Sleep is a basic necessity that modulates the immune system, regulates homeostasis, and improves some cognitive functions; moreover, it contributes to the adjustment of physiological functions through hormonal secretion and anabolic stimulation.1,2

At intensive care units (ICU), due to their characteristics and the high severity of their patients’ conditions, there are more sleep alterations, and patients suffer from both qualitative and quantitative deprivations.3 Studies conducted on the incidence of sleep alteration in the critically ill patient report a prevalence of 22% to 61%. Sleep pattern alterations are observed in the critically ill patient and they consist in a predominance of the stage N1 – light sleep and stage N2 – muscle tone reduction of the sleep cycle with a decrease or absence of stages N3-deep, restorative sleep, and REM sleep. The patients often experience awakenings, a high index of arousal, and periods of daytime sleep (40%–50% of the sleep can occur during the day)4–6; patients rarely complete a full sleep cycle.7,8

Assessing the sleep of patients admitted at the ICU is a complex process,9–11 and this situation requires tools to evaluate it objectively; but these tools like polysomnography and actigraphy are not available in all ICUs. They also require trained researchers for their interpretation. That is why, in order to assess sleep and the factors that interrupt it, there is the alternative of assessment through subjective methods based on questionnaires. This process is less expensive. In contrast, questionnaires, even when they have been validated through comparison with objective evaluation methods, they have not been submitted for psychometric studies.12–14

These psychometric studies assess the characteristics that every measuring tool should have, which necessarily includes controlling its precision by studying its reliability and validity. Validity refers to the degree to which the tool measures what it intends to measure; reliability refers to the trustworthiness conferred to the data obtained from it and it is associated with the internal coherence or consistency and precision of the measures collected.15 The objective of the present study is to assess whether the modified Freedman questionnaire is suitable, due to its psychometric characteristics, for sleep assessment in critically ill patients.

A sample of n=129 patients was analyzed. The mean age was 60.61±13.88 years, range (25-85); 62% were males. The mean stay of the period were 10.31±14.41 days, range (3–117). Regarding the reason for ICU admission, 46.5% was for medical reasons, 23.3% for postoperative reasons, 18.6% due to coronary disease, and 11.6% for traumatic reasons.

The results of exploring the validity of the content of the questionnaire from the answers given by members of the expert group to the questions of variable pertinence and its definition, in most of the questions are values >3, indicative of a correct or very correct definition, except for the correct definition in nursing activities, drug administration, and pulse oximeter (Table 1).

Before conducting the factor analysis, sampling adequacy was explored. The results of the KMO sampling adequacy test were 0.751 (result close to the value of 1 considered as an excellent value) and those of Bartlett's test were a significant value (P<.001), which allowed conducting the factor analysis. Subsequently, we assessed the validity of the construct through factor analysis using the varimax rotation method (results are shown on Table 2). The table shows the correlation between the variables analyzed and the 4 factors extracted (sleep quality, sleepiness, sleep interruption due to environmental causes, and sleep interruption caused by the human factor).

The internal consistency of the questionnaire was assessed using Cronbach's α coefficient that indicated an overall P value of .933. Instrument stability was assessed using an ICC test–retest; the coefficients obtained an excellent correlation in all the items analyzed. Table 3 shows the results associated with the items that study the quality of sleep, daytime sleepiness, environmental activities, and factors that interfere with the 3 phases of sleep.

The Spanish Society of Intensive and Critical Medicine and Coronary Units (SEMICYUC) proposes a series of recommendations to improve safety and quality of clinical practice in critically ill patients. We should mention here that, regarding the issue at hand, the Sedation, Analgesia, and Delirium Working Group at SEMICYUC gives great importance to facilitating sleep and controlling environmental stimuli during the night. Sleep control and evaluation, together with other recommendations, are supposed to be preventive measure to adjust, on the one hand, the administration of sedative drugs and, on the other hand, to minimize states of delirium providing the necessary rest for the critically ill patient.23,24

This fact makes us consider the need to evaluate the quality of sleep as perceived by the patients during their stay at the ICU and figure out what tool would be the most suitable one to provide us with the necessary information on both the patient and the environment. We started on the thought that sending a questionnaire is not a easy process because it requires the patient's collaboration and correct cognitive function, and the systematized performance of interviewers25,26; that is why the first step this study took was to train the interviewers in how to implement the scale.

The initial study conducted by Freedman et al.17 did not assess the psychometric characteristics of the questionnaire, but used factor analysis to study the correlation between each item and the 4 factors set out by the authors (factor 1, sleep interruption secondary to interruptions by the staff and diagnostic tests; factor 2, quality of sleep; factor 3, daytime sleepiness; and factor 4, sleep interruption by environmental factors [light and noise]). They found that the variables that showed the greatest correlation were the ones associated with noise, light, and interventions performed by researchers. The results of our study show a similar correlation between the variables analyzed and the dimensions that correspond with the factors described by Freedman.

Regarding content validity, the members of the expert group assessed practically all the items in the questionnaire with scores ≥3, both to the question that referred to the formulation of the item and to the one that referred to its pertinence. It should be pointed out that the items “nursing activities”, “administration of drugs” and “pulse oximeter” were assessed more poorly since they did not reach the value of 3 regarding their formulation. This leads us to consider that the variables “nursing activities” and “administration of drugs” could be included in the variable “nursing care”. A variable that was better evaluated and includes all the nursing competences. The variable “pulse oximeter” is in a similar situation, that could be considered as “alarms”.

The evaluation of internal consistency conducted by Cronbach's alpha coefficient obtained a global result of 0.933. Based on the study conducted by Streiner,27 who claims that the value to accept a correct internal consistency is a Cronbach's alpha coefficient of ≥0.7, it can be said that the modified Freedman questionnaire shows results that guarantee their homogeneity and reliability.

Regarding the evaluation of the level of concordance, the ICC results showed an excellent level of concordance; all the items showed ICC values >0.75 with statistical significance (P<.001). It is true that in this study intra-observer concordance was not evaluated; the time scheduled between the assessment of observer A and that of observer B did not make the intra-observer evaluation process easier.

The results of this study indicate that the Freedman questionnaire modified by Gómez Sanz shows good psychometric characteristics, which leads us to think that it can turn out to be a reliable tool to evaluate the quality of sleep in the ICU patient with a RASS between −1 and +1, as well as the environmental and human factors that can interfere with sleep.

This study has received no public or private funding whatsoever.

None reported.