Cost-Effectiveness of Prolonged Exposure Therapy Versus Pharmacotherapy and Treatment Choice in Posttraumatic Stress Disorder (the Optimizing PTSD Treatment Trial): A Doubly Randomized Preference Trial

Related blog

Cost-Effectiveness of Prolonged Exposure Therapy Versus Pharmacotherapy and Treatment Choice in Posttraumatic Stress Disorder (the Optimizing PTSD Treatment Trial): A Doubly Randomized Preference Trial

Quang A. Le, PharmD, PhD; Jason N. Doctor, PhD; Lori A. Zoellner, PhD;and Norah C. Feeny, PhD

Posttraumatic stress disorder (PTSD) is a chronic and debilitating condition, with lifetime prevalence rates ranging from 8%-14% of the US population.1 Posttraumatic stress disorder also has a huge financial impact; Greenberg and colleagues2 reported that through work impairment, hospitalization, and health visits, PTSD was more costly than any other anxiety disorder. Among the 1.64 million veterans returning from the Operation Enduring Freedom and Operation Iraqi Freedom, it is estimated that approximately 300,000 individuals currently suffer from PTSD or major depression, potentially costing $4.0 to $6.2 billion in a 2-year time frame.3 Providing evidence-based treatment for those in need could reduce these costs by as much as 27%.3 These considerations highlight the substantial impact of PTSD and the need for cost-effective interventions.

A significant limitation of traditional randomized clinical trials is that strong preferences for (or against) one treatment may influence outcome, willingness to receive treatment, or both.4,5 For mental health treatments in particular, preference for treatment may moderate treatment efficacy. If preference is a moderator, then pharmacotherapy or psychotherapy may be maximally effective for those who prefer it and minimally effective for those who do not. To address this limitation in traditional randomized clinical trials, a doubly randomized preference trial design has been proposed. In a doubly randomized preference trial, participants are randomized either into a “randomization arm,” within which treatment is randomized, or into a “preference arm,” within which participants get to choose their treatment, thus revealing a preference for treatment. With data from a doubly randomized preference trial, the effect of choice on outcome may be separated from the effect of randomization to treatment.5

Preliminary and updated results from the Optimizing PTSD Treatment Trial (OPT) indicated that patients with PTSD responded well to prolonged exposure therapy or pharmacotherapy with sertraline.6,7 In addition, there were effects on clinical outcomes for patients who received their preferred treatment.6,7 To our knowledge, not a single study has either established methods or evaluated the cost-effectiveness of treatment preference from data derived from a doubly randomized preference trial design. In this study, we examined the cost-effectiveness of the treatment effect of prolonged exposure therapy, overall treatment preference, preference effect of choosing prolonged exposure therapy, and preference effect of choosing pharmacotherapy with sertraline from the US societal perspective.

METHOD

Study Design and Participants

The OPT trial was a hybrid efficacy-effectiveness trial designed to compare the efficacy of prolonged exposure therapy versus pharmacotherapy with sertraline in treating PTSD and the impact of choice on both short-term and longer term outcomes. The OPT was a 2-site study conducted at the University of Washington, Seattle, and Case Western Reserve University, Cleveland, Ohio. The OPT trial included patients between ages 18 and 65 years who were currently diagnosed with primary PTSD based on the Diagnostic and Statistical Manual of Mental Disorders, Fourth Edition (DSM-IV) criteria, with a minimum duration of 12 weeks since the traumatic event, and diagnosed using the PTSD Symptom Scale-interview version (PSS-I).8 The study was conducted from July 2004 to January 2009. It was approved by the relevant institutional review boards, and all patients were provided written informed consent. The study was registered on ClinicalTrials.gov (identifier: NCT00127673).

Patients were randomly assigned to either choice of treatment or no choice of treatment conditions, using a doubly randomized preference trial design (Figure 1). In the choice condition, patients chose either prolonged exposure therapy or pharmacotherapy with sertraline. In the no-choice condition, patients were randomly assigned to either prolonged exposure therapy or sertraline. Patients received 10 weeks of acute treatment. All willing patients were assessed by independent evaluators at pretreatment and posttreatment and at 3-, 6-, and 12-month follow-up.

