Back

Adaptation and validation of screening measures of anxiety (GAD-7), depression (PHQ-9), and post-traumatic stress disorder (PC-PTSD-5) for use in population-based epidemiological studies in Malawi, Africa.

Nzawa-Soko, R. H.; Stewart, R. C.; Kenala-Malava, J.; Myaba, J.; Msiska, M.; Makhalira, M.; Mthepheya, T.; Matchado, A. J.; Mkandawire, J.; Nkosi, T.; Nyanjagha, I.; Umar, E.; Nakanga, W. P.; Coombes, L.; Seward, N.; MacBeth, A.; McIntosh, A. M.; Crampin, A. C.

2026-07-22 public and global health
10.64898/2026.07.21.26358551 medRxiv
Show abstract

Population-based studies of common mental health conditions (anxiety, depression, post-traumatic stress disorder) require measures that are valid in the study context. We set out to validate the Generalised Anxiety Disorder-7 scale (GAD-7), Patient Health Questionnaire-9 (PHQ-9); and Primary Care PTSD Screen for DSM-5 (PC-PTSD-5) in the most widely spoken languages in Malawi (Chichewa and Chitumbuka). We undertook translation, adaptation, and piloting to produce final versions in both languages. We conducted criterion validation of the GAD-7, PHQ-9 and PC-PTSD-5 against reference diagnoses of DSM-5 generalised anxiety disorder, major/minor depressive episode, and PTSD respectively, using the Structured Clinical Interview for DSM-5 (SCID-5). A weighted sample of screened participants had SCID interview and this was adjusted for in the analysis. We recruited convenience samples of women and men from two sites: a rural Chitumbuka-speaking site where 342 were screened and 219 had SCID; and an urban Chichewa-speaking site where 458 were screened and 251 had SCID. In both languages, the measures had acceptable internal consistency (Cronbachs alpha [≥] 0.75). Regarding convergent validity, PHQ-9 and GAD-7 were highly correlated but PC-PTSD-5 was only weakly/moderately correlated with the other measures. In Confirmatory Factor Analysis, best fit for GAD-7 and PC-PTSD-5 was a 1-factor structure, and for PHQ-9 was a 2-factor structure; there was only partial measurement invariance between the 2 language versions of each measure. Area under the ROC curve (AUC) for GAD-7 detection of generalised anxiety disorder was: Chitumbuka 0.759 (95%CI: 0.634, 0.871); Chichewa 0.868 (95%CI: 0.812, 0.915). AUC for PHQ-9 detection of major depression was: Chitumbuka 0.634 (95%CI: 0.441, 0.869); Chichewa 0.843 (95%CI: 0.721, 0.927). AUC for PHQ-9 detection of minor-or-major depression was: Chitumbuka 0.751 (95%CI: 0.619, 0.865); Chichewa 0.801 (95%CI: 0.714, 0.879). AUC for PC-PTSD-5 detection of PTSD was: Chitumbuka 0.682 (95%CI: 0.519, 0.853); Chichewa 0.741 (95%CI: 0.604, 0.859). In conclusion, GAD-7 and PHQ-9 showed good/acceptable validity, although criterion validity of PHQ-9 for major depression in Chitumbuka was poor. PC-PTSD-5 showed limitations to its validity, indicating need for further development of PTSD measures in Malawi.

Matching journals

The top 4 journals account for 50% of the predicted probability mass.

50% of probability mass above

"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.