

After PRE1, each participant was randomly assigned to one of the experimental groups: the group receiving a paradoxical intervention or the group receiving a non-paradoxical intervention. Randomization was performed using the =RANDBETWEEN(1,2) command in Microsoft Excel, where 1 corresponded to the paradoxical treatment and 2 to the non-paradoxical treatment. A separate control group was not included, as we considered it ethically inappropriate to withhold treatment from participants. Instead, the 4-week no-treatment period was employed as a waitlist control, which was compared with the treatment period to assess differences.

Following the 4-week waiting period, participants were scheduled to complete the Insomnia Severity Index (ISI) for the second time, a measurement point hereafter referred to as PRE2. Immediately thereafter, the first treatment session was initiated.

The duration of both treatments was 6 weeks, followed by a final session in which the Insomnia Severity Index (ISI) was administered for the third time along with the Consumer Reports Effectiveness Scale (CRES-4). This measurement point will hereafter be referred to as POST.

Permission to carry out the research was obtained from the Ethical Committee for Scientific Research of the University of Valencia (protocol code 2024-PSILOG-3468732).

2.4. Treatment

Both treatments consisted of six weekly individual sessions of approximately 30 min each, conducted at the Faculty of Psychology, University of Valencia. All treatment sessions were delivered by a single trained clinician (the first author), which eliminated between-therapist variability. Therefore, additional therapist-standardization procedures were not applicable. The non-paradoxical treatment was Coping Imagery, while the paradoxical treatment was Paradoxical Therapy.

In Paradoxical Therapy, participants were encouraged to understand that falling asleep is not a voluntary process and that relinquishing control over sleep can facilitate its onset, with the main strategy being to intentionally try to remain awake as long as possible. In Coping Imagery, participants were guided to recognize the role of pre-sleep cognitions in sleep initiation and to replace maladaptive thought patterns with more adaptive imagery-based strategies.

During the first four sessions, both treatments addressed four common thought patterns that delay sleep onset: mental overactivity, reviewing the past, worries about the future, and concerns about the consequences of poor sleep—each with specific paradoxical or imagery-based exercises. The final two sessions were used to repeat the two strategies that each participant had found to be the most effective and to consolidate the skills learned throughout the treatment.

For a detailed description of the full treatment protocols, readers are referred to Gould [18].

2.5. Measures

2.5.1. Personal and Demographic Characteristics

An ad hoc questionnaire was employed to gather the participants’ personal and demographic information, i.e., age, nationality, average sleep-onset latency, duration of insomnia, and questions to determine whether the eligibility criteria for participating in the study were met.

2.5.2. The Insomnia Severity Index (ISI)

The Insomnia Severity Index [25] in its Spanish version [26] was administered to the participants to assess the severity of their insomnia at the different measurement time points. The Index is composed of 7 items, with the responses given using a five-point Likert scale (0 = no problem, 4 = very severe problem). Cronbach’s α was 0.82.

2.5.3. The Menopause Rating Scale (MRS)

The Menopause Rating Scale (MRS) [27] was administered in its Spanish version [28] to assess the severity of each participant’s menopausal symptoms prior to treatment. The scale has 11 items, with the responses given using a five-point Likert scale (0 = none, 1 = mild, 2 = moderate, 3 = severe, 4 = very severe). Cronbach’s α was 0.86.

2.5.4. The Consumer Reports Effectiveness Scale (CRES-4)

The Consumer Reports Effectiveness Scale (CRES-4) [29] (Spanish version [30]) was administered to assess participants’ satisfaction with the treatment received upon treatment completion. It consists of four items: a satisfaction question, a question to rate the degree of resolution of the problem, a question about emotional state before starting treatment, and finally, a question about emotional state when the treatment has finished. The response scale varies from 0 to 5 for the first two questions, and from to 0 to 4 for the last two questions. Cronbach’s α was 0.78.

2.6. Statistical Analysis

The data were analyzed using JAMOVI (Version 2.7) with the GAMLj module for linear models. Descriptive statistics and normality analyses (Shapiro–Wilk test) were first conducted. Variables meeting the assumption of normality (Insomnia Severity Index, Menopause Rating Scale, and Consumer Reports Effectiveness Scale) were analyzed using parametric tests. To evaluate treatment efficacy, a mixed repeated-measures ANOVA was performed with measurement points (PRE1, PRE2, POST) as the within-subject factor and treatment group (paradoxical vs. non-paradoxical) as the between-subject factor. Subsequently, a general linear model was conducted to test whether menopausal symptoms (MRS scores) influenced treatment efficacy, including their interaction with treatment group. Finally, an independent samples t-test was performed to examine group differences in perceived efficacy and treatment satisfaction (CRES-4 scores).