Prolonged Exposure Therapy

Patients who chose or who were randomized to prolonged exposure therapy received up to 10 weekly 90- to 120-minute treatment sessions by trained MA-, MS-, or PhD-level psychologists. As is standard, treatment procedures included education about common reactions to trauma, breathing retraining, prolonged (repeated) exposure to trauma memories, repeated in vivo exposure to situations the patient is avoiding due to trauma-related fear, and discussion of thoughts and feelings related to exposure exercises. Each session begins with a review of the homework assignment and presentation of the agenda for that session and ends with the assignment of homework utilizing a standardized prolonged exposure therapy manual.9 Treatment sessions were videotaped or audiotaped. Prolonged exposure therapy supervision occurred weekly at each site, including case discussion and tape review. Trained outside raters reviewed 10% of videotapes, assessing essential treatment components and protocol violations. Prolonged exposure therapy providers completed 90% of essential components. No protocol violations were observed. Prolonged exposure therapy sessions were also rated for therapist competence (eg, engaged in interactive exchange with client) on a 3-point scale (1 = inadequate, 3 = adequate or better). Overall therapist competence was very good (mean = 2.73, SD = 0.32).

• Giving patients a choice between prolonged exposure therapy and sertraline is a cost-effective shared decision-making strategy.
• In the absence of choice, or if the patient asks for guidance on which to choose, prolonged exposure therapy is cost-effective relative to sertraline.
• Recommending treatments that patients do not find favorable may be a barrier to successful treatment.

Pharmacotherapy With Sertraline

Patients who chose or who were randomized to sertraline were assigned a board-certified psychiatrist throughout the study who monitored clinical state and medication effects and offered general support. Patients were seen weekly up to 30 minutes of medication management sessions for up to 10 weeks, with the first session lasting 45 minutes. The goal was to achieve a maximum target dosage of 200 mg/d of sertraline, if tolerated and indicated using a standardized titration algorithm.10,11 Patients were assessed for response and side effects at each visit, and sertraline was also titrated to achieve the target dose. Treatment sessions were videotaped or audiotaped. The administration of sertraline was overseen by a medical director at each site. Integrity ratings were based on Marshall et al.12 Trained raters reviewed 10% of videotapes, assessing essential treatment components and protocol violations. For essential components, pharmacotherapy with sertraline providers completed 96%. No protocol violations were observed.

Health Utility Measures

The self-administered EuroQoL Questionnaire-5 dimensions (EQ-5D)13 was also prospectively collected to assess preference-based health-related quality of life measures of health status at pretreatment and follow-ups throughout the trial. The EQ-5D is one of the most widely used generic, preference-based, health-related quality of life measures to estimate population-based health utilities. The US population scoring system of the EQ-5D was based on the work of Shaw and colleagues.14 In cost-effectiveness analyses, more specifically in cost-utility analyses, the standardized outcome is measured in terms of quality-adjusted life-year (QALY), which is a combined measure of quantity (in years) and quality (in population-based health utility) of life.

Health Service Utilization and Costs

All costs were based on 2012 US dollars; costs before 2012 were adjusted using the Consumer Price Index, Medical Care Service component.15 Frequencies and durations of health service utilization was captured using the Cornell Services Index.16 At each follow-up, patients were interviewed to answer questions about their use of outpatient (categorized into general medical health, mental health, substance abuse, and professional supportive services), inpatient, and emergency department services as well as pharmacy and other supportive services. We estimated the total cost per patient for each treatment arm as a sum of both direct costs (intervention costs, costs of outpatient services, inpatient hospitalization, emergency department visits, pharmacy services, and nonmedical services) and indirect cost from productivity losses.

