Skip to main content
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2025 Aug 1.
Published in final edited form as: J Am Geriatr Soc. 2024 Jun 18;72(8):2319–2328. doi: 10.1111/jgs.18910

The effectiveness of cognitive behavioral therapy for insomnia on sleep outcomes in the context of pain among older adult veterans

Alexander J Erickson 1, Juan Carlos Rodriguez 2, Scott G Ravyts 3, Joseph M Dzierzewski 4, Constance H Fung 1,5, Monica R Kelly 1,5, Armand M Ryden 5,6, Gwendolyn C Carlson 7,8,9, Karen Josephson 1, Michael N Mitchell 1, Jennifer L Martin 1,5, Cathy A Alessi 1,5
PMCID: PMC11887479  NIHMSID: NIHMS2055113  PMID: 38888493

Abstract

Background:

Cognitive behavioral therapy for insomnia (CBT-I) is the gold-standard treatment for insomnia disorder in adults. Compared to young adults, older adults have increased risk for the development of conditions associated with chronic pain, which may impact the efficacy of CBT-I in improving insomnia symptoms in older adults. This study evaluated the effect of participant-rated pain on sleep-related outcomes of a supervised, non-clinician administered CBT-I program in older adult patients with chronic insomnia disorder.

Methods:

Secondary analysis was conducted using data from a randomized controlled trial among 106 community-dwelling older adult veterans (N = 106; mean age 72.1 years, 96% male, 78.3% White, 6.6% Hispanic, 5.7% African American) with chronic (≥3 months) insomnia disorder. Participants engaged in five sessions of manual-based CBT-I in individual or group format within one Department of Veterans Affairs healthcare system, provided by non-clinician “sleep coaches” who had weekly telephone supervision by behavioral sleep medicine specialists. Insomnia symptoms (Insomnia Severity Index), perceived sleep quality (Pittsburgh Sleep Quality Index), fatigue (Flinder’s Fatigue Scale), daytime sleepiness (Epworth Sleepiness Scale), and perceived pain severity (items from the Geriatric Pain Measure) were assessed at 4 time points: baseline, one-week posttreatment, 6-month follow-up, and 12-month follow-up. Mixed effects models with time invariant and time varying predictors were employed for analyses.

Results:

CBT-I improved insomnia symptoms, perceived sleep quality, fatigue, and daytime sleepiness among older veterans with chronic insomnia. Participant-reported pain was associated with greater improvements in insomnia symptoms following CBT-I. Pain did not affect improvements in other sleep-related outcomes (−0.38 ≤ b ≤ 0.07, p > 0.05). Between-subjects differences in pain, but not within-subject changes in pain over time, appeared to play a central role in insomnia symptom improvement at posttreatment, with individuals with higher-than-average pain showing greater insomnia symptom improvement (ISI score reduction; −0.32 ≤ b ≤ −0.28, p ≤ 0.005).

Conclusions:

Pain did not meaningfully hinder the effects of CBT-I on sleep outcomes. Among older veterans with chronic insomnia disorder, individuals with higher pain exhibited slightly greater improvement in insomnia than those with lower levels of pain. These findings suggest that experiencing pain does not impair treatment response and should not preclude older adults with insomnia from being offered CBT-I.

Keywords: cognitive behavioral therapy for insomnia, insomnia, older adults, pain, veterans

INTRODUCTION

Sleep is a central function involved in many aspects of optimal biological and psychological aging processes. The benefits of good quality sleep establish the importance of supporting sleep wherever possible, particularly through interventions for those with sleep disorders. Past research found prevalence rates of insomnia in older adults ranging from 30% to 48%, with older adults being at higher risk for insomnia than younger adults.1 Insomnia often occurs with co-morbid medical conditions that cause pain, for example some studies report that one in three older adults present with co-morbid insomnia and arthritic pain.2 The co-occurrence of poor sleep and pain has been explored in past studies, though evidence regarding the potential impact of pain on the effectiveness of behavioral treatment of insomnia is limited.

Though pain and pain-related conditions are not inherent in normal aging processes, 30% to 37% of older adults experience some form of chronic pain.3 Pain and sleep tend to share a bidirectional relationship, in that sleep quality and quantity are negatively impacted by pain and poor sleep exacerbates pain sensitivity, intensity, and other characteristics of pain.4 Given this overlap, treatment for insomnia disorder presents as an opportunity to optimize improvements in both insomnia disorder and pain.

