Spa therapy improves sleep quality: evidence from questionnaire and actigraphy

Abstract

Background

Sleep-wake rhythm is highly correlated with core body temperature, and bathing in the hot spring can effectively manipulate core body temperature to improve sleep quality. However, there is still a lack of experimental research on the influencing factors of utilizing hot spring facilities to positively affect sleep.

Methods

In this study, we aimed to explore the factors improving sleep quality via hot spring bathing in terms of both subjective (self-rating scales) and objective (actigraphy) measures, 33 insomniacs were recruited and randomly divided into the afternoon (AG) and evening bathing groups (EG).

Results

The results showed that after the intervention, subjective sleep quality was improved significantly in both groups, and the evening bathing group had a better performance. In contrast, neither group had a significant change in objective sleep measures. Moreover, we found that hot spring spa therapy is particularly effective in reducing pre-sleep thoughts inventory and shortening sleep onset latency, and it is more applicable to people with difficulty falling asleep.

Conclusion

This study identified the timing and individual's behaviour traits as the important factors, and these formed practical guidance for sleep improvement by spa therapy.

Keywords:

• Insomnia
• hot spring
• passive body heating
• actigraphy
• pre-sleep thoughts inventory

1. Introduction

Sleep disturbance ranks second on the list of epidemic mental illnesses in the world [1], which may lead to the decline of cognitive functions such as attention and memory, resulting in a decrease in efficiency and an increase in making mistakes in both learning and working situations. Long-term sleep disturbances will lead to weakened immunity and induce cardiovascular diseases, endocrine diseases, and other serious physical diseases [2]; it is also one of the high-risk factors for mental diseases such as depression, anxiety, and dementia. In addition, the previous study had shown that the growth of age becomes a risk factor for the higher incidence of sleep disturbance in adults [3].

Currently, the treatment of insomnia is mainly divided into pharmacological and non-pharmacological treatments. Pharmacological therapy takes effect quickly but is prone to resistance or dependence, and has poor healing effects. In recent years, Cognitive Behavioral Therapy for Insomnia (CBTI), a non-pharmacological therapy, has become the “first-line therapy” [4,5] for insomnia, but its clinical use has not been standardized, and the course of treatment is time-consuming and expensive. Therefore, exploring other easy-to-operate non-pharmacological treatments to improve sleep makes practical sense.

The switch between sleep and wakefulness exhibited close relation to the alteration of the circadian rhythm of body temperature [6]. Previous studies found that sleep onset latency (SOL) can be shortened by decreasing core temperature [7,8]. Since sleep disturbances are partly due to a decline in heat dissipation from the core to the periphery of the body as we age, interventions that enhance heat dissipation prior to sleep can improve sleep quality [9]. Therefore, passive body heating (PBH) is often recommended as a simple, low-cost non-pharmacological treatment for insomnia. It improves sleep quality via modifying the temporal distribution of body temperature, and causing reduced core temperature before sleep [10].

Studies have shown that PBH has a positive effect on sleep, mainly on the reduction in SOL. One study focused on young soccer players observed a shortened SOL in both subjective and objective indicators by increasing the distal skin temperature through a shower before bed [8]. Healthy female subjects shortened their average SOL for about 18 min after the treatment of a 40 °C water bath at 20 min before bed [11]. Raymann et al. [12] demonstrated the positive correlation between a four-minute reduction of SOL in healthy young people and the use of heated bed socks. All these studies showed that different types and durations of PBH can reduce SOL.

For the timing of PBH, Horne and Shackell found in young people that manipulating body temperature via a hot bath in the early afternoon triggered sleepiness and shortened the SOL [13]. Another study showed that the SOL was shortened when the body was passively heated at night, but there was no change in SOL after bathing in the morning [14]. These results demonstrate that the timing of PBH also has an important effect on SOL. In previous studies, the majority of PBH interventions are applied at night, especially before bedtime. According to a systematic review, bathing at about 1–2 h before bedtime is an essential factor in the reduction of SOL [15]. However, there is still a lack of research on the timing of hot springs on sleep quality improvement.

