Highlights

Poor sleep quality is associated with a decrease in quality of life and a higher risk of suffering pathologies. Sleep hygiene habits are recommended based on their relationship with sleep quality but sometimes that intervention has no significant changes in sleep quality.

This article shows the prevalence of poor sleep quality and its association with lifestyle habits in a population recruited in community pharmacies, in Salamanca (Spain).

A real association is established between some lifestyle habits and subjective sleep quality. New strategies focused on certain habits could be developed thanks to this study.

Introduction

Disturbed sleep has been linked to cardiovascular and metabolic problems^1,2 and shown to have a negative effect on mental illnesses³. Its high prevalence generates considerable costs for health systems⁴. In Spain, at least 20% of the population suffer symptoms of insomnia 3 nights per week, with a higher prevalence among women (23.9%) than among men (17.6%)⁵.

Subjective sleep quality (SQ) is the perception individuals have of how they sleep. It is based on a series of subjective evaluations by the patient and on objective sleep-related measurements. The latter include sleep latency, sleep duration, sleep efficiency, sleep disturbances, use of sleep medication and daytime dysfunction^6-8. To prevent poor SQ and its consequences, strategies other than the use of medication have been investigated (cognitive behavioral therapy, sleep hygiene, etc.). In the long term, these strategies appear to be preferable to hypnotics and have demonstrated patient improvement⁹.

Peter Hauri coined the term “sleep hygiene” in 1977 in reference to habits, activities and practices (to encourage or avoid) that need to be considered for achieving adequate SQ¹⁰. Several studies have examined the effect of numerous lifestyle habits on improving SQ^6,11-13. However, the sleep changes thus achieved are not always significant and improvement varies depending on the sleep component^14-15 and age^16,17.

Few studies on SQ in the population and the factors possibly associated with it have been conducted in Spain and, as far as we are aware, none has been conducted in the town of Salamanca.

The aim of this study was therefore to study the prevalence of poor SQ among the population of Salamanca and its association with specific lifestyle habits of its residents. In addition, given that many people regularly visit their community pharmacy, these establishments may be the ideal location for recruiting individuals who wish to participate in health-related studies. It is therefore of interest to see if this is the case.

Methods

Results

A total of 261 patients were included in the study. 58.6% were female and mean age was 54.8 (SD: 15.7) years old. Mean BMI was 25.6 (SD: 4.1) and 54.0% were overweight or obese (BMI≥25). Almost half (48.7%) of the study population were found to have poor SQ. As shown in Table 2, the mean PSQI score for good sleepers was 3.5 (SD: 3.9), and for poor sleepers 9.7 (SD: 3.2) (p<0.001). Table 3 shows sample distribution as regards PSQI components. Statistically significant differences between good and poor sleepers were found in all of them. Table 4 shows patient behaviour as regards the lifestyle habits studied.

The multivariate analysis revealed a statistically significant relationship between being a “poor sleeper” and age (OR=1.032; CI95%:1.003-1.061), but the relationship between SQ and sex was not found to be significant. Patients who slept accompanied had a 52.1% lower risk of being a “poor sleeper” than patients who slept alone (OR=0.479; CI95%: 0.255-0.901). As regards comorbidities among patients, those with urinary incontinence were at 20 times higher risk of being a “poor sleeper” than patients not affected by this condition (OR=20.001; CI95%:2.055-194.771), and the risk for patients suffering anxiety was 3.950 times higher than for patients not affected by this condition (OR: 3.950; CI95%:1.585-9.840). Patients with diabetes mellitus had a 3.405 times higher risk than non-diabetic patients (OR: 3.405; CI95%: 1.347-8.607). Arterial hypertension and drinking coffee were not found to be risk factors (OR = 0.535; CI95%: 0.234-1.226 and OR = 0.406; CI95%: 0.149-1.107 respectively); they were however kept in the final multivariate analysis to serve as confounding variables. The other comorbidities among patients showed no statistically significant relationship with SQ. Physical exercise was a protective factor against poor SQ (OR = 0.221; CI95%: 0.081-0.603) and risk factors for poor SQ were being disturbed at night (OR = 3.606; CI95%: 1.892-6.872), going to bed preoccupied (OR = 3.145; CI95%: 1.528-6.471), reading trivial material (OR = 2.475; CI95%: 1.208-5.070), and playing video games (OR = 3.639; CI95%: 1.326-9.984). The Hosmer-Lemeshow value for this model was 4.682 (p=0.791) and AUC was 0.829 (CI95%: 0.780-0.879) (Table 5).