Cognitive behavioral therapy for insomnia (CBT-I) stands as the gold-standard treatment for chronic insomnia disorder, with scientific literature highlighting significant improvements in insomnia symptoms that are longer lasting than pharmacological interventions.5 Based on a recent meta-analysis, CBT-I improves sleep efficiency, sleep onset latency, wake after sleep onset, total sleep time, and other sleep-related outcomes in older adults.6 Based on multiple meta-analyses, CBT-I has been shown to be effective in improving insomnia for individuals with comorbid insomnia disorder and pain, particularly in young to middle aged adults.7,8 Studies have also illustrated some evidence that CBT-I may improve insomnia disorder in older adults with comorbid pain conditions like osteoarthritis.9,10 However, most studies in this area describe pain in a categorical (e.g., presence of pain condition) or time-invariant way, whereas pain is often variable over time and in its severity. Few longitudinal studies examining CBT-I effectiveness in the context of pain have employed models that treat pain as a time-varying phenomena, which this study seeks to address. In a similar respect, because most studies do not operationalize pain longitudinally, few studies have examined the degree to which pain may or may not impact the efficacy of CBT-I treatment. In a recent study, we found evidence suggesting that pain severity and pain-related interference with activities did not hinder benefits of behavioral insomnia treatment to sleep-related outcomes like insomnia symptoms in women veterans with mild-to-moderate pain.11 This study builds on our previous work, and the larger literature more generally, through examining pain both longitudinally and on a wider spectrum of severity. We know of no prior studies that examined how pain impacts the effect of CBT-I on sleep outcomes in a sample of older adults, particularly older veterans with insomnia disorder.

Using a sample of older veterans with chronic insomnia disorder who participated in a randomized clinical trial comparing group or individual CBT-I to an attention control condition,10 the current study examined whether pain impacted CBT-I efficacy in improving insomnia symptoms and other sleep-related symptoms among older adults with chronic insomnia. Given prior literature illustrating the effectiveness of CBT-I in older adults with co-morbid insomnia and pain, we hypothesized that pain severity would not meaningfully impede the efficacy of CBT-I on insomnia symptoms. We also tested whether pain impacted CBT-I effects on other sleep-related symptoms, including subjective sleep quality, fatigue, and daytime sleepiness. We hypothesized that older adults would benefit from CBT-I in terms of improved sleep and associated symptoms, regardless of pain severity. To further build upon past literature, mixed effects modeling with time-varying predictors was used to account for between-person and within-person variability often seen in adults with pain.

METHODS

Participants

Secondary analysis was conducted using data from a randomized clinical trial performed in a sample of community dwelling older veterans with chronic insomnia. The trial evaluated sleep-related outcomes of individual or group-based CBT-I administered by non-clinicians supervised weekly by behavioral sleep medicine specialists, compared to a sleep education control group.12 For the current study, the 106 participants who completed baseline assessment and were assigned to receive CBT-I were included in analyses. There was some relatively low attrition in the original study sample, with the posttreatment sample including 97 participants (91.5%), the 6-month follow-up sample including 92 participants (86.8%), and the 12-month follow-up including 89 participants (84%). Across the measurement periods, the reasons for attrition included refusal to complete assessments (n = 9), withdrawal from the study (n = 3), participant death (n = 3), and lost to follow-up (n = 2). A majority of the sample was cisgender male (n = 102; 96%) and identified as White (n = 83; 78.3%). Mean age was 72.1 years, and participants, on average, had 14.6 years of education. On average, participants appeared to endorse moderate-to-high pain across study time points (m = 7.2), with no statistically meaningful differences –based on dependent samples t-tests– in pain severity between time points across the study (−0.5 < t < 1.2, p > 0.05). For additional sample characteristics, see Table 1.

TABLE 1.

Sample demographics and descriptive statistics (n = 106).

Variable Baseline N (%) or mean (SD)
Race
 African American 6 (5.7%)
 Hispanic/Latinx 7 (6.6%)
 Native Hawaiian/Pacific Islander 1 (0.9%)
 Asian American 3 (2.8%)
 White 83 (78.3%)
 Other 3 (2.8%)
 Unknown 3 (2.8%)
Identified gender
 Male 102 (96%)
 Female 4 (4%)
Treatment condition
 Individual CBT-I 54 (51%)
 Group CBT-I 52 (49%)
Pain conditions
 Osteoarthritis and/or degenerative arthritis 48 (45%)
 Rheumatoid arthritis 11 (10%)
 Other arthritis 6 (6%)
 Chronic low back pain 60 (60%)
Age, years 72.1 (7.9)
Education, years 14.6 (2.9)

Design

Procedures from the original study have been previously described, with sampling focused on identifying older adult veterans with insomnia disorder.9 After an initial eligibility determination, participants completed informed consent procedures and a baseline assessment containing questions and measures related to demographic factors, medical conditions (including pain as a secondary outcome), and sleep-related outcomes. Enrolled participants were then randomized (1:1:1) to group-based CBT-I, individual CBT-I, or an attention control condition (group-based general sleep education). All conditions were structured, manual-based, and involved five sessions. The CBT-I intervention included sleep restriction, stimulus control, cognitive interventions and strategies, and sleep hygiene administered by non-clinician “sleep coaches” who completed structured training and engaged in weekly telephone supervision (at least 1 h) with behavioral sleep medicine specialists. After completing the 5-session intervention, follow-up assessments were performed at one-week posttreatment, at 6-months follow-up (primary endpoint) and 12-months follow-up (to examine maintenance of treatment effects).