Intervention cost in the prolonged exposure therapy arm was estimated on the basis of cost of psychologist supervisory time17 spent for therapy sessions as well as the consulting time between and/or after treatment. In the pharmacotherapy arm, intervention cost included cost of psychiatrist supervisory time17 for medication management and clinical consultation and cost of sertraline using its wholesale acquisition cost (WAC) from the Drug Topics Red Book18 based on the total dose received during 10-week treatment. The numbers of outpatient clinic visits, inpatient hospitalizations, and emergency department visits were recorded at each follow-up. The average costs per visit for the health care services were derived from the Healthcare Cost and Utilization Project (HCUP) and Medical Expenditure Panel Survey (MEPS) statistics,19,20 Veterans Affairs Outpatient Average Cost Dataset for VA Care,21 and relevant published studies.17,22 Costs of outpatient, inpatient, and emergency department visit services were estimated by multiplying the number of visit(s) by the average costs per visit. Medication use, including names of medication with their dosages, duration of use, and prescribers’ information, was obtained at each follow-up visit. Average cost of pharmacy services was estimated by using the WAC price18 plus a pharmacy dispensing fee.23 Cost of nonmedical services was estimated by using the self-reported commuting time to and from the clinic and time required for caregiving and/or personal home aide services with the appropriate average hourly wages from the Bureau of Labor Statistics.17

Indirect cost was estimated as costs of productivity losses due to time spent in weekly treatment sessions and travel time to and from the clinic multiplied by the 2012 national average wage.17

Statistical Analyses

All analyses were performed on an intent-to-treat approach including all 200 PTSD patients. Incremental cost-effectiveness ratios (ICERs) were estimated from the US societal perspective in a period of 12 months. The ratio was calculated as the mean difference of total cost per patient divided by the mean difference of QALY between interventions and then compared with the societal willingness-to-pay amounts for a QALY (ie, the societal valuations of expenditures for medical interventions typically between ₤20,000-₤40,000 in the United Kingdom and $50,000-$100,000 or more in the United States24-28) to determine whether or not an intervention is cost-effective. We used generalized linear models to compare the estimated total costs per patient and total QALY in 12 months between interventions.29 The models included demographic characteristics, baseline self-reported health status (EQ-5D), medical comorbidities, medication use, and health services utilization.

The following formula was used to estimate ICERs: ICER = ([total cost in X] – [total cost in Y])/([total QALY in X] – [total QALY in Y]).

For the treatment effect of prolonged exposure therapy as compared to pharmacotherapy with sertraline, X = patients who received prolonged exposure therapy, and Y = patients who received pharmacotherapy with sertraline.

For the overall treatment-preference effect, X = patients who were randomized to the choice arm, and Y = patients who were randomized in the no-choice or randomization arm.

For the preference effect of choosing prolonged exposure therapy, X = patients who received their preferred prolonged exposure therapy, and Y = patients who received prolonged exposure therapy but preferred pharmacotherapy with sertraline.

For the preference effect of choosing pharmacotherapy with sertraline, X = patients who received their preferred pharmacotherapy with sertraline, and Y = patients who received pharmacotherapy with sertraline but preferred prolonged exposure therapy.

Sensitivity Analyses

To quantify uncertainty around the ICERs, we examined the joint density of cost and QALY differences. In addition to base-case analyses for the most likely scenarios, we performed 1-way sensitivity analysis to test the robustness of the ICERs using the low (−25%) and high (+25%) estimates of the costs per visit for outpatient, inpatient, and emergency department services. Further, the 95% confidence intervals (CIs) for the total costs and QALYs were estimated using a nonparametric bootstrapping method with 5,000 replications.30

Missing Data

All missing values were imputed by using a multiple imputation method with STATA release 12.1.31 In the imputation models, we included 1-period lagged values of the relevant costs being imputed, demographic characteristics, baseline self-reported health status (EQ-5D), medical comorbidities, medication use, and health services utilization. We created 20 imputation data sets, and the imputation procedures were carried out separately for each intervention.

RESULTS

Table 1 briefly reports patients’ demographics at baseline. In the overall sample (N = 200), patients were primarily female (76.0%), white (65.0%), middle aged (37.5 years, SD = 11.3), and not college educated (66.5%), with approximately half earning less than $20,000 per year (51.5%). In all, 31.0% reported adult sexual assault, 24.0% reported childhood assault, 22.5% reported adult assault (nonsexual assault), 13.5% reported an accident or natural disaster, 6.5% reported death of or violence to a loved one, and 2.5% reported combat or war as the primary trauma from which PTSD was assessed. The mean (SD) time elapsed since the traumatic event was 11.97 (12.69) years.