3. Results

3.1. Descriptive Analysis

Descriptive analyses of the results are shown in

Table 1. Descriptive statistics of study variables.

	Age	Mean Latency (min)	Duration (Years)	MRS	ISI_PRE1	ISI_PRE2	ISI_POST	CRES-4
N	26	26	26	26	26	26	22	22
Missing	0	0	0	0	0	0	4	4
Mean	52.8	75.8	17.2	14.8	19.8	20.7	12.8	220
Median	52.0	70.0	10.0	14.0	20.0	20.0	13.0	220
Standard deviation	6.60	26.2	18.2	5.91	3.07	3.17	5.04	33.8
Minimum	40	40	1	5	13	13	3	145
Maximum	64	120	60	25	26	27	22	275
Skewness	−0.19	0.34	1.35	0.22	−0.31	0.09	0.02	−0.29
Std. error of skewness	0.45	0.45	0.45	0.45	0.45	0.45	0.49	0.49
Kurtosis	−0.76	−0.93	0.39	−0.70	−0.03	0.63	−0.31	−0.52
Std. error of kurtosis	0.88	0.88	0.88	0.88	0.88	0.88	0.95	0.95
Shapiro–Wilk W	0.96	0.88	0.76	0.94	0.96	0.95	0.97	0.96
Shapiro–Wilk p-value	0.507	0.008	<0.001	0.192	0.560	0.372	0.854	0.545

MRS, Menopause Rating Scale; ISI, Insomnia Severity Index; CRES4, Consumer Reports Effectiveness Scale.

. As we can see, 26 participants started the study and 22 completed the post-treatment assessments; 4 participants decided to discontinue the treatment before finishing it. Among the dropouts, 2 belonged to the non-paradoxical group (they dropped out after the fourth and the first session, respectively) and 2 belonged to the paradoxical group (both dropped out after the first session).

Regarding the distribution of the variables, most of them did not significantly deviate from normality according to the Shapiro–Wilk test. This is consistent with the skewness and kurtosis data, where the values between −2 and +2 allowed us to assume normality [31,32]. This information was considered when deciding on the most appropriate statistical tests for subsequent analyses.

3.2. Efficacy

A mixed repeated-measures ANOVA was conducted on the three time points (PRE1, PRE2, and POST) and two factors (paradoxical intervention group and non-paradoxical intervention group) based on Insomnia Severity Index scores (see

Table 2. Post hoc comparison group x moment.

Moment	Group	Moment		Group	Mean Difference	pbonferroni
PRE1	Paradoxical	-	PRE1	Non-paradoxical	−0.364	1
		-	PRE2	Paradoxical	−1.364	0.795
		-	PRE2	Non-paradoxical	−1.091	1
		-	POST	Paradoxical	8.727	<0.001
		-	POST	Non-paradoxical	4.455	0.229
	Non-paradoxical	-	PRE2	Paradoxical	−1	1
		-	PRE2	Non-paradoxical	−0.727	1
		-	POST	Paradoxical	9.091	<0.001
		-	POST	Non-paradoxical	4.818	0.079
PRE2	Paradoxical	-	PRE2	Non-paradoxical	0.273	1
		-	POST	Paradoxical	10.091	<0.001
		-	POST	Non-paradoxical	5.818	0.039
	Non-paradoxical	-	POST	Paradoxical	9.818	<0.001
		-	POST	Non-paradoxical	5.545	0.015
POST	Paradoxical	-	POST	Non-paradoxical	−4.273	0.656

). No significant differences were found between PRE1 and PRE2 scores in the paradoxical group (p = 0.795) or in the non-paradoxical group (p = 1.000). Significant differences were observed between PRE2 and POST scores in the paradoxical group (p < 0.001) and in the non-paradoxical group (p = 0.015). No significant differences were found between the POST scores of the two groups (p = 0.606). However, although both groups improved significantly from PRE2 to POST, the mixed model indicated a greater overall reduction in the paradoxical group (B = −4.676, p = 0.034; see

Table 3. Parameter estimates from the general linear model including MRS as a covariate.

