Usefulness of PHQ-9 in primary care to determine meaningful symptoms of low mood: a qualitative study

Abstract

BACKGROUND: Self-administered questionnaires, such as the Patient Health Questionnaire (PHQ-9), are regularly used in clinical practice to guide prescribing or to measure recovery and response to treatment. There are concerns that patients are not all interpreting the questionnaire items in the same way. Cognitive interviewing is a research technique that identifies 'interpretative measurement error' (IME). IME is distinct from traditional components of measurement error, such as not reading the question as worded, or recording answers inaccurately. AIM: To use cognitive interviewing techniques to explore patterns in answer mapping and comprehension of the PHQ-9 questionnaire to ascertain whether the measure captures meaningful symptoms of low mood. DESIGN AND SETTING: Qualitative study using cognitive interviewing techniques and card sorting in six GP practices in Bristol. METHOD: The study recruited 18 participants at the point of entry to a longitudinal primary care depression cohort study, PANDA (the indications for Prescribing ANtiDepressants that will leAd to a clinical benefit). Participants were interviewed 2, 4, and 6 weeks after their baseline visit. Cognitive interviews were digitally recorded. Analysis used the digital audio file, rather than verbatim transcripts, as it retained important features needed for analyses. RESULTS: Cognitive interviewing revealed that items on the PHQ-9 are interpreted in a range of ways, that patients often cannot 'fit' their experience into the response options, and therefore often feel the questionnaire is misrepresenting their experience of meaningful symptoms of low mood. CONCLUSION: The PHQ-9 may be missing the presence and/or intensity of certain symptoms that are meaningful to patients. Clinicians should adopt caution when using it.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 58 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.