Add To Cart

Brief Interventions for Anxiety Disorders with Children and Adults

Selected Readings

Test | Table of Contents

Additional Readings

Generalized anxiety disorder. (eng; includes abstract) By Gliatto MF, American Family Physician [Am Fam Physician], ISSN: 0002-838X, 2000 Oct 1; Vol. 62 (7), pp. 1591-600, 1602; PMID: 11037076 (3800 words)

Patients with generalized anxiety disorder experience worry or anxiety and a number of physical and psychologic symptoms. The disorder is frequently difficult to diagnose because of the variety of presentations and the common occurrence of comorbid medical or psychiatric conditions. The lifetime prevalence is approximately 4 to 6 percent in the general population and is more common in women than in men. It is often chronic, and patients with this disorder are more likely to be seen by family physicians than by psychiatrists. Treatment consists of pharmacotherapy and various forms of psychotherapy. The benzodiazepines are used for short-term treatment, but because of the frequently chronic nature of generalized anxiety disorder, they may need to be continued for months to years. Buspirone and antidepressants are also used for the pharmacologic management of patients with generalized anxiety disorder. Patients must receive an appropriate pharmacologic trial with dosage titrated to optimal levels as judged by the control of symptoms and the tolerance of side effects. Psychiatric consultation should be considered for patients who do not respond to an appropriate trial of pharmacotherapy. (Am Fam Physician 2000;62:1591-600,1602.)
Anxiety disorders, such as generalized anxiety disorder (GAD), panic disorder and social phobia, are the most prevalent psychiatric disorders in the United States,[1] and patients with these disorders are more likely to seek treatment from a primary care physician than from a psychiatrist.[2] Patients with anxiety disorders are more likely than other patients to make frequent medical appointments, to undergo extensive diagnostic testing,[ 3] to report their health as poor and to smoke cigarettes and abuse other substances.[ 4] Anxiety disorders, particularly panic disorder, occur more frequently in patients with chronic medical illnesses (e.g., hypertension, chronic obstructive pulmonary disease, irritable bowel syndrome, diabetes) than in the general population.[ 5] Conversely, patients with anxiety disorders are more likely than others to develop a medical illness,[ 6, 7] and the presence of an anxiety disorder may prolong the course of a medical illness.[ 2] Patients with anxiety disorders have higher rates of mortality from all causes.[ 4]
Characteristics of Generalized Anxiety Disorder
The definition of GAD has changed over time. Originally, little distinction was made between panic disorder and GAD. As panic disorder became better understood and specific treatments were developed, GAD was defined in the Diagnostic and Statistical Manual of Mental Disorders, 3d ed. (DSM-III)[ 8] as a disorder without panic attacks or symptoms of major depression. This definition had little reliability,[9] and current diagnostic criteria (Table 1)[10] for GAD in the Diagnostic and Statistical Manual of Mental Disorders, 4th ed. (DSM-IV)10 emphasize the psychic component (e.g., the worry) rather than the somatic (e.g., muscle tension) or autonomic symptoms (e.g., diaphoresis, increased arousal). In addition to the DSM-IV framework, the symptoms of GAD can be conceptualized as being contained in three categories: excessive physiologic arousal, distorted cognitive processes and poor coping strategies.[11] The symptoms associated with each of these categories are listed in Table 2.[ 10, 11] To make the diagnosis of GAD by DSM-IV[ 10] criteria, the worry and other associated symptoms must be present for at least six months and must adversely affect the patient's life (e.g., the patient misses work days or cannot maintain daily responsibilities).[ 10] The diagnosis can be challenging because the difference between normal anxiety and GAD is not always distinct [12] and because GAD often coexists with other psychiatric disorders (e.g., major depression, dysthymia, panic disorder, substance abuse).[13]
The lifetime prevalence of GAD is 4.1 to 6.6 percent,14 which is higher than that of the other anxiety disorders. The prevalence of GAD in patients visiting physicians' offices is twice that found in the community.[15] It is more prevalent in women than in men,[ 1] with the median age of onset occurring during the early 20s.[ 13] The onset of symptoms is usually gradual, although GAD can be precipitated by stressful life events. The condition tends to be chronic with periods of exacerbation and remission. [13]
The evaluation process for patients with anxiety is outlined in Figure 1.
Initial Assessment
After obtaining a patient history, the physician should try to categorize the anxiety as acute (or brief or intermittent) or persistent (or chronic).15,16 Acute anxiety lasts from hours to weeks (in contrast, panic attacks last for minutes) and is usually preceded by a stressor. Often, comorbid conditions (e.g., major depression) are not present.[15] Persistent anxiety lasts for months to years and can include what is called "trait anxiety." Trait anxiety can be viewed as part of a patient's temperament; for instance, a patient may say, "I've always been nervous, but I don't know about what."
Although there is usually not a precipitating stressor in most cases of persistent anxiety, a stressor can exacerbate the patient's baseline level of anxiety. This situation is called "double anxiety" (i.e., acute anxiety superimposed on persistent anxiety).[ 15] GAD is a form of persistent anxiety and can occur in patients with or without trait anxiety.