Estimated costs of health services utilization from the US societal perspective and total QALY during the 12-month period for the treatment effect of prolonged exposure therapy (prolonged exposure therapy vs pharmacotherapy with sertraline), overall effect of the treatment preference (choice of treatment vs no choice of treatment), preference effect of choosing prolonged exposure therapy, and preference effect of choosing pharmacotherapy with sertraline are presented in Tables 2 and 3.

For the treatment effect, the mean total cost of prolonged exposure therapy is slightly less than pharmacotherapy with sertraline ($7,033 vs $8,653), although this measure was not statistically significant. Similarly, for the overall effect of treatment preference, the mean total costs of both choice ($6,156) and no choice ($7,788) of treatment were not significantly different. Overall, 77.8%-86.7% of the total cost came from utilization of outpatient services, inpatient hospitalization, and emergency department visits. For the treatment effect, the QALY gain was significantly higher in the prolonged exposure therapy group than the pharmacotherapy with sertraline group (95% CI of the mean QALY difference ranged from 0.026 to 0.167, P < .01) (Table 2).

For the preference effect of choosing prolonged exposure therapy, the mean total costs of patients who received their preferred prolonged exposure therapy and those who received prolonged exposure therapy but preferred pharmacotherapy with sertraline were $4,497 and $9,441, respectively. For the preference effect of choosing pharmacotherapy with sertraline, the mean total costs of patients who received their preferred pharmacotherapy with sertraline and those who received pharmacotherapy with sertraline but preferred prolonged exposure therapy were $8,966 and $7,561, respectively. Similarly, 67.6%-88.2% of the total cost came from utilization of outpatient services, inpatient hospitalization, and emergency department visits (Table 3).

Using the base-case analyses for the most likely scenarios, low (−25%) and high (+25%) estimates of costs per visit for outpatient services, inpatient hospitalization, and emergency department visits, Table 4 reports mean incremental costs, incremental QALYs, 95% CIs, ICERs, as well as percentage of the bootstrap samples that resulted in a cost-effective strategy during a 12-month period for the treatment effect of prolonged exposure therapy, overall effect of treatment preference, preference effect of choosing prolonged exposure therapy, and preference effect of choosing pharmacotherapy with sertraline, adjusted for baseline and clinical characteristics.

For the treatment effect, prolonged exposure therapy was a dominant strategy (less costly but produced more QALYs) with the base-case analysis. The scatter plot (Figure 2A) shows that a majority of the bootstrap samples fell in the southeast quadrant, resulting in a dominant strategy for prolonged exposure therapy. At the willingness-to-pay amount of $100,000/QALY, the probability that prolonged exposure therapy is cost-effective compared with pharmacotherapy with sertraline was 93.2% (Figure 2A, Table 4).

For the overall effect of treatment preference, choice of treatment (ie, those who received the preferred treatment) was a dominant strategy with the base-case analysis. The scatter plot (Figure 2B) also indicates that a majority of the bootstrap samples were in the dominant strategy for prolonged exposure therapy. At the willingness-to-pay amount of $100,000/QALY, the probability that choice of treatment is cost-effective compared with no choice of treatment was 87.0% (Figure 2B, Table 4).

At the base-case analysis, the preference effect of choosing prolonged exposure therapy produced less cost (incremental cost was −$1,393) and fewer QALYs (incremental QALY was −0.048), thus resulting in an ICER of $29,021/QALY for the preference effect of choosing prolonged exposure therapy (Table 4). The preference effect of choosing pharmacotherapy with sertraline yielded more cost (incremental cost was $714) and more QALYs (incremental QALY was 0.027), resulting in an ICER of $26,444/QALY for the preference effect of choosing pharmacotherapy with sertraline (Table 4). In both cases, 1-way sensitivity analyses with low- and high-estimates of costs also resulted in close estimates as in the base-case analysis.

