OBJECTIVES: To explore the extent to which commonly used measures of specific outcomes in dementia are an appropriate proxy for quality of life in dementia. METHODS: This was a cross sectional study set in communities in London and Nottingham, comprising 101 people with dementia and their 99 main family caregivers. The main outcome measures were health related quality of life in dementia (measured by the DEMQOL-Proxy), cognition (Mini Mental State Examination), functional impairment (Barthel Index), behavioural and psychological symptoms in dementia (Neuropsychiatric Inventory; NPI), and carer mental health (General Health Questionnaire). RESULTS: On univariate analysis, decreased quality of life was statistically significantly correlated with higher levels of behavioural and psychological disturbance (NPI total score and its agitation, depression, anxiety, disinhibition, and irritability subscales); younger age of the person with dementia; and poorer mental health of the carer. Quality of life was not statistically significantly associated with cognition or carer age. In a multivariate model, psychological and behavioural disturbance and patient age remained statistically significantly associated with quality of life. Carer mental health was no longer statistically significantly associated, and cognition and functional limitation remained statistically insignificant. CONCLUSIONS: These data suggest that quality of life in dementia is complex, and that simple proxy substitutions of discrete measures such as cognition or function are likely to miss important factors.