Mental Illness and Psychotropic Drug Use Among Prescription Drug Overdose Deaths: A Medical Examiner Chart Review
Objective: Between 1999 and 2006, there was a 120% increase in the rate of unintentional drug overdose deaths in the United States. This study identifies the prevalence of mental illness, a risk factor for substance abuse, and chronic pain among prescription drug overdose deaths in West Virginia and ascertains whether psychotropic drugs contributing to the deaths were used to treat mental illness or for nonmedical purposes.
Method: In 2007, we abstracted data on mental illness, pain, and drugs contributing to death from all unintentional prescription drug overdose deaths in 2006 recorded by the West Virginia Office of the Chief Medical Examiner. Decedent prescription records were obtained from the state prescription drug monitoring program.
Results: Histories of mental illness and pain were documented in 42.7% and 56.6% of 295 decedents, respectively. Psychotropic drugs contributed to 48.8% of the deaths, with benzodiazepines involved in 36.6%. Benzodiazepines contributing to death were not associated with mental illness (adjusted odds ratio [AOR] = 1.1; 95% CI, 0.6–1.8), while all other psychotropic drugs were (AOR = 3.9; 95% CI, 2.0–7.6). Of decedents with contributory benzodiazepines, 46.3% had no prescription for the drug.
Conclusions: Mental illness may have contributed to substance abuse associated with deaths. Clinicians should screen for mental illness when prescribing opioids and recommend psychotherapy as an adjunct or an alternate to pharmacotherapy. Benzodiazepines may have been used nonmedically rather than as a psychotropic drug, reflecting drug diversion. Restricting benzodiazepine prescriptions to a 30-day supply with no refills might be considered.
J Clin Psychiatry 2010;71(4):491–496
© Copyright 2010 Physicians Postgraduate Press, Inc.
Submitted: July 28, 2009; accepted October 5, 2009 (doi:10.4088/JCP.09m05567blu).
Corresponding author: Robin L. Toblin, PhD, Office of Research and Evaluation, Federal Bureau of Prisons, 320 First St NW, Washington, DC 20534 (email@example.com).
Deaths resulting from unintentional drug overdoses have risen to epidemic proportions in the United States, increasing by 120% during the years 1999–2006.1 This increase has largely been attributed to prescription drugs, rather than to illicit drugs. Most of the problem is due to deaths from opioid analgesics2; these deaths have increased in parallel with the number of prescriptions for such drugs.3 While studies have focused on the role of chronic pain, drug diversion, and “doctor shopping” in prescription drug overdose deaths,4,5 few studies of these deaths have focused on mental health problems and the medications used to treat such problems.
Previous research has found that those who have mental disorders are more likely to have drug use disorders (ie, drug abuse and drug dependence).6 Approximately 15%–20% of persons with mental disorders have a drug use disorder at some point in their lifetimes.6,7 People who report having a diagnosis of major depressive disorder within the past year are 3.7 times more likely to report drug dependence.6 In addition, people with mental disorders are twice as likely as those without mental disorders to use prescribed opioids.8
Other epidemiologic studies have examined the mental disorders of those persons with drug use disorders. One study found that over half (53%) of persons with drug use disorders had non–substance abuse mental disorders, or 4.5 times the odds of persons without drug disorders.7 Another epidemiologic survey found that 32% of drug abusers had mood disorders, while 25% had anxiety disorders9; the study also found that nonmedical use of opioids was strongly associated with mood disorders, anxiety disorders, and personality disorders.10 Further, research regarding prescription drug overdose deaths suggests that a large proportion of decedents had histories of mental illness.4
In general, studies of deaths from drug overdoses have supported such self-reported associations.11 In New Jersey, for example, 24.4% of persons who died of unintentional drug overdoses had histories of mental health risk factors.5 However, the prevalence of mental illness may be higher for deaths involving prescription drugs, some of which are prescribed for mental disorders. Studies of deaths involving both illicit drugs and prescription opioids have reported a high likelihood of finding prescription psychotropic drugs (ie, drugs prescribed to treat mental disorders) on postmortem toxicology reports,12 even though such drugs are rarely responsible for overdose deaths by themselves.4
One aspect of the association between mental illness and psychotropic drugs that has not been examined is whether psychotropic drugs involved in prescription drug overdoses have been prescribed to treat mental illness or are being used nonmedically as part of a substance use disorder. Benzodiazepine sedatives, for example, are prescribed for anxiety disorders, but they are also often combined with drugs that produce a euphoric effect, such as heroin or prescription analgesics, and are addictive in their own right when misused.13 Therefore, one aim of this study was to determine the prevalence of mental illness and psychotropic drug involvement in a population that had died of prescription drug–related overdoses and the overlap between the two.