Measures

Self-reported pain

Pain was measured with a 7-item subscale from the Geriatric Pain Measure (GPM). These items included five items on a dichotomous yes/no scale (e. g., “Do you have pain that never completely goes away”), and two items on a rating scale from 0 to 10 (e.g., “On a scale from 0 to 10, with 0 meaning no pain and 10 meaning the worst pain you can imagine, how severe has your pain been on average?”). Original psychometric studies of the GPM short-form indicated evidence of good internal consistency and validity.13 We conducted confirmatory factor analysis (CFA) to provide statistical evidence that the GPM items used represented a single construct (pain severity; see Supplementary Materials for CFA results). Items that had factor loadings <0.7 were not included in the total score. Based on the CFA results, we used four GPM items representing pain severity in forming a composite score, including: “Do you have pain several times a week,” “Do you have pain every day,” “On a scale of zero to ten, with zero meaning no pain and ten meaning the worst pain you can imagine, how severe is your pain today (0–10)?,” and “In the last seven days, on a scale of zero to ten, with zero meaning no pain and ten meaning the worst pain you can imagine, how severe has your pain been on average (0–10)?” These four items were summed together to form a total score representing perceived pain severity, with possible total scores ranging from 0 to 22.

Insomnia Severity Index

The Insomnia Severity Index (ISI) was used to assess insomnia symptom severity. All seven items contained within the ISI use a four-point Likert-style scale, with some items rating perceived symptoms (e.g., “Please rate the current [i.e., last 2 weeks] severity of your insomnia problem[s]: 1. Difficulty falling asleep”), and additional items rating other areas of experience regarding insomnia (e.g., “How worried/distressed are you about your current sleep problem?”). Total ISI scores were calculated by summing all ISI items together, with a total possible score of 0 to 28. Past psychometric literature on the ISI has shown evidence of good internal consistency (0.71 < α < 0.91), in addition to convergent validity with measures of sleep quality such as the Pittsburgh Sleep Quality Index (PSQI) and sleep diaries.14,15

Pittsburgh Sleep Quality Index

Utilized as a measure of global sleep quality, the PSQI is an 18-item questionnaire that includes items with multiple response scales, including open-field response entries (e.g., “During the past month, what time have you usually gone to bed at night?”) and fixed choice frequency items (e.g., “During the past month, how often have you had trouble sleeping because you wake up in the middle of the night or early morning,” with response choices ranging from “not during the past month” to “three or more times a week”). The total score of the PSQI ranges from 0 to 21. Evidence of test–retest reliability (α = 0.87) and convergent validity of the PSQI has been established for patients with insomnia.16,17

Flinders Fatigue Scale

The Flinders Fatigue Scale (FFS) is a seven item measure of daytime fatigue associated with insomnia.18 Containing multiple response scales, the FFS largely has participants rate the degree that fatigue impacts them and their day-to-day living (ex. “Over the past 2 weeks, was fatigue a problem for you?”), with responses ranging from 0 (Not at all) to 4 (Extremely). Participants also rate the frequency of their fatigue experience from 0 (0 days) to 4 (7 days), as well as the typical times of day that fatigue is experienced (e.g., early morning, etc.). Total FFS scores can range from 0 to 31, with higher scores representing greater fatigue. Initial studies of the FFS illustrated good internal consistency (0.86 ≤ α ≤ 0.91) and evidence of discriminant validity with the Epworth Sleepiness Scale (ESS).18

Epworth Sleepiness Scale

The ESS was employed to measure daytime sleepiness. The ESS is an 8-item measure that has participants rate their likelihood of nodding off or falling asleep in provided situations (e.g., “Sitting and reading”) from zero (“would never nod off”) to three (“high chance of nodding off”). An ESS total summed score can range from 0 to 24. Studies have illustrated evidence of test–retest reliability within sleep-focused clinical trials,19 as well as evidence of validity in multiple contexts and populations.20

