Validity of the short-form five-item Problem Area in Diabetes questionnaire as a depression screening tool in type 2 diabetes mellitus patients

Abstract

Aims/introduction: Depression is prevalent in diabetes patients and associated with poor outcomes, but is currently underdiagnosed, with no firm consensus on screening methods. We evaluated the validity of the short-form five-item Problem Areas in Diabetes (PAID-5) questionnaire as a screening tool for depression, comparing it with the Beck Depression Inventory-II (BDI-II) and nine-item Patient Health Questionnaire (PHQ-9).

Materials and methods: A total of 208 English-speaking adults with type 2 diabetes, recruited from outpatient clinics, completed the BDI-II, PHQ-9 and PAID-5 questionnaires in English. Cronbach's α was used for internal reliability. Convergent validity was examined with BDI-II and PHQ-9. Receiver operating characteristics analyses were used to identify optimal PAID-5 cut-offs for the diagnosis of depression.

Results: All three screening tools were highly reliable, with BDI-II, PHQ-9 and PAID-5 having a Cronbach's α of 0.910, 0.870 and 0.940, respectively. There was a good correlation between BDI-II and PHQ-9, with a correlation co-efficient (r) of 0.73; and a moderate correlation between PAID-5 and PHQ-9, and PAID-5 and BDI-II, with r of 0.55 and 0.55 respectively (P values <0.01). An optimal PAID-5 cut-off ≥9 corresponded to both a BDI-II cut-off >14 (sensitivity 72%, specificity 784%, area under the curve 0.809) and a PHQ-9 cut-off >10 (sensitivity 84%, specificity 74%, area under the curve 0.806). Using a PAID-5 cut-off ≥9, the prevalence of depressive symptoms was 36.1%.

Conclusions: Depressive symptoms are prevalent in people with type 2 diabetes, with the degree of distress significantly related to the severity of depressive symptoms. PAID-5 is a valid and reliable screening tool, and a score ≥9 could prompt further confirmation for depression.

Keywords: BDI-II; PAID; PHQ9; T2DM; Type 2 Diabetes mellitus; depression; depressive symptoms; distress; psycometric.

Figure 1

Relationship between the Beck Depression Inventory‐II (BDI‐II) score and nine‐item Patient Health Questionnaire (PHQ‐9) score in the study population showing a strong correlation between the BDI‐II and the PHQ‐9 (r = 0.73; P < 0.01). The relationship between the BDI‐II score and the five‐item Problem Areas in Diabetes (PAID‐5) score in the study population showing a moderate correlation between the PAID‐5 and the BDI‐II (r = 0.55; P < 0.01). The relationship between the PHQ‐9 score and the PAID‐5 score in the study population showing a moderate correlation between the PAID‐5 and the PHQ‐9 (r = 0.55; P < 0.01).

Figure 2

Receiver operating characteristic curves show the diagnostic performance of the five‐item Problem Areas in Diabetes in discriminating between respondents with significant depressive symptoms as defined by (a) Beck Depression Inventory‐II (BDI‐II) cut‐off of ≥14 and (b) nine‐item Patient Health Questionnaire (PHQ‐9) cut‐off of ≥10 from those without. Receiver operating characteristic curve plots the true positive rate (sensitivity) against the false positive rate (1‐specificity) for different cut‐offs. The curved line represents the area under the receiver operating characteristic (AUC) for the five‐item Problem Areas in Diabetes, whereas the straight diagonal line represents chance (AUC of 0.5). CI, confidence interval.

Figure 3

The prevalence of individuals with depression according to diagnosis by five‐item Problem Areas in Diabetes (PAID‐5) criterion ≥9, Beck Depression Inventory‐II (BDI‐II) criterion ≥14 and nine‐item Patient Health Questionnaire (PHQ‐9) criterion ≥10.