Keywords:
Sleep - Pain - Aged - Surveys and Questionnaires - Health Impact Assessment
Palavras-chave:
Sono - Dor - Idoso - Inquéritos e Questionários - Avaliação do Impacto na Saúde
INTRODUCTION
Sleep disorders and chronic pain are two major health issues with high incidence in
older adults[1],[2],[3]. Approximately 40% of older adults suffer from sleep disturbances causing a reduction
in functional status and quality of life, besides being responsible for an increase
of clinical and psychiatric comorbidities in this population[4]. Sleep disturbances (or sleep-wake disorders) involve problems with the quality,
timing, and amount of sleep, which result in diurnal impairment and distress, and
impairment in functioning. Sleep disturbances include sleep-related breathing disorders
such as central, obstructive or mixed sleep apnea, insomnia, parasomnias, narcolepsy,
REM sleep behavior disorder, excessive sleepiness, circadian rhythm disorders, and
restless leg syndrome, amongst more than 80 sleep disturbances. Similarly, chronic
pain is a prevalent and extremely debilitating condition, affecting more than 50%
of community-dwelling older individuals and more than 80% of nursing home residents[2],[5],[6]. Chronic pain is also a risk factor for premature death and accelerated cognitive
decline[7].
There is a complex bidirectional relationship between chronic pain and sleep disorders,
which has important implications for clinical management of older patients[8],[9],[10]. It is estimated that 40 to 88% of the geriatric population has coexistences of
both conditions with various etiologies; for instance, migraine, musculoskeletal pain
(e.g. chronic low back pain, arthralgia), fibromyalgia, osteoarthritis, and irritable
bowel syndrome[11],[12],[13]. Moreover, both conditions can be predictors of frailty in older adults[14].
Chronic pain affects sleep in different ways, including increased insomnia, more fragmented
sleep, and shorter sleep duration[15],[16]. A recently published meta-analysis estimated a pooled prevalence of 44% of sleep
disorders in patients with chronic nonmalignant pain, the most common being insomnia
(72%), restless legs syndrome (32%), and obstructive sleep apnea (32%)[17]. Objective polysomnographic measures in individuals with chronic pain have demonstrated
increased sleep onset latency, time awake after sleep onset, number of awakenings,
light sleep, number of stage-shifts, respiratory-related events and periodic limb-movements,
and diminished sleep efficiency and sleep duration[17].
Despite the growing evidence of the sleep-pain association, sleep assessments are
not routinely made in older patients with chronic pain, a measure that could help
to ensure more effective treatment of pain conditions. The assessment of sleep should
involve both subjective and objective measures, but this can be complex and is not
always feasible. Concerning subjective sleep, there are some assessment tools which
have been shown to have reliability and validity such as the Pittsburgh Sleep Quality
Index (PSQI)[18], the Epworth Sleepiness Scale (ESS)[19], the Functional Outcomes of Sleep Questionnaire (FOSQ)[20], the Insomnia Severity Index (ISI)[21], the Insomnia Symptom Questionnaire (ISQ)[22], the Insomnia Impact Scale (IIS)[23], the Jenkins Sleep Evaluation Questionnaire (JSEQ)[24], and the Sleep Scale from the Medical Outcomes Study (MOS)[25]. However, none of these were specifically developed to assess “sleep and pain”,
and neither are they specific for older adults.
To assess sleep disturbances associated with chronic pain there are very few available
tools. Only two of them have been widely used: i) the Pain and Sleep Questionnaire
Three-item Index (PSQ-3), a direct measure of the impact of chronic pain on sleep[26], and ii) the Chronic Pain Sleep Inventory (CPSI), a 5 items tool using a 100 mm
visual analogue scale[27]. However, there is no guidance on which of these existing self-reported sleep-pain
measures best capture the aspects of sleep disturbance that are most pertinent to
older adults with chronic pain. Moreover, the two assessment tools were designed for
the general population and not specifically for the older population. In respect of
geriatric patients, the reasons for different types of pain and sleep disturbances
are related to changes specifically associated with the aging process. It is important
to have a tool that can be easily applied and provide the most comprehensive and detailed
assessment of sleep-pain possible.