There were previous studies showed that hot spring bathing has a positive effect on promoting sleep [16–18]. During bathing in a hot spring, the water can gently stimulate the peripheral nerves of the body, reducing their excitability and reflexively strengthening the inhibitory process of the cerebral cortex to achieve a sedative and analgesic effect; at the same time, the bath can promote the metabolism of substances, improve blood circulation on the surface of the body, so that people relax physically and mentally, and thus improve their sleep [16]. Hayasaka and his colleague found that people who took regular baths showed better sleep quality [17]. Goto et al. considered going to a hot spring once a month did not change an individual’s overall sleep status, but if the questionnaires were administered immediately after bathing, there might be an improvement in subjects’ sleep quality [18]. However, the number of studies on hot spring bathing improves sleep quality is limited and the findings of the promotion effects still lack consistency. Most of the subjects recruited by previous studies were healthy adults, and few insomniacs were considered [15]. While a previous study had found differences in the sleep structure of people with insomnia compared to healthy ones [19], it is of great significance to conduct research on the insomnia group. What’s more, previous studies neglected the individual factors affecting the improvement of sleep quality, and the difference between subjective and objective effects of hot spring treatment. Considering most of these studies were based on subjects’ self-rating scales and self-reports, which are susceptible to subjectivity and social expectations, the combination of a more objective measurement could be useful. In this study, we chose actigraphy, a portable and long-term monitoring recording instrument which has been proven to have a good detection accuracy to detect the objective sleep quality of the subjects. Actigraphy provides an objective physiological assessment via non-invasive use of motion accelerometers to detect motor activity and transfer the identified activity counts into epochs to report sleep and wakeful states according to some algorithms, and this method has been proven to moderate level of agreement with polysomnography in distinguishing between sleep and wakeful states, so it can be a reasonable alternative to measure sleep [20].

Additionally, patients with insomnia often complain of having uncontrollable unwanted thoughts before bedtime, and these worrisome thoughts affect their sleep greatly [21]. Sufficient evidence showed that: compared to people who sleep well, insomniacs have more unpleasant bedtime thoughts. Furthermore, insomnia is associated with more sleep-disrupting thoughts, counterfactual processing, worries, maladaptive thought control strategies, covert monitoring, and cognitive arousal [6]. Later, a study found that pre-sleep cognitive activities can lead to an increase in SOL [22]. Thus, although related studies have been less involved in passive body heating interventions, regarding the calming and relaxing effects of hot springs, we speculated that hot spring baths may be accompanied by a reduction in intrusive thinking in subjects during the reduction of their SOL.

This study focused on the effect of hot spring bathing on improving the subjective and objective sleep quality of people with insomnia and the effect of time selection on the sleep-promotion. We randomly divided the subjects into the afternoon (AG) and evening bathing groups (EG) to compare the effects; we used actigraphy and various questionnaires to collect both subjective and objective measures. Also, considering the prominent effect of passive body heating on SOL in previous studies and possible individual differences in bathing, we collected chronotype indicators of subjects to investigate in-depth the effect of hot spring intervention on sleep quality improvement in specific populations. Furthermore, we collected data on subjects’ pre-sleep thinking to explore the effect on subjects’ intrusive thinking. In this study, we hypothesized that: (1) hot spring baths have improved the subjective and objective sleep quality of the insomnia group, and EG with bathing time closer to bedtime has a better effect on sleep promotion; (2) hot spring recuperation is more effective for people who have trouble in falling asleep; (3) hot spring baths reduce bedtime intrusive thinking in people with insomnia, and the bedtime thinking may positively correlate with the length of subjects’ SOL.

2. Methodology

2.1. Subjects

The experiment was reviewed by the Ethics Committee of the Faculty of Psychology, Southwest University, and participants were recruited from the community in Beibei District, Chongqing via advertisements. A total of 33 (7 males and 26 females) people with insomnia were finally recruited after screening, all of whom met the following conditions: (1) aging between 45 and 69 years old; (2) physically qualified by physical examination to meet the physical standards for hot spring bathing: (3) not taking medication for insomnia or other psychiatric disorders withinone week before the trial; (4) no psychiatric disorders affecting sleep and no other sleep disorders, such as obstructive sleep apnea (OSA); (5) insomnia occurring and lasting for more than 15 days with an Insomnia Severity Index (ISI) score higher than 8.

