Issue Navigator

Volume 10 No. 02
Earn CME
Accepted Papers

Scientific Investigations

Impact of Brief Cognitive Behavioral Treatment for Insomnia on Health Care Utilization and Costs

Christina S. McCrae, Ph.D.1; Adam D. Bramoweth, Ph.D.2; Jacob Williams, M.S.1; Alicia Roth, M.A.1; Caterina Mosti, B.S.3
1Department of Clinical & Health Psychology, University of Florida, Gainesville, FL; 2VISN 4 Mental Illness Research, Education and Clinical Center (MIRECC) VA Pittsburgh Healthcare System, Pittsburgh, PA; 3Department of Psychology, Drexel University, Philadelphia, PA


Study Objectives:

To examine health care utilization (HCU) and costs following brief cognitive behavioral treatment for insomnia (bCBTi).


Reviewed medical records of 84 outpatients [mean age = 54.25 years (19.08); 58% women] treated in a behavioral sleep medicine clinic (2005-2010) based in an accredited sleep disorders center. Six indicators of HCU and costs were obtained: estimated total and outpatient costs, estimated primary care visits, CPT costs, number of office visits, and number of medications. All patients completed ≥ 1 session of bCBTi. Those who attended ≥ 3 sessions were considered completers (n = 37), and completers with significant sleep improvements were considered responders (n = 32).


For completers and responders, all HCU and cost variables, except number of medications, significantly decreased (ps < 0.05) or trended towards decrease at post-treatment. Completers had average decreases in CPT costs of $200 and estimated total costs of $75. Responders had average decreases in CPT costs of $210. No significant decreases occurred for non-completers.


bCBTi can reduce HCU and costs. Response to bCBTi resulted in greater reduction of HCU and costs. While limited by small sample size and non-normal data distribution, the findings highlight the need for greater dissemination of bCBTi for several reasons: a high percentage of completers responded to treatment, as few as 3 sessions can result in significant improvements in insomnia severity, bCBTi can be delivered by novice clinicians, and health care costs can reduce following treatment. Insomnia remains an undertreated disorder, and brief behavioral treatments can help to increase access to care and reduce the burden of insomnia.


McCrae CS; Bramoweth AD; Williams J; Roth A; Mosti C. Impact of brief cognitive behavioral treatment for insomnia on health care utilization and costs. J Clin Sleep Med 2014;10(2):127-135.

Chronic insomnia impacts approximately 8% to 10% of adults in the United States population1 and carries substantial economic burden. Direct costs (i.e., costs directly related to treatment) have been estimated at $13.9 billion annually in the United States, with total treatment costs estimated at $77 to $92 billion.2 When indirect costs (i.e., lost productivity, insomnia-related accidents, increased medical problems) are included, annual costs exceed $100 billion.3 Evidence from outside the U.S. reveals a similar pattern. Specifically, a study in Quebec, Canada, found that the combined direct and indirect per-person costs of insomnia syndrome (insomnia complaint ≥ 3 nights/ week with > 1 month duration, dissatisfaction with sleep, and daytime impairment; prescription sleep medication ≥ 3 nights/ week) was approximately $5,010; with $293 attributed to direct costs and $4,717 in indirect costs.4 Those with insomnia symptoms (complaint ≥ 3 nights/week; prescription sleep medication < 3 nights/week or over-the-counter medication > 1 night/ week) also incurred significant costs: $160 in direct costs and $1,271 in indirect costs.

Direct comparisons of individuals with and without insomnia further highlight the economic impact of insomnia. In a study comparing age-matched individuals with and without insomnia, the average health care related costs incurred by young adults with insomnia was more than $900 greater on average over a 6-month period than the costs incurred by young adults without insomnia. For older adults with insomnia, that figure rose to more than $1,100 on average.5 Moreover, Sarsour et al. found that average health care costs were significantly greater for individuals with moderate to severe insomnia ($1,739) compared to those without insomnia ($1,013), representing an approximate 72% increase in health care costs.6


Current Knowledge/Study Rationale: Cognitive behavioral treatment for insomnia (CBTi) is a highly efficacious brief treatment for chronic insomnia. While significant health care utilization and financial costs are linked to chronic insomnia, there is limited research investigating the potential of CBTi to reduce health care utilization and costs.

Study Impact: Patients who completed and responded to brief CBTi had significantly reduced estimated health care costs following treatment. Brief CBTi has the potential to decrease health care utilization and costs related to insomnia.

Individuals with insomnia demonstrate increased rates of health care utilization (HCU), which contributes to increased insomnia-related health care costs. Compared to good sleepers, individuals with insomnia have more visits to general practitioners and medical specialists, twice as many hospitalizations, and use more medications than good sleepers.7,8 The common co-occurrence of insomnia with other conditions is an important consideration when examining HCU and associated costs as the rates of insomnia occurring in the context of other common medical and psychiatric conditions are much higher than that in the general population.9,10 For example, an epidemiological study revealed that approximately 38% of individuals with a comorbid condition exhibited insomnia compared to only 8% of individuals without a comorbid condition.9 Individuals with comorbid insomnia and major depressive disorder have demonstrated significantly greater direct health care costs (i.e., greater number of overall outpatient physician visits, depression-related visits, and a greater number of antidepressant prescriptions) than individuals with major depressive disorder alone.10 A recent study by Bramoweth and colleagues provides support for increased costs related to comorbid insomnia.11 Specifically, they found significantly greater estimated health care costs for young adults with insomnia comorbid with a medical or psychiatric disorder compared to those with insomnia alone and good sleepers.11 Two other recent studies also found evidence for increased costs linked to insomnia among comorbid samples. Sarsour et al. found people with insomnia (per the Insomnia Severity Index) and a psychiatric diagnosis had the highest health care costs.6 A study by Tian et al. found patients with major depressive disorder who were taking medication for insomnia had higher health care costs than those who were not taking such medication and a healthy control group.12