There is, therefore, a lack of validated measures of sleep disturbances in chronic
pain conditions designed specifically for older adults. The creation of such a measure
could help in the management of pain in aging, and optimize quality of life and functioning.
METHODS
The development and validation of the instrument to assess the impact of pain on sleep
in older adults involved three steps, following a model previously described by Kline[28]: (1) the establishment of the theoretical background/literature review, (2) a qualitative
study involving the development of the tool items and the establishment of face validity,
and (3) a quantitative study that involved testing the instrument with older adults
with chronic pain as part of the validation process.
Ethical aspects
All procedures performed in studies involving human participants were in accordance
with the ethical standards of the institutional and/or national research committee.
The study protocol was approved by the Institutional Research Ethics Committee of
the Universidade Federal de São Paulo (UNIFESP, CEP # 072353-2014). This article does
not contain any studies with animals performed by any of the authors.
Theoretical evidence and elaboration of the tool items
Using the process suggested by Kline (1995) for the development of an instrument[28], we followed three steps as outlined. i) We gathered theoretical evidence through
a careful thematic bibliographic review, which served to clarify the nature and range
of the content related to the target construct, namely, an instrument to assess pain
and sleep disorders in older adults. ii) We elaborated the tool items, taking into
consideration the terms and language known to the geriatric population. In this stage,
we analyzed some reliable and validated questionnaires related to sleep assessment,
some involving assessment of both pain and sleep (PSQI, ESS, MOS, PSQ-3, and CPSI).
The tool included seven items, demonstrating “content validity” (proved to be representative
of the universe of related content). iii) The tool was then analyzed by a committee
of five experts in neurology, psychiatry, gerontology, family medicine, and sleep
medicine related to the subject. A minimal consensus by 80% of the experts was the
criterion required to retain an item (face validity). The committee of experts stated
that the older population with chronic pain conditions and sleep disturbances is clinically
heterogeneous and complex, thus making the management of this entangled relationship
(sleep-pain) challenging; the two conditions must be dissociated in clinical practice.
Also, medications prescribed to treat pain should not affect sleep, and vice-versa,
as sleep disturbances may have an adverse effect on the course of chronic pain conditions
in older adults.
This process resulted in the production of the “Sleep Assessment Instrument for Older
Adults with Pain” (SAIOAP) ([Figure 1]) — a tool consisting of seven items with “yes” or “no” answers, grouped according
to the broad dimensions of sleep: initiating sleep (item 1), maintaining sleep (items
2 and 3), physical discomfort such as tiredness, exhaustion and fatigue (item 4),
self-perception of sleep (item 5), daytime sleepiness (item 6), and medications used
to sleep (item 7). In addition, four sub-items related to sleep latency, duration,
and efficiency (in items 1 and 2) were also established for a qualitative analysis.
Figure 1 Sleep Assessment Instrument for Older Person with Pain (SAIOAP).
To evaluate the new tool, we carried out a pilot test in a sample of 15 older individuals
with chronic pain. The questionnaires were completed by an investigator who read the
questions and wrote down the patients’ answers. None of the test subjects found any
difficulties in answering the questions so no adjustments to the tool were necessary.
Sample size calculation
The sample size was calculated for 95% power, an effect size of 0.602 with Cohens’
d of 0.50, and an α error of 0.05. For the subjective parameters (PSQI, GPM, GEAP,
ADLs, and IADLs as outcomes) the required sample size was 44 participants. The sample
size calculation was made using G*Power software (version 3.1.9.2, Franz Faul, Utah,
USA, www.ncss.com.).
Validation of Sleep Assessment Instrument for Older Person with Pain
The validation process was also carried out according to the method described by Kline
and collaborators[28]: a) reliability analysis (internal consistency); b) validation; c) standardization
of the application (implementing) process, assessment, and interpretation of the tool.