Figure 1. Algorithm for evaluating patients with complaints of anxiety. (GAD=generalized anxiety disorder)
Patients with GAD present with a wide variety of symptoms and range of severity.[ 12] Some patients may emphasize a special symptom (e.g., insomnia) and not report other symptoms that are usually associated with GAD.[ 12] Some patients may not complain of anxiety or specific worries but present with exclusively somatic symptoms such as diarrhea, palpitations, dyspnea, abdominal pain, headache or chest pain.[ 2] These patients warrant a full medical evaluation because there may be no indication that GAD is the etiology. Conversely, physicians should include a psychiatric disorder in the differential diagnosis when symptoms are vaguely described, do not conform to known pathophysiologic mechanisms, persist after a negative work-up and are not resolved by reassurance. Patients with this clinical profile should be asked as early in the evaluation as possible about worries, "nerves," or anxiety, acute or chronic stressors, and about the presence of symptoms listed in Tables 110 and 2.[ 10, 11]
Evaluation for Medical Disorders
Nonpsychiatric disorders must be ruled out before GAD can be diagnosed as the etiology of a patient's complaint of anxiety. Neurologic and endocrine diseases, such as hyperthyroidism and Cushing's disease, are the most frequently cited medical causes of anxiety.[17] Other medical conditions commonly associated with anxiety include mitral valve prolapse, carcinoid syndrome and pheochromocytoma.[ 17] It must be stressed that these conditions, although often cited, are clinically uncommon, and all patients who present with the complaint of anxiety do not require an extensive work-up to exclude these other medical conditions. Medications such as steroids, over-the-counter sympathomimetics, selective serotonin reuptake inhibitors (SSRIs), digoxin, thyroxine and theophylline may cause anxiety.[ 17] The physician should also inquire about the use of herbal products or vitamins. If possible, all exogenous substances should be removed before initiating treatment of patients with GAD.
Evaluation for Substance Abuse
Use of and withdrawal from addictive substances can cause anxiety. The physician should inquire about the patient's use of alcohol, caffeine, nicotine and other commonly used substances (including those given by prescription). Corroborative history from family members may be necessary.[ 17]
Evaluation for Other Psychiatric Disorders
The evaluation for other psychiatric disorders is probably the most challenging aspect of patient evaluation because the symptoms of GAD overlap with those of other psychiatric disorders (e.g., major depression, substance abuse, panic disorder) and because these disorders often occur concomitantly with GAD. Generally, GAD should be considered a diagnosis of exclusion after other psychiatric disorders have been ruled out. Because of the overlap and comorbidity associated with other psychiatric disorders, some authors have questioned whether GAD is a distinct entity and have posited that it is a variant of panic disorder or major depression.[ 18, 19]
Symptoms of GAD can occur before, during or after the onset of the symptoms of major depression or panic disorder.[ 12] Some patients have symptoms of anxiety and depression but do not meet the full criteria as delineated in DSM-IV,[ 10] for GAD, panic disorder or major depression. In these cases, the term "mixed anxiety-depressive disorder" can be applied, although it is not yet part of the official nosology. Despite the confusion, anxious patients should be asked about the symptoms of panic attacks and the neurovegetative symptoms (including suicidal ideation) that are associated with major depression. When more than one psychiatric condition exists, an attempt should be made to determine which disorder occurred first. Some distinguishing characteristics of GAD, panic disorder and major depression are listed in Table 3.[ 10, 19]
Obsessive-compulsive disorder and social phobia should also be considered in the evaluation for comorbid disorders. The key symptom of obsessive-compulsive disorder is recurring, intrusive thoughts or actions. Social phobia is characterized by intense anxiety provoked by social or performance situations.[ 10] Because patients with GAD may present with mostly somatic complaints, somatization disorder is also a consideration. The distinguishing feature of this disorder is chronic, multiple physical complaints that involve several organ systems. Patients with exclusively GAD have a much more limited range of physical complaints.
Psychologic Treatments
Nonpharmacologic modalities should be the initial treatment for patients with mild anxiety[ 4] to address the three categories of symptoms of GAD (Table 2).[ 10, 11] Relaxation techniques and biofeedback are used to decrease arousal.[ 11] Cognitive therapy helps patients to limit cognitive distortions by viewing their worries more realistically, enabling them to make better plans to manage their anxiety. In cognitive therapy, patients may be taught to record their worries, listing evidence that justifies or contradicts the extent of their concerns.[20] Patients also learn that "worrying about worry" maintains anxiety and that avoidance and procrastination are not effective ways to solve problems.[ 20] A small number of studies have found that cognitive therapy is more effective than behavior therapy[ 21], psychodynamic psychotherapy[ 22] and pharmacotherapy,[ 23, 24] but more research needs to be conducted before the superiority of cognitive therapy is firmly established.[ 20] Patients with personality disorders, those with chronic social stressors and those who expect little benefit from psychologic treatments do not respond well to psychotherapeutic techniques for the treatment of anxiety.[ 25] Often, psychotherapy and pharmacotherapy are necessary. Family members should participate in the treatment plan of patients with GAD. Initially they can provide additional historical information and contribute to the formulation of the treatment plan. Because patients with this disorder tend to be vigilant for signs of danger and may misinterpret information,[ 26] family members can provide another perspective on the patient's problems. In addition, family members should be included in efforts to help the patient develop problem-solving skills and can also help decrease the social isolation of patients with GAD by providing structured activities to promote interaction with others and lessen rumination about problems. Another source of structured activity includes day programs provided by community mental health centers. If substance abuse is prominent, the patient should be referred to a rehabilitation facility. Patients and their families can be referred to the Anxiety Disorders Association of America (301-231-9350) for further information about available resources.
Pharmacologic Treatment
Pharmacologic therapy should be considered for patients whose anxiety results in significant impairment in daily functioning. Study results have not revealed an optimal duration of pharmacologic treatment for patients with GAD.[27] While 25 percent of patients relapse within one month of discontinuing drug therapy and 60 to 80 percent relapse within one year,[16] patients treated for at least six months have a lower relapse rate than those treated for shorter time periods.[ 16]
Benzodiazepines. The anxiolytics most frequently used are the benzodiazepines[ 4] (Table 4).[ 12, 28] All of the benzodiazepines are of equal efficacy[ 29] and all act on the g-aminobutyric acid (GABA)/benzodiazepine (BZ) receptor complex, causing sedation, problems in concentrating and anterograde amnesia.[ 4] Benzodiazepines do not decrease worrying per se, but act to lower anxiety by decreasing vigilance and by eliminating somatic symptoms such as muscle tension. Tolerance to the sedation, impaired concentration and amnesia effects of these drugs develop within several weeks,[ 27] although tolerance to the anxiolytic effects occurs much more slowly--if at all.[ 4] Benzodiazepine therapy can begin with 2 mg of diazepam, or its equivalent, three times daily. The dosage can be increased by 2 mg per day every two to three days until the symptoms abate, side effects develop4 or a daily dose of 40 mg is reached.[ 28] Diazepam does not have to be used as the initial agent; several alternatives are available. In the elderly patient, benzodiazepine therapy should begin at the lowest possible dosage and increased slowly.[ 30]
Agents with short half-lives, such as oxazepam (Serax), are easily metabolized and do not cause excessive sedation. These agents should be used in the elderly and in patients with liver disease. They are also suitable for use on an "as-needed" basis. Agents with long half-lives, such as clonazepam (Klonopin), should be used in younger patients who do not have concomitant medical problems. In addition to improved compliance, the longer acting agents offer several advantages: they can be taken less frequently during the day, patients are less likely to experience anxiety between doses and withdrawal symptoms are less severe when the medication is discontinued. When a benzodiazepine is prescribed, the patient should be warned about driving and operating heavy machinery while taking this medication.