Data analysis plan

Analyses were conducted using Stata (v. 17.0). Hypotheses were tested using mixed effects models (MEMs). Original study analyses illustrated non-inferiority between the two treatment groups (individual vs group CBT-I), so treatment group was not included as a predictor of interest. Traditional models were chosen as initial “null” models to ensure model stability prior to bifurcating pain factors into within-person and between-person components. Initial traditional mixed effect models were conducted treating pain as a time independent predictor, with models containing treatment phase (i.e., baseline, posttreatment, 6-month follow-up, 12-month follow-up), pain, and a treatment phase by pain interaction term predicting outcomes of interest. If traditional models demonstrated statistical significance, planned additional analyses involved time-varying models, with pain treated as a time varying predictor. Time varying models contained the following variables: study phase, within-person (WP) effects of pain, between-persons (BP) effects of pain, and two-way interaction terms between pain components and study phase predicting outcomes of interest. These additional analyses built off initial models to better characterize if pain-related effects or interaction effects were associated with pain differences between participants in the study or with individual changes in pain over time within participants. Statistical tests of interaction terms within time-varying models (8 tests) represent the primary analysis of interest. A centering approach was used to separate WP and BP effects of pain, which was informed by simulation studies that demonstrated this technique being optimal for use in small samples compared to other approaches.21

RESULTS

Results of the pain and sleep-related measures across study timepoints are shown in Table 2. All models presented below were statistically significant, and model statistics and p-values can be found in Table 3 (traditional models) and Table 4 (time-varying predictor models). Figures of significant interactions can be found in Figure 1.

TABLE 2.

Pain and sleep-related measures across study timepoints.

Baseline (n = 106) Mean (SD) Posttreatment (n = 97) Mean (SD) 6-Month (n = 92) Mean (SD) 12-Month (n = 89) Mean (SD)
GPM 6.7 (5.5) 7.6 (5.6) 7.1 (5.3) 7.3 (5.5)
ISI 11.7 (5.3) 6.1 (4.5) 5.5 (4.8) 6.4 (5.3)
PSQI 9.4 (3.5) 5.5 (3.6) 6.0 (3.8) 6.4 (3.9)
FFS 9.8 (7.6) 5.0 (6.4) 5.8 (6.7) 6.5 (7.2)
ESS 5.4 (4) 4.5 (3.9) 4.0 (3.8) 4.0 (4.0)

Abbreviations: ESS, Epworth Sleepiness Scale; ESSrange, 0–24; FFSrange, 0–31; FSS, Flinders Fatigue Scale; GPM, Geriatric Pain Measure Composite Score; GPMrange, 0–22; ISI, Insomnia Severity Index; ISIrange, 0–28; PSQI, Perceived Sleep Quality Index; PSQIrange, 0–21; SD, standard deviation.

TABLE 3.

Traditional MEMs testing moderation of sleep outcomes after treatment by pain severity.

ISI Coef (SE) PSQI Coef (SE) FSS Coef (SE) ESS Coef (SE)
Study phase
 Posttreatment −3.7*** (0.9) −3.2*** (0.6) −4.3*** (1.2) −1.0 (0.5)
 6-Month follow-up −5.2*** (0.9) −3.5*** (0.6) −3.5** (1.2) −1.2* (0.6)
 12-Month follow-up -4.2*** (0.9) -2.5*** (0.6) -3.0* (1.3) -1.8** (0.6)
Pain 0.3*** (0.1) 0.1* (0.1) 0.4*** (0.1) 0.1 (0.1)
 Posttreatment by pain −0.3** (0.1) −0.1 (0.1) −0.1 (0.1) 0.0 (0.1)
 6-Month follow-up by Pain −0.2 (0.1) −0.0 (0.1) −0.1 (0.1) −0.0 (0.1)
 12-Month follow-up by pain −0.2 (0.1) −0.1 (0.1) −0.1 (0.1) 0.1 (0.1)

Note: Results show coefficient (Coef ) with standard errors (SE) in parentheses; Baseline study phase serves as the reference group.

Abbreviations: ESS, Epworth Sleepiness Scale; FSS, Flinders Fatigue Scale; ISI, Insomnia Severity Index; MEMs, mixed effects models; PSQI, Perceived Sleep Quality Index.

***

p < 0.001

**

p < 0.01

*

p < 0.05.

TABLE 4.

Time-varying predictor MEMs testing moderation of sleep outcomes after treatment by pain severity.

ISI Coef (SE) PSQI Coef (SE) FLINDERS Coef (SE) ESS Coef (SE)
Study phase
 Posttreatment −5.8*** (0.5) −4.0*** (0.4) −5.2*** (0.7) −0.9** (0.3)
 6 Month follow-up −6.3*** (0.5) −3.5*** (0.4) −4.2*** (0.7) −1.5*** (0.4)
 12 Month follow-up −5.3*** (0.5) −3.0*** (0.4) −3.4*** (0.7) −1.5*** (0.4)
PainWP 0.3* (0.1) 0.1 (0.1) 0.4* (0.2) 0.2* (0.1)
 Posttreatment by PainWP −0.2 (0.2) −0.2 (0.1) 0.1 (0.3) −
 6 Month follow-up by PainWP −0.4 (0.2) −0.1 (0.1) −0.2 (0.3) −
 12 Month follow-up by PainWP −0.1 (0.2) −0.0 (0.1) −0.1 (0.3) −0.1 (0.1)
PainBP 0.4*** (0.1) 0.2** (0.1) 0.5** (0.1) −0.0 (0.1)
 Posttreatment by PainBP −0.3** (0.1) −0.1 (0.1) −0.2 (0.2) 0.1 (0.1)
 6 Month follow-up by PainBP −0.1 (0.1) 0.0 (0.1) −0.1 (0.2) 0.0 (0.1)
 12 Month follow-up by PainBP −0.2 (0.1) −0.1 (0.1) −0.0 (0.2) 0.1 (0.1)