2.2. Procedures

Figure 1 shows the whole process of the experiment: the first 3 days are the baseline period, each subject wore an actigraphy device from the time they arrived at the experiment site to the time they left the site, which last for 10 days, and completed the morning and evening sleep diaries after waking up and before going to sleep. The subjects all completed the total scales on the third day of arrival, and from the fourth day onwards, they took a daily hot spring bath of half an hour for 7 days. The timing of baths depends on their group. Subjects completed their last sleep diary in the morning of the 10th day, returned the actigraphy devices, filled out some of the scales for the second time (Figure 1(A)), and left the experimental site.

Figure 1. Flow chart of the experiment of bathing in a hot spring. (A) The recording time point for all 10 days; (B) the between-group designs of afternoon and evening groups; (C) the schedule of the 30 min bathing.

During the 7-day spa therapy period, subjects in AG and EG were required to receive a daily hot spring bathing around 16:00–17:00 and 20:00–21:00, respectively (Figure 1(B)). The bathing temperature of both groups was at least 38 °C and the water level was required to be up to the chest, and the total duration was 30 minutes, with a break of at least 3 minutes every 7 minutes (Figure 1(C)).

2.3. Scales and measurements

The 10-day sleep diary was used to assess the subjective sleep time and wake-up time. In order to compare the difference between before and after treatment, the diary data from day 1–3 were defined as the baseline period, and the data from day 8 to 10 were defined as the PBH period. We used many other scales to measure the mental states of the subjects (general demographic variables, depression, anxiety, thinking, fatigue, personality traits, sleep quality, etc.) during the baseline period and after the PBH period. The scales were filled on day 3 (baseline period) and day 10 (PBH period), respectively. Our instruction asks the subject to answer based on their habits during the past three days only.

Scales include: Pittsburgh Sleep Quality Index (PSQI): Measurement of seven parameters including (1) sleep quality, (2) sleep onset latency, (3) sleep duration, (4) sleep efficiency, (5) sleep disturbance (6) use of sleep medication, and (7) daytime dysfunction, which are effective tools for sleep quality investigation of individuals in general, and for sleep quality evaluation of patients with sleep disturbances and mental disorders [23]; Insomnia Severity Index (ISI): Insomnia self-assessment scale to determine the severity of the patient’s insomnia [24]; Glasgow Content of Thoughts Inventory (GCTI): Assessing the content, characteristics, and intrusive thinking of adults at bedtime is particularly applicable to assessing the cognitive processes of patients with insomnia, where cognition may influence the development of treatment plans and treatment outcomes for patients [21]; Munich ChronoType Questionnaire (MCTQ): Assessing the chronotype (both workdays and work-free days) and additional, optional modules on substance use, light exposure and food intake, mainly in the field of scientific research to study how chronotype is related to age, gender and external environment [25,p.12]. Here we used three parameters: MCTQ-MSFsc, MCTQ-SLF, and MCTQ-SLW. MCTQ-MSFsc refers to the chronotype; MCTQ-SLF refers to the sleep latency on work-free days, and MCTQ-SLW refers to the sleep latency on workdays; Self-Rating Depression Scale (SDS); Self-Rating Anxiety Scale (SAS); sleep diary, etc. We use these scales to explore the changes in sleep quality and its related factors in people with insomnia. Three of these scales: ISI, PSQI, and GCTI were measured twice, both at baseline (the third day) and after the intervention (the 10th day).

2.4. Actigraphy

All subjects were given actigraphy devices to assess changes in their objective sleep quality. The wGT3X-BT wireless triaxial actigraphy was used to measure sleep parameters, movement steps, and light. For the sleep analysis in this study, the Cole-Kripke sleep staging algorithm [26] was chosen to obtain the time of falling asleep and the time of wakefulness as objective sleep indices to be compared with subjective reports.