A next logical step in research examining insomnia-related HCU and health care costs is to examine the potential for treatment to reduce those costs. One study found that among Medicaid claims, the majority of the cost of insomnia treatment was for prescription drugs, with zolpidem and trazodone the most prescribed medications.8 Pharmacotherapy is the most common approach to treating insomnia13; to date, at least two studies have examined its impact on health care costs with mixed results. Jhaveri et al. demonstrated that several pharmacological interventions for insomnia (e.g., eszopiclone, indiplon, ramelteon, low-dose trazodone, zaleplon, zolpidem, and zolpidem extended-release) produced sizeable yet variable cost savings when accounting for sleep onset latency, wake after sleep onset, and sleep efficiency.14 For example, when using total sleep time as the efficacy measure, mean per-patient per-year cost savings linked to low-dose trazodone, zolpidem, and zolpidem extended-release were $1,022, $1,158, and $2,573, respectively.14 In another study, Snedecor and colleagues found minimally increased costs ($495 vs. $428; costs of medication, office visits, absenteeism, and loss of productivity) following treatment with eszopiclone (5 months) compared to placebo.15 However, eszopiclone users had an improvement over baseline in Quality Adjusted Life Years (QALY) while placebo had a decrease, and eszopiclone users had greater estimated productivity versus placebo at 6 months post-baseline ($689 vs. $333). Even though overall per-person costs were increased, eszopiclone was considered a cost-effective treatment.15 Costs and cost-savings linked to medication are complicated by the inconsistent duration of medication use in clinical trials, even more so in the real world. A brief review of the literature found duration of clinical trials of sedative hypnotics range from short (2 weeks16,17) to moderate (1-3 months1820) to long (6 months15).

Cognitive behavioral treatment for insomnia (CBTi) is not as widely used as pharmacotherapy, but represents an attractive alternative for treating chronic insomnia. The potential for CBTi to reduce HCU and health care related costs has not yet been examined. This is unfortunate, because CBTi has been shown to be efficacious and effective in a variety of populations, including adults, older adults, and several comorbid patient populations (fibromyalgia,21 chronic pain,22 breast cancer,23 and cardiovascular disease). A meta-analytic review found CBTi to be more effective at reducing sleep onset latency than hypnotic medication.24 CBTi was equally as effective as hypnotic medication in improving other sleep-related outcomes (e.g., wake time after sleep onset, number of awakenings, sleep quality), with both approaches achieving an effect size (Cohen's d) greater than 0.80.24 Unfortunately, broader adoption of CBTi has been hindered by several factors, including the relatively limited number of clinicians trained to provide it and the fact that the length of traditional protocols (6-10 sessions) is not well-matched to the resources available in primary care settings where the majority of insomnia cases are seen.2527 However, Manber and colleagues have described the successful training of mental health clinicians without sleep specialization to deliver CBTi in the VA healthcare system.28

In response to this, shortened protocols (i.e., ≤ 4 sessions) have been developed, and research has shown them to be efficacious and effective in treating insomnia in adults of all ages (see McCrae,29 Buysse,30 or Edinger and Sampson,31 for review). For example, Edinger and Sampson demonstrated that four sessions of CBTi resulted in clinically significant improvements in wake after sleep onset, total wake time, and sleep efficiency compared to sleep hygiene recommendations alone.31 Two other studies have found that brief behavioral treatment of insomnia (BBTi; 2 in person, 2 phone calls) resulted in clinically significant sleep improvements compared to information only and sleep hygiene education only controls.32 When the sleep outcomes found for brief CBTi studies were matched with previous research done with a typical 6 sessions, the results were comparable. In addition to sleep outcomes, brief behavioral protocols have been shown to produce significant improvements in several secondary outcomes. For example, Buysse and colleagues also found moderate effect sizes for other outcomes following brief BBTi, including improved anxiety, depression, and quality of life.30

HCU and other health related cost reductions represent other potentially important secondary outcomes that, unfortunately, have yet to be examined. Given evidence of brief behavioral protocols' impact on sleep, mood, and quality of life outcomes, demonstration of its ability to also reduce HCU and health care costs is particularly important. Such demonstration may provide additional impetus for efforts aimed at broader dissemination of this highly effective and efficacious approach to treating chronic insomnia.

The main objective of the present study was to examine HCU and health care costs following a brief CBTi (bCBTi) protocol, delivered in an outpatient behavioral sleep medicine clinic based in an accredited sleep disorders center within an academic medical center. Specifically, we measured HCU and costs—number of physician office visits, costs related to office visits (CPT costs), number of medications, and estimated health care costs and utilization (total costs, outpatient costs, and number of primary care)—over a 6-month period prior to and following delivery of a manualized bCBTi to patients with chronic insomnia. We conducted two types of comparisons based upon: (1) treatment completion (completers vs. non-completers), and (2) treatment response (responders vs. non-completers). For these comparisons, we hypothesized that patients who completed and/or responded to bCBTi would exhibit significantly greater decreases in their number of physician office visits, costs related to office visits, number of medications, and estimated health care costs than patients who did not complete treatment.


Patients and Procedure

Eighty-four outpatients treated for insomnia at the Insomnia and Behavioral Sleep Medicine (IBSM) Clinic at the University of Florida & Shands Sleep Disorders Center in Gainesville, FL between 2005 and 2010 served as participants. Patients referred to the clinic were seen for an intake session that included an in-depth clinical sleep history interview and a standard battery of questionnaires (see Measures). Criteria for insomnia were consistent with the International Classification of Sleep Disorders, Second Edition,33 the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition, Text Revision (DSM-IV-TR)34 and published research diagnostic criteria (RDC)35 and included: (1) complaint of difficulty initiating or maintaining sleep, or non-restorative sleep, for at least one month; (2) the sleep disturbance (or associated daytime fatigue) causes clinically significant impairment or stress; (3) the sleep disturbance does not occur exclusively in the presence of another sleep disorder; (4) the sleep disturbance does not occur exclusively in the presence of another mental disorder; and (5) the disorder is not due to the direct effect of substance use. All patients who met these criteria were offered bCBTi. Given that people with insomnia are approximately 10 times more likely to be depressed and 17 times more likely to be anxious than those without insomnia,36 patients with symptoms of anxiety and/or depression were offered bCBTi, unless their symptoms were severe enough to warrant immediate, alternative treatment (e.g., active suicidal ideation or intent, psychosis). Additionally, participants with comorbid medical or psychiatric disorders deemed too severe (e.g., active psychosis) for bCBTi and/or if bCBTi was contraindicated (e.g., active mania, uncontrolled seizures) were not offered treatment. In those cases, appropriate referrals were made.