Note: Results show coefficient (Coef) with standard errors (SE) in parentheses; Baseline study phase serves as reference group.

Abbreviations: BP, Between Persons Differences in Pain Severity; ESS, Epworth Sleepiness Scale; FSS, Flinders Fatigue Scale; ISI, Insomnia Severity Index; PainWP, Within Person Change in Pain Severity; PSQI, Perceived Sleep Quality Index.

***

p < 0.001

**

p < 0.01

*

p < 0.05.

FIGURE 1.

FIGURE 1

Trends of insomnia symptom change (ISI) by study phase across general pain severity and between-subject differences in pain severity (GPM). Provided figures represent statistically meaningful interactions between pain severity variables and study phase from baseline to posttreatment. Figure colors outside of legend represent confidence interval overlap. 12M, 12-month follow-up; 6M, 6-month follow-up; BL, baseline; GPM, Geriatric Pain Measure; ISI, Insomnia Severity Index (ISIrange = 0–28); PT, posttreatment.

Insomnia symptoms

Within the traditional model, baseline pain severity was a significant predictor of baseline insomnia symptoms (b = 0.32, p > 0.001), with ISI scores increased (i.e., worse) by 0.32 for every unit higher in pain severity at baseline Within the traditional model, study phase was a significant predictor of insomnia symptoms (−3.73 ≤ b ≤ −5.19, p < 0.001), with ISI scores, on average, reducing (improving) by 3.7 to 5.2 for individuals endorsing no pain. A significant pain by study phase interaction was illustrated for posttreatment (b = −0.28, p = 0.004), with ISI reduction (improvements) from baseline to posttreatment increasing as pain severity increases.

In the time varying covariate model, study phase remained a significant predictor of insomnia symptoms (−5.30 ≤ b ≤ −6.31, p < 0.001). Within-person changes in pain severity (b = 0.28, p < 0.041) predicted ISI, with increases in pain severity over time within participants being associated with worsened ISI scores. Similarly, between-person variability in pain severity (b = 0.39, p < 0.001) predicted ISI, in that ISI scores worsened the higher participant average pain severity scores were compared to their peers. The interaction between between-persons pain levels and study phase significantly predicted ISI at posttreatment (b = −0.32, p = 0.005), with greater reductions (improvements) in ISI for individuals who exhibited higher (worse) levels of pain compared to their peers of lower pain levels. Figure 1 presents interaction effects of both the traditional model (depicted on the left) and the time-varying covariate model (depicted on the right), illustrating greater reduction (improvement) in ISI scores (worse) with higher pain ratings.

Perceived sleep quality

Baseline pain severity was a significant predictor of baseline perceived sleep quality (b = 0.12, p = 0.026), with baseline PSQI scores worsening by 0.12 for every unit increase in pain severity at baseline. Within the traditional model, study phase was a significant predictor of perceived sleep quality (−2.52 ≤ b ≤ −3.45, p < 0.001). No significant interaction between pain and study phase was found regarding perceived sleep quality, indicating that pain did not significantly moderate CBT-I treatment effects on PSQI scores.

In the time varying covariate model, study phase remained a significant predictor of perceived sleep quality (−2.99 ≤ b ≤ −3.98, p < 0.001). Within-person changes in pain severity did not uniquely predict PSQI scores, though between-person differences in pain severity did predict PSQI scores, with individuals who had higher pain than the sample average having worse PSQI scores (b = 0.21, p = 0.007). No significant interactions were found between pain factors and study phase regarding perceived sleep quality.

Perceived fatigue

Within the traditional model, pain severity was a significant predictor of perceived fatigue at baseline (b = 0.43, p < 0.001), with FFS scores increasing (i.e., worse) by 0.43 for every unit increase in pain severity at baseline. Study phase was a significant predictor of perceived fatigue (−3.02 ≤ b ≤ −4.3, p’s ≤ 0.016). No significant interaction between pain and study phase was found regarding perceived fatigue levels.