As the actigraphy could not collect the sleep data of the day before the subject arrived, hence only data from days 2 to 3 was used to calculate the sleep parameter for the baseline period. The activity indexes of each of these periods were averaged and recorded as the activity indicators after the intervention.

2.5. Statistical analysis

In this experiment, a mixed experimental design of 2 (treatment effect of hot spring bathing: the baseline period, the PBH period) × 2 (timing of hot spring bathing: AG vs. EG) was adopted. The treatment effect was the within-subject factor and the timing was the between-subject factor, and repeated measures ANOVA is utilized here. The scale scores were statistically analyzed using SPSS 24.0 software. For the post-hoc effect, paired-samples t-test and independent samples t-test were conducted. In addition, for the effect of individual differences, Pearson correlation was used.

3. Results

3.1. General demographics and psychological indicators

The general demographic and psychological indicators of the baseline status of the subjects the in AG and EG groups are shown in Table 1.

Table 1. General demographics and psychological indicators.

Variables	Afternoon group (AG)	Evening Group (EG)	t (χ^2)	p
Age (years)	57.30 ($\pm 7.92$)	59.19 ($\pm 6.05)$	−0.77	.448
Gender (male, female)	(2, 15)	(5, 11)	1.87	.171
BMI	22.19 (±1.40)	21.74 (±2.29)	0.67	.507
SAS	44.35 (±2.32)	42.69 (±2.39)	0.50	.621
SDS	38.94 (±1.78)	36.44 (±1.84)	0.98	.335
MCTQ-MSFsc	3.55 (±1.15)	3.05 (±1.40)	1.15	.259
MCTQ-SLF	47.94 (±27.79)	63.75 (±47.84)	−1.17	.251
MCTQ-SLW	49.71 (±27.30)	68.44 (±52.40)	−1.30	.203

Values are means ± SE. SDS: Self-Rating Depression Scale; SAS: Self-Rating Anxiety Scale; MCTQ: Munich Chronotype Questionnaire; MCTQ-MSFsc: Chronotype; MCTQ-SLF: sleep latency on work-free days; MCTQ-SLW: sleep latency on workdays.

There were no statistically significant differences between AG and EG in terms of age, gender, and body mass index (BMI). There were no statistically significant differences between the two groups on the SAS and SDS scales. For the subjects’ circadian rhythms, we used the MCTQ to examine the chronotype, sleep onset latency, and other indicators on workdays and work-free days, and there were no differences between the two groups.

3.2. Changes in sleep and mental health indicators

The changes in sleep and mental health indicators are displayed in Table 2 and Figure 2.

Figure 2. Changes in ISI, GCTI, PSQI, and sub-indexes of PSQI before and after hot spring bathing. ISI: Insomnia Severity Index; GCTI: Glasgow Content of Thoughts Inventory; PSQI: Pittsburgh Sleep Quality Index. **p < 0.01; ***p < 0.001.

Table 2. Sleep and mental health indicators for 2 groups in baseline and PBH periods.

Variables	Groups
	Afternoon		Evening
	Baseline	PBH	Baseline	PBH
ISI	12.471 (±0.97)	10.59 (±0.98)	13.06 (±1.00)	10.19 (±1.01)
GCTI	59.94 (±2.35)	51.59 (±2.63)	58.31 (±2.42)	49.81 (±2.71)
PSQI	11.82 (±0.84)	8.12 (±0.81)	12.75 (±0.80)	7.81 (±0.83)
PSQI-SQ	3.35 (±0.17)	2.36 (±0.18)	3.25 (±0.17)	2.31 (±0.19)
PSQI-SL	2.53 (±0.15)	1.47 (±0.25)	2.56 (±0.16)	1.69 (±0.26)
PSQI-ST	1.77 (±0.24)	1.24 (±0.25)	2.19 (±0.24)	0.81 (±0.23)
PSQI-SE	1.82 (±0.27)	1.12 (±0.26)	2.06 (±0.28)	1.19 (±0.26)
PSQI-SD	0.41 (±0.12)	0.41 (±0.14)	0.38 (±0.13)	0.38 (±0.14)
PSQI-DSD	1.65 (±0.24)	1.53 (±0.22)	1.75 (±0.25)	1.19 (±0.22)