This study was approved by the University of Florida (UF) Health Science Center Institutional Review Board (IRB-01) as a retrospective data/record review study. The study included a waiver of informed consent and a waiver of authorization to obtain protected health information. Eighty-four charts were reviewed for the current study.

Patients had to complete at least one session of bCBTi to be included. The study sample's demographic and health characteristics are provided in Table 1.

Demographic and health characteristics of the total sample and by treatment outcome group for outpatients with chronic insomnia treated with brief cognitive behavioral treatment for insomnia (bCBTi) between 2005-2010 in a behavioral sleep medicine clinic.


table icon
Table 1

Demographic and health characteristics of the total sample and by treatment outcome group for outpatients with chronic insomnia treated with brief cognitive behavioral treatment for insomnia (bCBTi) between 2005-2010 in a behavioral sleep medicine clinic.

(more ...)

Health Care Utilization

The patients' charts were reviewed to obtain the following health care utilization (HCU) variables: number of office visits at UF/Shands, costs related to office visits (CPT costs), number of medications, and Chronic Disease Score (CDS).37 All HCU variables were measured during the 6 months before treatment began and the 6 months after treatment ended. The number of physician office visits and medications were determined by each patient's electronic medical records at UF/Shands. Only appointments at UF/Shands were available as part of the chart review process. Costs related to office visits were determined by the CPT code of each visit. CPT codes were linked to Medicare relative value payments in the state of Florida (not Fort Lauderdale or Miami) using 2011 reimbursement values (

CDS is an algorithm that estimates health care costs based on medications.37 Weights are given to different medication classes that represent different chronic diseases. Each medication class is associated with a cost based on regression models using automated data from a health maintenance organization. Medications are classified by American Hospital Formulary System (AHFS) categories. Additionally, gender and 10-year age groups are weighted and associated with costs. The CDS represents 6 months of health care costs and provides 3 estimates of HCU— total costs, outpatient costs, and number of primary care visits. Different weights for each medication class, gender, and 10-year age group are used to determine the 3 HCU estimates. CDS has previously been used to calculate HCU in an insomnia sample.11

Treatment Protocol

Patients with insomnia who opted to pursue treatment were asked to complete 2 weeks of daily sleep diaries to establish their baseline sleep behaviors prior to their first treatment session. Patients also completed daily sleep diaries throughout treatment, beginning the first night of treatment until their final appointment. The sleep diary information was entered into a spreadsheet at the beginning of each session. The spreadsheet was used to create graphs depicting the patients' total sleep time and sleep efficiency (see Measures for definitions). The therapist reviewed these graphs with the patient to provide feedback, troubleshoot treatment adherence, and monitor therapy progress. Therapy consisted of brief, multi-component CBTi, and followed a treatment manual developed by the first author. Components of the treatment included sleep education, sleep hygiene, stimulus control therapy, sleep restriction, a brief (10-min) relaxation exercise, and cognitive therapy. An accompanying patient workbook was also used. To prevent patients from working ahead in the workbook, the workbook chapter corresponding to the topics to be covered during a particular session was handed out at the beginning of each session. Treatment was delivered by clinical psychology graduate students and predoctoral interns in UF's American Psychological Association accredited predoctoral and internship programs, respectively. All therapy was supervised by the first author, a licensed clinical psychologist who is also certified in behavioral sleep medicine.

Treatment generally lasted 4 to 6 sessions, depending on the individual needs of the patients, and occurred weekly. Session one consisted of general education about sleep, sleep hygiene,38 and stimulus control.39 Sleep hygiene recommendations were: (1) Avoid caffeine after noon; (2) Avoid exercise within 2 h of bedtime; (3) Avoid nicotine within 2 h of bedtime; (4) Avoid alcohol within 2 h of bedtime; and (5) Avoid heavy meals within 2 h of bedtime. Stimulus control recommendations were: (1) Don't use your bed or bedroom for anything (any time of the day) but sleep (or sex); (2) If you don't fall asleep within 15-20 minutes, leave the bed and do something in another room. Go back to bed only when you feel sleepy again. Clock watching with regard to this rule is not recommended. If you don't fall asleep within 20 minutes upon returning to bed, repeat this instruction as many times as needed; (3) If you wake up during the night and don't fall back to sleep within 20 minutes, follow rule 2 again; and (4) Avoid napping. Session 2 implemented sleep restriction, which involved tailoring the amount of time spent in bed to the participant's reported total sleep time. A time in bed (TIB) prescription was determined by adding 30 min to the participant's average total sleep time at baseline. If this was < 5 h, the TIB prescription was set at 5 hours. The therapist and participant worked together to set regular bed- and wake-times to help the participant follow the TIB prescription. The TIB prescription was reviewed at each session subsequent to session 2 and adjusted based on the average sleep efficiency demonstrated the previous week. If the average sleep efficiency was > 90% and the patient reported daytime fatigue/sleepiness, 30 minutes were added to the TIB prescription. If the average sleep efficiency was < 85%, 30 minutes were subtracted from the TIB prescription. Whenever a change was made in the TIB prescription, the therapist and patient discussed how to adjust bed and wake times to accommodate the new prescription. Session 3 involved a 10-min passive relaxation technique. An audio version of the therapist administering the relaxation exercise was given to the participant for daily home practice use (before bedtime and once during the day). Sessions 4 and 5 focused on cognitive therapy, which helped the patients identify and challenge dysfunctional beliefs and common worries about sleep as well as other thought patterns that interfered with sleep. The final session, often session 4 or 5, reviewed previously learned materials and established a relapse prevention plan.

Treatment Completion

Patients were classified as completers if they attended ≥ 3 sessions of bCBTi. If they only attended 1 or 2 sessions, they were classified as non-completers.