Use of benzodiazepines in therapeutic dosages does not lead to abuse, and addiction is rare.[ 12] The benzodiazepines most likely to be abused are those that are rapidly absorbed such as diazepam, lorazepam (Ativan) and alprazolam (Xanax).[ 27] Alprazolam should be prescribed only for patients with panic disorder. When abused, benzodiazepines are usually abused with other substances, particularly opiates.[31] The patients most likely to abuse benzodiazepines are those who have a previous history of alcohol or drug abuse, and those with a personality disorder.[ 28]
All benzodiazepine therapy can lead to dependence; that is, withdrawal symptoms occur once the medications are discontinued. Withdrawal symptoms include anxiety, irritability and insomnia, and it can be difficult to differentiate between withdrawal symptoms and the recurrence of anxiety. Seizures occur rarely during withdrawal.[ 28] Withdrawal symptoms tend to be more severe with higher dosages, with agents that have short half-lives, with rapidly tapered dosages, and in patients with current tobacco use or with a history of illegal drug use.[ 15] The risk of dependence increases as the dosage and the duration of treatment increases, but it can occur even when appropriate dosages are used continuously for three months.
Withdrawal symptoms begin six to 12 hours after the last dose of an agent with a short half-life and 24 to 48 hours after the last dose of an agent with a longer half-life.[ 27] In patients who have taken a benzodiazepine for more than six weeks, the dosage should be decreased by 25 percent or less per week to prevent withdrawal symptoms.[ 4] Patients may experience rebound anxiety (akin to the rebound hypertension that occurs when some antihypertensives are discontinued) once the tapering process is completed, but this is transient and ends within 48 to 72 hours.[ 28] Once the rebound anxiety ends, a patient may re-experience the original symptoms of anxiety, referred to as recurrent anxiety.
Although few controlled studies support the long-term use of benzodiazepines, GAD is a chronic disorder, and some patients will require benzodiazepine therapy for months to years.[ 9] Generally, patients who present with acute anxiety or those with chronic anxiety who undergo a new stressor ("double anxiety") should receive benzodiazepine therapy for several weeks.[ 13] Patients may be less tolerant of anxiety that recurs when the benzodiazepine is discontinued12 and, if necessary, it may have to be resumed indefinitely. Patients who use benzodiazepines chronically tend to be elderly, to be in psychologic distress and to have multiple medical problems.[ 30]
Other Medications. Buspirone (BuSpar) is the drug often used in patients with chronic anxiety and those who relapse after a course of benzodiazepine therapy.[ 13] It is also the initial treatment for anxious patients with a previous history of substance abuse.13 Buspirone appears to be as effective as the benzodiazepines in the treatment of patients with GAD,[ 32] and its use does not result in physical dependence or tolerance.[ 15]
Unlike the immediate relief of symptoms that occurs with benzodiazepine therapy, buspirone's onset of action takes two to three weeks.[ 31] Therefore, patients should be informed of the expected delay in relief of symptoms. Buspirone has an opposite effect of the benzodiazepines in that it treats the worry associated with GAD rather than the somatic symptoms.[ 31] However, buspirone may not be as effective in patients who have been treated with a benzodiazepine during the previous 30 days.[ 13]
The initial dosage of buspirone is 5 mg three times daily with a gradual dose titration until symptoms remit or the maximum dosage of 20 mg three times daily is reached.[ 15] If the dosage is titrated too quickly, headaches or dizziness may occur.[ 15] In patients who are taking benzodiazepines at the time of the initiation of buspirone, tapering the benzodiazepine should not begin until the patient reaches a daily dosage of 20 to 40 mg of buspirone.[ 15]
In addition to the GABA/BZ complex, research has shown that GAD involves several neurotransmitter systems, including that of norepinephrine and serotonin.[ 9] Pharmacologic agents that affect these neurotransmitters, such as the tricyclic antidepressants and the SSRIs, have been studied in patients with GAD who do not respond to the benzodiazepine therapy or buspirone. Imipramine (Tofranil) has been shown to be effective in controlling the worrying that is associated with GAD,[ 33] but whether it is as effective as benzodiazepines or buspirone in those patients who have anxiety without depressive symptoms has not been determined.[ 15] Imipramine also has anticholinergic and antiadrenergic side effects that limit its use. Desipramine (Norpramin) and nortriptyline (Pamelor) can be used as alternatives.
Trazodone (Desyrel) is a serotonergic agent, but because of its side effects (sedation and, in men, priapism), it is not an ideal first-line agent.[ 9] Daily dosages of 200 to 400 mg are reported to be helpful in patients who have not responded to other agents.[ 9] Nefazodone (Serzone) has a similar pharmacologic profile to trazodone, but it is better tolerated and is a good alternative.[ 9, 31] Paroxetine (Paxil), an SSRI, has also been studied as a treatment for patients with GAD,[ 34] but the trial was small, as has been the case with most of the antidepressants under investigation. Venlafaxine SR (Effexor) is the first medication to be labeled by the U.S. Food and Drug Administration as an anxiolytic and as an antidepressant; thus, it can be used for the treatment of patients with major depression or GAD, or when they occur comorbidly.[ 35]
Antihistamines are not potent anxiolytics. Although some antipsychotic drugs have sedating properties, they should rarely be used as therapy for patients with GAD.[ 12] Beta-adrenergic agents are useful for the treatment of patients with performance anxiety (a type of social phobia) because they lower heart rate and decrease tremulousness; they do not however, decrease the worrying or other somatic symptoms associated with GAD.[ 12]
Two herbal remedies that are often used for the treatment of anxiety are Valeriana officinalis (valerian), a root extract, and a beverage made from the root of Piper methysticum (kava-kava). Both have sedating properties, but kava has worrisome side effects that include synergy with alcohol and benzodiazepines,[ 36] dyskinesias and dystonia,[ 37] and dermopathy.[ 38] Valerian root has been reported to cause delirium and cardiac failure if abruptly discontinued.[ 39] Further studies are needed before these herbal products can be recommended as therapeutic agents for persons with GAD.
Consultation with a psychiatrist should be considered if a patient with GAD does not respond to an appropriate course of benzodiazepine or buspirone therapy. A psychiatrist can help clarify the diagnosis, determine if a comorbid psychiatric disorder is present and determine which comorbid disorder should be treated as the primary disorder. A psychiatrist can also make recommendations about therapy, including the addition of psychotherapy and changes in medications.
This article exemplifies the AAFP 2000 Annual Clinical Focus on mental health.
Address correspondence to Michael F. Gliatto, M.D., Department of Psychiatry, Philadelphia Veterans Affairs Medical Center, 38th and Woodland Ave., Philadelphia, PA 19104. (e-mail: [email protected]) Reprints are not available from the author.
Information from Your Family Doctor: What You Should Know About Generalized Anxiety Disorder
What is generalized anxiety disorder?
Generalized anxiety disorder (GAD) causes people to be worried or tense most of the time. Sometimes, they think something terrible will happen even though there's no reason to think that it will. They may also worry about health, money, family or work. They may feel tense without knowing why. GAD usually starts when people are in their early 20s. About 10 million adults in the United States have this disorder. Women are more likely to have it than men.
How do I know if I have GAD?
Most people worry and these occasional worries are normal. This doesn't mean that you have GAD. You may have GAD if you can't stop worrying and relax. As a rule, if you have GAD, you worry so much that it interferes with your day-to-day life, and it happens more days than not. Here are other signs of GAD:
Trouble falling or staying asleep
Muscle tension
Trouble concentrating, or your mind goes blank
Getting tired easily
Restlessness, or feeling "keyed up" or on edge
If you feel tense most of the time and have some of these symptoms, talk to your doctor. Your doctor will probably examine you and ask some questions to make sure that something else isn't causing your symptoms. Sometimes certain kinds of medicine may cause GAD. Your thyroid gland could be too active or you may be depressed. If your doctor doesn't find any other reason for your symptoms, you may need to be treated for GAD.
How is GAD treated?
If you have GAD, you must learn ways to cope with your anxiety and worry. You'll probably need some counseling to help you figure out what's making you so tense. Also, you may need to take some medicine to help you feel less anxious. Your doctor will be able to recommend the treatment that will be best for you. Patients with GAD can get better. If you take medicine for generalized anxiety disorder, you may be able to stop taking it in the future. To learn more about GAD, you can visit the Web site of the Anxiety Disorders Education Program at
This handout provides a general overview on this topic and may not apply to everyone. To find out if this handout applies to you and to get more information on this subject, talk to your family doctor.




