The aim was to investigate factors associated with depression and social function, two outcomes identified as important by people with multiple sclerosis (MS) and to identify underlying dimensions of psycho-social well-being that may be useful as outcome measures. People with MS in eight randomly selected health authorities/boards in England and Scotland completed a postal questionnaire relating to preferences and needs for their health and social care, along with the Beck Depression Inventory and the SF-36. Responses to 10 of the original items were subjected to factor analysis. These and other explanatory variables were entered into multivariable regression models for the two outcomes. The factor analysis resulted in three dimensions representing different aspects of psycho-social well-being; one of these (representing autonomy) was associated with improvements in both outcomes, as was the SF-36 emotional role limitation score. Three other SF-36 dimensions and lack of contact with a health professional in the last year were related just to social function. The regression models emphasise the value of enabling autonomy and self-reliance amongst people with MS, as well as more general measures of emotional health. The present work identifies specific questions that could be used to measure pivotal aspects of an individual's psycho-social well-being. While these findings warrant replication for people with MS, they may have relevance to those with other long-term illnesses.