Values are means ± SE. ISI: Insomnia Severity Index; GCTI: Glasgow Content of Thoughts Inventory; PSQI: Pittsburgh Sleep Quality Index; PSQI-SQ: sleep quality; PSQI-SL: sleep latency; PSQI-ST: sleep time; PSQI-SE: sleep efficiency; PSQI-SD: sleep disturbance; PSQI-DSD: daytime dysfunction.

Repeated measures ANOVA was conducted on the scores of the ISI, PSQI, and GCTI scales respectively, with the following results:

ISI was significantly lower after the PBH (M = 10.39, SD = 1.67) compared with the baseline period (M = 12.77, SD = 1.53), F(1,33)=7.65, p=.009; the main effect of the group was not statistically significant, nor was the interaction effect of treatment and timing.

As for PSQI, the main effect of treatment was statistically significant, subjects obtained significantly lower scores after PBH (M = 7.96, SD=.58) than the baseline (M = 12.28, SD = 0.60), F(1, 33)=30.09, p<.001; the main effect of timing was not statistically significant, nor was the interaction effect of the two.

The repeated measures ANOVA results for CGTI were similar to the above two scale scores. GCTI scores were significantly lower after the PBH period (M = 50.70, SD = 1.89) compared with the baseline period (M = 59.13, SD = 1.69), F(1,33)=17.99, p<.001; the main effect of timing was not statistically significant, nor was the interaction effect.

The results of the repeated measures ANOVA of the average scores on the ISI, GCTI, PSQI, and PSQI sub-indicators for the two groups of subjects at baseline and after PBH are shown in Figure 2.

For the PSQI sleep quality sub-index scores, the results showed a statistically significant difference between the subjects’ scores at PBH and baseline; the main effect of timing was not statistically significant; the interaction effect was not statistically significant. Similar to the results of this sub-indicator, the scores of sleep onset latency, sleep time, and sleep efficiency index showed similar results that there were significant main effects of treatment; non-significant main effects for timing and their interaction.

For sleep quality, the main effect of treatment F(1, 31)=23.48, p<.001; for sleep time, F(1, 31)=17.83, p<.001; for sleep efficiency, F(1, 31)=17.83, p<.001. Different from the sub-indicators above sleep disturbance and daytime dysfunction of PSQI were not statistically significant in main effects and interaction effects.

In a post-hoc analysis, a paired-samples t-test of the scores of each sub-indicator for subjects in each group at baseline and after PBH revealed that: in both afternoon and evening bathing groups, subjects’ scores on sleep quality (t = 3.23, p<.05), sleep onset latency (t = 4.24, p<.001), and sleep efficiency (t = 2.14, p<.05) indicators were significantly lower. Besides, sleep time and daytime dysfunction sub-indicators were significantly reduced after hot spring bathing.

3.3. Subjective changes in sleep dairy

Changes in subjective reports of sleep time and wake-up time in both groups are shown in Figure 3.

Figure 3. Subjective reporting of changes in sleep and wake-up times for subjects in the afternoon and evening bathing groups, where the grey dots and thin lines = sleep time for the afternoon group; black dots and thin lines = sleep time for the evening group; grey dots and thick lines = wake-up time for the afternoon group; black dots and thick lines = wake-up time for the evening group. The shaded area is the time window of baseline and passive body heating (PBH) periods.

Repeated measures ANOVA showed that: the main effect of treatment(F(1, 31)=0.24, p=.63), the main effect of timing (F(1, 31)=0.42, p=.52) and the interaction (F(1, 31)=1.09, p=.31) were not statistically significant. The results illustrated that sleep time didn’t change significantly through hot spring bathing, nor was there a statistically significant difference between subjects in the afternoon and evening bathing groups. So it could not be assumed that the objective sleep time varies with the bathing intervention.