Differentiating DSM-IV Anxiety and Depressive Disorders in the General Population: Comorbidity and Treatment Consequences. By: Ohayon, Maurice M; Shapiro, Colin M; Psych, MRC; Kennedy, Sidney H. Canadian Journal of Psychiatry, Mar2000, Vol. 45 Issue 2, p166, 6p, 3 charts; (AN 3204143)

The comorbidity of psychiatric disorders can have a major bearing on therapeutic choice, quality of life, and course of illness ( 1-4). Yet our knowledge of the profile and occurrence of concomitant mental disorders remains limited, even though certain surveys assessing the prevalence of psychiatric disorders in the general population have revealed high rates of association between some types of psychiatric disorders. For example, in the Epidemiological Catchment Area (ECA) survey, 54% of subjects with a lifetime history of mental illness were found to have received multiple diagnoses ( 5). Similarly, the National Comorbidity Survey (NCS) reported a rate of 56% ( 6). In clinical studies, high rates of comorbidity were found between depressive and anxiety disorders. For example, Katerndahl and Realini reported a rate of 26.7% of social phobia in their outpatient sample of 243 individuals with depression ( 7). Pini and others reported that about one-third of the 87 studied patients with a bipolar or unipolar depression also had a panic disorder. Another one-third had concomitantly a generalized anxiety disorder, 20% of the bipolar and 14.2% of the patients with unipolar depression also had an obsessive-compulsive disorder ( 8). In a study using consecutive clinic attendees without known psychiatric disorder, Stein and others found a co-occurrence of anxiety and depressive disorders in 19.2% of the subjects ( 9). However, what was found in these clinical studies remains to be reproduced in the general population. This is most important because epidemiological studies in the general population are done with a large number of subjects not interviewed by clinicians. Rather, interviews are performed by interviewers using paper-pencil or computerized questionnaires as diagnostic tools. Ideally, an epidemiological study of mental disorder comorbidity should utilize a diagnostic tool capable of positive and differential diagnosis that follows a reasoning process conforming as closely as as possible to that of a physician. To date, however, the tools employed have not afforded this possibility. Furthermore, the use of sophisticated computer algorithms, has been limited to positive diagnoses built after the end of the study. An important limitation in the use of a posteriori algorithms to reach a diagnosis is that key symptoms are not necessarily part of the disorder investigated. Consequently, crucial information is missed regarding the characteristics of individuals with comorbid conditions and how physicians deal with them. In the natural interview, the clinician explores a series of symptoms in a specific context. Even if a symptom is already described in another context, the possibility that it might be related to the disorder under investigation is not discarded, and the clinician is able to ask supplementary questions. Lay interviewers lack the necessary competence to do this. The risk of error would, in any event, probably be greater with lay interviewers than with the use of a posteriorialgorithms. One of the advantages of using expert systems in epidemiological studies lies precisely in their ability to simulate the clinical interview even if conducted by a lay interviewer. Diagnostic decision trees built during the interview respect the positive and differential diagnosis process. The lay interviewer is blind to the process and simply reads out the questions to the interviewee. Accurate recognition of mental disorders is an important issue for the treatment and follow-up of patients. The comorbidity of anxiety and depressive disorders is a particularly glaring case in point: when anxiety symptomatology is present during depressive episodes, it can obscure the underlying depression and lead to ineffective treatment. The same problem occurs with sleep disorders, and distinguishing between primary and secondary sleep disorders may help to find the most appropriate treatment. The purpose of the present study was threefold: 1) to assess the prevalence of anxiety and mood disorders in the general population using an expert system that performs differential diagnosis in real time during interviews; 2) to estimate the prevalence of the comorbidity of mood, anxiety, and sleep disorders in terms of DSM-IV differential diagnoses; and 3) to investigate medical consultations and treatment prescribed by physicians.


