Improvement in Post-traumatic Stress Symptoms Following an Internet-based Cognitive-Behavioral Intervention for Insomnia: An Open Trial among Rural Appalachian Women

Abstract

Appalachian women face significant health disparities and have limited access to health care. Mental health conditions and treatment-seeking are stigmatized in Appalachian communities. Appalachian women may benefit from web-based interventions targeting less stigmatized health complaints (e.g., insomnia), while simultaneously yielding benefit in associated mental health conditions including symptoms of post-traumatic stress disorder (PTSD). In this study, 37 trauma-exposed adult women aged 45 and older from rural Appalachian Kentucky completed a six-session online self-administered cognitive behavioral therapy for insomnia (CBT-I) intervention and completed measures of PTSD symptoms, insomnia, and depression at pre- and post-treatment. Participants reported a significant reduction in PTSD symptoms from pre-to post-intervention, and this remained significant after adjusting for severity of insomnia and depression pre-treatment. Pending replication in a randomized controlled trial, web-based CBT-I may offer an adjunctive mental health treatment option that circumvents cultural stigmas and reduces PTSD symptoms for trauma-exposed Appalachian women.

The Appalachian region of the Eastern United States represents the largest geographic hot spot for insufficient sleep in the country, with 25–58% of adults not getting enough sleep the majority of nights in the past month.¹ Women in Appalachia experience disproportionate risk for insufficient sleep and insomnia due to clinical risk factors including mental health difficulties, prescription sleep-medication use, and gendered social stressors, risks that increase with age.^2–4 Appalachian women also must navigate logistical barriers to accessing general health and specialty mental health services (e.g., lack of providers, transportation barriers).⁵ Though many of these barriers are shared across rural areas, women in Appalachia belong to an established minority subculture, with unique psychological service needs related to ethnic and cultural diversity that extend beyond rurality.⁶ Notably, substantial cultural stigma around having mental health difficulties and seeking associated care, concerns regarding quality and privacy of available care, and the tendency toward self-reliance are factors that specifically prevent Appalachian women from seeking mental health treatment, even if care is accessible.^4,6 Internet-based interventions represent one tool that may help some Appalachian women overcome many of these barriers. Most Appalachian households (86%) have access to one or more computer devices, and 78% have broadband (high-speed) internet.⁷ Rural residents, including those in Appalachia, increasingly use and are comfortable with personal technology.^8,9 Though access problems persist, greater reliance on telehealth and web-based interventions in Appalachia have been identified as necessary to address inequitable distribution of health providers in the region.^5,9 Appalachian women may also benefit from web-based, self-administered interventions that target less stigmatized health complaints (such as sleep problems), yet yield additional benefit in associated mental health conditions.

Psychotherapeutic interventions are a first-line treatment approach for sleep problems and are generally favored over commonly prescribed pharmacological interventions due to risks inherent in the latter (e.g., psychomotor impairment, addiction potential).¹⁰ One highly effective psychotherapeutic intervention for sleep is cognitive-behavioral therapy for insomnia (CBT-I).¹⁰ Cognitive-behavioral therapy for insomnia also demonstrates therapeutic benefits for various mental health difficulties including depression, anxiety, and post-traumatic stress disorder (PTSD).¹¹ Traditionally, CBT-I has been administered to individuals in person. However, internet-based CBT-I is increasingly considered a more accessible option¹² that is effective in reducing insomnia symptoms for up to 12 months.¹³

Compared with women in non-Appalachian areas, Appalachian women may be disproportionately likely to experience trauma.¹⁴ Exposure to trauma prospectively predicts the onset of sleep problems,¹⁵ which, in turn, can contribute to both the development and maintenance of PTSD symptoms.¹⁶ Trauma-related sleep problems (e.g., difficulty falling or staying asleep, restless sleep) and nightmares are primary symptoms of PTSD.¹⁷ Notably, trauma-exposed women report sleep problems and nightmares at higher rates than men with trauma histories.¹⁸ In-person CBT-I has been shown to ameliorate sleep symptoms and mitigate PTSD symptoms in trauma-exposed individuals.^16,18 Other modalities for intervention administration, such as web-based delivery, have yet to be explored in trauma-exposed individuals.