DISCUSSION

To our knowledge, this is the first study to have evaluated not only the cost-effectiveness of the treatment effect of cognitive-behavioral therapy (CBT) versus pharmacotherapy in patients with PTSD but also the cost-effectiveness of the overall treatment preference and the preference effects of choosing the preferred treatments from the US societal perspective. In general, prolonged exposure therapy and being given a choice of treatment were each more effective (produced more QALYs) at a lower cost than pharmacotherapy with sertraline and when given no choice of treatment (ie, received assigned treatment), respectively. In addition, the ICERs of patients who preferred and received prolonged exposure therapy versus those who did not prefer but received prolonged exposure therapy and of patients who received their preferred pharmacotherapy with sertraline versus those who received pharmacotherapy with sertraline but did not prefer pharmacotherapy with sertraline were $29,021/QALY and $26,444/QALY, respectively. Our study indicates that prolonged exposure therapy and receipt of preferred treatment in PTSD patients are cost-effective strategies relative to pharmacotherapy with sertraline and the receipt of assigned treatment, respectively.

Previous studies32-35 have showed similar estimates for the average annual cost in patients with PTSD. Of the annual total cost, the majority (67.6%-88.2%) was due to utilization of outpatient services, inpatient hospitalization, and emergency department visits. Interestingly, patients with PTSD who received prolonged exposure therapy had significantly higher health-related quality-of-life measures than those who received pharmacotherapy with sertraline. We also included the low and high estimates of major components of the total cost (outpatient services, inpatient hospitalization, and emergency department visits) in our analyses to examine the robustness of the results. Overall, the results were somewhat less sensitive to low and high cost estimates and rather consistent with the base-case analysis.

There have been few cost-effectiveness studies on treatments and/or preference treatments of PTSD. Using population-based estimates, Issakidis et al36 found that evidence-based optimal care for PTSD (also including other anxiety disorders such as panic disorder with and without agoraphobia, social phobia, and generalized anxiety disorder) had both clinical and economic benefits against current care from the Australian health care’s perspective. More recently, Gospodarevskaya and Segal37 used Markov models to assess the cost-effectiveness of CBT, combination of CBT and pharmacotherapy with sertraline, and nondirective counseling for treatment of PTSD in sexually abused children in Australia. This study showed that combination of CBT and sertraline was the most cost-effective therapy (ICER = 17,520 Australian dollars/QALY) in PTSD sexually abused children.37 Joesch et al38 examined the cost-effectiveness of treatment preference in all anxiety disorders (including panic, generalized anxiety, social anxiety, and PTSD) in terms of the Coordinated Anxiety Learning and Management (CALM) treatment model, which allows choice of treatment modality and provision of additional treatment when needed, based on a primary care effectiveness randomized controlled trial in the United States.39 The study suggested that the CALM treatment model was cost-effective as opposed to usual care treatment in a population of patients with anxiety disorders when the willingness-to-pay amount is set at $50,000/QALY.38 It is consistent with our findings regarding the treatment preferences for PTSD, ie, giving patients an opportunity to choose their preferred treatment tends to be more cost-effective than assigning treatment to patients. It should be noted that the CALM study was not a doubly randomized preference trial and, accordingly, is unable to support strong causal statements about costs of patient preference.

The OPT trial utilized a methodologically sound doubly randomized preference design. Randomized efficacy trials make an assumption of equipoise; that is, treatments are generally perceived as equally desirable. If this is not the case, inferences are then limited to part of the population willing to be randomized, potentially excluding an important subset.5 Notably, for treatments that require strong motivation or effort on the part of participation (eg, prolonged exposure therapy), when randomization to such a treatment that is not desired occurs, randomization is undermined and complicates causal inferences.5 A doubly randomized preference trial helps to address this problem and is useful in studying the causal effect of choosing treatment rather than being randomly assigned, and thus can provide important information for assessing effectiveness of an intervention under real-world conditions.

Interpretation of these findings should be done with consideration of some additional limitations. First, study patients were largely female. Therefore, the results might not generalize to the general population of patients with PTSD, the veteran population in particular, as gender might play an important role in health services utilization. Second, we relied on structured interviews and patient self-report at follow-up for most measures, which might not reflect the true costs of health services utilization as if collected directly through medical records or hospital cost accounting systems. Finally, as with large-scale clinical psychotherapy and pharmacotherapy trials, some patients did not complete treatment or had missing follow-up data. This limitation was addressed by using an intent-to-treat approach and a widely accepted multiple-imputation method.