Similarly, repeated-measures ANOVA was performed on the subjects’ wake-up time, the results indicated that the main effect of treatment was statistically significant F = 4.72, p<.05, subjects’ wake-up time was about 0.4 h earlier with the treatment; but the wake-up time did not differ significantly for timing, F = 0.53, p=.47; marginal significance was shown in the interaction effect, F = 4.03, p = 0.05, after treatment, the wake-up time of the afternoon group was about 0.35 h earlier than the baseline period, while the wake-up time of the evening group did not change significantly with the number of days of bathing.

Repeated measures ANOVA on the subjects’ sleep duration (wake-up time - sleep time) revealed that the main effect of treatment was non-significant, F = 1.37, p=.25; nor was the main effect of timing, F = 1.17, p=.28; but the interaction effect between the bathing intervention and the bathing time point was statistically significant, F = 5.92, p<.05. Along with the bathing days, the subjects in the afternoon group slept about 0.4 h less than in the baseline period while in the evening group they slept about 0.14 h more than in the baseline period.

In post-hot analysis, independent sample t-tests were conducted after dividing the subjects’ sleep time, wake-up time, and sleep duration into baseline and PBH periods according to the time point in which the hot spring intervention was performed. And the results showed:

In the baseline period, there was no statistically significant difference between the subjects in the afternoon group and the subjects in the evening group in all indicators; in the PBH period, there was no statistically significant difference between the subjects in the afternoon group and the subjects in the evening group in the sleep time, the wake-up time and the sleep duration were marginally significant between the subjects in the afternoon and evening groups, the sleep duration was about 0.57 h longer in the evening group than in the afternoon group, p=.065.

3.4. Objective changes in daily routine

As shown in Figure 4, the subject’s objective activity indicators were recorded by actigraphy, which recorded the subject’s sleep time and wake-up time during the latter 9 days.

Figure 4. Objective activity indicators were recorded by actigraphy devices, where the grey dots and thin lines = sleep time for the afternoon group; black dots and thin lines = sleep time for the evening group; grey dots and thick lines = wake-up time for the afternoon group; black dots and thick lines = wake-up time for the evening group. The shaded area is the time window measured by the scales.

Due to missing data from the actigraphy for 1 subject in the afternoon group, only 16 subjects in the afternoon group and 16 subjects in the evening group are analyzed. Repeated measures ANOVA was performed on the sleep time and wake-up time reported by the actigraphy devices, respectively.

For sleep time, wake-up time, and sleep duration, all the main effects of treatment, the main effect of timing, and the interaction effect of both were not statistically significant. Specifically, for the sleep time, treatment F(1, 30)=1.12, p=.30, timing F(1, 30)=0.94, p=.34, the interaction effect of the treatment and timing F(1, 30)=0.12, p=.73; for the wake-up time, treatment F(1, 30)=0.14, p=.71, timing F(1, 30)=1.14, p=.29, the interaction effect of the treatment and timing F(1, 30)=0.20, p=.66; for the sleep duration, treatment F(1, 30)=0.01, p=.93, timing F(1, 30)=0.29, p=.60, the interaction effect of the treatment and timing F(1, 30)=0.32, p= .57.

3.5. Effect of individual differences on the effect of sleep promotion

To identify the important individual traits that influence spa therapy on sleep promotion, we investigate the correlation between sleep quality and chronotype. The difference in total PSQI between the baseline and the PBH reflects the amount of sleep quality improvement of the subject, which is denoted as dPSQI. Pearson’s correlation analysis was performed on each factor of the MCTQ, and the correlation was found between the "sleep latency on workdays (SLW)" and "sleep latency on work-free days (SLF)" of the MCTQ. As shown in Figure 5, there is a moderate positive correlation between dPSQI and SLW (r = 0.516, p<.01); also, there is a moderate positive correlation between dPSQI and SLF (r = 0.505, p<.01).