Sleep Diaries

All patients completed a daily sleep diary for 14 consecutive days prior to starting treatment as well as throughout treatment. Sleep diaries provided subjective estimates of sleep onset latency (SOL; time from initial lights-out until sleep onset), wake after sleep onset (WASO; time spent awake after initial sleep onset until last awakening), time in bed (TIB; time from bedtime to final out-of-bed time), total sleep time [TST; computed by subtracting total wake time (SOL + WASO) from time in bed], and sleep efficiency percentage (SE = TST/TIB × 100%). Sleep diaries are a cost-effective method for monitoring sleep patterns and are recommended for the assessment of insomnia.35,40

Patients who completed treatment (≥ 3 sessions; see above) were further classified as responders if their sleep diaries from their last week of treatment indicated: (1) reduced symptoms—SOL and WASO—by 50% compared to baseline, or (2) if baseline SE was < 85%, SE increased by 10% points at end of treatment (e.g., SE 78% → 88%). If not, they were classified as non-responders.

Beck Depression Inventory II

The BDI-II is a 21-item self-report measure with questions scored 0 to 3, with higher responses indicating more severe depressive symptoms. Scores on the 21 items are summed to produce a total score range of 0-63. Scores < 14 indicate minimal depression, while scores ≥ 20 indicate clinically significant depression.41 The BDI-II has a high internal consistency and high test-retest reliability (α = 0.93).

State-Trait Anxiety Inventory, Trait Scale

The STAI-Trait scale is a 20-item self-report measure of tension, apprehension, and physiological signs of stress.42 Items are scored 1-4 and are summed for a total score range of 20-80. Higher scores indicate greater anxiety severity. The STAI-Trait scale has good internal consistency (α = 0.72-0.96) and good test-retest reliability (α = 0.82-0.94).

Statistical Analysis

Data was inspected for extreme outliers (> 5 standard deviations above the mean) for the dependent variables, and this data was excluded from the final analyses. A preliminary inspection of the data revealed a relatively small sample of patients who qualified as treatment responders and a non-normal distribution of data across the outcome variables. Standard normalizing transformations were not effective in correcting the relatively negatively skewed patterns observed in the outcome variables. As is recommended for research with small patient samples,43 and data that is not normally distributed, nonparametric methods were used to analyze treatment outcomes. In this study, we utilized the related samples Wilcoxon signed-rank test to analyze the total difference in HCU pre and post-bCBTi. To examine between group differences, Mann Whitney U tests were used. Effect sizes were calculated using the following formula: Z/√ (# observations over 2 time points). The relative strength of these effect sizes were defined as: r = 0.1 to 0.29 as small, 0.3 to 0.49 as medium, and ≥ 0.5 as large.44


Treatment Completion and Response

Forty-seven patients were classified as non-completers, and 37 were classified as completers. Of those who completed treatment, 32 were further classified as treatment responders and 5 as treatment non-responders. The non-responders group was not included as a separate group in any subsequent analyses, because the small size of the group prohibited drawing any meaningful conclusions from such analyses. Instead, between group analyses compared treatment completers to non-completers, and treatment responders to non-completers.

Group Comparisons

There were no significant group differences between completers, responders, and non-completers in demographic variables including age, gender, race/ethnicity, education, BMI, or number of health problems. In addition, there were no significant group differences in BDI-II score or STAI score among the 3 groups. Between group differences at pre-treatment on HCU measures were assessed. The only variable with a significant difference at pre-treatment was the CDS estimated number of primary care visits between completers and non-completers (p = 0.05), and responders and non-completers (p = 0.03).

Impact on Health Care Utilization

Treatment Completion

Among treatment completers, there was a significant decrease in CPT costs (Mean difference [MΔ] = $195.86, SD = $569.51; p < 0.05, r = 0.26), number of observed office visits (MΔ = 0.37, SD = 4.02; p < 0.05, r = 0.28), and CDS estimated total costs (MΔ = $72.16, SD = $235.81; p < 0.05, r = 0.28). There were also trends toward decreased CDS estimated outpatient costs (MΔ = $38.84, SD = $116.88; p = 0.07, r = 0.25), and CDS estimated number of primary care visits (MΔ = 0.21, SD = 0.61; p = 0.07, r = 0.25). There were no differences in the total number of medications. There were no significant decreases for treatment non-completers (see Table 2). There were no significant between-group differences, and a comparison of relative change from pre-treatment between completers and non-completers did not reveal any significant differences.

Within-groupa differences in health care utilization (HCU) outcome variables at six months pre-treatment (Pre-Tx) and post treatment (Post-Tx) with brief cognitive behavioral treatment for insomnia (bCBTi) by treatment outcome group.


table icon
Table 2

Within-groupa differences in health care utilization (HCU) outcome variables at six months pre-treatment (Pre-Tx) and post treatment (Post-Tx) with brief cognitive behavioral treatment for insomnia (bCBTi) by treatment outcome group.

(more ...)

Treatment Response

Among treatment responders, there was a significant decrease in CPT costs (MΔ = $210.39, SD = $488.06; p < 0.05, r = 0.30) and the number of observed office visits (MΔ = 0.89, SD = 4.47; p < 0.05, r = 0.27). There were also trends toward decreased CDS estimated total costs (MΔ = $54.23, SD = $228.82; p = 0.06, r = 0.28), CDS estimated outpatient costs (MΔ = $22.72, SD = $95.88; p = 0.08, r = 0.25), and CDS estimated number of primary care visits (MΔ = 0.21, SD = 4.47; p = 0.08, r = 0.25). There were no observed differences in the total number of medications (see Table 2). There were no significant between-group differences, and a comparison of relative change from pre-treatment between responders and non-completers did not reveal any significant differences.


