OBJECTIVE: This paper describes the process employed to adapt the Problem Gambling Severity Index (PGSI) for use with Indigenous Australian populations. METHODS: This study comprised a two-stage process: an initial consultation with Indigenous health workers, informing the textual and conceptual adaptation of items, followed by trial of the adjusted instrument with Indigenous community members (n=301). RESULTS: Internal reliability was demonstrated: Australian Indigenous Problem Gambling Index (AIPGI) Cronbach's alpha α = 0.92 (Original PGSI, α = 0.84). Item-rest correlations confirmed that responses to items were consistent and related to the total score of remaining items. The AIPGI could predict gambling severity based on gambling frequency, when controlling for age and gender (OR=1.28, 95%CI 1.17-1.40). CONCLUSIONS: The adapted instrument is accessible to a cross-section of Indigenous Australians and has demonstrated properties of reliability and validity. An extended trial is needed to test the application of the instrument to a broader Indigenous audience and to further explore and confirm psychometric properties of the adapted instrument. IMPLICATIONS: This study introduces a culturally adapted tool for measuring rates of disordered gambling among Indigenous Australians.
OBJECTIVE: This paper describes the process employed to adapt the Problem Gambling Severity Index (PGSI) for use with Indigenous Australian populations. METHODS: This study comprised a two-stage process: an initial consultation with Indigenous health workers, informing the textual and conceptual adaptation of items, followed by trial of the adjusted instrument with Indigenous community members (n=301). RESULTS: Internal reliability was demonstrated: Australian Indigenous Problem Gambling Index (AIPGI) Cronbach's alpha α = 0.92 (Original PGSI, α = 0.84). Item-rest correlations confirmed that responses to items were consistent and related to the total score of remaining items. The AIPGI could predict gambling severity based on gambling frequency, when controlling for age and gender (OR=1.28, 95%CI 1.17-1.40). CONCLUSIONS: The adapted instrument is accessible to a cross-section of Indigenous Australians and has demonstrated properties of reliability and validity. An extended trial is needed to test the application of the instrument to a broader Indigenous audience and to further explore and confirm psychometric properties of the adapted instrument. IMPLICATIONS: This study introduces a culturally adapted tool for measuring rates of disordered gambling among Indigenous Australians.