In the time varying covariate model, study phase remained a significant predictor of perceived fatigue (−3.39 ≤ b ≤ −5.18, p < 0.001). Both within-person changes in pain severity (b = 4.1, p = 0.033) and between-person variability in pain severity (b = 0.46, p = 0.001) predicted FFS scores. No significant interaction between pain and study phase was found in regard to perceived fatigue levels.

Daytime sleepiness

Within the traditional model, study phase was a significant predictor of daytime sleepiness symptoms at 6-month and 12-month follow-up (−2.01 ≤ b ≤ −2.98, p ≤ 0.045). Neither pain, nor the interaction between pain and study phase, were significant predictors of daytime sleepiness symptoms.

In the time varying covariate model, study phase remained a significant predictor of daytime sleepiness (−0.93 ≤ b ≤ −1.51, p’s ≤ 0.001). Only within-person changes in pain severity predicted ESS scores (b = 0.19, p = 0.039), with worsening pain over time within participants being associated with higher levels of daytime sleepiness. No significant interaction between pain and study phase was found regarding perceived fatigue levels.

DISCUSSION

This study examined the role of patient-reported pain on CBT-I treatment outcomes in older adult veterans with chronic insomnia disorder, over half of whom experienced some form of chronic pain. Consistent with past literature, prior to treatment, higher levels of baseline participant-reported pain were associated with worse baseline insomnia symptoms, worse baseline perceived sleep quality, and more baseline fatigue.4 Following CBT-I, insomnia symptoms, perceived sleep quality, and fatigue improved up to the follow-up periods. Pain did not meaningfully reduce the effectiveness of CBT-I. In regard to insomnia symptoms, individuals with higher levels of pain appeared to have more pronounced benefits than those with lower levels of pain. This more pronounced reduction in insomnia compared to other sleep outcomes is likely associated with the treatment focus of CBT-I, with insomnia symptoms being the primary outcome for improvement compared to other sleep outcomes. These findings generally highlighted that treatment effects were robust to pain experienced by participants.

Time-varying covariate models highlighted additional insights to trends seen in traditional models. Fluctuation in pain severity was associated with worsened baseline insomnia symptoms, fatigue, and daytime sleepiness, with individuals who had greater within-person fluctuation in pain generally having worse baseline sleep-related symptoms compared to individuals with more stable pain severity. Having higher levels of pain in comparison to the group average was also associated with higher baseline insomnia and fatigue as well as poorer baseline sleep quality. Altogether, individuals with the worst sleep symptoms at baseline were those who had pain that was higher and more unstable than their peers. Consistent with the traditional models, participants generally improved across sleep outcomes, with pain generally not playing a role in the efficacy of CBT-I outside of its impact on insomnia symptoms. Of note, based on the time varying model, slightly more improvements in insomnia symptoms following CBT-I were seen in participants with higher levels of pain severity than the study average, not individuals with higher levels of fluctuation in their pain severity. Overall, results support the use of CBT-I for older individuals experiencing pain of varying severities and stability, with individuals who are in moderate to high pain appearing to benefit slightly more in regard to insomnia symptom improvement; a presentation of pain commonly seen in older veterans.

Prior evidence suggests that veterans are 40% more likely to have severe pain than the general population, and 70% of veterans with chronic pain also have a diagnosis of insomnia disorder; both factors being detrimental to healthy and optimal aging.22,23 Our findings have larger implications in the management of insomnia in older adults who suffer from pain, in that the presence of pain should not stop older adults with insomnia from receiving CBT-I. Through CBT-I, older adults with pain and insomnia can find potential relief from life-hindering sleep-related symptoms in a short span of sessions.

This study is not without limitations, which present future researchers with opportunities to strengthen the depth and impact of these findings. Though this study was able to delve into the unique role that between- and within-persons variation in pain can have on sleep symptoms, this deeper exploration is limited by the time span between measurement points. Individual day-to-day pain experience likely presents with more variability that one can see when only looking at measures spaced out over months, and future research should consider collecting data at more frequent time points (e.g., daily or weekly ratings) and applying intensive longitudinal modeling approaches to better detect the role of this variability on sleep changes during treatment. In addition, the pain scale used in this study was brief and unidimensional (pain severity); future research is needed to replicate and expand these findings using multi-dimensional pain metrics subjected to more rigorous psychometric testing. Future studies should also seek to specifically sample from populations with pain conditions to increase generalizability to samples with diverse pain etiologies (e.g., fibromyalgia, diabetic neuropathy), given that this study conducted secondary data analysis on clinical trial data where pain conditions were not the primary focus. Also, these findings should be replicated in samples with different cultural identities and backgrounds to ensure generalizability to other populations, particularly older women who represented a small number of participants in the parent trial. Though six statistical tests (i.e., time-varying interaction terms) were of primary focus in this study, sample size limitations and the number of statistical tests conducted may impact analysis results, warranting replication in future studies. Lastly, though attrition was generally minor in this study, attrition bias in results, particularly in regard to missing data in follow-up periods as a consequence of attrition, cannot be ruled out.