Given the gap between population need and accessible therapeutic modality, the present study is noteworthy as it describes proof-of-concept findings from a single-group pre-test post-test pilot trial of web-based CBT-I in a sample of trauma-exposed Appalachian women (N = 37). Data were collected as part of a larger project^22,23 that evaluated the feasibility and acceptability of an internet-based CBT-I intervention called “Sleep Healthy Using the Internet” (SHUTi)¹⁹ in middle-aged and older Appalachian women with insomnia. This high-risk population was targeted because of the need for accessible, efficacious insomnia treatment in a region with limited health care access.

Sleep Healthy Using the Internet has demonstrated efficacy in the reduction of depression symptoms comorbid with insomnia in other populations^13,20,21 and was found in the parent project^22,23 to be a feasible and acceptable insomnia intervention. However, no other study, to our knowledge, has evaluated the potential of SHUTi to influence PTSD symptoms. Thus, the aim of the present analysis is to examine whether PTSD symptoms improved in trauma-exposed Appalachian women with insomnia participating in SHUTi. Consistent with previous findings demonstrating that in-person CBT-I leads to improvement in both insomnia and PTSD symptoms,¹⁶ we hypothesized that PTSD symptoms would decrease following SHUTi’s six-session online self-administered CBT-I intervention. A self-administered, non-pharmacologic intervention, such as SHUTi or another web-based CBT-I, could be highly beneficial in this high-risk population by allowing indirect access to mental health care.

Methods

Participants.

Participants were recruited from Appalachian Kentucky using a variety of methods including paper/electronic flyers, social media outlets, local health clinics, and ResearchMatch.org. Some participants were recruited via snowball sampling. We recruited the sex (women) and age group (45+) most likely to report difficulty falling and staying asleep.^24,25 Eligible participants were English speakers, reported difficulty falling or staying asleep for two or more nights each week for at least three months, currently or previously used prescription and/or over-the-counter sleep aids, and had regular internet access. Self-reported exclusion criteria were obstructive sleep apnea, schizophrenia, dementia, Alzheimer’s disease, Cushing’s disease, or bipolar disorder with psychosis. A total of 46 women were enrolled in the parent project.^22,23 Of these, 44 (95.7%) reported a history of at least one potentially traumatic event at baseline. Trauma-exposed participants who completed all six SHUTi sessions and post-treatment assessments were included in subsequent analyses (n = 37). Women were 55 years old on average (SD = 6.3). Most women in the present sample self-reported their race as White (94.6%), either did not attend college or did not obtain a bachelor’s degree (73.0%), were employed full-time (75.7%), and were married (75.7%). The most common potentially traumatic event endorsed by participants was learning about serious illness, injury, or death of a loved one (91.9%) followed by seeing or handling dead bodies other than at a funeral (45.9%); experiencing a serious accident (35.1%); being a victim of a crime (e.g., robbery, mugging; 29.7%); witnessing someone be seriously injured or killed (27.0%); experiencing a natural disaster (24.3%); experiencing a sexual assault or sexual abuse (21.6%); experiencing the death of a spouse, romantic partner, or child (13.5%); experiencing a physical assault or physical abuse (10.8%); experiencing a close friend or family member be murdered or killed by a drunk driver (8.1%); or experiencing some other highly distressing event (27.0%). Participants reported a mean of 3.62 different potentially traumatic events (SD = 2.23; range = 1–10).

Procedure.

Full study procedures are described elsewhere.^22,23 Briefly, enrolled participants (N = 46) self-administered six once-weekly modules over nine weeks as part of SHUTi.¹⁹ Sleep Healthy Using the Internet consists of six 45-minute modules in which participants are presented with new material on insomnia, sleep restriction, stimulus control, cognitive restructuring, sleep hygiene, and relapse prevention. Participants also review and receive feedback on the previous core’s assignments, and this is supplemented by interactive features and daily prompts reminding them to complete a sleep diary. Before and after the SHUTi intervention, participants completed an online assessment in REDCap²⁶ and a semi-structured qualitative interview.^22,23 Participants were compensated with a $50 gift card at each timepoint. The study was approved by the University of Kentucky Institutional Review Board, and all participants provided informed consent.

