Kreyol Distress Idioms (KDI) Scale

Information about Measure
First Name Bonnie
Last Name Kaiser
Affiliation University of California San Diego
Other means of contacting author (e.g., website,, ResearchGate)
Mental health assessment tool that was adapted/developed/validated Kreyol Distress Idiom (KDI) Scale
Mental health condition assessed Other: Multiple idioms of distress
Idiom of distress included, if any Multiple
Lifestage of interest Adult (General)
Age range (age – age) 18+
Country or countries where tool was developed/adapted/validated Haiti
Language(s) of the adapted/developed/validated tool Kreyol/Haitian Creole
Clinical or community sample? Community
Subpopulation in which tool was developed/validated (e.g., tool was developed and tested among middle-class women)? Rural communities in the Central Plateau
Development procedures Locally developed
If validated, what was the gold standard?
Description of other development procedures, if applicable “Development of the Kreyòl Distress Idioms (KDI) screening tool was based on
ethnographic data collection and pilot testing. We identified potential idioms of distress using multiple qualitative methods: (1) participant observation with four case studies identified by community members to be experiencing mild to moderate mental distress; (2) approximately 60 hours of observant participation in hospital, clinic, and mobile clinic settings; (3) 31 in-depth interviews with community leaders, religious and traditional healers, and healthcare providers; and (4) 11 focus-group discussions with lay community members (see Supplemental Table 1). From these various forms of data collection, we identified 43 potential idioms of interest, representing somatic, emotional, cognitive, and psychological experiences that seem to indicate mild to moderate mental ill-health or associated sequelae.We then discussed these 43 idioms with two Haitian clinicians and in a focus-group discussion with lay community members. Following these conversations, idioms were removed from the list if they seemed to be caused solely by physical illnesses, were thought to be nearly universal experiences (i.e., not specific enough), or were not well understood. Additionally, when multiple idioms seemed to represent very similar experiences, only the best understood item was kept. The remaining 17 items were developed into a screening tool that was piloted among 97 participants. Based on the pilot study, four more items were removed because they were infrequently endorsed and seemed to be potentially stigmatizing concepts. See Kaiser et al. (2013) for full description of sampling strategy, idiom selection, and pilot testing.” (Kaiser et al. 2015)
Cronbach’s alpha .89
Other information about tool (e.g., additional psychometrics [NPV, PPV, Youden’s index, diagnostic odds ratio]) Principal component analysis extracted four components: (1) lack of control over thoughts and behaviors, (2) worry and rumination, (3) somatic indicators of anxiety, and (4) fatigue and foreshortened future. The KDI correlated with the BAI (r = 0.67) and BDI (r = 0.52). Factors associated with KDI score included female gender, older age, alcohol consumption, traumatic exposures, and having a household member with mental distress.
Links to development/adaptation/validation studies and/or previous studies using the tool In the original development article (Kaiser et al., 2013), 17 idioms of distress were tested; however, the final screener consists of 13 items
Kaiser, B. N., Kohrt, B. A., Keys, H. M., Khoury, N. M., & T., A.-R. (2013). Strategies for assessing mental health in Haiti: Local instrument development and transcultural translation. Transcultural Psychiatry, 50(4), 532–558.
Kaiser, B. N., Kohrt, B. A., Wagenaar, B. H., Kramer, M. R., McLean, K. E., Hagaman, A. K., … & Keys, H. M. (2015). Scale properties of the Kreyòl Distress Idioms (KDI) screener: association of an ethnographically-developed instrument with depression, anxiety, and sociocultural risk factors in rural Haiti. International Journal of Culture and Mental Health8(4), 341-358.
Notes when administering the tool Tool was designed and tested to be administered verbally among community samples. Response options correspond to a Likert scale (0-5), with all items equally weighted in generating sum scores