Universal Depression Screening to Improve Depression Outcomes in Primary Care: Sounds Good, but Where Is the Evidence?

The 2016 depression screening recommendation statement by the U.S. Preventive Services Task Force (USPSTF) (1) endorses screening for depression in the general adult population, including pregnant and postpartum women. The recommendations are similar to statements published by the USPSTF in 2002 and 2009. The recommendations were based on a review of studies reporting the benefits of depression screening in settings with adequate support systems (2). The review also concluded that the magnitude of any harm of depression screening is small to none.

Improved detection of depression in primary care is intuitively appealing. Many people with depression do not actively seek professional help, and only about half of individuals with depression in primary care are diagnosed as such (3). Screening instruments have been shown to have greater accuracy than unaided clinician diagnoses. For example, the sensitivity (.80) and specificity (.92) of the nine-item Patient Health Questionnaire (PHQ-9), a commonly used screening instrument (4), are much higher than the sensitivity (.50) and specificity (.81) of primary care providers’ diagnoses (3), suggesting that universal depression screening would improve detection of otherwise undetected cases.

In 2013, in stark contrast to the USPSTF recommendations, the Canadian Task Force on Preventive Health Care (CTFPHC) recommended against routinely screening for depression among adults, including at-risk groups (5). The CTFPHC reviewers failed to identify any studies evaluating the benefits of depression screening. Although the report did not identify any direct evidence for harm, the authors expressed concerns about the unnecessary treatment of false-positive cases and the opportunity cost of depression screening, given that the time could be used for other primary care activities.

The differences between the two recommendations are puzzling. The authors of the recommendations had access to the same literature. Yet, they came away with very different conclusions from their reviews of this literature. The question is, Why?

The Evidence Base

A careful reading of the USPSTF and CTFPHC recommendations and the supporting literature reviews suggests that each task force addressed a somewhat different question. The CTFPHC inquired about the evidence that screening for depression improved depression outcomes. Only studies that examined the effects of “routine screening as a normal part of care” in comparison with a no-screen study arm were included (6). The CTFPHC reviewers found no such studies.

In contrast, the USPSTF reviewers asked whether primary care depression “screening programs” resulted in improved outcomes and whether sending screening results to providers improved outcomes (2). The reviewers found only one study by Williams and colleagues that directly addressed the first question (7) and eight studies that addressed the second question.

The study that addressed the first question randomly assigned patients to usual care or to one of two methods of case finding: using a single question about mood or using a 20-item instrument. Screening results were reported to the physicians (7). The gold standard for clinical depression in the study was a semistructured interview conducted after the physician visit. Screening increased the likelihood of receiving a diagnosis of depression, but not counseling, filling prescriptions for antidepressants, or mental health referrals. The study also failed to find evidence of “consistent benefit” in a three-month follow-up (7).

The CTFPHC reviewers did not include the Williams and colleagues’ (7) study in their review of evidence. They argued that “the goal of this study was to determine the difference between two screening tools and all participants underwent a diagnostic interview which meant that the participants were preselected for depression and not an asymptomatic population” (6). This decision is somewhat puzzling because the study by Williams and colleagues (7) does not report excluding patients based on the results of the diagnostic interview.

The eight studies in the USPSTF review that examined the benefits of sending results to providers screened all patients and randomly sent the results of screening to providers in the intervention arm. Improved outcomes were found in only four of these studies (2). The description and results of these studies were included in the systematic review that accompanied the recommendation statement (Tables 18–21) and in an appendix to the report (Appendix D, Tables 16–19) (2). In each of these eight studies, depression screening was only one element of a package of enhanced depression care that involved other staff, further training, and—often—additional services. The authors of the USPSTF’s systematic review admitted that it was “impossible to isolate the effects of screening alone” based on these data (2). Accordingly, the USPSTF recommended that screening should be implemented “with adequate systems in place to ensure accurate diagnosis, effective treatment, and appropriate follow-up”(1).

Referring to the studies that included depression screening as one component of a package of services, the CTFPHC writers argued that it was not clear that screening was a “necessary component” of these enhanced care packages. Therefore, even in integrated staff–assisted programs in which support services are available, the CTFPHC did not recommend universal screening.