This epidemiological study was carded out from March 1996 to January 1997 in the Toronto Metropolitan Area. Toronto is the largest city in Canada, with approximately 3 138 415 inhabitants aged 15 years or over. A representative sample of this population was constituted using a 2-stage sampling design. First, a random sample of telephone numbers was drawn based on the population distribution of the Toronto Metropolitan Area, using the first 3 digits of the telephone numbers to identify the location of target households. Second, a controlled selection method was applied to limit the within-sampling-unit non-coverage error. Under the Kish method used ( 10), the household member to be interviewed is randomly selected according to 8 selection tables based on the age, sex, and number of residents in the household. Interviewers explained the goals of the study to potential participants before soliciting their verbal consent to proceed. Excluded from the study were subjects who did not speak sufficient English or who suffered from a hearing or speech impairment or an illness which precluded an interview. Individuals who refused to participate or who withdrew before completing at least one-half of the interview were classified as refusals. Phone numbers were dropped and replaced only after a minimum of 10 unsuccessful dial attempts made at different times and on different days, including weekends. An added-digit technique, wherein the last digit of a telephone number is increased by 1, was employed to control for unlisted telephone numbers ( 11). As a result, the final sample consisted of 13.8% unlisted numbers. The participation rate (72.8%) was calculated by dividing the number of completed interviews (n = 1832) by the number of eligible telephone numbers, which comprised all residential numbers not meeting any of the exclusion criteria (N = 2516). Five percent of the final sample consisted of individuals who initially refused to participate in the study but agreed when contacted a second time. Interviews were conducted by telephone using the Sleep-Eval Knowledge-Based System. They were performed by 30 university students who were inexperienced in psychiatric assessment but had received special training on the use of Sleep-Eval. The mean duration of interviews was 40.4 plus and minus 20.0 minutes. Interviewers were monitored daily by 2 supervisors to ensure that questions were asked correctly and data entered properly.