To complete the validation, a sample of 100 individuals was randomly selected from
patients undergoing clinical follow up at the Geriatrics and Gerontology Department
of the Universidade Federal de São Paulo. These patients complied with the inclusion
criteria of being aged 60 years or older and presenting chronic pain (≥6 months).
Individuals with cancer pain and cognitive impairment were excluded. All participants
signed an informed consent form.
A semi-structured questionnaire was used to collect information on socio-demographic
data (age, gender, race, marital status and education), self-perception of health,
comorbidities, and regularly used medications, including those prescribed for sleep
disorder (e.g., benzodiazepines and “z drugs”, such as Zolpidem). Based on the medication
information, we classified the participants according to the number of drugs regularly
used, and defined polypharmacy as the use of 5 to 9 different drugs, and excessive
polypharmacy as the use of ≥10 different drugs.
Functional capacity was assessed using the Katz activities of daily living (ADLs)
scale[28] to evaluate the abilities to undertake activities such as feeding, bathing, dressing,
and leisure. The Lawton instrumental activities of daily living (IADLs) scale[29] was used to evaluate activities that are not necessary for fundamental functioning,
but important for independent living, such as cleaning and maintaining the house and
managing money. In addition, information was collected on self-perception of health
status and categorized as excellent, good, regular, and poor.
The multidimensional aspects of pain were evaluated, including duration, frequency,
intensity, localization, affectivity, functional impact, and others. The verbal numerical
scale (VNS) of pain was applied to assess the intensity of pain[30] and the geriatric pain measure (GPM) was applied to assess the multidimensional
aspects through domains of pain (sensory-discriminative, motivational-affective, and
cognitive evaluative)[31]. According to the physiopathogenesis, pain was classified as nociceptive, neuropathic,
or mixed, and according to its localization, as muscle, joint, nerve, or other pain.
To assess affectivity, an important aspect related to pain, we used the Geriatric
Emotional Assessment of Pain (GEAP), which indicates little or no pain-induced depression
(scores 0 to 5), moderate pain-induced depression (scores 6 to 9), and severe pain-induced
depression (scores ≥10)[32],[33]. For the subjective assessment of sleep, we used the Brazil Portuguese version of
the PSQI, which assesses sleep quality with a sensitivity of approximately 80% and
specificity of approximately 68.8%[18].
Lastly, the participants were asked to answer the items of SAIOAP; we established
a standard method of application to ensure uniformity in the use of the instrument.
Statistical data analysis
Data were analyzed by means of descriptive and parametric statistics using Statistical
Package for the Social Sciences version 17 (SPSS Inc. Released 2008. SPSS Statistics
for Windows, Version 17.0. Chicago: SPSS Inc.), Minitab 16 (Minitab 16 Statistical
Software 2010. Computer software. State College, PA: Minitab, Inc. (www.minitab.com)),
and Microsoft Excel (Microsoft Corporation, 2010). To evaluate the internal consistency
of the instrument, Cronbach’s alpha and total-item correlation were calculated. The
associations between SAIOAP and PSQI, GPM, GEAP, ENV, comorbidity, and medication
were analyzed using Pearson’s correlation coefficient, and those associations related
to quantitative variables, such as age range, functionality, pain characterization,
self-perception of health, comorbidity, medication, GEAP, and PSQI were analyzed using
the analysis of variance test (ANOVA). The Receiver Operating Characteristic (ROC)
methodology was used to assess the ability of SAIOAP to predict pain in older adults
with sleep disturbance. Significance was defined as p<0.05.
RESULTS
The development process described above resulted in the production of a final instrument
that could be easily understood by the participants. It was very easy to apply, taking
approximately five minutes.
A sample of 100 older adults with an average age of 83.1 years old and that was predominantly
female (86.9%) was used in the validation process. The data from baseline measurements
are shown in [Table 1].