The differences in these recommendations may also be partly the result of the USPSTF’s view regarding the availability of support services in the current U.S. health care system. The USPSTF authors stated that depression support services are “now much more widely available and accepted” than in 2009, when the previous version of the USPSTF recommendation statement was published (1). However, surveys of primary care practices in the United States show that only a small proportion of practices use care management processes for depression (for example, use of nurse care managers, patient educators, and patient registries), with little change between 2006 and 2013 (8). These findings contradict USPSTF’s assessment of availability of staff-assisted depression support in U.S. care settings. Furthermore, the USPSTF recommendation did not specify the nature of the required support services.

Moreover, the two task forces differed in their views on potential harms of screening. Whereas the USPSTF recommendation statement mainly focused on harms associated with antidepressant medications, the CTFPHC statement voiced concerns about the increase in the number of false-positive cases. The impact of this concern on the CTFPHC’s recommendations was likely amplified by the task force’s conclusion that screening is not beneficial, which may have tipped the benefit-harm balance toward net harm.

There is some evidence that universal screening would increase the prevalence of diagnosis and treatment of depression in primary care settings (9,10). For example, incentivizing depression screening for primary care patients with diabetes and coronary heart disease under the British Quality Outcome Framework (QOF) of 2006–2013 led to an increase in diagnosis and treatment of depression among these patients (9,10). Doubts about the benefits of depression screening led to the withdrawal of the incentive from the QOF in subsequent years. Surprisingly, the results of this natural experiment were not discussed in the two task force reports or their respective reviews of the literature (1,2,5,6).

One can speculate about the reasons why universal depression screening may not produce the benefits that many of its supporters hope to achieve. Patients who seek treatment for depression, and are diagnosed by physicians, are often more symptomatic and distressed than patients whose depression is not clinically detected and is detected only via screening (11). There is evidence that treatment of depression is more efficacious among patients with more severe symptoms (12). Furthermore, individuals who seek mental health treatment tend to have a more positive attitude toward treatment and greater motivation to follow treatment recommendations (11). These factors may lead to different treatment outcomes among patients who are diagnosed clinically and those who are detected via depression screening.

Failing to identify convincing direct evidence for the clinical benefits of depression screening, USPSTF relied on indirect evidence of benefit to advocate for universal screening, including evidence that screening instruments can identify depression and that treatment with net clinical benefit is available for persons with depression.

Other USPSTF screening recommendations for physical health problems have similarly relied on indirect evidence supporting the benefits of early diagnosis and treatment (http://www.uspreventiveservicestaskforce.org/BrowseRec/Index/browse-recommendations). For example, in its review of evidence for the 2015 USPSTF recommendations for type 2 diabetes screening, the task force identified only two long-term randomized studies that directly examined the benefits of screening; neither reported any benefits in terms of reduced mortality. Instead, the task force based its recommendation on studies that had established the benefits of lifestyle changes in preventing or delaying progression to diabetes among at-risk individuals. The 2015 USPSTF recommendations for blood pressure monitoring among adults ages 18 and older identified only one randomized trial that reported beneficial effects of blood pressure screening. The sample of that study, however, comprised adults ages 65 and older.

Future Research

Based on the previous USPSTF and other recommendations, many integrated care systems have already adopted universal depression screening. Data from these care systems will provide some of the evidence needed for future screening recommendations. However, there remains a need for evidence from randomized controlled trials comparing depression screening in the context of usual care against usual care without depression screening as well as trials examining the added benefit of depression screening in collaborative and enhanced depression care programs (13). These future studies should recruit from settings with high and low prevalence of detection and treatment of depression, given that the impact of the screening may vary across these settings.

Future research also needs to examine the benefits of selective use of depression screening. There is some evidence that selective depression screening may improve the match between diagnosis of mood disorders and antidepressant medication prescriptions (14). Depression measures may also be useful for tracking response to treatment (11). Consistent with this view, the 2009 British National Institute for Health and Care Excellence guidelines recommended selective use of screening measures in the context of a stepped care model for depression (15).

Conclusions

The concerns raised here are not new and have been voiced by others in response to previous versions of the USPSTF recommendations for universal depression screening (11,13). Future research may indeed prove the clinical benefits of universal depression screening. The data available at this time, however, do not support these recommendations. Pending further evidence, the time and effort devoted to universal depression screening during the brief primary care visits may be more profitably used for other primary care activities, including improving the care of patients with clinically identified depression.