Little is known about the performance of outcome instruments with consumers of different cultural backgrounds. We report a study of using routine outcome measures with Indigenous adult mental health consumers, whereby mental health clinicians used two leading Australian measures with consumers who self-identified as Indigenous. Mean scores on most items were different from the predominantly non-Indigenous national averages; some, particularly those assessing accommodation and occupation, were worse while others (those assessing depression) were better. Scores were consistently worse when more informants were involved in the assessment. A set of good practice principles for the use of standardized measures with Indigenous consumers was developed. They recommend using informants whenever possible, ratings should objectively reflect underlying disadvantage and behaviours that are not sanctioned or accepted within the local culture, but not phenomena that are socially or culturally accepted. Despite the advice and training, most assessments did not involve additional informants. The (mainly non-Indigenous) clinicians mostly indicated no additional difficulty in using the outcome measures with the Indigenous consumers, although clinicians reported greater difficulty in completing them when additional informants were present. In 52% of assessments, clinicians felt that the scales reflected the consumer's problems reasonably, and in 22%, well. Perceived validity was positively associated with consumer engagement. The study provides qualified support for the practice of using the HoNOS and LSP for routine outcome assessment with Indigenous consumers. The main provisos are that clinicians need to maximize the engagement of consumers and try to ensure involvement of additional informants to reduce the chances of invalid assessments.