Despite the mentioned limitations, the current study has several strengths. First, it is the first economic analysis that has examined both the treatment effect of prolonged exposure therapy compared with pharmacotherapy with sertraline and the treatment-preference effects of received preferred therapies versus received nonpreferred therapies in patients with chronic PTSD. Second, the study was based on a relatively large, diverse, and clinical complex sample from a methodologically sound doubly randomized preference trial; thus, our results would be less susceptible to potential bias or confounding. Third, we evaluated not only the point estimate results but also the uncertainty (95% CI estimation) and variation (low- and high-cost estimate analyses) in incremental costs and QALYs.

In conclusion, our findings indicate that giving patients with PTSD a choice of treatment appears to be cost-effective. When choice is not possible, prolonged exposure therapy may provide a cost-effective option as compared to pharmacotherapy with sertraline.

Drug names: sertraline (Zoloft and others).

Author affiliations: Department of Social and Administrative Sciences, Western University of Health Sciences, Pomona, California (Dr Le); Department of Clinical Pharmacy, Pharmaceutical Economics and Policy/Leonard D. Schaeffer Center for Health Policy and Economics, University of Southern California, Los Angeles (Dr Doctor); Department of Psychology/Center for Anxiety and Traumatic Stress, University of Washington, Seattle (Dr Zoellner); and Department of Psychological Sciences, Case Western Reserve University, Cleveland, Ohio (Dr Feeny).

Author contributions: As principal investigators, Drs Zoellner and Feeny designed, coordinated, and collected data for the clinical trial. Drs Le and Doctor performed statistical analyses, interpreted results, and drafted the manuscript of this study. All authors reviewed and approved the final manuscript. Dr Le had full access to all the data in this study and had final responsibility for the decision to submit the manuscript for publication.

Potential conflicts of interest: Dr Doctor served as principal consultant for Precision Health Economics. Drs Le, Zoellner, and Feeny report no conflicts of interest.

Funding/support: This research was made possible by grants R01MH066347 and R01MH066348 from the National Institute of Mental Health (“Effectiveness of PTSD Treatment: Prolonged Exposure Therapy vs Zoloft”). Dr Le was supported by a New Investigator Award from American Association of Colleges of Pharmacy (AACP). Sertraline was supplied by Pfizer, Inc, New York, New York.

Role of the sponsor: The sponsor of the study had no role in study design, data collection, data analysis, data interpretation, or writing of the manuscript. Pfizer, Inc, had no role in study design, data collection, data analysis, data interpretation, or writing of the manuscript.

Previous presentation: Primary findings of this study were presented in part at the Annual Meeting of the Society of Medical Decision Making; October 19, 2012; Phoenix, Arizona.

REFERENCES

1. Kessler RC, Sonnega A, Bromet E, et al. Posttraumatic stress disorder in the National Comorbidity Survey. Arch Gen Psychiatry. 1995;52(12):1048-1060. PubMed doi:10.1001/archpsyc.1995.03950240066012

2. Greenberg PE, Sisitsky T, Kessler RC, et al. The economic burden of anxiety disorders in the 1990s. J Clin Psychiatry. 1999;60(7):427-435. PubMed doi:10.4088/JCP.v60n0702

3. Eibner C, Ringel JS, Kilmer B, et al. The cost of post-deployment mental health and cognitive conditions. In: Tanielian T, Jaycox LH, eds. Invisible Wounds of War: Psychological and Cognitive Injuries, Their Consequences, and Services to Assist Recovery. Santa Monica, CA: RAND; 2008.

4. Torgerson DJ, Sibbald B. Understanding controlled trials: what is a patient preference trial? BMJ. 1998;316(7128):360. PubMed doi:10.1136/bmj.316.7128.360

5. Long Q, Little RJ, Lin XH. Causal inference in hybrid intervention trials involving treatment choice. J Am Stat Assoc. 2008;103(482):474-484. doi:10.1198/016214507000000662

6. Feeny NC, Zoellner LA. Anxiety treatment—new research findings for the clinician. Talk presented at: 2010 American Psychiatric Association Annual Meeting Symposium 12; May 22, 2010; New Orleans, LA.