CONCLUSION

This study sought to expand our current understanding of how pain levels influence treatment effects of CBT-I on sleep outcomes in older adults. Based on the results, pain did not meaningfully reduce the effect of CBT-I on sleep outcomes. Regardless of pain, older adult veterans experienced significant improvements in insomnia symptoms, perceived sleep quality, fatigue, and daytime sleepiness. Those with worse pain at baseline achieved a slightly greater degree of improvement. Though pain can impact an older adult’s life in many ways, it does not need to stand in the way of older adults getting better sleep through CBT-I, a potent and effective treatment regardless of age and pain-related symptoms.

Supplementary Material

Erickson et al 2024 - Supp materials

Key points

  • Insomnia and pain are often co-morbid in older adults.

  • Cognitive behavioral therapy for insomnia (CBT-I) is a gold standard treatment for insomnia disorder and is effective in older adults.

  • Co-occurring pain does not interfere with the effectiveness of CBT-I in older adult patients.

Why does this paper matter?

Good sleep is an important part of healthy and optimal aging. CBT-I provides a treatment path for older adults with insomnia disorder, even if they have co-morbid chronic pain. Through CBT-I, older patients can see improvements in sleep—and many other areas of functioning impacted by sleep—using this non-pharmacological, time-limited intervention.

Funding information

National Institutes of Health, Grant/Award Numbers: K23 HL157754, K24HL143055; VA Greater Los Angeles Healthcare System GRECC; Office of Academic Affiliations, Department of Veterans Affairs; U.S. Department of Veterans Affairs Health Services Research & Development, Grant/Award Numbers: CDA 20–227, IIR 08–295, RCS 20–191

Footnotes

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

SPONSOR’S ROLE

The sponsor had no role in the design, methods, subject recruitment, data collections, analysis, and preparation of the paper.

FINANCIAL DISCLOSURE

Data used was from a study funded by VA Health Services Research and Development (HSR&D; PI: Alessi; IIR 08-295). Dr. Erickson is supported by the VA Office of Academic Affiliations through the Advanced Fellowship Program in Geriatrics. Dr. Carlson is supported by VA HSR&D CDA 20-227. Dr. Kelly is supported by NIH/NHLBI K23 HL157754. Dr. Martin is supported by NIH/NHLBI K24HL143055 and VA HSR&D Research Career Scientist Award RCS 20–191. Dr. Alessi is supported by VA HSR&D and NIH. Support was also provided by the VA Greater Los Angeles Geriatric Research, Education and Clinical Center.

SUPPORTING INFORMATION

Additional supporting information can be found online in the Supporting Information section at the end of this article.