Table 1
Summary of the main characteristics of the participants (n=100).
|
Qualitative variables
|
|
n
|
%
|
p-value
|
|
Age (years)
|
60–69
|
3
|
3.0
|
<0.001
|
|
70–79
|
24
|
24.2
|
<0.001
|
|
≥80
|
72
|
72.2
|
Ref.
|
|
Gender
|
Female
|
86
|
86.9
|
<0.001
|
|
Male
|
13
|
13.1
|
Ref
|
|
Race
|
Black
|
13
|
13.1
|
<0.001
|
|
White
|
59
|
59.6
|
Ref.
|
|
Brown
|
26
|
26.3
|
<0.001
|
|
Asian
|
1
|
1.0
|
<0.001
|
|
Marital status
|
Married
|
30
|
30.3
|
<0.001
|
|
Separated/divorced
|
7
|
7.1
|
<0.001
|
|
Widowed
|
54
|
54.5
|
Ref.
|
|
Single
|
8
|
8.1
|
<0.001
|
|
ADLs
|
1–2
|
1
|
1.0
|
<0.001
|
|
3–4
|
4
|
4.0
|
<0.001
|
|
5–6
|
94
|
94.9
|
Ref.
|
|
IADLs
|
09–15
|
6
|
6.1
|
<0.001
|
|
16–20
|
14
|
14.1
|
<0.001
|
|
21–25
|
30
|
30.3
|
0.006
|
|
26–27
|
49
|
49.5
|
Ref.
|
|
Self-perception of health
|
Excellent
|
5
|
5.1
|
<0.001
|
|
Good
|
26
|
26.3
|
<0.001
|
|
Regular
|
61
|
27.3
|
Ref.
|
|
Poor
|
7
|
7.1
|
<0.001
|
|
Polypharmacy
|
No polypharmacy
|
20
|
20.2
|
<0.001
|
|
Polypharmacy
|
52
|
52.5
|
Ref.
|
|
Severe polypharmacy
|
27
|
27.3
|
<0.001
|
|
Pain type
|
Nociceptive
|
77
|
77.8
|
Ref.
|
|
Neuropathic
|
7
|
7.1
|
<0.001
|
|
Mixed
|
15
|
15.2
|
<0.001
|
|
Pain frequency
|
Continuous
|
46
|
46.5
|
Ref.
|
|
Intermittent
|
45
|
45.5
|
0.887
|
|
Occasional
|
8
|
8.1
|
<0.001
|
|
Pain localization
|
Muscle
|
26
|
26.3
|
<0.001
|
|
Joint
|
82
|
82.8
|
Ref.
|
|
Nerve
|
20
|
20.2
|
<0.001
|
|
Others
|
2
|
2
|
<0.001
|
|
Pain intensity (VNS)
|
Mild
|
4
|
4.0
|
<0.001
|
|
Moderate
|
36
|
36.4
|
<0.001
|
|
Severe
|
59
|
59.6
|
Ref.
|
|
Pain by GPM
|
Mild
|
7
|
7.1
|
<0.001
|
|
Moderate
|
60
|
60.6
|
Ref
|
|
Severe
|
32
|
32.3
|
<0.001
|
|
GEAP
|
No or mild depression
|
43
|
43.4
|
Ref
|
|
Moderate depression
|
37
|
37.4
|
0.385
|
|
Severe depression
|
19
|
19.2
|
<0.001
|
|
PSQI
|
Good sleeper
|
43
|
43.4
|
0.065
|
|
Poor sleeper
|
56
|
56.6
|
Ref.
|
Table 1
Continuation.
