Development of a brief screening method for identification of depression in older adults in Sub-Saharan Africa

Molly Howarth-Maddison , Editruda Gamassa , Ssenku Safic , Damas Andrea , Sarah Urasa , Richard W. Walker , William K. Gray ORCID Icon, Irene Haule , Catherine L. Dotchin & Stella-Maria Paddick
Publication year: 


To develop a brief, culturally appropriate screening tool for identifying late life depression (LLD), for use by non-specialist clinicians in primary and out-patient care settings in sub-Saharan Africa (SSA).


Depressive disorders are a leading contributor to the global health burden. LLD is common and cases will increase as populations’ age, particularly in low- and middle-income countries (LMICs), such as those in SSA. A chronic mental health workforce shortage and the absence of culturally adapted LLD screening tools to aid non-specialist clinicians have contributed to a significant diagnostic gap.


A systematic random sample of older people attending general medical clinics were interviewed using a 30-item LLD questionnaire, developed utilizing a Delphi consensus analysis of items from the Geriatric Depression Scale, Patient Health Questionnaire-2 and questions developed from a study of lay conceptualisations of depression in Tanzania. The items were assessed for validity against blinded DSM 5 diagnosis of depression by a research doctor. Factor and item analysis were then used to refine the questionnaire.


The 12-item Maddison Old-age Scale for Identifying Depression (MOSHI-D) was developed. It has good internal consistency (Cronbach’s α = 0.820) and construct and criterion validity (AUROC = 0.880).


On initial evaluation, the MOSHI-D showed good internal validity. It should be easy for non-specialists to administer. External validation and further refinement will be conducted. A culturally-appropriate LLD screen may improve mental health care integration into existing healthcare settings within SSA and facilitate greater patient access to care, in accordance with current WHO strategy.