In addition, little is known about how persons with mental illness and a history of chronic pain come to use prescription painkillers. Therefore, a second aim of the present study was to describe the association between mental illness and chronic pain among drug overdose deaths. Our hope was to identify associations among mental illness, psychotropic drugs, and pain to help guide prevention efforts in reducing prescription drug overdose deaths.
Case Identification and Data Sources
All decedents were West Virginia residents who died within the state in 2006. West Virginia was selected because it experienced the largest increase (675%) in the unintentional drug overdose mortality rate among all US states between 1999 and 2004.1 Decedent records were reviewed if the death certificates listed the underlying cause of death as unintentional drug poisoning (International Classification of Diseases, 10th Revision [ICD-10] codes X40–X44). All cases were identified via an electronic database of vital records at the Health Statistics Center of the West Virginia Department of Health and Human Resources using the appropriate ICD-10 codes. Identifying information was cross-referenced with the case logbook and electronic database of the investigations of the Office of the Chief Medical Examiner (OCME) to identify decedent records. We excluded all decedents certified without benefit of autopsy as well as those decedents in whom the only drugs contributing to death were illicit or over-the-counter drugs. In West Virginia, the chief toxicologist routinely screens all potential victims of drug overdose for illicit and prescription drugs. Trained and certified death investigators conduct scene-of-death investigations and write reports on all drug-related cases. Subsequently, OCME staff forensic pathologists review these reports, medical records, decedents’ prescription records, autopsy and toxicology findings, and other evidence to determine which factors, including drugs, are contributory to death.
In late 2007, we abstracted information from autopsy reports, toxicology reports, death-scene investigation reports, death certificates, and copies of medical records in the OCME files. The West Virginia Board of Pharmacy also provided decedent prescription histories available through the state’s Controlled Substances Monitoring Program (CSMP), which, since its inception in 2003, has maintained electronic records of all Schedule II, III, and IV, and prescribed Schedule V controlled substances dispensed by all pharmacies licensed by the State of West Virginia.
We characterized the decedents in terms of sociodemographic factors, history of chronic pain, and history of mental illness. Demographic information on the decedents, including sex, age, marital status, and highest level of education, was collected from death certificates. The OCME death scene investigation reports, medical records, and autopsy reports were also used to determine past medical histories and comorbidities.
A history of mental illness (hereafter referred to as mental illness) was derived from diagnoses recorded in medical records and information reported to death scene investigators. Specific diagnoses could not be independently verified with criteria from the Diagnostic and Statistical Manual of Mental Disorders, 4th Edition (DSM-IV) due to the study population of decedents. Thus, diagnoses were recorded as they appeared, even when mutually exclusive disorders were listed. Disorders were grouped into major diagnostic categories. However, 2 individual disorders, major depressive disorder and bipolar disorder, were recorded separately because they are strongly associated with substance abuse and chronic pain and have different pharmacotherapies.
A history of pain was identified from medical records and death scene investigation records. Pain was classified by anatomic system as decided before our investigation. The nature and location of the pain was used to classify the pain as subjective—the source of the pain was likely to be taken from the patient’s own statement; objective—the pain was likely to be apparent from physical findings; and unclassifiable. Subjective pain included that of fibromyalgia, headaches, neck/back, and neuralgia. Objective pain included that of cancer, infection, recent dental work, and trauma. Unclassifiable pain included arthritis, gastrointestinal, genitourinary, heart/chest, and orofacial pain.
We also assessed the proportion of cases that had at least 1 psychotropic drug contributing to death, as determined by the OCME. Psychotropic drugs were prescription drugs usually used to treat mental disorders. We categorized psychotropic drugs into drug class on the basis of the drug reference vocabulary established by the Drug Alert and Warning Network.12 Drug classes included benzodiazepines, antidepressants, and “other psychotropics,” which included anticonvulsants and antipsychotics, among others. We used information from both the CSMP and the OCME to determine whether decedents using benzodiazepines had prescriptions for those drugs.