Sleep-Eval is an expert system specially designed to administer questionnaires and conduct epidemiological studies in the general population. It includes a nonmonotonic, level-2 inference engine endowed with a causal reasoning mode that attempts to simulate the reasoning process of a psychiatrist. The causal reasoning mode enables the Sleep-Eval system to formulate a series of diagnostic hypotheses based on the data provided by a respondent. The nonmonotonic, level-2 inference engine examines these hypotheses and confirms or rejects them through further questions and deductions. The system formulates initial diagnostic hypotheses based on the responses to a standard set of questions put to all participants and allows concurrent diagnoses in accordance with the DSM-IV classification. The differential process is based on a series of key rules allowing or prohibiting the co-occurrence of 2 diagnoses. For example, if a respondent were to meet the full criteria for both generalized anxiety and major depressive disorder, additional questions would need to be asked to determine whether generalized anxiety occurred exclusively in the course of major depressive disorder. If so, only the diagnosis of major depressive disorder would be given. The system terminates the interview once all diagnostic possibilities are exhausted. The design of the expert systems questionnaire ensures that any decision regarding the presence of a symptom is based on the interviewees' responses rather than on the interviewer's judgement. This approach has been proven to yield better agreement between lay interviewers and psychiatrists on the diagnosis of minor psychiatric disorders ( 12). Questions are selected and phrased by the system. As they appear on a computer monitor, the interviewer simply reads them out and enters the responses. Examples and instructions on quoting the answers are provided. Questions can be closed-ended (for example, yes-no, present-absent-unknown, five-point scale) or open-ended (for example, name of illness, duration). Further details on the methodology and on the Sleep-Eval system can be found elsewhere ( 13, 14). The system has been tested in various contexts: in clinical psychiatry, kappas between the diagnoses of 4 psychiatrists and those of the system ranged from 0.44 with 1 psychiatrist to 0.78 with 2 psychiatrists (n = 114 cases) ( 15, 16). Another study involved 91 forensic patients. The kappa between diagnoses obtained by the system and those given by psychiatrists was 0.44 for specific psychotic disorders (mainly schizophrenia) ( 17). In a study performed in the general population (n = 150), the diagnoses obtained by 2 lay interviewers (inexperienced in sleep and psychiatric assessments) using Sleep-Eval were compared with those obtained by 2 clinician psychologists. A kappa of 0.85 was obtained in the recognition of sleep problems and of 0.70 for insomnia disorders. In another study performed in 2 sleep disorders centers (Stanford, USA; Regensburg, Germany) the diagnoses of the Sleep-Eval system were compared with that of the sleep specialist. Overall agreement on any sleep-breathing disorder was 96.9% (kappa 0.94). More than one-half of the patients were diagnosed with obstructive sleep apnea syndrome (OSAS); the agreement rate for this specific diagnosis was 96.7% (kappa 0.93) with no significant difference between the 2 sites ( 18).


A weighting procedure was applied to correct for disparities in the geographical, age, and sex distribution between the sample and the Toronto area population as per the 1991 Canadian census. Results are based on weighted n-values. Percentages for target variables are given with 95% confidence intervals (95%CI). Bivariate analyses were performed using the chi-square test with Yates's correction or Fisher's exact test when n-values were less than 5. Reported differences were significant at the 0.05 level or less.


After weighting, the sample comprised 48.3% men and 51.7% women, ranging in age from 15 to 90 years. Most of the respondents were white (73.7%). Black respondents made up 5.7% of the sample, and Asians 7.8%. One-half of the sample was married (49.9%) and one-third (35.5%) was single. One-half (53.1%) had 11 to 13 years, and one-fifth (20.7%) had fewer than 11 years of schooling.

Overall Results

Overall, 13.2% (n = 242) of the sample had either a mood disorder (n = 127;6.9%) or an anxiety disorder (n = 170;9.3%) at the time of interview. As Table 1 shows, the prevalence was higher among women (16.5%) than among men (9.7%), for an odds ratio of 1.8 (P < 0.001).
Similarly, the prevalence of depressive disorders (major depressive disorder, single episode or recurrent, and dysthymic disorder) was higher among women (9.0%) than among men (4.7%), as was the prevalence of anxiety disorders. Panic disorder, agoraphobia, and posttraumatic stress disorder accounted for the significant difference between sexes. Rates for bipolar disorder (bipolar disorder type I or II and cyclothymic disorder) were comparable for men and women.
As indicated in Table 2, the highest rates of anxiety or depressive disorders were found in the youngest age group ( 15-24years). However, only the oldest age group (65 years and over) posted a rate significantly lower than that for each of the other age groups. Of the 127 respondents with a mood disorder, 55 (43%) were also diagnosed with an anxiety disorder, 23 (18%) other subjects had an anxiety disorder occurring only during the course of their mood disorder. This represents 3% of the entire sample. The most common co-occurring anxiety disorders were panic disorder and posttraumatic stress disorder (Table 3). The comorbidity of anxiety and mood disorders was higher among women (4.2%) than among men (1.7%; P < 0.001), for an odds ratio of 2.6. Differences across age groups were not significant. Insomnia symptoms are often observed in individuals with mood or anxiety disorders. In our sample, 2 symptoms differentiated respondents with both an anxiety and a mood disorder from those with a mood disorder alone and those with an anxiety disorder alone:
  1. Disrupted sleep (occurring at least 3 nights per week with difficulty resuming sleep) was reported by 54.9% of respondents with both types of disorders, compared with 40.4% of those with only a mood disorder, 30.8% of those with only an anxiety disorder, and 15.1% of the rest of the sample (P < 0.0001).
  2. Early morning awakenings were reported by 46.3% of respondents with both disorders, compared with 34.4% of those with only a mood disorder, 27.4% of those with only an anxiety disorder, and 10.4% of the rest of the sample (P < 0.0001 ).