|
Quantitative variables
|
Mean
|
SD
|
Range
|
|
Age (years)
|
83.13
|
±7.21
|
64–98
|
|
Education (years)
|
3.84
|
±3.14
|
0–15
|
|
Comorbidity (number)
|
5.71
|
±2.29
|
2–11
|
|
Medication (number)
|
7.93
|
±3.04
|
2–14
|
|
Pain duration (months)
|
9.15
|
±9.74
|
0.25–50
|
|
Pain intensity (VNS)
|
7.10
|
±1.95
|
3–10
|
|
Pain by GPM
|
61.40
|
±19.41
|
4.7–99.9
|
|
GEAP
|
6.86
|
±4.59
|
0–21
|
|
PSQI
|
6.54
|
±3.93
|
1–18
|
|
SAIOAP
|
1.96
|
±1.55
|
1–7
|
ADLs: activities of daily living; IADLs: instrumental everyday activities; SD: standard
deviation; VNS: verbal numeric scale; GPM: geriatric pain measure; GEAP: Geriatric
Emotional Assessment of Pain; PSQI: Pittsburgh Sleep Questionnaire Index; SAIOAP:
Sleep Assessment Instrument for Older Person with Pain.
We analyzed the internal consistency of SAIOAP to test its reliability. The Cronbach’s
alpha for the whole instrument was 0.602, indicating a moderate level of reliability
([Table 2]). The total-item correlation showed that all values had a score higher than 0.4,
indicating good homogeneity between the items of the SAIOAP (suggesting that it was
not necessary to delete any items) ([Table 2]).
Table 2
Internal consistency of Sleep Assessment Instrument for Older Person with Pain (SAIOAP)
measure with Cronbach alpha and corrected total-item correlation.
|
SAIOAP
|
Corrected total-item correlation
|
|
Item 1
|
0.520
|
|
Item 2
|
0.533
|
|
Item 3
|
0.589
|
|
Item 4
|
0.580
|
|
Item 5
|
0.496
|
|
Item 6
|
0.613
|
|
Item 7
|
0.598
|
|
Cronbach alpha (Total)
|
0.602
|
SAIOAP: Sleep Assessment Instrument for Older Person with Pain.
We observed statistically significant correlations between the SAIOAP and sleep quality,
pain (both in respect of intensity and multidimensional aspects), independence in
respect of ADLs, comorbidity, medication consumption, and depression ([Table 3]). The strongest correlation observed was with sleep quality (r=68.3%; p<0.001),
indicating convergent construct validity of the SAIOAP.
Table 3
Correlation between Sleep Assessment Instrument for Older Person with Pain (SAIOAP)
and principal variables by Pearson’s correlation.
|
Correlation (r)
|
p-value
|
|
Sleep quality (PSQI)
|
68.3%
|
<0.001
|
|
Pain intensity (VNS)
|
28.5%
|
0.004
|
|
Multidimensional aspects of pain (GPM)
|
40.5%
|
<0.001
|
|
Depression (GEAP)
|
46.1%
|
<0.001
|
|
Comorbidity (number)
|
28.2%
|
0.005
|
|
Medication (number)
|
32.6%
|
0.001
|
PSQI: Pittsburgh Sleep Quality Index; VNS: verbal numeric scale; GPM: geriatric pain
measure; GEAP: geriatric emotional of pain.
The ROC curve analysis with the PSQI demonstrated that the SAIOAP had sensitivity
of 73.2% and specificity of 79.1% (with an area under the curve of 0.798, p=0.046).
The best cutoff point found for the SAIOAP was 1.5, therefore, two or more scores
from the SAIOAP could produce a total score that predicted sleep disturbances in older
adults with pain ([Figure 2]).
Figure 2 ROC curve for the performance of the Sleep Assessment Instrument for Older Person
with Pain (SAIOAP) according to the Pittsburgh Sleep Questionnaire Index.
DISCUSSION
In the current study we developed and validated an assessment tool designed specifically
for older adults with chronic pain. This is the first instrument designed to measure
the pain-sleep relationship in this population. The SAIOAP is quick and easy to apply,
and considers important sleep dimensions, namely, sleep onset and maintenance, physical
discomfort, diurnal repercussions of sleep such as excessive daytime sleepiness, self-perception
of health status, and medication used for sleep. The SAIOAP was applied in a sample
of older adults whose pain intensity was mainly moderate or severe, and the average
duration of experiencing pain was about nine years. In addition, the participants
frequently presented pain-induced depression (more than 50% of the sample), which
suggests that pain had an important affective impact in our sample. The presence of
an intrinsic relationship between pain and depression is significant, and both conditions
should be simultaneously treated[17]. This is in line with the suggestion by some authors that sleep problems need to
be considered in a broader context[25].