This study found changes in HCU and costs in patients who received bCBTi delivered by predoctoral clinical psychology trainees in an outpatient behavioral sleep medicine clinic located within an academic medical center. Of the six HCU and cost variables examined, all but total number of medications either significantly decreased or trended toward a decrease from pre- to post-treatment. Strength of the findings (significant or trending) varied as a function of treatment completion and response. There were no significant changes in any of the variables for patients who did not complete the treatment (attended only 1 or 2 sessions). While between group comparisons were not significant, this clinic-based study may have lacked both sufficient power and adequate control (i.e., non-completers received some treatment) for detecting such differences. Despite these limitations, the findings of this study provide promising preliminary evidence of decreases in HCU and costs for both treatment completers and responders following bCBTi.

On average, treatment completers exhibited a significant decrease of almost $200 dollars in CPT-based health care costs, almost $75 in CDS estimated total health care costs, and approximately one-third of a primary care visit in the 6-month post-treatment period compared to the 6-month pretreatment period. Both CDS estimated outpatient costs and CDS estimated number of primary care visits trended toward significance. Not surprisingly, decreases in health care utilization and costs were larger in patients who responded to treatment. On average, treatment responders exhibited a significant decrease of almost $210 in CPT-based health care costs and almost 1 observed visit (i.e., 0.89 of a visit). CDS estimated total costs, outpatient costs, and primary care visits trended toward significance. In this small sample, it is worthwhile to note that the majority of treatment completers also achieved a treatment response, 86%. Due to the small sample size (n = 5), it was not possible to draw conclusions about the treatment completers who were non-responders.

For both treatment completers and responders, however, the measures of HCU and costs are likely an underestimate of true change. CPT-related costs represent an estimate of the number and type of office visits and are determined by the length and complexity of the appointment (e.g., cardiology exam vs. primary care annual visit). When the number of office visits decrease, or the complexity of visits decrease, costs to the patient will decrease, including direct costs such as co-pays and indirect costs such as travel, and time away from work. Because of the limitations of what CPT actually measures—Medicare reimbursement of medical procedures—it is an underestimate of actual cost-savings to the patient and may better represent a reduction in number of visits and procedures. CPT-related costs do not account for any out of pocket costs for which the patient may be responsible, again lending support that the CPT variable underestimates the direct health care costs of patients.

This underestimation holds true for CDS estimates of HCU. Estimated total health care costs were significantly reduced preto post-bCBTi for completers with a trend towards reduction for treatment responders. The CDS weights of chronic illness have not been adjusted for inflation of health care costs; as costs for chronic disease have increased over the years, so should the weights of the CDS variables. In the current study, an increase of weights for the CDS variables would likely result in even greater estimated costs during each six-month period and potentially a greater reduction of estimated health care costs from pre- to post-treatment. However, it is possible that a constant increase in HCU would result in the same amount of change from pre- to post-treatment.

While both CPT and CDS measures likely underestimate HCU, the difference in magnitude of the two measures is due to what is actually being measured. CPT measures Medicare reimbursement costs of medical procedures and office visits; CPT is a direct cost of health care reimbursement on the provider side and may indirectly measure patient costs. CDS is an estimate of total costs, outpatient costs, and the number of primary care visits based on an algorithm that includes costs linked to chronic illness, based on medication and costs associated with gender and age. CDS is an estimated cost of HCU for an individual, but does not distinguished between patient, provider, direct, and indirect costs. An adjustment for inflation, as well as updates to the medications linked to chronic illness, would allow CDS to provide a more accurate estimate and would likely be more closely correlated to CPT-based costs.

In addition to overall monetary costs, office visits were significantly reduced from pre- to post-CBTi for both treatment completers and responders. This variable is most valuable if interpreted at a health care system and/or population level. Less than one office visit reduction per person over six months may not be a meaningful change at the individual level for one 6-month period. By extrapolating the reduction of office visits to all patients who complete treatment, and then further extrapolating at the system and population level over longer periods of time, the cost-savings become much more meaningful in terms of health economics and health care services. Our findings are consistent with previous research that medication to treat insomnia can reduce HCU and costs, although the costs of the medication can cancel out any savings.14 While the costs of bCBTi may also cancel out any savings in the short term, the benefit of bCBTi over medication is that once bCBTi is complete, the costs of treatment cease. Cost of the diagnostic interview ($125.12) and four 50-minute sessions of bCBTi ($83.32/session × 4 sessions) was approximately $460 in our clinic (based on Florida Medicare relative value payments for 2011). When taking medication, the costs continue throughout treatment and may rise if medication changes, especially from generic to brand name, and as dose and frequency of use increases. Our results are also supported by a recent study using predictive modeling that found the treatment of insomnia, including pharmacotherapy and behavioral therapy, to be cost-effective and cost-saving.45

It is well established that people with insomnia have greater costs than good sleepers. Direct costs of insomnia for older adults have been estimated at $1,143 over a six-month period,5 while indirect costs (e.g., work absenteeism, lost productivity) may exceed $4,000 for those diagnosed with insomnia and $1,200 for those exhibiting insomnia symptoms.4 Furthermore, individuals with insomnia have been shown to incur greater lost productivity costs to employers, including decreased performance and on-the-job safety.46 Individuals with chronic insomnia and a comorbid medical or psychiatric condition have greater health care costs than those with chronic insomnia alone.11 The current study demonstrated that bCBTi can significantly decrease HCU in patients who complete and respond to treatment. Given our means of data collection, we were not able to measure change in indirect costs, but it is plausible that indirect costs, which are often greater than direct costs, would have been reduced as well. The results of this study suggest that the utilization of health care services by individuals with chronic insomnia7 can be reduced to some extent by delivery of bCBTi.