Difficulty initiating sleep was reported by one-half of the respondents with both disorders (50.6%), compared with 41.1% of those with a mood disorder alone, 23.2% of those with only an anxiety disorder, and 9.7% of the rest of the sample (P < 0.0001). Rates for nonrestorative sleep were comparable for respondents with both disorders (46.4%) and those with only a mood disorder (48.2%), but were much lower for those with only an anxiety disorder (25.5%) and those without these 2 disorders (10.3%; P < 0.0001). As a result, more than one-half of the respondents (56.7%) with both a mood disorder and an anxiety disorder reported being moderately or severely sleepy during the day. This was the case for 37.6% of those with only a mood disorder and 19.8% of those with only an anxiety disorder (P < 0.0001).

Medical Consultations

Respondents were asked whether they had consulted a physician in the past year and whether they had consulted a physician for a psychological or nervous problem in the past year and in their lifetime. Overall, 79.0% of the sample had consulted a physician at least once in the past year. Respondents with an anxiety or a mood disorder reported slightly more frequent consultations (83.8%) than did the rest of the sample (78.2%; P < 0.05). About 5% of respondents with both disorders or a mood disorder alone had consulted a psychiatrist, compared with 1.4% of respondents with an anxiety disorder alone and 0.6% of the rest of the sample (P < 0.005). Among respondents who consulted a physician at least once in the past year, those with both disorders did so significantly more frequently than all the other groups, averaging 8.7 consultations compared with 5 for those with a mood disorder alone. 5 for those with an anxiety disorder alone, and 3.7 for the rest of the sample (F [ 3,1337] = 12.292; P < 0.0001). Medical consultations in the past year for a psychological or nervous problem were reported by 5.7% of the sample. The highest rate for this type of consultation was found among respondents with both an anxiety and a mood disorder (49.4%), followed by those with a mood disorder alone (36.4%), and those with an anxiety disorder alone (13.6%; P < 0.0001). Lifetime consultations for a psychological or nervous problem were reported by 10.5% of the sample. Rates were comparable for respondents with both an anxiety and a mood disorder (48.1%) and those with a mood disorder alone (50.4%), but lower for those with an anxiety disorder alone (27.9%; P < 0.0001).


Use of antidepressants was more prevalent among respondents with both an anxiety and a mood disorder and among those with a mood disorder alone (13.8% and 18.2%, respectively). The rate was 4.3% among respondents with an anxiety disorder alone and 1.8% in the rest of the sample (P < 0.0001). The most commonly used antidepressants were fluoxetine (36.4%), amitriptyline (20%), and sertraline (14.5%). Anxiolytics were more likely consumed by respondents with both an anxiety and a mood disorder (9.8%) than by those with a mood disorder alone (4.2%), those with an anxiety disorder alone (3.0%) or the rest of the sample (1.3%; P < 0.0001). About one-half (51.5%) of the anxiolytic consumers used lorazepam and 24.2% used diazepam.
Use of hypnotics was highest among respondents with both an anxiety and a mood disorder and among those with a mood disorder alone (5.6% and 4.3%, respectively). The rate was 1.9% among respondents with an anxiety disorder alone and 0.8% in the rest of the sample (P < 0.0001). The most common hypnotics were temazepam and zopiclone, each accounting for 25% of the consumption. Overall, about one-third of the respondents (34.5%) with both an anxiety and a mood disorder received some form of treatment from their physicians. About 30% of those with a mood disorder alone and 11.7% of those with an anxiety disorder alone also received some treatment. Nonpharmacological treatment accounted for less than 10% of physician prescriptions.