We observed adequate psychometric proprieties for the SAIOAP. Foremost, content and
face validity were demonstrated, respectively, by including universally representative
content and by experts assessing if the content looked appropriate “on the face of
it” (face validity). The other psychometric properties of the SAIOAP were shown to
be good. The Cronbach’s alpha did not reach 0.7, which is generally considered the
cut off for good internal consistency, and this may be related to the small number
of items in the instrument. All items in the scale had scores above 0.4 which indicated
very good discriminant validity[34]. Nevertheless, the total-item correlation indicated that the assessment tool had
uniform items (values all above 0.4). That is, all 7 items had moderate correlation
with SAIOAP total scores[35]. In the validation process, a statistically significant correlation was verified
between the SAIOAP and several factors, particularly sleep quality (PSQI). There was
also a statistically significant and positive association of SAIOAP with the uni-
and multidimensional aspects of pain, pain-induced depression, comorbidities, number
of medications used, and self-perception of health. The correlation with sleep quality
determined the construct validity of the SAIOAP.
Our finding that sleep quality in our sample was considered poor is in accordance
with the literature. Older adults tend to present poor sleep quality, with studies
in the literature showing that rates of sleep disorders elderlies with pain range
from 40 to 88%[36],[37]. However, poor quality of sleep is not found in all older adults, and it should
not be assumed to be “a normal finding”. There are many highly functional older adults
that have a restorative sleep, even though it is often of objectively poorer quality
compared with younger adults[38],[39]. In addition, in older adults there may be age-related sleep changes such as changes
in circadian rhythm (phase advance) and sleep architecture changes (e.g., shortened
sleep duration, frequent awakenings, and decreased slow wave sleep). Furthermore,
an increased frequency of daytime naps and time spent awake during the night could
also negatively influence sleep. A lack of physical activity and sedentary behavior
could also predispose individuals to more pain, increased pain perception, and reduced
sleep consolidation. Contrary, pain might also predispose individuals to less activity
and sedentary behavior. Therefore, although sleep disorders have multifactorial causes,
sleep disorders in older adults are often attributed to pain[3]. Again, the homeostatic mechanisms of sleep become less robust with normal aging
and added to changes in hormone secretion profiles can negatively impact the sleep-pain
relationship[38],[39].
Sleep disturbance is a crucial component in the conceptual model for sleep-related
problems in patients with pain. The SAIOAP was shown to have both sensitivity and
specificity in relation to predicting pain in older adults with sleep disturbance,
thus this tool can be used in clinical practice to screen, diagnose, and monitor the
pain-sleep relationship in this specific clinical population and evaluate the effectiveness
of treatments.
There are some limitations to be considered. Sleep analysis could be better determined
using objective polysomnographic or actigraphy examination, but this was not possible
due to the high cost and was not therefore included in the analysis. In addition,
other factors such as reproducibility and predictive validity were not fully explored
and will require further investigation.
The results of the present study show that the SAIOAP should be used as a direct measure
of the impact of sleep in older patients with pain. This is an easy-to-apply, detailed
assessment tool for the pain-sleep relationship in the older population, and the study
has shown the tool to be a reliable and valid for the screening of sleep disturbances
in older adults with pain. This tool could be used to identify older adults with sleep
disturbances and pain, facilitating prompt targeted treatment in relation to both
conditions.
The SAIOAP is an easy and short tool for researchers and health professionals to directly
evaluate the impact of co-occurring pain and sleep disturbance on older adults, something
that is not possible with the few currently validated measures.