First, we calculated the prevalence of mental illness and specific disorders as well as classes of contributory psychotropic drugs. We then used multiple logistic regression to examine the association between mental illness and psychotropic drugs, controlling for age and sex, which were found to be significantly associated with mental illness in univariate models. We ran multiple logistic regressions to determine if each drug class was associated with mental illness.
Next, we compared decedents with and without mental illness, using multiple logistic regression to examine associations with demographic factors. Except for cases in which age was the variable of interest, age was used as a continuous variable. Consistent with census methodology,14 we limited the analysis of highest level of education attained to decedents aged ≥ 25 years. For analyses that involved cell sizes less than 5, the Fisher exact test was used.
We also calculated the associations between demographic variables and having at least 1 psychotropic drug contributory to death, having a benzodiazepine as the only psychotropic drug, and having at least 1 psychotropic drug that was not a benzodiazepine. We examined benzodiazepines independently due to their propensity for abuse.
We used available information on prescription history to further characterize how benzodiazepines were being used. We conducted a Cochrane-Armitage test for trend in the association between mental illness and the number of benzodiazepines prescribed. We also identified the proportion of decedents who had been prescribed benzodiazepines that contributed to death.
Finally, controlling for sex and age, we conducted a series of multiple logistic regressions to look at the association between mental illness and varied types of pain by comparing decedents with each pain type to decedents with no pain: a history of any pain; subjective, unclassifiable, or objective pain; and specific subtypes of pain. All analyses were conducted by use of SAS software, Version 9.1 (SAS Institute Inc, Cary, North Carolina).
The study included 295 decedents who met inclusion criteria. Men composed 67.1% of the study population. More than half of the decedents (53.2%) were 35–54 years of age, while 38.6% were aged 18–34 and the remainder (8.1%) were aged 55 or older. Roughly half of the decedents (51.6%) had a high school education, a third (31.0%) had not graduated from high school, and the remainder (17.3%) had some college. Marital status was about evenly split between never married (34.9%), married (32.9%), and divorced or widowed (32.2%).
Click figure to enlarge
Click figure to enlarge
Of these 295 decedents, 126 (42.7%) had a history of mental illness other than substance abuse (Table 1). Mood disorders and anxiety disorders were most prevalent. A history of pain was found in 56.6% of decedents. Opioid analgesics contributed to 93% of the deaths. Psychotropic drugs contributed to 48.8% of all deaths (Table 1).
Decedents with mental illness had over twice the odds of having a psychotropic drug contribute to death and were nearly 4 times as likely to have had an antidepressant or “other” psychotropic drug contribute to death, compared to decedents without mental illness. However, benzodiazepines contributing to death were not associated with mental illness (Table 2).
Mental illness was associated with female sex and greater age, while no associations were identified between contributory psychotropic drugs and any of the demographic variables examined (Table 3). There was no association between demographics and benzodiazepines as the only contributory psychotropic drug or between demographics and having at least 1 contributory antidepressant or “other” psychotropic (not shown).
Of decedents with contributory benzodiazepines, 53.7% had prescriptions for all benzodiazepines; the remaining 46.3% had at least 1 benzodiazepine for which they did not have a valid prescription. Regardless of whether benzodiazepines were involved in the decedent’s fatal overdose, the number of different benzodiazepines prescribed in the year prior to death (eg, diazepam versus alprazolam) was significantly associated with mental illness in a dose-response fashion (test for trend: Z = −5.3, P < .001). The odds ratios were 2.5, 3.5, and 5.2 for 1, 2, and 3 or more unique benzodiazepines prescribed in the year prior to death, respectively.
Click figure to enlarge
Most decedents (56.6%) had a history of pain. Decedents with both subjective and unclassifiable pain were more likely to have mental illness (Table 4). The association with mental illness varied by pain subtype. Neck and back pain, headaches, and arthritis were all associated with mental illness, with odds ratios of 4.6, 4.5, and 3.6, respectively.
Click figure to enlarge
We found that nearly one-half of the people dying from a prescription drug overdose in West Virginia in 2006 had a history of mental illness during their lifetime and that over one-half had a history of chronic pain. Psychotropic drugs contributed to nearly one-half of the overdoses in this study population, usually in combination with an opioid; of these, about three-fourths of those deaths involved a benzodiazepine. We found that psychotropic drugs were associated with mental illness, but, for benzodiazepines, specifically, there was no association with mental illness. Mental illness was associated with sex, age, and both subjective and unclassifiable pain; therefore, we were surprised that psychotropic drugs, commonly used to treat mental health problems, were not associated with these same demographic or pain variables. These findings suggest that benzodiazepines are being diverted and used nonmedically.