Measures.

Trauma history.

History of potentially traumatic events was assessed using the Trauma History Questionnaire (THQ)²⁷ a 24-item self-report index of potentially traumatic events experienced throughout one’s lifetime, such as natural and manmade disasters, accidents, crime experiences, and physical or sexual assault. The THQ has demonstrated fair-to-excellent test–retest reliability over two to three months, as well as strong construct and cultural validity.²⁷

PTSD symptoms.

A four-item version of the PTSD Checklist for DSM-5 (PCL-5) was used to assess past-month PTSD symptoms.²⁸ The PCL-5 was originally created as a 20-item psychometrically valid self-report tool to measure PTSD symptoms based on criteria defined in the Diagnostic and Statistical Manual of Mental Disorders, 5th edition,^17,29 and has since been adapted into the shorter version used in this study. During pre- and post-intervention surveys, participants were prompted to recall their most distressing traumatic event and indicate how much they had been bothered by the following reactions in the past 30 days: 1) repeated, disturbing, and unwanted memories of the stressful experience; 2) avoiding external reminders of the stressful experience; 3) having strong negative beliefs about yourself, other people, or the world; and 4) feeling jumpy or easily startled. Responses were coded on a five-point Likert-type scale ranging from 0 (not at all) to 4 (extremely). Total scores ranged from 0 to 16, with higher scores indicating more severe PTSD symptoms. A cut score of 10 or higher on the four-item version of the PCL-5 achieves optimal sensitivity (.76) and specificity (.52) for predicting probable PTSD.^28,29 In the present sample, the measure demonstrated acceptable to good internal consistency, Cronbach’s α = .80 (pre-treatment) and .78 (post-treatment).

Insomnia.

Past two-week insomnia severity was measured using the seven-item Insomnia Severity Index (ISI).³⁰ A total score is derived by summing responses to all items rated on a five-point Likert-type scale, with higher scores representing more severe insomnia. Possible scores range from 0–28 with scores above 10 indicating clinically meaningful insomnia.³¹ Scores on the ISI display strong convergent validity with other self-report measures of sleep^30,31 and acceptable to good internal consistency in this sample α = .76 (pre-treatment), α = .82 (post-treatment).

Depression.

Past-week depression symptoms were assessed using the 10-item Center for Epidemiologic Studies Depression Scale-Revised (CESD-R-10).³² Participants rate the frequency of depression symptoms on a four-point scale. Possible scores range from 0–30, with higher scores representing more severe depression. Scores on the CESD-R-10 demonstrate good convergent and divergent validity³³ as well as good internal consistency in the present sample at pre-treatment α = .80. Reliability for scores on the CESD were less than optimal at post-treatment (α = .64).

Data analytic approach.

Zero-order correlations among continuous variables at pre-treatment and change from pre- to post-treatment were examined. Given skewness observed in pre- and post-treatment PTSD symptom scores, descriptive statistics included examination of medians/interquartile ranges as well as frequencies of primary study variables. A paired-samples t-test was conducted to examine change in PTSD after the SHUTi intervention, and a one-way repeated measures analysis of covariance (ANCOVA) was used to examine change in PTSD symptoms after adjusting for baseline insomnia and depression.

Results

Pre-treatment, PTSD symptoms were positively correlated with both insomnia (r = 0.49, p = .002) and depression severity (r = .61, p < .001). At pre-treatment 8.1% of participants (n = 3) were above the clinical cutoff for a probable PTSD diagnosis (i.e., score of ≥ 10 on the PCL-5), and 2.7% scored above this cutoff at post-treatment (n = 1). Figure 1 displays the median and interquartile range for scores on the PCL-5 at pre- and post-treatment. A paired-samples t-test demonstrated that PTSD symptoms decreased significantly from pre- to post-treatment, t(36) = −2.81, SE = 0.65, p = .008, and the effect size was small to moderate (Cohen’s d = −0 .46). Results from the one-way repeated measures ANCOVA found that PTSD symptoms still reduced significantly when adjusting for pre-treatment insomnia and depression severity, F(1,34) = 4.06, p = .05. The effect size for PTSD symptom improvement was moderate in this adjusted model (partial η² = .11). Change in PTSD symptoms was positively correlated with change in insomnia (r = .53, p < .001) and depression severity (r = .33, p = .045).