Figure 5. Correlation of the sleep quality improvement (dPSQI) with MCTQ indicators. Where (A) is a scatter plot of dPSQI and sleep latency on workdays (SLW); (B) is a scatter plot of dPSQI and sleep latency on work-free days (SLF). Grey dots = afternoon group, black dots = evening group.

4. Discussion

This study used a community-based population with difficulty in falling asleep to conduct an experimental study of the sleep promotion effect of hop spring. There are three main findings: first, hot spring bathing can improve subjective sleep quality, and the effects of improving sleep are influenced by time selection, but the intervention effects are not shown on objective sleep; second, hot spring bathing is particularly effective in reducing pre-sleep thinking and shortening sleep onset latency. Third, spa therapy is more suitable for people who have difficulty falling asleep.

GOTO et al. found that bathing was associated with good sleep quality [18]. Regarding the relationship between the use of hot spring facilities and sleep quality, they concluded that going to a hot spring bathing once a month would not change a person’s overall sleep status, but if administer questionnaires to subjects after bathing, their sleep quality may have improved consistently compared to a normal situation without this intervention. This experiment complemented their study by exploring the effects of continuous hot spring bathing for 7 days on sleep quality: subjects’ ISI scale scores decreased by 2 points and PSQI scores decreased by 4.32 after a week of bathing [18]. This result is consistent with our result, and it depicts that hot spring bathing can lead to an improvement in subjective sleep quality and a decrease in the insomnia severity for insomnia patients. Subjective sleep quality was significantly better, with shortened sleep onset latency and higher sleep efficiency in both groups one week after hot spring bathing compared to the baseline period. The results of this experiment suggest that spa therapy can significantly improve sleep quality, mainly by shortening sleep latency and improving sleep quality. Such results are consistent with previous observations that older adults slept better after evening baths [27]. But our results differ from the non-significant results on sleep shown by a study of having before-bedtime foot-baths in patients with sleep disorders [28]. This contradiction may indicate that controlling body temperature by proximal heating in people with sleep disorders may improve sleep quality more than distal heating.

For the comparison of treatment timing in the two groups, the daytime dysfunction was significantly enhanced, and sleep time was earlier in the evening group, while no such significant changes were observed in the afternoon group. According to subjects’ sleep diaries, there was no significant difference in sleep time, wake-up time, and sleep duration between subjects in the afternoon group and subjects in the evening group at baseline. While at the PBH period, subjects in the evening group slept approximately 0.54 h longer than the afternoon group at this period. In addition to this, the bath time point and bathing intervention in the evening and afternoon groups had significant interaction effects on subjective wake-up time and sleep duration: the evening group had a constant wake-up time and approximately 0.4 h longer sleep duration than the baseline period, while the afternoon group had an earlier wake-up time and shorter sleep duration than the baseline period. All of these results suggest that the effect of hot spring bathing on improving sleep is influenced by the time selection with evening bathing promoting sleep more than afternoon bathing, and choosing a time closer to bedtime can improve more sleep sub-indicators. Since the duration of the hot spring bath in the evening group was approximately 2.5 h relative to the sleep time, it consists to the study revealed that a bath of 10 min or more within 2–3 h before bedtime can shorten the sleep latency by 6 min, but if the bath is taken within 1 h or more than 3 h before bedtime it does not help with sleep [29]. It also considers the sleep-promoting time window mentioned in the review: Before bedtime passive body heating by warm shower or bath to improve sleep [15]. In addition, one study described more immediate (after 2 weeks) and substantial long-term (after 4–6 weeks) improvement in sleep when bathed twice weekly in the early evening (18:00–19:00) than an early afternoon (14:00–15:00) in a small group of dementia patients [30]. These consistent results can re-emphasize the important role of timing of the bath for sleep improvement and that such a time window applies to people with sleep disorders.

It is noteworthy that the significant differences in subjective sleep did not extend further to objective indicators of sleep. In our objective sleep and wake times measured by actigraphy devices, subjects showed an overall trend of positive changes in each indicator after the hot spring bath, but the changes were not significant nor did they reflect a significant difference between the two groups. This may be related to group characteristics of the insomnia population: sleep-disordered populations tend to show a tendency to underestimate sleep duration and overestimate sleep latency, waking after falling asleep [31], sleep instability [32], and errors in time estimation [33] that affects the perception of sleep, resulting in subjective and objective differences.