To our knowledge, this is the first study to show a reduction in HCU and costs following bCBTi. Results of this study indicate that not only was a relatively brief intervention effective in reducing insomnia severity, but it may also be capable of making a significant economic impact at the individual and health care systems level. These findings are consistent with previous research on effectiveness of brief behavioral insomnia treatments.30,32,47 Furthermore, a significant treatment response may not be necessary to impact HCU and costs—treatment completers saw a reduction—although a treatment response may augment the cost-savings. Furthermore, the intervention was delivered by clinical psychology doctoral students and predoctoral interns, only some of whom had a background and previous training in insomnia and behavioral sleep medicine interventions. This lends support that bCBTi may be an easily disseminated intervention with minimum training required.32,48

There were several limitations associated with this research study. This sample was small and consisted of primarily middle-aged and older Caucasian adults, many with several comorbid medical conditions such as hypertension, cardiovascular disease, and diabetes. The prevalence of comorbid medical conditions may have been an important confounding variable in accurately estimating change in HCU and costs related to bCBTi; however, in the real world, comorbid disorders are unavoidable and contribute to HCU. Most insomnia cases are comorbid in nature.49 Including patients with comorbid disorders increases ecological validity, as the patient sample is representative of the types of insomnia patients seen in primary care settings— patients with multiple medical and/or psychiatric comorbidities. It is possible that younger populations may see greater HCU reductions post-treatment; conversely, younger persons in better health likely have lower health care utilization and costs in the first place. Thus, it is important to study the use of bCBTi protocols and to examine health care costs in comorbid samples, such as the one used in this study, despite the methodological challenges they pose. Also, a more diverse sample would be needed to make more generalizable conclusions.

Another issue with the current study, as well as studies of HCU and costs in general, is working with monetary data that is rarely normally distributed. As mentioned in the statistical analysis section above, standard normalizing transformations were not effective with our negatively skewed data. As a result of the small sample size and the non-normal distribution of our outcome variables, it was necessary to utilize nonparametric analyses. Results, therefore, show large standard deviations, often larger than the means, due to the negative skewness even with significant outliers removed. However, this reporting of HCU data is consistent with other studies of insomnia and costs.11,12,14,50 Additionally, while significant differences in costs pre- and post-treatment were identified, effect sizes were small. The study was not adequately powered to find meaningful change in HCU, and effect size was influenced by the high variability (i.e., large standard deviations) of the cost data. With the use of nonparametric analyses, it can be more difficult to ascertain the magnitude of the effect being observed. Future research would benefit from a larger sample in order to overcome these limitations.

It is also important to note that this study was a retrospective chart review, and not a randomized controlled trial (RCT). This study used an existing clinical sample and as a result, did not use the extensive, screening protocol often used in RCTs. Also, comparison of completers and responders to non-completers was likely an inadequate control, as the non-completers filled out two weeks of sleep diaries and completed one or two sessions. The use of patients from a specialty clinic housed within an academic medical center is also limiting, because the impact of bCBTi on HCU may differ in other settings, and HCU and cost information were not available for any care received outside our system. Nonetheless, this study provides valuable preliminary data on the ability of bCBTi to reduce HCU and costs. Multiple measures of cost would be beneficial, such as direct costs (e.g., CPT, co-pays), indirect costs (e.g., travel, missed work), estimated costs (e.g., CDS), and cost-effectiveness metrics (e.g., QALY) that can be compared against industry-established standards. Additionally, because data was extracted from patient medical records within an academic medical setting, inconsistencies among patients and missing data limited the statistical analyses performed.

Another limiting factor was that the CDS variable was an estimate of HCU that relied on an algorithm based on medications listed in the patients' electronic medical records. Although the CDS as a measure of HCU is a validated method,37 it is unlikely that all medical conditions and associated costs of a patient can be accurately accounted for by medication alone. Also, not all medications may have been included in the records as many patients see providers from different health care systems. Additionally, the CDS algorithm does not include many over-the-counter medications. More detailed information, such as cost information from insurance plans, co-pays, travel time and travel costs, time away from work, and impact of insomnia on productivity (e.g., absenteeism and presenteeism) are needed for more accurate measurement. As such, the variables reported in this study are likely an underestimate of true HCU and costs. With additional cost information, either a cost-effectiveness analysis and/or cost-utility analysis could be performed to better estimate the economic value of implementing brief cognitive behavioral interventions for insomnia.

Predoctoral level clinicians delivered the treatment in this study. Thus, it is possible that bCBTi delivered by a doctoral-level clinician, or a more experienced behavioral sleep medicine clinician, may lead to greater reduction of HCU and costs. However, if treatment outcome is similar between doctoral-level and predoctoral clinicians, other health care providers (e.g., physician assistants, social workers, nurse practitioners/ nurses) may be well-suited to provide brief, cost-effective treatment as well. However, it is important to recognize that the brief treatment protocol used was developed by and supervised by the first author, a doctoral level clinician with over 16 years' experience delivering behavioral and cognitive behavioral treatments for insomnia. Thus, supervision may be another important component to consider when studying the impact of bCBTi on HCU and costs.

Lastly, the time intervals used to estimate change in HCU and costs may not have been sufficient to see the full economic benefits of treatment. It is plausible that reductions in HCU following treatment completion and response to brief cognitive behavioral treatment for insomnia are not fully realized until well beyond six months. For example, those with medical and/or psychiatric comorbidities might be motivated by their success with bCBTi to pursue treatment of their other medical conditions, which may result in further reduction of costs long term. CBTi has been shown to reduce the use of medication27 and psychiatric symptoms,51 and these factors, not measured in the current study, would likely contribute to even further reductions of HCU and costs following bCBTi. However, it is also possible that given longer periods of time post-treatment, HCU will return to pre-treatment levels. It is important to note that no standard time frame for measuring change in costs has been established, and this area of research, especially in insomnia, is relatively new. One study of costs pre- and post-treatment for seasonal affective disorder measured costs at four weeks before treatment initiation and four weeks before treatment ended.52 A review of the literature found that health care costs are generally measured in 6-month or 12-month time frames.46,14,53 Future research should aim to measure changes after longer time intervals to better evaluate the impact of bCBTi on HCU and costs.


This was not an industry supported study. The authors have indicated no financial conflicts of interest. This study did not involve off-label or investigational use of any pharmaceutical or medical treatment. Work was performed at the University of Florida, Gainesville, FL


Dr. Bramoweth began work on this study as a predoctoral intern at the University of Florida (UF). Thus, his work was partially supported by the Department of Clinical and Health Psychology at UF. He continued this work as a postdoctoral fellow, and as a result, his effort on this study was also partially supported by the VISN 4 Mental Illness Research, Education and Clinical Center (MIRECC, Director: D. Oslin; Pittsburgh Site Director: G. Haas), VA Pittsburgh Healthcare System. The contents do not represent the views of the Department of Veterans Affairs or the United States Government.

