Screening for depression in hospitalized medical patients
Depression among hospitalized patients is often unrecognized, undiagnosed, and therefore untreated. Little is known about the feasibility of screening for depression during hospitalization, or whether depression is associated with poorer outcomes, longer hospital stays, and higher readmission rates. We searched PubMed and PsycINFO for published, peer-reviewed articles in English (1990-2016) using search terms designed to capture studies that tested the performance of depression screening tools in inpatient settings and studies that examined associations between depression detected during hospitalization and clinical or utilization outcomes. Two investigators reviewed each full-text article and extracted data. The prevalence of depression ranged from 5% to 60%, with a median of 33%, among hospitalized patients. Several screening tools identified showed high sensitivity and specificity, even when self-administered by patients or when abbreviated versions were administered by individuals without formal training. With regard to outcomes, studies from several individual hospitals found depression to be associated with poorer functional outcomes, worse physical health, and returns to the hospital after discharge. These findings suggest that depression screening may be feasible in the inpatient setting, and that more research is warranted to determine whether screening for and treating depression during hospitalization can improve patient outcomes. Journal of Hospital Medicine 2017;12:118-125. © 2017 Society of Hospital Medicine
© 2017 Society of Hospital Medicine
The purpose of this study was to describe the feasibility and performance of depression screening tools in inpatient medical settings, as well as associations between depression diagnosed in the inpatient setting and clinical and utilization outcomes. The median rate at which depression was detected among inpatients was 33%, ranging from 5% to 60%. Studies from several individual hospitals indicated that depression can be associated with higher healthcare utilization, including return to the hospital after discharge, as well as worse clinical outcomes. To detect undiagnosed depression among inpatients, screening appears feasible. Depression screening instruments generally exhibited good sensitivity and specificity relative to comprehensive clinical evaluations by mental health professionals. Furthermore, several self-administered and brief instruments had good performance. Prior authors have reported that screening for depression among inpatients may not be particularly burdensome to patients or staff members.41 The studies we reviewed used diverse screening instruments. Further research is needed to determine which tools are preferable in which patient populations, and to confirm that brief instruments are adequate for screening. The GDS is widely used, and many patients hospitalized in the United States fall into the geriatric group. The PHQ has been validated for self-administration and is widely used among outpatients42; it may be more suitable for younger populations. We found that several abbreviated versions of these and other screening instruments have exhibited good sensitivity and specificity among inpatients. However, many of the studies excluded patients with cognitive impairment or communication barriers. For individuals with auditory impairment, the Brief Assessment Schedule Depression Cards (BASDEC) might be an option. Used in 2 studies, the BASDEC involves showing patients a deck of 19 easy-to-read cards. The time required to administer the BASDEC is modest.15,23 Sets of smiley face diagrams might also be suitable for some patients with communication barriers or cognitive impairment. An ineligible study among stroke survivors found that selecting a sad face had a sensitivity of 76% and specificity of 77% relative to a formal diagnostic evaluation for depression.43 In considering the instruments that may be most suitable for inpatients, the role of somatic symptoms is also important because these can overlap between depression and the medical conditions that lead to hospitalization.44–46 Prior investigators found, for example, that 47% of Beck Depression Inventory (BDI) scores were attributable to somatic symptoms among patients hospitalized after myocardial infarction, whereas 37% of BDI scores were attributable to somatic symptoms among depressed outpatients.47 Future research is needed to determine the significance of somatic symptoms among inpatients, including whether they should be considered during screening, add prognostic value, or warrant specific treatment. In addition, although positive and negative predictive values were generally high among the screening instruments we evaluated, confirming the diagnosis of depression with a thorough clinical assessment is likely to be necessary.44,45DISCUSSION
Despite the high prevalence of depression, associations with suboptimal outcomes, and the good performance of screening tools to date, screening for depression in the inpatient setting has received little attention. Prior authors have questioned whether hospital-based screening is an efficient and effective way to detect depression, and have raised valid concerns regarding false-positive diagnoses and unnecessary treatment, as well as a lack of randomized controlled trials.7,48,49 Whereas some studies suggest that depression is associated with greater healthcare utilization,3,4 little information exists regarding whether screening during hospitalization and treating previously undiagnosed depression improves clinical outcomes or reduces healthcare utilization.
Several important questions remain. What is the pathophysiology of depressed mood during hospitalization? How often does depressed mood during hospitalization reflect longstanding undiagnosed depression, longstanding undertreated depression, an acute stress disorder, or a normal if unpleasant short-term reaction to the stress of acute illnesses? Do the manifestations and effects of depressed mood differ among these situations? What is the prognosis of depressed mood occurring during hospitalization, and how many patients continue to have depression after recovery from acute illness; what factors affect prognosis? In a small sample of hospitalized patients, nearly 50% of those who had been depressed at intake remained depressed 1 month after discharge.50 Given that most antidepressant medications have to be taken for several weeks before effects can be detected, what, if any, approach to treatment should be taken? More research is needed on the effectiveness and cost-effectiveness of diagnosing and treating depression in the inpatient setting.
This work has several limitations. We found relatively few studies meeting eligibility criteria, particularly studies assessing clinical and utilization outcomes among depressed inpatients. Among the screening tools that were studied in the hospital setting, the highly diverse instruments and modes of administration precluded a quantitative synthesis such as meta-analysis. Prior meta-analyses on specific screening tools have focused on outpatient populations.51–53 Furthermore, we did not evaluate study quality or risk of bias.
In conclusion, screening for depression in the inpatient setting via patient self-assessment or assessment by hospital staff appears feasible. Several brief screening tools are available that have good sensitivity and specificity relative to diagnoses made by mental health professionals. Limited evidence suggests that screening tools for depression may be ready to integrate into inpatient care.41 Yet, although depression appears to be common and associated with worse clinical outcomes and higher healthcare utilization, more research is needed on the benefits, risks, and potential costs of adding depression screening in the inpatient healthcare setting.