Discussion

One of the questions that this study aimed to clarify was whether community pharmacies may be an appropriate place to make assessments. The short period of time required to recruit the sample size, the large number of pharmacies and their distribution across the region, pharmacist training and patient trust in their pharmacist make community pharmacies an ideal place for conducting epidemiological studies, such as research into SQ, that require a large proportion of the population to be recruited without investing major effort.

This study revealed a high prevalence of patients with poor SQ (48.7%) among the population of Salamanca. Prevalence is higher than that found in other studies carried out in Spain using the same SQ measurement method (PSQI)^16,20. However, these studies were conducted in other areas of the country, and it was not possible to establish the reasons for this difference in prevalence.

There are, unsurprisingly, considerable differences in the SQ dimensions measured using PSQI. Good sleepers perceived their sleep as “fairly good” or “very good”, whilst poor sleepers perceived it as “fairly poor” or “very poor”. Similarly, better values for sleep latency, sleep duration and sleep efficiency were obtained among good sleepers than among poor sleepers. Poor sleepers experienced more sleep disturbances, more daytime dysfunction and took more sleep medication. In this last respect, a comment must be made here: it is logical to think that patients who needed medication were those who had problems sleeping. However, after one month (the period over which SQ was measured using PSQI), patients taking such medication continued to report that they were suffering poor SQ. It may therefore be stated that the therapeutic strategy prescribed for treating sleep issues was ineffective, and that the pharmacist should have suggested that the patient saw their general practitioner for treatment review²¹.

Madrid-Valero et al. found significant differences in SQ in the Spanish population depending on age and gender¹⁶. These results match those found in this study as regards age, but not gender. Gender-related differences may have been eclipsed in the multivariate analysis by the effect of other study variables, because the bivariate analysis (Table 1) shows a clear difference in SQ according to gender.

Caffeine is widely consumed around the world. Yet despite this, evidence of the extent to which it affects sleep is not clear and depends on the amount of caffeine consumed and the tolerance and sensitivity of the person consuming it. Although high doses (several cups of coffee) shortly before going to bed may hinder sleep, there is no evidence in the non-clinical population that low or moderate consumption hinders sleep²². This may account for why, in this study, no statistically significant relationship is found between SQ and consuming caffeine a few hours prior to going to bed. In any case, this variable is shown to be a confounding variable in the presence of AH, and for this reason it was kept in the logistic model.

By contrast, the consumption of high doses of nicotine whether in the form of cigarettes, tablets or patches has been associated with sleep disorders²³. However, the acute nervousness caused by stopping smoking has negative effects on sleep and, in the long term, there is no evidence that stopping smoking is beneficial for improving SQ²². Our study found no correlation between SQ and tobacco consumption. This could be explained by the low number of smokers in the study sample. Given that there are increasingly fewer smokers in Spain²⁴, this may become a fairly insignificant risk factor in the near future.

The consumption of alcohol has also been shown to cause sleep disturbances, and recommendations range from avoiding high consumption of alcohol to not drinking at all before going to bed¹⁰. In addition, subsequent studies found that in people who are not alcohol-dependent, drinking alcohol (even a small amount) before going to bed may affect their sleep that night, although this will also depend on individual tolerance²². The majority of the study population either do not drink alcohol at all or only very occasionally, and therefore although they might experience poor sleep on a given night, this does not appear to be sufficient reason for them to describe their SQ as poor over the preceding month. This would account for the fact that the questions relating to alcohol consumption in this study show no relationship with SQ.

