PHQ-9 and GAD-7 as Screeners

In an era of ‘measurement-based care,’ the PHQ-9 and GAD-7 have become industry standards. If you work in large healthcare systems, these screeners may be mandatory parts of your workflow. Many clinicians see the PHQ-9 and GAD-7 as intuitive and easy to use, supposedly producing quick snapshots about the diagnostic criteria for depression and anxiety.

In reality, these measures can perpetuate errors in your clinical decision-making. The PHQ-9 and GAD-7 are indeed sensitive screeners. Yet, they often fall short as tools for differential diagnostics. There are high-quality alternatives for clinicians who want more precise mental health assessments.

The Problem of "General Distress"

The most significant hurdle with the PHQ-9 and GAD-7 is their overlap. You have likely witnessed this in your own practice: scores on both instruments tend to rise and fall together. Furthermore, scores can be elevated due to any number of factors, not just depression or anxiety (e.g., PTSD, OCD, psychosis). That’s because the PHQ-9 and GAD-7 both pull for responses colored by general distress—that overall sense of demoralization and hopelessness about life that isn't specific to any one disorder. Thus, distress-related ‘noise’ can drown out the ‘signal’ you might need to rule out one disorder versus another.

Clinical Consequences

The PHQ-9 and GAD-7 come from a tradition that previously assumed mental illnesses are discrete diseases. Those assumptions are baked into how these screeners operate.

It’s becoming increasingly clear that depression and anxiety don't work that way; instead, they are messy collections of several different symptom spectrums. This can explain how two clients with the same PHQ-9 and GAD-7 total scores might differ radically in terms of their actual symptoms.

• Consider “anhedonia,” a core symptom of depression, reflecting difficulty experiencing joy and pleasure. While the PHQ-9 does have one anhedonia item, the total score is so heavily influenced by distress that anhedonia gets lost in the shuffle. Anhedonia is an inclusion criterion for several specialized depression interventions (e.g., positive affect treatment). This makes the PHQ-9 insufficient if you want to be confident about recommending the right treatment.

• A similar issue can happen with PTSD and the GAD-7. A person with PTSD would likely elevate the GAD-7 even if they did not meet criteria for a separate anxiety disorder. However, the GAD-7 items do not ask about any of the symptoms specific to PTSD. If a clinician administered the GAD-7 and stopped there, they would not have enough information to confidently refer someone for specialized trauma-focused psychotherapy.

Technical Issues

Beyond these clinical and conceptual issues, there are more basic problems with these tests and their interpretation schemes:

• Floor Effects: Both measures are heavily skewed, with most people in the general population producing scores close to zero. From a statistical perspective, this makes it difficult for clinicians to reliably detect subtle changes in symptoms.

• Expert Consensus vs. Population Norms: The "mild, moderate, and severe" labels used for interpreting PHQ-9 and GAD-7 scores were largely decided by expert consensus. Without direct access to the population normative data, clinicians have to take it on faith that "mild" depression on the PHQ-9 is equivalent in severity to "mild" anxiety on the GAD-7.

• The Liability Problem: The final item of the PHQ-9 is about thoughts of death and self-injury, which obligates clinicians to do follow-up safety evaluations. Frustratingly, that PHQ-9 item does not capture much information about the nature or severity of those thoughts. As the clinician, you are left in the dark about your client’s true risk level until you perform a separate assessment.

Improving Your Measurement

In my private practice, I lean on newer instruments that assume mental health symptoms are more like traits on spectrums rather than binary indicators of discrete diagnoses. If you as a clinician are looking to refine your assessment processes, here are some of the tools I find to be useful:

1. Patient-Reported Outcomes Measurement Information System (PROMIS)

PROMIS measures are a suite of free screeners developed to assess many experiences relevant for mental health (e.g., mood, anger, pain, sleep, role functioning, etc.). They serve similar roles as the PHQ-9 and GAD-7, but incorporate more modern statistical approaches.

• Benefits for You: PROMIS measures provide estimates of how common or uncommon a particular score is in the general population. This makes it easier to directly compare symptom levels across scales. In addition, PROMIS measures do not have as pronounced of floor effects.

2. Inventory of Depression and Anxiety Symptoms (IDAS)-II

This is a longer questionnaire designed to produce comprehensive profiles of recent mood and anxiety symptoms. The IDAS-II test authors used methods to isolate generalized distress from other symptoms. As such, it provides a more focused assessment of experiences relevant for anhedonia, OCD, PTSD, and bipolar disorders.

• Benefits for You: The IDAS-II was designed to minimize the overlap problem of the PHQ-9 and GAD-7. Users get dimensional symptom scores without being beholden to traditional diagnostic categories. Similar to PROMIS, there are normative data to interpret scale elevations.

3. Depressive Symptom Index-Suicidality Subscale (DSI-SS)

For suicide risk assessment, I often use the DSI-SS. It’s brief—only four questions—but its response format maps onto established risk assessment frameworks. Clinicians can use the data to get feedback about the relative seriousness of their client's ideation, which can be used to guide safety planning.

• Benefits for You: The DSI-SS does not remove the need for interview-based risk assessments, but it gives more nuanced hints about the current level of risk than the PHQ-9.

4. Minnesota Multiphasic Personality Inventory (MMPI)-3

For a truly comprehensive look at mental health, there is no substitute for longer tests like the MMPI-3. This most recent version is one of the most sophisticated tools available for informing differential diagnostics. Rather than just asking about the last two weeks, it assesses how clients typically experience the world.

• Benefits for You: The MMPI instruments are supported by a huge research literature, which can be used to help interpret the test. The MMPI-3 is too impractical for weekly monitoring, but it can provide deeper insights than brief screeners.

Updating Your Clinical Toolkit

The PHQ-9 and GAD-7 aren't "bad" tests, but they are blunt instruments. If you want your assessment and interventions to be targeted, these screeners will only get you so far. Clinicians who want to go further need tools that capture the subtle, trait-like variations underlying the experiences of depression and anxiety.

By using population-referenced, dimensionally-focused tests, you can gain additional insights about how to customize your clinical interventions in ways informed by your assessment data.