We are grateful to Drs McQueen & St John-Smith for their response, which highlights that our study raises the question of the purpose of mental healthcare.

We agree that the effect we showed is small but we believe it is more meaningful than that shown by other study designs. Our data comprised repeated measures at monthly intervals over 7 months, and we demonstrated temporal precedence in the relationship between patient-rated unmet need and quality of life – reduction in the former precedes improvement in the latter. Cross-sectional studies more easily demonstrate apparent associations, which prove on further investigation to be spurious.

The analysis controlled for baseline symptomatology (assessed using the Brief Psychiatric Rating Scale) and diagnosis, and found no evidence of a mediating role for psychiatric illness. Furthermore, our use of random-effects regression models controlled for further unmeasured individual characteristics that are stable over time. Our finding of a modest but robust effect is meaningful and therefore clinically important, especially when combined with other small effects. Further research into determinants of quality of life will provide other levers of change for improvement, which are unlikely to be staff-rated symptomatology (Reference Lasalvia, Ruggeri and SantoliniLasalvia et al, 2002).

We agree that interventions to improve mental health will have an impact on patient-rated unmet need, which in turn (as we demonstrate) will improve quality of life. However, the advantage of identifying a modest but robust causal relationship is that it highlights the importance of a more comprehensive approach to meeting needs. Mental healthcare that focuses exclusively on treating psychiatric illness can risk neglecting the importance of other consequences of mental ill health, such as discrimination in travel (Driver and Vehicle Licensing Agency, 2005), insurance (Association of British Insurers, 2003) and debt (Reference Meltzer, Singleton and LeeMeltzer et al, 2002). Mental health services that also address a wide range of health and social needs (as, for example, assessed in our study by the Camberwell Assessment of Need) are more likely to improve quality of life.