Table 1 Procedure for linguistically translating and culturally adapting assessment instruments in the sites in Uganda and Rwanda

A—Initial expert translation

Each of the instruments were translated from English to standard Swahili by a staff member at the Department of Languages at Makerere University, Kampala, Uganda.

Group translation from English by a team of six bilingual (Kinyarwanda-English) members of the CoSTAR research team. Team members noted items or concepts considered difficult to either understand and/or translate (e.g. terms and phrases that lacked obvious equivalents in standard Kinyarwanda). Proof-read by 4 external members fluent in Kinywarwanda.

B—Review by local mental health experts and practitioners inside and outside the refugee settings, and cultural-competent members of the CoSTAR Refugee Community Advisory Groups (RCAG) to assess comprehensibility, acceptability and relevance

The translated instrument produced in ‘Step A’ were reviewed and the standard Swahili translation of the instruments were appropriately amended to Congolese Swahili.

The translated instruments produced in ‘Step A’ (along with the associated notes) were reviewed and appropriate Kinyarwanda translations for phrasing and concepts that had been noted as challenging were proposed.

C—Instrument adaptation workshop (attendees included RCAG members, representatives of the MHPSS implementing non-governmental organization, Community Based Sociotherapy Rwanda staff, and members of the local CoSTAR research teams)

Data collected as part of a rapid qualitative study exploring the problems experienced by the communities and the tasks they undertake as part of their routine functioning was presented in Kinyarwanda (Rwanda) or Kiswahili (Uganda) to the workshop attendees (see Chiumento and colleagues (10) for details of this study).

Two groups (both featuring bilingual speakers of Kinyarwanda/Congolese Kinyarwanda-English in Rwanda, and Congolese Swahili-English in Uganda) were formed from the workshop attendees. Each group focused on two different assessment instruments (Group 1: PHQ-9 & SASCAT; Group 2: MSPSS & CDES).

The groups were provided with the local language versions of the assessment instruments produced in ‘Step B’ alongside the original English language versions. The groups reviewed and compared each of the items in the assessment instruments in the two language versions. Issues with the translation were recorded by a designated ‘scribe’ on a Translation Monitoring Form (46) which drew specific attention to:

 Comprehensibility (i.e. is the translation understandable in the local language?)

 Acceptability (i.e. would the respondent be uncomfortable honestly answering the question posed?)

 Relevance (i.e. is this question relevant in the local culture?), and Completeness (i.e. would translating the items back into English convey the same concepts and ideas as the original?)

There was also space on the forms for the groups to record further comments if required. In addition, the group working on the CDES was asked if any items should be omitted, or additional items added considering the findings from our rapid qualitative study (10). Based on group consensus through the Translation Monitoring Form prompts, each group produced agreed translated versions of each assessment instrument.

The English language versions of the assessment instruments were removed.

Two members (including the scribe) from Group 1 presented the versions of the PHQ-9 and SASCAT to Group 2 by reading the items aloud and making the written copy available to view. This served as a cross-check for any issues relating to comprehensibility, acceptability, relevance and completeness. Notes about any additional changes that were agreed by consensus were made. Similarly, two members of Group 2 (including the scribe) presented the versions of the MSPSS and CDES to Group 1 by reading the items aloud and making the written copy available to view. Any additional amendments agreed by consensus were noted.