Jacob Williams and Caterina Mosti's efforts on this study were partially supported by awards (NIH/NIAMS, AR055160 and AR055160-S1; Christina S. McCrae, PhD, PI). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NIAMS.

Alicia Roth's efforts on this study were partially supported by awards (NIH/NHLBI, HL087831 and HL087831-S1; Christina S. McCrae, PhD, PI). The content is solely the responsibility of the authors and does not necessarily represent the official views of the NHLBI.

The authors offer special thanks to Jason G. Craggs, PhD (database design and development), the predoctoral clinical psychology trainees who provided the brief cognitive behavioral treatment for insomnia, and the patients seen in UF's Insomnia and Behavioral Sleep Medicine clinic.



Ancoli-Israel S, Roth T, authors. Characteristics of insomnia in the United States: results of the 1991 National Sleep Foundation Survey. I. Sleep. 1999;22 Suppl 2:S347–53. [PubMed]


Martin S, Aikens J, Chervin R, authors. Toward cost-effectiveness analysis in the diagnosis and treatment of insomnia. Sleep Med Rev. 2004;8:63–72. [PubMed]


Stoller M, author. Economic effects of insomnia. Clin Ther. 1994;16:873–97. [PubMed]


Daley M, Morin CM, LeBlanc M, Gregoire JP, Savard J, authors. The economic burden of insomnia: direct and indirect costs for individuals with insomnia syndrome, insomnia symptoms, and good sleepers. Sleep. 2009;32:55–64. [PubMed]


Ozminkowski RJ, Wang S, Walsh JK, authors. The direct and indirect costs of untreated insomnia in adults in the United States. Sleep. 2007;30:263–73. [PubMed]


Sarsour K, Kalsekar A, Swindle R, Foley K, Walsh JK, authors. The association between insomnia severity and healthcare and productivity costs in a health plan sample. Sleep. 2011;34:443–50. [PubMed]


Leger D, Guilleminault C, Bader G, Levy E, Paillard M, authors. Medical and socio-professional impact of insomnia. Sleep. 2002;25:625–9. [PubMed]


Roy AN, Smith M, authors. Prevalence and cost of insomnia in a state Medicaid fee-for-service population based on diagnostic codes and prescription utilization. Sleep Med. 2010;11:462–9. [PubMed]


Taylor DJ, Mallory LJ, Lichstein KL, Durrence HH, Riedel BW, Bush AJ, authors. Comorbidity of chronic insomnia with medical problems. Sleep. 2007;30:213–8. [PubMed]


Asche CV, Joish VN, Camacho F, Drake CL, authors. The direct costs of untreated comorbid insomnia in a managed care population with major depressive disorder. Curr Med Res Opin. 2010;26:1843–53. [PubMed]


Bramoweth AD, Taylor DJ, authors. Chronic insomnia and health care utilization in young adults. Behav Sleep Med. 2012;10:106–21. [PubMed]


Tian H, Abouzaid S, Gabriel S, Kahler KH, Kim E, authors. Resource utilization and costs associated with insomnia treatment in patients with major depressive disorder. Prim Care Companion CNS Disord. 2012;14(5).


Rieu-Werden ML, Sutton-Skinner KM, Romeo SA, Fromson JA, Birnbaum RJ, authors. Underutilization of nonpharmacologic interventions for insomnia by healthcare providers. In: New Orleans, LA: Annual Meeting of the American Psychiatric Association, 2010.


Jhaveri M, Seal B, Pollack M, Wertz D, authors. Will insomnia treatments produce overall cost savings to commercial managed-care plans? A predictive analysis in the United States. Curr Med Res Opin. 2007;23:1431–43. [PubMed]


Snedecor SJ, Botteman MF, Bojke C, Schaefer K, Barry N, Pickard AS, authors. Cost-effectiveness of eszopiclone for the treatment of adults with primary chronic insomnia. Sleep. 2009;32:817–24. [PubMed]


Uchiyama M, Hamamura M, Kuwano T, Nishiyama H, Nagata H, Uchimura N, authors. Evaluation of subjective efficacy and safety of ramelteon in Japanese subjects with chronic insomnia. Sleep Med. 2011;12:119–26. [PubMed]


Hajak G, Hedner J, Eglin M, et al., authors. A 2-week efficacy and safety study of gaboxadol and zolpidem using electronic diaries in primary insomnia outpatients. Sleep Med. 2009;10:705–12. [PubMed]


Krystal AD, Durrence HH, Scharf M, et al., authors. Efficacy and safety of doxepin 1 mg and 3 mg in a 12-week sleep laboratory and outpatient trial of elderly subjects with chronic primary insomnia. Sleep. 2010;33:1553–61. [PubMed]


Krystal AD, Lankford A, Durrence HH, et al., authors. Efficacy and safety of doxepin 3 and 6 mg in a 35-day sleep laboratory trial in adults with chronic primary insomnia. Sleep. 2011;34:1433–42. [PubMed]


Ancoli-Israel S, Krystal AD, McCall WV, et al., authors. A 12-week, randomized, double-blind, placebo-controlled study evaluating the effect of eszopiclone 2 mg on sleep/wake function in older adults with primary and comorbid insomnia. Sleep. 2010;33:225–34. [PubMed]


Edinger JD, Wohlgemuth W, AD K, Rice J, authors. Behavioral insomnia therapy for fibromyalgia patients: a randomized clinical trial. Arch Intern Med. 2005;165:2527–35. [PubMed]


Currie S, Wilson K, Pontefract A, deLaplante L, authors. Cognitive-behavioral treatment of insomnia secondary to chronic pain. J Consult Clin Psychol. 2000;68:407–16. [PubMed]


Quesnel C, Savard J, Simard S, Ivers H, Morin C, authors. Efficacy of cognitive-behavioral therapy for insomnia in women treated for nonmetastatic breast cancer. J Consult Clin Psychol. 2003;71:189–200. [PubMed]


Smith MT, Perlis ML, Park A, et al., authors. Comparative meta-analysis of pharmacotherapy and behavior therapy for persistent insomnia. Am J Psychiatry. 2002;159:5–11. [PubMed]


Lichstein KL, Morin CM, authors. Treatment of Late-Life Insomnia. Thousand Oaks, CA: Sage, 2010.