Although mental illness was common in our study, its prevalence is comparable to the lifetime prevalence in the United States as a whole.15 The prevalence of mood disorders, including major depressive disorder and bipolar disorder, was higher than national figures. Anxiety disorders and sleep disorders were lower in study decedents than in national data, and psychotic disorders were equivalent. The increased rate of mood disorders in this study is not surprising in light of the fact that substance use disorders and mood disorders are highly correlated.15
There was a significant association between contributory psychotropic drug use and mental illness. This association was strong for antidepressant and other psychotropic drug classes, but it did not hold for benzodiazepines, which constituted the largest proportion of psychotropic drugs. However, the significant dose-response relationship between the number of prescriptions for different benzodiazepines and mental illness suggests that some benzodiazepines were being used medically. Perhaps those with multiple prescriptions had a specific mental illness requiring more complex treatment, whereas those with complaints of nonspecific anxiety received a prescription without a diagnosable illness. Nonetheless, the widespread use of benzodiazepines by people without mental illness and the finding that 46.3% of people did not have prescriptions for at least 1 benzodiazepine that contributed to death suggest that abuse of benzodiazepines was involved. This finding is consistent with the known use of benzodiazepines by drug abusers to moderate the effects of cocaine, heroin, or opioids and the Drug Enforcement Administration’s opinion that benzodiazepines are among the most commonly abused drugs in West Virginia.16
While we found that female decedents in this study were nearly twice as likely as male decedents to have mood and anxiety disorders, consistent with national findings in the general population,15 we found no association between sex and psychotropic drug involvement in the overdoses, a finding not consistent with data demonstrating that women are more likely to be prescribed psychotropic drugs than men in the United States.17–19 Because men composed two-thirds of the study population, this lack of association is very likely due to the high prevalence of nonmedical use and diversion of psychotropic drugs in this population and the known association between male sex and drug abuse.4,20 Although the literature does not suggest that increasing age is a risk factor for mental illness, we found that, in our population, increasing age was associated with mental illness but not with psychotropic drug involvement. However, there may be differences between a decedent population and the general population.
The association we found between mental illness and pain is consistent with the literature. The World Health Organization suggests that people with chronic pain have a 4-fold increase in anxiety and depressive disorders.21 Among decedents dying of drug overdoses, however, we found that mental illness was associated with only certain types of pain. There are several potential explanations for this finding. People without a physically verifiable cause of pain may have been more likely to have been labeled as mentally ill. Alternatively, people with mental illness may have been more likely to seek care and complain of some types of pain. Underlying symptoms of mental illness tend to exacerbate the intensity of and focus upon one’s pain.22 Finally, given the established association between substance abuse disorders and other mental illnesses, individuals with such disorders might use subjective pain complaints as a mechanism to obtain prescription drugs for nonmedical purposes. The comorbidity of substance abuse and other mental disorders generates such complaints when patients realize they have a high likelihood of obtaining narcotic painkillers.
The limitations of this study are the result of substantial uncertainties inherent in the nature of death investigation systems and CSMP prescription data collection. First, the diagnosis of mental disorders and reporting of pain were based upon the OCME records, which were not always consistent among data sources. For instance, medical records were available for only 43% of decedents, and not all death-scene investigation reports addressed mental illness. Further, some diagnoses were based on reports by friends and family and, therefore, may not have met true criteria for diagnosis. In addition, listing of death-scene medications and statewide CSMP data collection were incomplete, and CSMP data were limited to scheduled drugs; thus, no reliable analyses of the prescription history for antidepressants and other psychotropic drugs could be conducted.
Clinicians should assess pain patients for a history of both drug abuse and mental illness. Currently, clinicians are told to use caution in their prescribing of opioids to people with a history of substance abuse.23 This guidance should also apply for prescribing to patients with a history of mental illness, particularly if patient reports of pain have no physical findings. Chronic pain has high comorbidity with anxiety and mood disorders. Clinicians should consider referring patients for psychotherapy because it has been shown to be an important adjunct in the treatment of chronic pain24 and may reduce the need for opioid analgesics. Further, cognitive-behavioral therapy has been shown to be highly effective in treating adult anxiety disorders,25 and such therapy may reduce the need for benzodiazepines. Clinicians should be cautious when prescribing opioids and benzodiazepines concurrently for chronic pain, even in patients without mental illness, unless there is a specific medical indication.23 Judicious prescribing of opioid analgesics and benzodiazepines not only might help prevent abuse in patients but also might reduce the ambient level of drugs subject to abuse in the community.