REFERENCES

  • 1.Patel D, Steinberg J, Patel P. Insomnia in the elderly: a review. J Clin Sleep Med 2018;14(6):1017–1024. doi: 10.5664/jcsm.7172 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 2.McCurry SM, Von Korff M, Vitiello MV, et al. Frequency of comorbid insomnia, pain, and depression in older adults with osteoarthritis: predictors of enrollment in a randomized treatment trial. J Psychosom Res 2011;71(5):296–299. doi: 10.1016/j.jpsychores.2011.05.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3.Dahlhamer J, Lucas J, Zelaya C, et al. Prevalence of chronic pain and high-impact chronic pain among adults—United States, 2016. MMWR Morb Mortal Wkly Rep 2018;67(36):1001–1006. doi: 10.15585/mmwr.mm6736a2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Finan PH, Goodin BR, Smith MT. The association of sleep and pain: an update and a path forward. J Pain. 2013;14(12):1539–1552. doi: 10.1016/j.jpain.2013.08.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5.Mitchell MD, Gehrman P, Perlis M, Umscheid CA. Comparative effectiveness of cognitive behavioral therapy for insomnia: a systematic review. BMC Fam Pract 2012;13:40. doi: 10.1186/1471-2296-13-40 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 6.Huang K, Li S, He R, et al. Efficacy of cognitive behavioral therapy for insomnia (CBT-I) in older adults with insomnia: a systematic review and meta-analysis. Australas Psychiatry 2022;30(5):592–597. doi: 10.1177/10398562221118516 [DOI] [PubMed] [Google Scholar]
  • 7.Selvanathan J, Pham C, Nagappa M, et al. Cognitive behavioral therapy for insomnia in patients with chronic pain—a systematic review and meta-analysis of randomized controlled trials. Sleep Med Rev 2021;60:101460. doi: 10.1016/j.smrv.2021.101460 [DOI] [PubMed] [Google Scholar]
  • 8.Enomoto K, Adachi T, Fujino H, Kugo M, Tatsumi S, Sasaki J. Comparison of the effectiveness of cognitive behavioral therapy for insomnia, cognitive behavioral therapy for pain, and hybrid cognitive behavioral therapy for insomnia and pain in individuals with comorbid insomnia and chronic pain: a systematic review and network meta-analysis. Sleep Med Rev 2022;66: 101693. doi: 10.1016/j.smrv.2022.101693 [DOI] [PubMed] [Google Scholar]
  • 9.McCurry SM, Zhu W, Von Korff M, et al. Effect of telephone cognitive behavioral therapy for insomnia in older adults with osteoarthritis pain: a randomized clinical trial. JAMA Intern Med 2021;181(4):530–538. doi: 10.1001/jamainternmed.20209049. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10.Vitiello MV, McCurry SM, Shortreed SM, et al. Cognitive-behavioral treatment for comorbid insomnia and osteoarthritis pain in primary care: the lifestyles randomized controlled trial. J Am Geriatr Soc 2013;61(6):947–956. doi: 10.1111/jgs.12275 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 11.Erickson AJ, Ravyts SG, Dzierzewski JM, et al. Pain in your day? Get sleep treatment anyway! The role of pain in insomnia treatment efficacy in women veterans. J Sleep Res 2024;e14147. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 12.Alessi C, Martin JL, Fiorentino L, et al. Cognitive behavioral therapy for insomnia in older veterans using nonclinician sleep coaches: randomized controlled trial. J Am Geriatr Soc 2016; 64(9):1830–1838. doi: 10.1111/jgs.14304 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13.Ferrell BA, Stein WM, Beck JC. The Geriatric Pain Measure: validity, reliability and factor analysis. J Am Geriatr Soc 2000; 48(12):1669–1673. doi: 10.1111/j.1532-5415.2000.tb03881.x [DOI] [PubMed] [Google Scholar]
  • 14.Morin CM, Belleville G, Belanger L, Ivers H. The Insomnia Severity Index: psychometric indicators to detect insomnia cases and evaluate treatment response. Sleep. 2011; 34(5):601–608. doi: 10.1093/sleep/34.5.601 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15.Bastien CH, Vallieres A, Morin CM. Validation of the Insomnia Severity Index as an outcome measure for insomnia research. Sleep Med 2001;2(4):297–307. doi: 10.1016/s1389-9457(00)00065-4 [DOI] [PubMed] [Google Scholar]
  • 16.Backhaus J, Junghanns K, Broocks A, Riemann D, Hohagen F. Test–retest reliability and validity of the Pittsburgh Sleep Quality Index in primary insomnia. J Psychosom Res 2002;53(3): 737–740. doi: 10.1016/s0022-3999(02)00330-6 [DOI] [PubMed] [Google Scholar]
  • 17.Buysse DJ, Reynolds CF 3rd, Monk TH, Berman SR, Kupfer DJ. The Pittsburgh Sleep Quality Index: a new instrument for psychiatric practice and research. Psychiatry Res 1989;28(2):193–213. doi: 10.1016/0165-1781(89)90047-4 [DOI] [PubMed] [Google Scholar]
  • 18.Gradisar M, Lack L, Richards H, et al. The Flinders Fatigue Scale: preliminary psychometric properties and clinical sensitivity of a new scale for measuring daytime fatigue associated with insomnia. J Clin Sleep Med 2007;3(7):722–728. [PMC free article] [PubMed] [Google Scholar]
  • 19.Rosenberg R, Babson K, Menno D, et al. Test-retest reliability of the Epworth Sleepiness Scale in clinical trial settings. J Sleep Res 2022;31(2):e13476. doi: 10.1111/jsr.13476 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 20.Schokman A, Bin YS, Naehrig D, Cheung JMY, Kairaitis K, Glozier N. Evaluation of psychometric properties of patient-reported outcome measures frequently used in narcolepsy randomized controlled trials: a systematic review. Sleep. 2022;45(10): 1–16. doi: 10.1093/sleep/zsac156 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21.Ligthart-Smith A A Comparison of Linear Mixed Models That Include Time-Varying Covariates. AMSI Higher Ed; 2016:1–9. [Google Scholar]
  • 22.Goulet JL, Kerns RD, Bair M, et al. The musculoskeletal diagnosis cohort: examining pain and pain care among veterans. Pain. 2016;157(8):1696–1703. doi: 10.1097/j.pain.0000000000000567 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Colvonen PJ, Almklov E, Tripp JC, Ulmer CS, Pittman JOE, Afari N. Prevalence rates and correlates of insomnia disorder in post-9/11 veterans enrolling in VA healthcare. Sleep. 2020;43(12): 1–19. doi: 10.1093/sleep/zsaa119 [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Erickson et al 2024 - Supp materials

RESOURCES