7. Youngstrom EA, Feeny NC, Zoellner LA, et al. Doubly randomized preference trial in PTSD-prolonged exposure versus sertraline (Symposium 5032). Talk presented at: American Psychological Association Annual Convention; August 3, 2013; Honolulu, HI.

8. Foa E, Riggs DS, Dancu CV, et al. Reliability and validity of a brief instrument for assessing post-traumatic stress disorder. J Trauma Stress. 1993;6(4):459-473. doi:10.1002/jts.2490060405

9. Foa EB, Hembree EA, Dancu CV. Prolonged Exposure (PE) Manual-Revised Version. Philadelphia, PA: University of Pennsylvania; 2002.

10. Brady K, Pearlstein T, Asnis GM, et al. Efficacy and safety of sertraline treatment of posttraumatic stress disorder: a randomized controlled trial. JAMA. 2000;283(14):1837-1844. PubMed doi:10.1001/jama.283.14.1837

11. Davidson J, Pearlstein T, Londborg P, et al. Efficacy of sertraline in preventing relapse of posttraumatic stress disorder: results of a 28-week double-blind, placebo-controlled study. Am J Psychiatry. 2001;158(12):1974-1981. PubMed doi:10.1176/appi.ajp.158.12.1974

12. Marshall RD, Beebe KL, Oldham M, et al. Efficacy and safety of paroxetine treatment for chronic PTSD: a fixed-dose, placebo-controlled study. Am J Psychiatry. 2001;158(12):1982-1988. PubMed doi:10.1176/appi.ajp.158.12.1982

13. Kind P. Guidelines for value sets in economic and on-economic studies using EQ-5D. In: Brooks R, Rabin R, Charro FD, eds. The Measurement and Valuation of Health Status Using EQ-5D: A European Perspective. Amsterdam, the Netherlands: Kluwer Academic Publishers; 2003:29-41. doi:10.1007/978-94-017-0233-1_4.

14. Shaw JW, Johnson JA, Coons SJ. US valuation of the EQ-5D health states: development and testing of the D1 valuation model. Med Care. 2005;43(3):203-220. PubMed doi:10.1097/00005650-200503000-00003

15. Bureau of Labor Statistics. Consumer price index (CPI). http://www.bls.gov/cpi/. Accessed January 9, 2014.

16. Sirey JA, Meyers BS, Teresi JA, et al. The Cornell Service Index as a measure of health service use. Psychiatr Serv. 2005;56(12):1564-1569. PubMed doi:10.1176/appi.ps.56.12.1564

17. Bureau of Labor Statistics. May 2012 National Occupational Employment and Wage Estimates United States. http://www.bls.gov/oes/current/oes_nat.htm. Accessed January 9, 2014.

18. Drug Topics Red Book. Montvale, NJ: Thomson Healthcare; 2012.

19. Stranges E, Levit K, Stocks C, et al. State Variation in Inpatient Hospitalizations for Mental Health and Substance Abuse Conditions, 2002-2008: Statistical Brief #117. Healthcare Cost and Utilization Project. http://www.hcup-us.ahrq.gov/reports/statbriefs/sb117.pdf. Accessed January 9, 2014.

20. Machlin S, Chowdhury S. Expenses and Characteristics of Physician Visits in Different Ambulatory Care Settings, 2008. Statistical Brief #318. Agency for Healthcare Research and Quality, Rockville, MD. http://meps.ahrq.gov/mepsweb/data_files/publications/st318/stat318.shtml. Updated March 2011. Accessed January 9, 2014.

21. Phibbs CS, Scott JY, Flores NE, et al. HERCs Outpatient Average Cost Data Sets for VA Care: Fiscal Years 2010 Update. Palo Alto, CA: Health Economics Resource Center; 2011. http://www.herc.research.va.gov/methods/methods_cost_ac.asp. Updated April 22, 2013. Accessed January 9, 2014.