Another lifestyle habit highly recommended for improving SQ is physical exercise, although it should not be taken too near bedtime¹⁰. However, more recent pieces of research agree that physical exercise improves SQ²⁰, even when taken shortly before going to bed²². It has also been shown to improve SQ in insomnia patients²⁵. In our initial hypothesis, we posited that exercise was a predictive factor for SQ, and in this respect the results obtained matched the data expected.

Naps have been widely studied in scientific literature as having a possible effect on sleep, although there does not seem to be any clear link between daytime napping and broken nights²². The data obtained in this study support this observation. It must nevertheless be borne in mind that the question was framed in terms of naps lasting longer than one hour, so this result must be interpreted with caution.

Another of the most widely-made recommendations for achieving good SQ was to reduce stress levels before going to sleep, which is why individuals are encouraged not to do any activities that make them nervous or worried, but rather some kind of relaxing activity before going to bed¹⁰. Stress has indeed been found to cause problems getting to sleep and staying asleep²⁶. This is why study participants were asked about activities such as reading things that they found important or not, watching TV, listening to the radio, etc. All such activities may favour or hinder proper sleep. Listening to relaxing music does not have the same effect as watching a football match. In any case, most of these activities were found not to have any relationship with SQ in this study, not even those relating to use of new technologies. This may be due to the highly general nature of the questions, which allowed subjects to use their own judgement when interpreting what they considered important or stressful. On the other hand, very few participants stated that they did these types of activities, which made it difficult to determine whether their effects on SQ were significant. Only playing video games and reading material that the participant considered trivial were found to be risk factors for poor SQ.

In addition to the foregoing, new technologies must be considered today. Their negative effect on SQ has not been clearly demonstrated²⁷. It must be emphasised that the majority of research looking at the use of technology and its relationship with sleep only analyses how often it is used before going to sleep²⁷, not why it is used²⁸. This study investigated the frequency and intensity of technology use-related activities but even so, in our model only playing video games was found to have a significant bearing on SQ. Playing video games before going to sleep appear to cause a state of excitation not conducive to sleep; this might not be caused by other activities such as reading emails or news on social networks. However, the study by Carter et al²⁹ showed that the use of technology among children and young people did have a considerable bearing. This fact was perhaps not observed in our study given the different age of the study population.

Several limitations must be considered for our study results to be correctly interpreted. Firstly, it must be borne in mind that given the type of study design used (cross-sectional), no causal relationship can be established between the variables analysed and poor SQ. However, the results do provide an overview of the current situation with regard to SQ in the population of Salamanca. It would be of interest to conduct prospective studies that enable causal relationships between lifestyle habits and poor SQ to be identified. Secondly, conducting this study in community pharmacies entails a potential selection bias, since people who visit such establishments usually have health issue; only in a few cases was this not so. Caution must therefore be exercised when extrapolating results.

In future research, it may be of interest to conduct longitudinal studies that enable causal relationships to be identified between specific lifestyle habits and poor SQ. Moreover, studying each lifestyle habit separately as regards time of day, frequency and intensity and comparing between groups may well clear up many of the existing doubts surrounding the lifestyle habits mentioned and poor SQ. This study shows that SQ can be assessed in community pharmacies. However, further research is needed to explore whether pharmaceutical interventions can improve SQ in the population.

Conclusion

Studying SQ was considered important because it is a factor that directly affects quality of life among the population. In the Spanish town of Salamanca, almost half of the study population stated that they suffered poor SQ. Although numerous lifestyle habits that may relate to poor SQ were studied, a statistically significant relationship was only found for a few of them. It was shown that physical exercise is a factor that protects against poor SQ, and being disturbed at night, going to bed preoccupied, reading trivial material and playing video games were risk factors for poor SQ. Other risk factors identified were Type II diabetes, anxiety and urinary incontinence. This notwithstanding, other studies with more tailored designs are required to demonstrate causality between the predictive variables analysed and SQ. In addition, community pharmacies appear to be a suitable place for conducting these types of study, given that they are found all over Spain and the people visiting them are willing to participate in activities suggested by their pharmacist.