Edinger JD, Wohlgemuth WK, Radtke RA, Marsh GR, Quillian RE, authors. Cognitive behavioral therapy for treatment of chronic primary insomnia: a randomized controlled trial. JAMA. 2001;285:1856–64. [PubMed]


Dolan DC, Taylor DJ, Bramoweth AD, Rosenthal LD, authors. Cognitive-behavioral therapy of insomnia: a clinical case series study of patients with co-morbid disorders and using hypnotic medications. Behav Res Ther. 2010;48:321–7. [PubMed]


Manber R, Carney C, Edinger J, et al., authors. Dissemination of CBTI to the non-sleep specialist: protocol development and training issues. J Clin Sleep Med. 2012;8:209–18. [PubMed]


McCrae CS, Dautovich ND, Dzierzewski JM, authors; Sateia M, Buysse D, editors. Short-term and group interventions for insomnia. Insomnia: diagnosis and treatment. New York: Informa Healthcare, 2010.


Buysse DJ, Germain A, Moul DE, et al., authors. Efficacy of brief behavioral treatment for chronic insomnia in older adults. Arch Intern Med. 2011;171:887–95. [PubMed Central][PubMed]


Edinger JD, Sampson WS, authors. A primary care “friendly” cognitive behavioral insomnia therapy. Sleep. 2003;26:177–82. [PubMed]


McCrae CS, McGovern R, Lukefahr R, Stripling AM, authors. Research Evaluating Brief Behavioral Sleep Treatments for Rural Elderly (RESTORE): a preliminary examination of effectiveness. Am J Geriatr Psychiatry. 2007;15:979–82. [PubMed]


American Academy of Sleep Medicine. International Classification of Sleep Disorders. Diagnostic and Coding Manual. 2nd ed. Westchester, IL: American Academy of Sleep Medicine, 2005.


American Psychatric Association. Diagnostic and Statistical Manual of Mental Disorders. 4th ed. text revision. Washington, DC: American Psychatric Association, 2000.


Edinger JD, Bonnet MH, Bootzin RR, et al., authors. Derivation of research diagnostic criteria for insomnia: report of an American Academy of Sleep Medicine Work Group. Sleep. 2004;27:1567–96. [PubMed]


Taylor DJ, Lichstein KL, Durrence HH, Reidel BW, Bush AJ, authors. Epidemiology of insomnia, depression, and anxiety. Sleep. 2005;28:1457–64. [PubMed]


Clark DO, Von Korff M, Saunders K, Baluch WM, Simon GE, authors. A chronic disease score with empirically derived weights. Med Care. 1995;33:783–95. [PubMed]


Riedel BW, author; Lichstein KL, Morin CM, editors. Sleep hygiene. Treatment of late-life insomnia. Thousand Oaks, CA: Sage, 2000.


Bootzin R, author. A stimulus control treatment for insomnia. In: Paper presented at: 80th Annual Convention of the American Psychological Association,1972.


Schutte-Rodin S, Broch L, Buysse D, Dorsey C, Sateia M, authors. Clinical guideline for the evaluation and management of chronic insomnia in adults. J Clin Sleep Med. 2008;4:487–504. [PubMed]


Beck AT, Steer RA, Brown G, authors. Manual for the Beck Depression Inventory-II. San Antonio, TX: Psychological Corporation, 1996.


Spielberger CD, Gorsuch RL, Lushene R, Vagg PR, Jacobs GA, authors. Manual for the State-Trait Anxiety Inventory. Palo Alto, CA: Consulting Psychologists Press, 1983.


Blair RC, Higgins JJ, authors. A comparison of the power of the paired sample t-test to that of Wilcoxon's signed-ranks test under various population shapes. Psychol Bull. 1985;97:119–28.


Cohen J, author. A power primer. Psychol Bull. 1992;112:155–9. [PubMed]


O'Keeffe KM, Gander PH, Scott WG, Scott HM, authors. Insomnia treatment in New Zealand. N Z Med J. 2012;125:46–59.


Rosekind MR, Gregory KB, Mallis MM, Brandt SL, Seal B, Lerner D, authors. The cost of poor sleep: workplace productivity loss and associated costs. J Occup Environ Med. 2010;52:91–8. [PubMed]


Germain A, Moul D, Franzen P, et al., authors. Effects of a brief behavioral treatment for late-life insomnia: preliminary findings. J Clin Sleep Med. 2006;2:403–6. [PubMed]


Troxel WM, Germain A, Buysse DJ, authors. Clinical management of insomnia with brief behavioral treatment (BBTI). Behav Sleep Med. 2012;10:266–79. [PubMed Central][PubMed]


McCrae CS, Lichstein KL, authors. Secondary insomnia: diagnostic challenges and intervention opportunities. Sleep Med Rev. 2001;5:47–61. [PubMed]


Shahly V, Berglund PA, Coulouvrat C, et al., authors. The associations of insomnia with costly workplace accidents and errors: results from the America Insomnia Survey. Arch Gen Psychiatry. 2012;69:1054–63. [PubMed]


Ulmer CS, Edinger JD, Calhoun PS, authors. A multi-component cognitive-behavioral intervention for sleep disturbance in veterans with PTSD: a pilot study. J Clin Sleep Med. 2011;7:57–68. [PubMed]


Cheung A, Dewa C, Michalak EE, et al., authors. Direct health care costs of treating seasonal affective disorder: a comparison of light therapy and fluoxetine. Depression Res Treat. 2012;2012:628434.


Konig HH, Leicht H, Bickel H, et al., authors. Effects of multiple chronic conditions on health care costs: an analysis based on an advanced tree-based regression model. BMC Health Serv Res. 2013;13:219. [PubMed Central][PubMed]