				95% Confidence Intervals
Names	Effect	Estimate	SE	Lower	Upper	β	df	t	p
(Intercept)	(Intercept)	7.862	1.021	5.717	10.007	0.008	18	7.700	<0.001
Group1	Non-paradoxical— Paradoxical	−4.676	2.042	−8.967	−0.386	−0.898	18	−2.290	0.034
MRS	MRS	−0.206	0.166	−0.555	0.144	−0.252	18	−1.237	0.232
Group 1 × MRS	(Non-paradoxical— Paradoxical) × MRS	0.273	0.333	−0.426	0.972	0.334	18	0.821	0.422

SE, Standard Error; df, degrees of freedom.

for parameter estimates, including MRS as a covariate).

3.3. Influence of Menopause

As shown in Table 3, MRS did not significantly predict changes in Insomnia Severity Index scores (B = −0.206, p = 0.232). In addition, the interaction between group and MRS was not significant (B = 0.273, p = 0.424).

3.4. Participants’ Perceived Effectiveness

To examine differences in perceived treatment effectiveness between the paradoxical and non-paradoxical groups, independent-samples t tests were conducted. Levene’s test indicated that the assumption of homogeneity of variances was not violated, F(1, 20) = 3.62, p = 0.072. Based on this result, the t test was interpreted using the row assuming equal variances.

The results of the independent-sample t test revealed no statistically significant differences in CRES-4 scores between the paradoxical and non-paradoxical groups (t(20) = 0.62, p = 0.541).

4. Discussion

The results suggest that both interventions were effective in the treatment of sleep-onset insomnia in women aged 40 to 70. Importantly, no differences were observed between pre- and post-waiting list assessments, indicating that the mere passage of time did not exert a significant therapeutic effect. This finding is particularly relevant, as we deliberately chose not to include a no-treatment control group, considering that such a procedure is ethically questionable. By including all participants in a one-month waiting list and evaluating the severity of their insomnia before and after this period, we were able to assess the influence of time alone, without treatment, on their recovery process. In this case, the passage of time without treatment proved neither beneficial nor detrimental for participants.

In contrast, statistically significant improvements were observed from pre- to post-treatment in both the paradoxical intervention group (p < 0.001) and the non-paradoxical intervention group (p = 0.015). These findings indicate a significant reduction in insomnia severity following treatment. The absence of improvement during the waiting-list period, combined with the presence of significant changes during the treatment period, strengthens the interpretation that the observed benefits are attributable to the interventions rather than to external factors such as the passage of time.

Furthermore, no significant differences were observed between the two groups at the post-treatment level (p = 0.656). However, the general linear model including menopausal symptoms (MRS) as a covariate revealed a significant advantage for the paradoxical intervention over the non-paradoxical approach (p = 0.034). Moreover, 63.6% (n = 7) of the participants who received paradoxical intervention showed a >8-point reduction in their ISI scores from pre- to post-treatment, a level of improvement considered clinically significant according to Morin et al. [33]. In contrast, only 27.3% (n = 3) of the participants who received the non-paradoxical intervention improved by more than 8 points. This indicates that, once potential variability associated with menopausal symptoms was controlled, participants in the paradoxical group demonstrated greater improvements in insomnia severity. Both interventions produced significant pre–post treatment reductions in insomnia, yet the paradoxical intervention appears to provide a more consistent clinical benefit.

These results are consistent with those reported by Gould [18] in 1988, although the present study introduces some noteworthy novelties.

-

First, the findings suggest that the intervention originally proposed by Gould may still be effective today. This is particularly relevant given that sleep dynamics and habits have changed considerably over the past decades, raising the possibility that treatments effective in the past might not retain their efficacy in the current context. For example, recent evidence indicates that screen use substantially reduces total sleep time [20]. This highlights the emergence of new factors in recent years that may pose challenges for insomnia treatments developed in the previous century.

-

Second, whereas Gould [18] examined the efficacy of paradoxical intention in both men and women aged 19 to 80, the present study focused specifically on women during and after the climacteric stage—a population in which sleep problems are particularly prevalent [21].

Moreover, these findings are in line with the results of review articles on paradoxical interventions, such as the meta-analysis conducted by Peluso and Freund [1] and the scoping review by Viguer, Díaz, and Martín [15], both of which concluded that paradoxical interventions are particularly indicated for the treatment of sleep-onset insomnia. It is also worth highlighting the study conducted by Ascher and Schotte [34], in which they evaluated the efficacy of a paradoxical intervention in the treatment of speaking anxiety. They concluded that paradoxical intervention was more effective than the non-paradoxical treatment when speaking anxiety included a recursive component, suggesting that the efficacy of such techniques may depend on whether the anxious symptomatology is also recursive. Recursive anxiety is defined as anxiety that generates further anxiety; in other words, when the onset of the anxious symptoms themselves triggers an escalation of anxiety in the individual. This is particularly relevant in the present study, as sleep-onset insomnia could also involve a recursive component: the worry about not being able to sleep may produce sleep-onset insomnia. Anxiety is generating more anxiety. The results of this pilot study support the hypothesis proposed by Ascher and Schotte [34] and underscore the importance of examining the efficacy of paradoxical intention in other disorders characterized by recursive anxiety.