States can also employ administrative tools to address the problem of abuse of psychotropic drugs. West Virginia and other states with benzodiazepine abuse problems should consider treating benzodiazepines, currently Schedule IV drugs, like Schedule II drugs by specifically restricting them to a 30-day supply with no refills. States with prescription drug abuse problems should also consider expanding their use of prescription drug monitoring program data for case management and surveillance. With these data, clinicians and pharmacists could monitor their patients’ prescriptions, medical examiners could review the drug use history of decedents, and public health agencies could better identify provider characteristics and high-risk populations for prescription drug abuse as well as the impact of prevention measures.
Drug names: alprazolam (Xanax, Niravam, and others), carbamazepine (Carbatrol, Equetro, and others), diazepam (Diastat, Valium, and others), hydroxyzine (Vistaril and others) quetiapine (Seroquel), topiramate (Topamax), zolpidem (Ambien, Edluar, and others).
Author affiliations: Epidemic Intelligence Service, Office of Workforce and Career Development (Drs Toblin, Logan, and Hall), and National Center for Injury Prevention and Control (Drs Toblin, Paulozzi, and Logan), Centers for Disease Control and Prevention, Atlanta, Georgia; and West Virginia Department of Health and Human Resources, Charleston (Drs Hall and Kaplan).
Potential conflicts of interest: None reported.
Funding/support: None reported.
Disclaimer: The contents of this article are solely the responsibility of the authors and do not necessarily represent the official views of the Centers for Disease Control and Prevention.
Previous presentation: Data from this article were presented at the 58th Annual Epidemic Intelligence Service Conference; April 20–24, 2009; Atlanta, Georgia.
Acknowledgment: We thank Edna Wong McKinstry, MD, Epidemiology Elective Program, Centers for Disease Control and Prevention, for her assistance with data collection; James C. Kraner, PhD, West Virginia Department of Health and Human Resources, for access to the toxicology data; and Danae Bixler, MD, West Virginia Department of Health and Human Resources, and Alex E. Crosby, MD, National Center for Injury Prevention and Control, Centers for Disease Control and Prevention, for their contributions to the conceptualization and oversight of the overall investigation and data collection. Drs McKinstry, Kraner, Bixler, and Crosby report no potential conflicts of interest relevant to this article.
1. Centers for Disease Control and Prevention, National Center for Health Statistics. Compressed Mortality File 1999–2006. CDC WONDER Online Database, compiled from Compressed Mortality File 1999–2006 Series 20 No. 2K, 2009. http://wonder.cdc.gov/cmf-icd10.html. Accessed May 13, 2009.
2. Fingerhut LA. Centers for Disease Control and Prevention. Increases in Poisoning and Methadone-Related Deaths: United States, 1999–2005. http://www.cdc.gov/nchs/data/hestat/poisoning/poisoning.htm. Verified February 5, 2010.
