U.S. flag

An official website of the United States government

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

O'Connor EA, Whitlock EP, Gaynes B, et al. Screening for Depression in Adults and Older Adults in Primary Care: An Updated Systematic Review [Internet]. Rockville (MD): Agency for Healthcare Research and Quality (US); 2009 Dec. (Evidence Syntheses, No. 75.)

Cover of Screening for Depression in Adults and Older Adults in Primary Care

Screening for Depression in Adults and Older Adults in Primary Care: An Updated Systematic Review [Internet].

Show details


Scope and Purpose

We conducted this systematic review to aid the United States Preventive Services Task Force (USPSTF) in updating its 2002 recommendation for adult depression screening in primary care. We focused on gaps in evidence identified by the previous review1 and on integrating relevant research published in the interim. Questions that the USPSTF judged to have strong, coherent evidence in the previous review are not re-addressed here. Specifically, we did not update evidence regarding the accuracy of screening instruments for identifying depressed adults and older adults in primary care, nor treatment of adult depression with antidepressants or psychotherapy. We updated direct evidence that primary care depression screening programs improve health outcomes and examined evidence for the efficacy of depression treatment in older adults and evidence for the harms of screening and adverse events from antidepressant treatment in adults and older adults.

Condition Definition

The term “depression” is not a specific term for a single diagnostic condition. Depressive disorders generally consist of major depressive disorder (MDD), dysthymia, and minor depression, but not other conditions that include depressive features, such as bipolar disorder. The American Psychiatric Association2 specifies diagnostic criteria for the different depressive disorders, each of which requires a minimum number of symptoms to be present and significant distress or impairment (Table 1). MDD is the most serious diagnosis, and is given to a person who meets criteria for major depressive disorder without manic or hypomanic features or a psychotic disorder. Dysthymia is similar to MDD but is generally longer-lasting and less severe. A variety of terms are used for people with depressive symptoms but whose depression does not meet criteria for MDD or dysthymia, such as subthreshold depression, subsyndromal depression, and minor depression. Some studies use the DSM-IV definition of minor depression (developed for research rather than clinical purposes), but many define these patients idiosyncratically so that it is very difficult to compare rates across studies.

Table 1. Primary DSM-IV depression disorders, criteria for adults.

Table 1

Primary DSM-IV depression disorders, criteria for adults.

Prevalence and Burden of Disease/Illness

Depressive disorders are common in community and primary care patients. The estimated lifetime prevalence of MDD is approximately 13.2 percent, with a 12-month prevalence of 5 to 7 percent in community-dwelling adults.3–5 Prevalence Estimates in community-dwelling older adults are much lower (from 1 to 5 percent,6 with an average of 1.8 percent).7 A different study found that approximately one to 2.5 percent of older adults are likely to experience a first episode of depression over the course of one year.8 In primary care settings, the prevalence of MDD ranges from 5 to 13 percent9–12 in adults and from 6 to 9 percent in older adults.13,14 For MDD patients receiving treatment in primary care settings, depressive symptoms and severity are equivalent to that seen in MDD patients treated in psychiatric settings; of note, approximately 43 percent of such primary care patients report suicidal ideation within the prior week.15,16

Considering other depressive disorders (e.g., dysthymia, and subthreshold depressive disorders) increases the prevalence. Twelve-month prevalence of dysthymia is estimated at 1.5 to 1.6 percent3,5 in community-dwelling younger (aged 18 to 54) and older (aged ≥ 55) adults.5 Dysthymia in primary care settings is estimated from 2 to 4 percent.17 It is difficult to reliably estimate the prevalence of subthreshold disorders in the US due to the wide range of definitions used. In primary care settings, prevalence of broadly defined subthreshold disorders is about 9 percent in adults18 and 10 percent in older adults.

