Screening for Depression in a Cardiac Rehabilitation Program

The PHQ-9 is a reasonable case-finding instrument, but in cardiac populations the somatic items inflate scores: fatigue, sleep disturbance, and appetite change are common in heart failure, post-surgical recovery, and beta-blocker use. A cutoff of 10 will produce many false positives, and because prevalence of depression in cardiac rehab is elevated (roughly 15-25%), the positive predictive value may be moderate but the absolute number of false positives at 300 patients/year is nontrivial. A defensible approach: (1) retain the PHQ-9 but treat it as a two-stage process - a positive screen triggers a brief structured clinical interview (or the PHQ-2 followed by the full PHQ-9, then interview) rather than automatic referral; (2) consider an instrument less confounded by somatic symptoms or use the cognitive-affective subscale as a supplementary index, with the caveat that subscale use is not well validated for diagnosis; (3) explicitly instruct nurses to attribute items to cardiac causes where appropriate and document this; (4) define the referral pathway before screening begins, because a screen without a pathway is an ethical problem. The medical director's demand for 'a number' should be met with a clear metric - e.g., number screened, number interviewed, number engaged in treatment - while explaining that a single cutoff number without a confirmatory step will misclassify patients.

The design should be stepped and capacity-matched. Step 1: universal PHQ-2/PHQ-9 at intake with nurse-administered scripted attribution instructions. Step 2: Dr. Lin reviews all positives within a defined window and conducts a brief interview (in person or telehealth) to distinguish depression from cardiac-symptom overlap, delirium, and adjustment reactions. Step 3: triage - mild/subthreshold to a brief behavioral activation or problem-solving intervention delivered in group or telehealth format (feasible at 0.4 FTE if group-based); moderate to a shared-care model with the patient's primary care provider and a warm handoff; severe or suicidal to same-day psychiatric/emergency pathway, which must be established in writing before screening begins. Dr. Lin should also use a collaborative-care registry so that patients who do not respond are tracked rather than lost. On the nurses' concern: the fear is legitimate and reflects a real systems problem, not resistance. Dr. Lin should reframe - a screen is not a diagnosis, and the protocol specifies exactly what the nurse does and does not do (administer, score, hand off). Training should include a script for responding to a positive screen, and a clear escalation contact. 'Opening a can of worms' is accurate in the sense that unmet need will surface; the answer is to build the pathway, not to avoid the screen. She should also note that identifying and treating depression in cardiac rehab is associated with improved adherence and outcomes, which is a program-level argument the medical director will hear.

The 12% engagement rate is a measure of the referral pathway, not of the screen's validity. A screen that identifies need and a system that fails to convert identification into care are different problems. Dr. Lin should analyze the funnel: how many positives were interviewed, how many were offered an appointment, how many were offered an appointment within a feasible window, and what barriers existed (transportation, scheduling, stigma, cost, competing rehab sessions). Common culprits include referral to an offsite clinic, appointment lag, and lack of warm handoff. She should propose embedding behavioral health contact into the rehab visit itself (same-day brief intervention or telehealth from the rehab site), which typically improves conversion substantially. She should also present the data as a quality-improvement finding: the screen is doing its job; the pathway needs redesign. Calling the screen a failure risks abandoning a valid case-finding tool because of an implementation problem. She should recommend a PDSA cycle with a specific target (e.g., 50% of positives receive a same-week behavioral health contact) and re-measure, and she should distinguish program evaluation from research so the IRB question is settled.