It has been found that a reduction in unwanted pre-sleep cognitive activities can reduce the sleep latency of subjects [22], we collected data on subjects’ pre-sleep thinking before and after hot spring recuperation by the GCTI scale and found that hot spring bathing significantly reduced intrusive cognitive activities in people with insomnia: subjects’ GCTI scores were reduced by 8.50 in the evening group and by 8.35 in the afternoon group. This suggests that hot spring recuperation had a favorable effect on the insomnia group and also significantly reduced the intrusive pre-sleep cognition of the subjects. Combined with the PSQI sleep onset latency, we suggest that the reduction of sleep onset latency and the reduction of pre-sleep thinking in subjects by hot spring baths are related and that the reduction of pre-sleep thinking may alleviate the difficulty in falling asleep in insomnia patients to some extent. However, other views suggest that stopping or suppressing pre-sleep thinking may be the cause of insomnia [34]. Taken together, it appears that the further relationship between sleep onset latency and intrusive pre-sleep cognition after hot spring bathing remains uncertain, and whether stopping pre-sleep thinking can be a therapy for insomnia remains unknown.

Another important finding of this study is the prominent effect of spa therapy on people who have difficulty falling asleep. Insomnia is known to be the most common sleep disorder in clinical practice and patients with insomnia often have difficulty initiating and/or maintaining sleep or early awakening, often associated with poor sleep quality and a sense of daytime dysfunction [35]. Previous studies on hot spring interventions on sleep have reported that bathing could significantly promote sleepiness, slow-wave sleep, and non-rapid eye movement (NREM) stage IV sleep, thereby improving sleep quality [36]. Results from a population-based study suggested that a hot spring bath may be an effective intervention for improving sleep and is suitable as a therapeutic intervention for patients with insomnia [17]. The current experiment find by correlation analysis that the longer the SOL (time from going to bed to falling asleep) on workdays or/and non-workdays, the more likely people were to have improved sleep quality after the hot spring bathing intervention. Combined with the existing studies on the effect of bathing on the sleep latency of subjects [11,15,18], hot spring recuperation may be more applicable to groups with difficulty falling asleep, and this finding has operational implications for the development of targeted programs for hot spring bathing.

There are some limitations of our current study: first, the actigraphy devices have technical limitations in the objective sleep monitoring of subjects, and their sleep staging accuracy greatly depends on the algorithm. Future studies could use Polysomnography (PSG) to explore finer sleep structure and a deeper understanding of its electrophysiological mechanism such as the presence or absence of spindle wave loss coupling, etc. Second, the study did not follow up on the sleep of the subjects for a while after the hot spring bath, and it was not possible to collect information about the long-term delayed effect of the intervention; (3) The sample of subjects in the experiment was not large enough, only 33 subjects, the gender was not balanced enough as there were more female subjects; (4) The experiment lacks a control group without hot spring baths, i.e. lacks a comparison between the post-bath subjects and the subjects who did not undergo the bathing intervention. In future experiments, an additional control group can be set up based on larger sample size and a balance of sex ratio to improve the above defects of this experiment.

5. Conclusion

In this study, both subjective (self-rating scales) and objective (actigraphy) measures were utilized to investigate the influence of hop spring spa therapy. After the intervention, subjective sleep quality was improved in both groups, though EG has better effects. In contrast, neither group had an obvious change in objective sleep quality. Moreover, hot spring spa therapy was particularly effective in reducing pre-sleep thoughts inventory and shortening sleep onset latency, and it was more applicable for people who had difficulty falling asleep. This study identified the timing and individual’s behavior traits as the important factors, and these formed practical guidance for sleep improvement by spa therapy.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the National Key Research and Development Program of China (2021YFC2501500) and the National Natural Science Foundation of China (31971028).