Depression has been ranked as a leading cause of Years of Life Lived with a Disability (YLD) for persons 15 years and older19 and the third leading cause of loss in quality-adjusted life years (QALY) in older adults.20 In addition to its impact on the depressed person, depression is often associated with a drastic loss of productivity at work and home21,22 and impairment in relationships and social functioning.21 Depression in parents is associated with behavioral and emotional difficulties in their children.23–25 Depression may increase the risk of physical disability,26–29 medical conditions (such as coronary heart disease and diabetes mellitus),30,31 other mental health conditions, and mortality.30–34 In one study, the increased mortality seen in depressed older adults was comparable to that seen in patients with emphysema or heart disease.33 Depression is also a major risk factor for suicide. Suicide mortality among patients treated for depression is estimated at 59 per 100,000 among an insured population.35

Depression’s economic burden is substantial and includes individual costs (suffering, treatment side effects, possible suicide, health care and medication fees, work disability, and lost earnings); costs to family and friends (informal care-giving, time off work, career burden); employer costs (contributions to treatment and care, reduced productivity); and costs to society (costs of mental health and general medical care, reduced productivity, and loss of lives).36 In 2000, combined direct and indirect costs of depression in the US were estimated at 83.1 billion dollars, 31.5 billion in direct costs and the remaining in indirect, mostly workplace costs.37 A 2006 study of depression costs in Europe indicated that 87 percent of the costs associated with depression were indirect costs, such as losses due sickness-related absence from work.38 Studies of primary care patients have found that health care costs are higher in depressed patients than non-depressed in many categories, including primary care visits, medical specialty visits, lab tests, pharmacy costs, inpatient medical costs, and mental health visits.31,39 A 1999 study found that while many high utilizers were depressed, their providers often did not recognize their depression.40

Natural History

Depression is a chronic disease characterized by partial remissions and recurrences in most of those who recover fully.41 While depression can occur in people of any age,17,21 the average age of onset is in the mid-twenties.42 Cumulative Kaplan-Meier curves for age-at-onset show fairly low risk until the early teens, with subsequent risk rising in a roughly linear fashion.21

Many of the treated depression cases are managed in primary care; roughly one third to one half of non-elderly adults 21,43 and almost two thirds of older adults44 who are treated for depression are treated in primary care. Large-scale studies of patients initiating treatment for depression indicate that about half to two-thirds of patients achieve remission within a year,9,45–47 although remission may require up to 4 adequate treatment trials.46 Patients seen in primary care with depressive disorders whose depression was not recognized by providers do about as well: 50 to 60 percent of these patients are also likely to recover from their depression.9,47,48 This may be partially due to the fact that patients with unrecognized depression often have less severe symptomatology and impairment.9,47,49 With provision of evidence-based treatment, recovery rates for identified depressed patients in primary care are equivalent to similarly depressed patients treated in psychiatric settings.50

Lower rates of recovery have been seen in population-based studies of depression in the community. Two large Canadian epidemiological databases were used as a basis for developing a depression prognosis calculator, where rate of recovery is estimated from length of current episode.51 The calculator52 estimated the 12-month recovery rate for patients whose depression episodes lasted for 12 weeks to be 27 percent, which is considerably lower than the approximately half to two-thirds of patients reported to recover in treatment settings. However, this rate is similar to 28 to 33 percent rate for a single antidepressant trial in the Sequenced Treatment Alternatives to Relieve Depression (STAR*D) study, a large-scale, community-care based effectiveness study of depression treatment.53

Older adults have similar or slightly lower recovery rates than younger adults, possibly due to more frequent medical co-morbidities.54 A meta-analysis of depression in older adult community and primary care populations found that after 2 years an average of 33 percent of patients were categorized as “well,” 21 percent had died, and 33 percent still met criteria for full depression. The remaining participants were described as experiencing partial remission or other mental health disorders, such as dementia.55 This study did not describe recovery rates specifically in treated populations. Similarly, a naturalistic study reporting on the course of depression in older adult primary care patients found that 39 percent of patients with diagnoses of either major or minor depression were in complete remission one year later under usual care conditions56 and 25 percent still met full criteria for major or minor depression.