This epidemiological study confirms the high prevalence of anxiety and mood disorders in the general population. The comorbidity of mood and anxiety disorders was found in 3 % of the entire sample. This means that 43.7% of subjects with a mood disorder also have an anxiety disorder. This is close to figures reported by the ECA and the NCS studies ( 5, 6). The point prevalence is also comparable to that of recent epidemiological surveys in North America. The ECA survey set the one-year prevalence rate of major depressive episode at 3.7% ( 19). The NCS found the past-month prevalence of major depressive episode, as per DSM-III-R criteria, to be 3.8% for men and 5.9% for women ( 20). Using the same diagnostic tool as in the NCS, the Ontario Health Supplement survey reported a one-year prevalence rate of major depressive episode of 4.1% ( 21).
Although other studies have investigated psychiatric comorbidity, none have applied the DSM-IV differential diagnosis process during the interview, most likely due to the need for skilled interviewers with training in psychiatry. Needless to say, such large-scale epidemiological surveys would cost an exorbitant amount. Consequently, interviewers are usually lay persons trained to use a specific tool. Surveys conducted to date have been limited in 2 respects: first, as a result of a posteriori differential diagnosis, and second, as a result of the diagnostic classification method applied. Consequently, prevalence rates for specific disorders and the comorbidity rates for certain disorders seen thus far tend to be inflated. The use of a posteriori diagnostic algorithms allows a thorough investigation of the disorders under study, but the results have been shown to differ considerably from the clinical practice where only the most significant diagnosis is considered with all other manifestations treated as just a part of the diagnosed disorder. This is a traditional approach with paper-pencil or computerized questionnaires. In the latter, predetermined diagnostic trees can be implemented in the software, but the apparent reasoning is just an artifice. These computerized questionnaires do not have the capacity to adjust their reasoning and to explore other diagnostic paths. In expert systems, the decision trees are built during the interview, seeking the optimal way to achieve a diagnosis with respect to positive and differential diagnostic indications provided by the referent classification. The prevalence rates of anxiety disorders in our survey were found to be lower for social phobia, generalized anxiety disorder, specific phobia and obsessive-compulsive disorder than those reported in the ECA study ( 19) and in other studies using DSM-III or DSM-III-R diagnostic criteria. This is most likely due to changes that have occurred in the definitions of these disorders over time. For example, the DSM-III-R and the DSM-IV have an additional criterion not present in the DSM-III stipulating that specific or social phobia has to interfere with the daily life of the individual or provoke marked distress. In our study, this criterion decreased the prevalence of these disorders by nearly 50%. The definition of generalized anxiety disorder has also changed considerably in DSM-III-R compared with DSM-III (that is, the duration criterion was increased from 1 to 6 months) and even more so in the DSM-IV (that is, additional symptoms were changed from the presence of at least 6 of 18 symptoms to the presence of 3 of 6 symptoms). The other possible explanation for these lower rates lies in the tool we used. Often, respondents met all the diagnostic conditions for a disorder except the differential diagnosis criteria (for example, the disorder did not occur exclusively in the course of a major depressive episode). In the case of social phobia, 16% of the sample initially reported a fear of social situations, but only 1.4% of the sample experienced distress or impaired functioning. Three subjects avoided social situations because they feared a panic attack, while 11 others feared them only during the occurrence of another mental disorder. Consequently, only 0.9% of the sample had a diagnosis of social phobia. The risk of a concomitant diagnosis of anxiety disorder is 5 to 15 times as high for individuals with a mood disorder, with the risk being highest for the combination of depressive disorder and posttraumatic stress disorder. Other epidemiological studies have reported a similarly high risk. A few years ago, Akiskal suggested the existence of a new type of disorder where individuals alternate between panic disorder and depressive disorder ( 22). In the present study, 14% of respondents with depressive disorder presented with a phasic disorder of this sort.
The increased prevalence of sleep disturbances in the comorbid group adds significant new findings to previous epidemiological reports and highlights the importance of exploring all facets of initiation, maintenance and termination of sleep. In this study, we found that 86.6% of the comorbid group had at least 1 insomnia complaint. This was 75.5% in the group of mood disorder alone and 49.2% in the group of anxiety disorder alone. High occurrences of insomnia in depressive disorders have also been reported by other studies ( 23), but few of them have attempted to elucidate the relationship between insomnia and mood and anxiety disorders or to apply differential diagnosis process to select the final diagnosis ( 24, 25). This study also revealed that less than one-third of respondents with a mood and/or anxiety disorder were being treated for a mental disorder by a physician. However, these same individuals were greater consumers of health care services. Most consulted a physician on average 5 times in the past year. On the other hand, individuals on medication diagnosed with a mood and an anxiety disorder consulted a physician 12 times on average in the past year. This suggests that these individuals are at least being closely monitored. Unfortunately, only 13% of them were treated with antidepressants, and under 9% with anxiolytics, suggesting that only the anxiety component was recognized and treated. More alarming is the fact that about 60% of cases are neither recognized nor treated. This situation is not specific to Toronto. It has been identified also in Montreal, in Edmonton, and in Europe ( 26-29). Given the large number of individuals with mental disorders who consult a general practitioner at least once a year, several educational programs have been launched around the world to help physicians recognize and treat mental disorders ( 30, 31). Unfortunately, many of these campaigns have met with limited success, mainly because such efforts must be continuous in order to provide physicians with information on the constellation of symptoms that characterize mental disorders.
In conclusion, this is the first attempt to apply the DSM-IV differential diagnosis process during the course of the interview in a general population survey. Results confirm the importance of differentiating among mood, anxiety, and sleep disorders and of identifying the pathological context in which they occur. The prominence of one disorder over the other was assessed, and this assessment revealed the importance of studying their interaction and cooccurrence carefully to ensure more accurate treatments and follow-up care by physicians.

Clinical Implications

Co-occurrence of mood and anxiety disorders involves insomnia complaints in nearly 90% of the cases.
Insomnia is mainly secondary to a mood or anxiety disorder.
About 60% of mental disorders remain unrecognized and untreated.

Table of Contents