Assessing the performance of a new depression screener for primary care (PC-SAD©)

Abstract

As many as 50% of patients with major depression seen in primary care settings are not diagnosed. To facilitate efficient identification of primary care patients with depression, we developed a new patient-administered depression screening instrument (PC-SAD

Introduction

About 7% of patients making a visit to a primary care doctor have a depressive disorder 1, 2, making depressive disorder one of the most prevalent conditions in primary care. The estimated total cost to society is as high as $40 billion [3] For affected individuals, depressive disorders cause tremendous suffering, disrupting families, friendships, and work relationships. Good treatment interventions 4, 5, 6, 7 and guidelines for treatment 8, 9 are available and are known to produce benefits up to a year after they are applied 4, 10, 11, 12, 13, 14, 15, but not necessarily over a longer term 16, 17.

Unfortunately, about half of patients are either not detected or accurately diagnosed 1, 8, 18, 19, 20, 21, 22. Although most patients with depression see a primary care provider at least once during the year [23], few see mental health specialists 5, 18, 24, 25. Some patients are able to access the mental health system directly, but doing so can be expensive, stigmatizing, and presupposes a recognition on the part of the patient of the need to do so. Patients in primary care consistently underutilize mental health treatments 18, 24, 25. Primary care physicians have the means to treat patients, but without screening tools that are accurate, easy to use and inexpensive, it is difficult to do so. Improving diagnostic rates for depressive disorders in primary care settings is a necessary first step in addressing an important public health issue 26, 27, 28. Although many screeners for depression have been developed, they are rarely used 29, 30. Without a reliable assessment of the individual symptoms, shortness by itself does not seem to suffice. Physicians need psychometrically sound instruments—which have excellent specificity and sensitivity—that also provide them with documented symptom data.

In 1997, in preparation for a primary care-based depression intervention trial, we began work on a new depression screening instrument. Based on a review of the literature and extensive conversations with primary care physicians about the barriers to effective office-based screening for depressive disorders, we determined that a practical screener needed to be:

• Patient administered rather than clinician administered 31, 32, 33, 34, 35

• Short—not a time burden to the patient or the office staff 36, 37, 38, 39

• Well-documented with proven good validity and reliability 34, 40

• Transparent—producing an accurate DSM-IV diagnosis [41], including which clinical symptoms of depression are present 34, 42, 43.

Regarding criterion 4, our concern was that a positive result from a screener without corresponding DSM-IV symptom information, may require further probing from the PCP and would be ignored. Screeners that do not provide the DSM-IV diagnosis make physician follow-up more complex and time consuming. Moreover, depression screeners that identify a patient based on simple cutoffs require a great deal of trust in the instrument. Providing DSM-IV symptoms keeps the physician as the locus of decision making and helps him or her distinguish false from true positive cases [44].

Few depression screeners meet all of the above criteria [45]. The Zung [46], the Beck Depression Inventory (BDI) 47, 48, and the CES-D 49, 50 do not produce a documented DSM-IV diagnosis. In addition, the CES-D has inadequate sensitivity and specificity 36, 51. Clinician-administered screeners 38, 39, 52, 53, 54 pose an impractical burden on the primary care physician. Some depression-severity instruments have been adapted as screeners [55], but are not suitable for primary care because they rely heavily on physical symptoms potentially unrelated to depression.

When we started our work, the screener that best met the above four criteria was the Inventory to Diagnose Depression (IDD) [42]. However, it takes 15–20 min for patients to complete—far too long for office-based screening. We, therefore, developed a new, shorter screener to meet all of our criteria.

This paper describes the screening instrument we developed, and compares its performance to the two other screening instruments that most closely met the four criteria, the IDD and the PRIME-MD-PHQ (PHQ) [56]. Our screener is called the Primary Care Screener for Affective Disorders or PC-SAD.

To maximize generalizability, our sample included both inpatients and outpatients from different practice settings (see Table 1). We compared instruments using a symptom-by-symptom approach to determine presence or absence of major depressive disorder (MDD) (i.e., met five or more out of nine criteria). We compared all of these screeners with a gold standard of a psychiatrist's diagnosis when such a diagnosis was available.