Interpretation Issues With the Patient Health Questionnaire Instructions
Margarita Panayiotou, Josip Razum, Gudrun Eisele, Shirley B. Wang, Eiko I. Fried, Zachary D. Cohen
Abstract
Importance: Versions of the Patient Health Questionnaire (PHQ) such as PHQ-2, PHQ-8, and PHQ-9, are among the leading assessment tools for depression in research and clinical practice. However, important questions remain about its validity, particularly whether responses reflect symptom frequency or the degree to which symptoms are bothersome. Objective: To investigate how participants respond to the PHQ, described as a severity measure by its developers, with instructions about being bothered by symptoms and response options focused on symptom frequency. Design, Setting, and Participants: This study used data from a general population sample collected via Amazon Mechanical Turk (MTurk) and a clinical sample with medium to high depression from the Operationalizing Digital PhenoTyping in the Measurement of Anhedonia (OPTIMA) study. Data were collected between 2022 and 2023. Main Outcomes and Measures: After completing the PHQ-8, participants' interpretation of instructions was assessed via 3 questions: (1) how they would respond to the PHQ sleep item in a hypothetical scenario where they overslept nearly every day but were comfortable with oversleeping; (2) whether they had based their earlier PHQ responses on symptom frequency, being bothered by symptoms, or both; and (3) how they would answer the PHQ in the future, based on these same 3 options. Results: The study sample consisted of a general population sample collected via MTurk (n = 503; mean [SD] age, 40.63 [13.62] years; 253 male, 245 female, 3 transgender, 2 other) and a clinical sample with medium to high depression from the OPTIMA study (n = 349; mean [SD] age, 33.44 [12.23] years; 120 male, 216 female, 5 transgender, 8 other). In the hypothetical oversleeping scenario, only 54.7% (n = 275; MTurk) and 15.5% (n = 53; OPTIMA) of participants interpreted the PHQ as instructed (ie, the frequency with which the problems bothered them). When asked how they had responded to the PHQ, only 21.3% (n = 107; MTurk) and 11.7% (n = 40; OPTIMA) of participants interpreted the instructions as instructed and only 22.3% (n = 112; MTurk) and 9.9% (n = 34; OPTIMA) reported they would do so in the future, indicating stability in their interpretation. The current study also found that the PHQ-8 validity varied depending on how participants interpret its instructions. Conclusions and Relevance: This study identified widespread misinterpretation of the PHQ instructions across community and clinical samples, raising doubts about its validity for both research and clinical decision-making.