3. Paulozzi LJ, Budnitz DS, Xi Y. Increasing deaths from opioid analgesics in the United States. Pharmacoepidemiol Drug Saf. 2006;15(9):618–627. doi:10.1002/pds.1276 PubMed
4. Hall AJ, Logan JE, Toblin RL, et al. Patterns of abuse among unintentional pharmaceutical overdose fatalities. JAMA. 2008;300(22):2613–2620. doi:10.1001/jama.2008.802 PubMed
5. Hempstead K. Manner of death and circumstances in fatal poisonings: evidence from New Jersey. Inj Prev. 2006;12(suppl 2):ii44–ii48. doi:10.1136/ip.2006.012583 PubMed
6. Hasin DS, Goodwin RD, Stinson FS, et al. Epidemiology of major depressive disorder: results from the National Epidemiologic Survey on Alcoholism and Related Conditions. Arch Gen Psychiatry. 2005;62(10):1097–1106. doi:10.1001/archpsyc.62.10.1097 PubMed
7. Regier DA, Farmer ME, Rae DS, et al. Comorbidity of mental disorders with alcohol and other drug abuse. Results from the Epidemiologic Catchment Area (ECA) Study. JAMA. 1990;264(19):2511–2518. doi:10.1001/jama.264.19.2511 PubMed
8. Sullivan MD, Edlund MJ, Zhang L, et al. Association between mental health disorders, problem drug use, and regular prescription opioid use. Arch Intern Med. 2006;166(19):2087–2093. doi:10.1001/archinte.166.19.2087 PubMed
9. Grant BF, Stinson FS, Dawson DA, et al. Prevalence and co-occurrence of substance use disorders and independent mood and anxiety disorders: results from the National Epidemiologic Survey on Alcohol and Related Conditions. Arch Gen Psychiatry. 2004;61(8):807–816. doi:10.1001/archpsyc.61.8.807 PubMed
10. Huang B, Dawson DA, Stinson FS, et al. Prevalence, correlates, and comorbidity of nonmedical prescription drug use and drug use disorders in the United States: results of the National Epidemiologic Survey on Alcohol and Related Conditions. J Clin Psychiatry. 2006;67(7):1062–1073. doi:10.4088/JCP.v67n0708 PubMed
11. Sorg MH, Greenwald M. Maine Drug-Related Mortality Patterns: 1997–2002. Orono, ME: University of Maine; 2002.
12. Substance Abuse and Mental Health Services Administration. Drug Abuse Warning Network (DAWN) Vocabulary List. http://dawninfo.samhsa.gov/drug_vocab/default.asp. Accessed September 2, 2007.
13. Drug Enforcement Administration. Drug Information. http://www.usdoj.gov/dea/concern/b.html#2. Accessed May 13, 2009.
14. US Census Bureau. Annual Estimates of the Population for the United States, Regions, States, and for Puerto Rico: April 1, 2000 to July 1, 2006 (NST-EST2006-01). http://www.census.gov/popest/states/NST-ann-est2006.html. Verified February 5, 2010.
15. Kessler RC, Berglund P, Demler O, et al. Lifetime prevalence and age-of-onset distributions of DSM-IV disorders in the National Comorbidity Survey Replication. Arch Gen Psychiatry. 2005;62(6):593–602. doi:10.1001/archpsyc.62.6.593 PubMed
16. Drug Enforcement Administration. DEA Appalachian Report: West Virginia 2007. Washington, DC: US Department of Justice; 2007.
17. Simoni-Wastila L. Gender and psychotropic drug use. Med Care. 1998;36(1):88–94. doi:10.1097/00005650-199801000-00010 PubMed
18. Empereur F, Baumann M, Alla F, et al. Factors associated with the consumption of psychotropic drugs in a cohort of men and women aged 50 and over. J Clin Pharm Ther. 2003;28(1):61–68. doi:10.1046/j.1365-2710.2003.00464.x PubMed
19. Goldney R, Bain M. Prevalence of psychotropic use in a South Australian population. Australas Psychiatry. 2006;14(4):379–383. PubMed
20. Inciardi JA, Surratt HL, Kurtz SP, et al. Mechanisms of prescription drug diversion among drug-involved club- and street-based populations. Pain Med. 2007;8(2):171–183. doi:10.1111/j.1526-4637.2006.00255.x PubMed
21. Gureje O, Simon GE, Von Korff M. A cross-national study of the course of persistent pain in primary care. Pain. 2001;92(1-2):195–200. doi:10.1016/S0304-3959(00)00483-8 PubMed
22. Bair MJ, Wu J, Damush TM, et al. Association of depression and anxiety alone and in combination with chronic musculoskeletal pain in primary care patients. Psychosom Med. 2008;70(8):890–897. doi:10.1097/PSY.0b013e318185c510 PubMed
23. Washington State Agency Medical Directors’ Group. Interagency Guideline on Opioid Dosing for Chronic Non-Cancer Pain. http://www.agencymeddirectors.wa.gov. Accessed November 3, 2008.
24. Turk DC, ed. Psychological Approaches to Pain Management: A Practitioner’s Handbook. 2nd ed. New York, NY: Guilford Press; 2002.
25. Hofmann SG, Smits JA. Cognitive-behavioral therapy for adult anxiety disorders: a meta-analysis of randomized placebo-controlled trials. J Clin Psychiatry. 2008;69(4):621–632. PubMed