22. Helvie CO. Efficacy of primary care in a nursing center. Nurs Case Manag. 1999;4(4):201-210. PubMed

23. Curtiss FR, Lettrich P, Fairman KA. What is the price benchmark to replace average wholesale price (AWP)? J Manag Care Pharm. 2010;16(7):492-501. PubMed

24. Eichler HG, Kong SX, Gerth WC, et al. Use of cost-effectiveness analysis in health-care resource allocation decision-making: how are cost-effectiveness thresholds expected to emerge? Value Health. 2004;7(5):518-528. PubMed doi:10.1111/j.1524-4733.2004.75003.x

25. Shiroiwa T, Sung YK, Fukuda T, et al. International survey on willingness-to-pay (WTP) for one additional QALY gained: what is the threshold of cost effectiveness? Health Econ. 2010;19(4):422-437. PubMed doi:10.1002/hec.1481

26. Hirth RA, Chernew ME, Miller E, et al. Willingness to pay for a quality-adjusted life year: in search of a standard. Med Decis Making. 2000;20(3):332-342. PubMed doi:10.1177/0272989X0002000310

27. Ubel PA, Hirth RA, Chernew ME, et al. What is the price of life and why doesn’ t it increase at the rate of inflation? Arch Intern Med. 2003;163(14):1637-1641. PubMed doi:10.1001/archinte.163.14.1637

28. Braithwaite RS, Meltzer DO, King JT Jr, et al. What does the value of modern medicine say about the $50,000 per quality-adjusted life-year decision rule? Med Care. 2008;46(4):349-356. PubMed doi:10.1097/MLR.0b013e31815c31a7

29. Glick H. Economic Evaluation in Clinical Trials. New York, NY: Oxford University Press; 2007.

30. Efron B, Tibshirani R. Cross-validation and other estimates of prediction error. In: Efron B, Tibshirani RJ, eds. An Introduction to the Bootstrap. New York, NY: Chapman & Hall; 1993. doi:10.1007/978-1-4899-4541-9

31. Multiple imputation. STATA Reference Manual, Release 12. College Station, TX: Stata Press; 2011.

32. Walker EA, Katon W, Russo J, et al. Health care costs associated with posttraumatic stress disorder symptoms in women. Arch Gen Psychiatry. 2003;60(4):369-374. PubMed doi:10.1001/archpsyc.60.4.369

33. Marciniak MD, Lage MJ, Dunayevich E, et al. The cost of treating anxiety: the medical and demographic correlates that impact total medical costs. Depress Anxiety. 2005;21(4):178-184. PubMed doi:10.1002/da.20074

34. Mills KL, Teesson M, Ross J, et al. The costs and outcomes of treatment for opioid dependence associated with posttraumatic stress disorder. Psychiatr Serv. 2005;56(8):940-945. PubMed doi:10.1176/appi.ps.56.8.940

35. Congressional Budget Office. The Veterans Health Administration’s treatment of PTSD and traumatic brain injury among recent combat veterans: February 2012. http://www.cbo.gov/sites/default/files/cbofiles/attachments/02-09-PTSD.pdf. Accessed January 9, 2014.

36. Issakidis C, Sanderson K, Corry J, et al. Modelling the population cost-effectiveness of current and evidence-based optimal treatment for anxiety disorders. Psychol Med. 2004;34(1):19-35. PubMed doi:10.1017/S003329170300881X

37. Gospodarevskaya E, Segal L. Cost-utility analysis of different treatments for post-traumatic stress disorder in sexually abused children. Child Adolesc Psychiatry Ment Health. 2012;6(1):15. PubMed doi:10.1186/1753-2000-6-15

38. Joesch JM, Sherbourne CD, Sullivan G, et al. Incremental benefits and cost of coordinated anxiety learning and management for anxiety treatment in primary care. Psychol Med. 2012;42(9):1937-1948. PubMed doi:10.1017/S0033291711002893

39. Roy-Byrne P, Craske MG, Sullivan G, et al. Delivery of evidence-based treatment for multiple anxiety disorders in primary care: a randomized controlled trial. JAMA. 2010;303(19):1921-1928. PubMed doi:10.1001/jama.2010.608