With respect to menopausal symptoms, no significant relationship was found between MRS scores and treatment efficacy. This suggests that the severity of menopausal symptomatology does not predict treatment outcomes, indicating that paradoxical and non-paradoxical interventions may be equally effective regardless of the level of menopausal symptom severity. This finding is particularly relevant given that the treatment of insomnia during menopause remains under-researched [22]. The scarcity of empirical evidence highlights the need for further research focused on this population where sleep disturbances are highly prevalent and often undertreated. The present study may represent an initial step toward developing a systematic line of research aimed at establishing a standardized protocol for the treatment of insomnia in menopausal women.

With respect to perceived efficacy and satisfaction with the treatment received, the CRES-4 scores indicated a generally good level of satisfaction with the interventions. The scale ranges from 0 to 300, and the mean and median scores in our sample were 220, placing them in the upper third of the scale. No significant differences were found between groups (p = 0.541), suggesting comparable levels of perceived efficacy and satisfaction across both interventions. This is noteworthy given that the ethics of paradoxical interventions have often been questioned [6]. However, if such interventions prove to be effective and yield satisfaction levels similar to those of non-paradoxical practices, the ethical debate may have limited grounds for continuation.

Regarding the limitations of the study, the most evident is the relatively small sample size, which constrains both the statistical power of the analyses and the generalizability of the findings. Although we incorporated a four-week no-treatment baseline period, which allowed for meaningful within-subject comparisons, the absence of an independent control group without treatment limits the ability to draw stronger causal conclusions about the intervention’s effects. The inclusion and exclusion criteria could also have been more precisely defined. Insomnia was identified through retrospective subjective self-report rather than a prospective method such as a 1–2-week sleep diary, which would have provided a more detailed characterization of sleep-onset latency. Likewise, while the physical condition exclusion criterion was intended to rule out cases in which insomnia was clearly secondary to physical discomfort, the study did not explicitly screen for specific sleep-related conditions such as Obstructive Sleep Apnea (OSA) or Restless Legs Syndrome (RLS). Psychiatric screening followed a pragmatic approach by excluding only individuals currently in treatment; however, not assessing untreated severe comorbidities (e.g., depression or anxiety) may have introduced additional variability. In addition, no objective measures of sleep, such as polysomnography or actigraphy, were employed. The inclusion of such measures could have strengthened the validity of the results. Instead, we relied on standardized self-report instruments, such as the Insomnia Severity Index, which may be subject to social desirability and other response biases. On the other hand, the results of this study do not allow for the evaluation of the intervention’s effectiveness in the medium or long term since no follow-up measure was included. Likewise, the influence of other variables—beyond menopausal symptoms—that could have affected the outcomes obtained was not assessed.

Despite these limitations, the findings of this pilot study are encouraging and underscore the need for a randomized clinical trial with a larger sample to further evaluate the efficacy of this technique in the treatment of sleep-onset insomnia in women during and after the climacteric stage. It would also be valuable to incorporate follow-up measures to assess long term effectiveness.

Following with future research lines, it would be interesting to examine the efficacy of paradoxical intention in the treatment of sleep-onset insomnia with recursive anxiety, that is, in patients whose worry about not being able to sleep is what keeps them awake. It seems reasonable to expect that precisely this type of patient would benefit the most from instructions such as ‘try to stay awake all night’. If they genuinely begin to wish to remain awake, the worry about not being able to sleep may eventually dissipate. In addition, future studies could include a minimal-intervention control group (e.g., psychoeducation) and objective sleep parameters allowing to increase the attribution of results and strengthen the validity of the findings.

In conclusion, both treatments demonstrated significant effects in addressing sleep-onset insomnia in women during and after the climacteric period, with the paradoxical intervention appearing to provide a more consistent clinical benefit. Despite the study’s limitations, the findings underscore the need to continue exploring paradoxical therapeutic mechanisms—particularly for insomnia—in larger samples.