Figure 1.

Box plot of PTSD symptom severity pre- and post-SHUTi.

Discussion

Appalachian women must navigate a host of psychosocial, cultural, and practical barriers to accessing health care, and seeking treatment for mental health conditions is stigmatized in Appalachian communities. This research describes preliminary proof-of-concept findings regarding changes in PTSD symptoms observed following participation in an open-label pilot trial of SHUTi, a self-guided, internet-based CBT intervention for insomnia in middle-aged and older adult women in Appalachia. To our knowledge, this study is the first to evaluate PTSD symptom improvement following online CBT-I in this population. Similar to prior findings, we found that PTSD symptoms were positively correlated with insomnia and depression severity prior to initiating the SHUTi treatment program, and that PTSD symptoms improved significantly over the course of the intervention period.^11,16 Symptom improvements for PTSD remain significant after adjusting for severity of insomnia and depression at baseline. It is notable that the adjusted model demonstrated a moderate effect for improvement in PTSD symptoms despite the non-intensive nature of the intervention (i.e., self-guided, web-based) and its focus on insomnia versus trauma-related issues. We also found that improvement in PTSD symptoms positively correlated with improvement in insomnia and depression severity. Although these findings are preliminary, and the lack of control group prohibits us from concluding that the SHUTi intervention specifically led to changes in PTSD symptoms, these results are consistent with prior controlled trials demonstrating that in-person CBT-I has the added benefit of PTSD symptom reduction in addition to improving the primary target of insomnia and other sleep-related difficulties among trauma-exposed populations.^11,16 Following replication using a randomized controlled trial, internet-based CBT-I may hold promise as an accessible and brief intervention for addressing insomnia as well as co-occurring mental health difficulties including PTSD and depression symptoms.^22,23 Such an approach may have potential within a stepped-care framework³² for addressing mental health difficulties through a less resource-intensive intervention for sleep, which is traditionally less stigmatized than mental health difficulties such as PTSD or depression. Future studies should assess whether engagement in SHUTi or other web-based interventions has potential to reduce perceived stigma surrounding seeking mental health treatment in this region. Given Appalachian women’s limited access to mental health services and reticence to engage with such services even when available, provision of evidence-based, accessible treatments is crucial for this population.^4,34 Feasibility of implementing web-based and other telehealth interventions should be explored given access to reliable internet is still limited for many in Appalachia, and a number rely on smartphones alone as their primary access point.^7,34 Web-based interventions are unlikely to replace face-to-face care, but these novel approaches may prove to be important tools to enhance health care access for some.

Several limitations warrant discussion. Despite reporting a history of trauma exposure, only a small portion of our sample reported PTSD symptoms above the recommended clinical cutoff. A thorough clinician-administered trauma and PTSD assessment should be conducted in future studies to increase clarity regarding diagnoses (e.g., PTSD versus adjustment disorder or persistent complex bereavement disorder). It will also be important to determine if potential benefits observed in this sample would generalize to a group with clinical PTSD, and particularly to those with chronic PTSD. In addition, power was limited by our small sample size, and the representativeness of the sample was limited by the presence of higher educational attainment and greater access to private health insurance in our sample compared with the region as a whole. However, all women in Appalachia—regardless of educational attainment and insurance—are members of a health disparity population due to health care provider shortages, sparse regional resources, and increased risk of morbidity and mortality.³⁵ Despite these limitations, the present findings offer the first evidence that internet-delivered CBT-I may hold promise as an accessible intervention for insomnia and co-occurring mental health difficulties, including PTSD symptoms, among trauma-exposed women in rural Appalachia.