Despite fairly high rates of recovery from a particular episode, depression is highly recurrent. A recent evidence-based NICE guideline noted that at least half of individuals diagnosed with depression will have a recurrence following their first episode of depression.57 This chance of recurrence increases with subsequent episodes. The STAR*D trial found that about half of patients who achieved remission relapsed during the subsequent year.46

Risk Factors for Depression

Individuals are at risk for depression across their entire adult life span. Consistently identified high-risk groups include: women;6,21,58 people with other psychiatric disorders, including substance misuse;6,21,59 people with a family history of depression;6 people with chronic medical diseases;60 and people who are unemployed or with lower socio-economic status.6,21,58,61 While the prevalence of MDD is lower in community-dwelling older adults than in younger adults, significant depressive symptomatology is associated with common life events in older adults, including medical illness, cognitive decline, bereavement, and institutional placement in residential or inpatient settings.6,62

Depression in Older Adults

Although MDD is somewhat less prevalent in older adults, depression is a significant public health issue is this age group. Older adults have the highest risk of suicide of all age groups. According to a 1992 NIH Consensus Development Panel on late-life depression, most of these suicidal patients were experiencing their first MDD episode, which had gone unrecognized and untreated.63 These patients are highly relevant to primary care clinicians because more than 50 to 75 percent of older adults who commit suicide have seen their medical doctor during the prior month for general medical care, and 39 percent are seen during the week prior to their death.64

Depression can be particularly difficult to identify in older patients, and much of the burden of depression diagnosis will fall to primary care providers. Diagnosis is complicated because medical conditions or medications can cause symptoms of depression, such as weight loss or appetite change, psychomotor retardation, loss of energy or fatigue, insomnia or hypersomnia, and difficulty concentrating. Further, the depressive symptoms of depressed mood and feelings of guilt tend to be less prominent in older depressed patients, whose primary complaints tend to be somatic.65 Depression in older adults is further complicated by the high levels of co-morbidity with medical conditions, including cancer, cardiovascular disease, neurological disorders, metabolic disturbances, arthritis, and sensory loss.63,66

Interventions and Treatment for Depression

Remission of most or all symptoms is the desired outcome of depression treatment67,68 and is associated with improved functioning in adults.69 Improvements in depressed mood may help reduce functional decline in older adults.70

Response to treatment is typically defined as a reduction of at least 50 percent in baseline symptom levels. Response without full remission, however, is associated with continuing impairments in psychosocial functioning, productivity, continued disabling symptoms and higher levels of health-care use, and higher rates of relapse, recurrence, and potentially suicide.53

Effective depression treatment in adults include pharmacotherapy and psychotherapy, delivered singly or in combination.71 These treatments are widely available for delivery by, or referral from, primary care providers. In samples limited to primary care patients, pharmacologic treatments, such as selective serotonin reuptake inhibitors (SSRIs) and tricyclic antidepressants (TCAs), are more effective than placebo,72 with some comparative evidence favoring SSRIs over TCAs.73–75 The reduced tolerability and increased toxicity in overdose of TCAs and other “first-generation” antidepressants has resulted in the use of SSRIs and other “second-generation” antidepressants for the majority of pharmacologic depression treatment delivered in primary care.76 By 2000, SSRIs accounted for 65 percent of all antidepressants prescribed in primary care and another 17 percent consisting of non-SSRI “second-generation” antidepressants (see Table 2 for listing and categorization of antidepressants).77 Indeed, between 1996 and 2001, the number of persons using these newer antidepressants increased from 7.9 million to 15.4 million, while the number using TCAs decreased from 2.3 million to 1.2 million.78

Table 2. List of antidepressants and their categorizations.

Table 2

List of antidepressants and their categorizations.

Remission rates at 12 to 14 weeks with SSRIs range from 2279 to 28 percent53 in naturalistic settings, and 3580 to 47 percent81 in research settings. Rates of symptom improvement without full remission are higher (47 to 63 percent).53,75,79,81 Remission rates for psychotherapy appear to be comparable to antidepressants. A systematic review of three-arm, intent-to-treat trials comparing antidepressants, psychotherapy, and a control condition reported 46 percent remission rate for antidepressants and 46 percent remission rate for psychotherapy after 10 to 16 weeks.82 Additionally, a 2001 Health Technology Assessment Report83 showed a 52.5 percent overall remission rate for trials comparing some form of psychotherapy with a control condition.

