I would like to suggest that Holland et al (BMJ 2012; 345:e4708) should not have been surprised with their conclusions; I would go one step further to suggest that I could have predicted their outcomes. Why? Whether it is presenting intricate genetic knowledge of disease risk to family members, or reducing the appeal of tobacco packaging to consumers, especially young people; (bmanews, 2012;18:8); it all boils down to one approach to behaviour change 'the expert model'. The expert model tells people what they should do, but what we must accept is it does not work, if it did we would all be Olympians!

What does work is in actual fact very simple. Firstly, we need to understand that all behaviour is motivated and the behaviour (desirable/undesirable) is motivated by the person's values. Whilst there is a realisation that there continues to be challenges in relation to behaviour change, not least in the area of smoking cessation; we must strive to adopt proven methods of engaging people in successful behavior change. There is a strong case to be made for policy makers and government advisors, to ensure that greater emphasis is placed on the development of practitioners who are skilled in values based behaviour change. Were this to happen I would be so bold as to suggest that outcomes of future research on behaviour change will be more favourable for Holland et al (BMJ 2012; 345:e4708).

Competing interests: None declared