Current Practice

Detection and Treatment of Depression in Primary Care

Current mental health screening rates may be as high as 74 percent in primary care, according to Healthy People 2010 midcourse review,84 although these estimates aren’t specific for depression screening. Once a primary care provider has identified a patient as depressed, almost 90 percent of providers recommend antidepressants, either alone or in combination with psychotherapy.85,86 Only 25 percent of patients receive follow-up visits meeting Healthcare Effectiveness Data and Information Set (HEDIS) criteria of three visits within the first 12 weeks,87 and among those patients who initiate antidepressant use, however, up to 40 to 67 percent discontinue use within 3 months87–89 in real-world settings. This is considerably higher than discontinuation rates reported in the context of clinical trials, where early treatment discontinuation rates range from 16 to 29 percent.72,75,90–95,46,96 A further synopsis of recent evidence on depression detection and treatment in primary care can be found in Appendix A.

Proportion of Depression Cases Missed in Primary Care

Although we found no recent evidence on the proportion of depression cases not detected in current primary care in the US, a study in a staff model health maintenance organization (HMO) in western Washington state in the early 1990’s suggests that 30 to 40 percent of cases may be missed by primary care providers.9 In this study, clinicians recognized 64 percent of patients with MDD, and those not recognized were more likely to have less severe depression and to be younger. Three major sets of evidence-based guidelines for recognition and treatment of depression in the US have been developed since this study was conducted, which may have, at minimum, increased awareness of depression and possibly improved the current depression recognition and treatment practices: Agency for Health Care Policy and Research (currently known as AHRQ),17,97 Veteran’s Health Administration (VA),98 and the American Psychiatric Association (APA).99 Thus, current recognition rates may be at least comparable to, or possibly higher than, those published in the 1995 study.

Rationale for Depression Screening

Mass screening in primary care may help clinicians identify missed depression cases and initiate appropriate treatment. Additionally, screening may help clinicians identify patients earlier in their course of depression. In both of these cases, it is presumed that usual care delivers effective treatment and that treating these patients would improve their depression and alleviate their suffering sooner or more thoroughly than if they had not been screened. Unlike other screening tests, screening all patients for depression, including those previously identified as depressed, may be useful since it might help identify ineffectively treated patients whose treatment needs modification.

Depression Screening Instruments

The previous USPSTF review71 found that there are reliable and valid depression screening instruments for adults. Screening instruments generally demonstrated sensitivity of 80 to 90 percent and specificity of 70 to 85 percent. The previous reviewers modeled the probability of MDD after a positive screening test using several sensitivity and specificity estimates in these ranges, and background population depression rates of 5, 10, and 15 percent. They estimated that between 12 percent and 50 percent of those screening positive would meet criteria for MDD, with most estimates falling between 24 and 44 percent. Thus, the majority of patients screening positive will not meet criteria for MDD, though some of these may still benefit from counseling or treatment. Clearly, screening instruments are not sufficient for diagnosing depression, but do indicate the need for more detailed follow-up by a clinician to determine whether the person meets diagnostic criteria for a depressive disorder, to explore other possible causes for depression (such as hypothyroidism or medication or substance use), and assess for co-existing psychiatric disorders.

Controversies about Depression Screening

Since the previous USPSTF-sponsored review,71 other reviewers have reached different conclusions about depression screening. Pignone et al. concluded that depression screening in primary care is effective in improving health outcomes, but only in the presence of other systems to ensure accurate diagnosis, treatment, and follow-up among patients screening positive.71 Other reviewers have concluded that screening does not improve health outcomes,100 but that care management systems for depressed patients significantly improve rates of depression remission.101 Commentators on these divergent review results have also been divided.102,103

Previous USPSTF Recommendations

In 2002, the U.S. Preventive Services Task Force (USPSTF) recommended screening adults for depression in clinical practices that have systems in place to assure accurate diagnosis, effective treatment, and follow-up (B recommendation).

The USPSTF concluded the evidence is insufficient to recommend for, or against, routine screening of children or adolescents for depression (I recommendation). Since this recommendation’s update is being conducted separately, this report only addresses adults.


  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (2.2M)

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...