Cases for change– different approaches to ‘what works’

In 2003, the former National Institute for Mental Health in England (NIMHE) published a broad review of ‘what works’ in adult mental health (Glasby et al, 2003). Cases for change in mental health comprised 10 reports summarising the evidence with regard to different sectors of the mental health system (primary care, community services, hospitals and so on) and a range of cross-cutting themes (partnership working, user involvement, discrimination and so forth). Adorned with a series of paintings by service user artists, the report was launched at a large event in Liverpool with actor Dean Sullivan (whose character Jimmy Corkhill in the long-running soap Brookside had developed a mental health problem). The report itself seemed well received and was one of the early products of the new National Institute.

Initially, the aim had been to produce a high-profile, evidence-based summary of ‘what works’ – particularly given recent NHS changes (which meant that some primary care-based commissioners would now have responsibility for mental health services without necessarily having a background in working with this user group). If mental health commissioners read only one document in preparation for their new responsibilities, NIMHE argued, they should read Cases for change – and this would tell them pretty much all they needed to know.

Although Cases for change was initially well received by the funders and those who attended the launch event, the peer-review journal articles that the research team subsequently drafted attracted a more mixed response. While some high-ranking journals were very positive indeed, others were equally negative (see Box 6.1 for a summary of some of the more scathing responses). In particular, critics felt that Cases was lacking in academic rigour and was unscientific (to put it mildly) – and several leading journals wouldn't touch it with a bargepole. Interestingly, other journals seemed to love it, and they accepted the same articles with virtually no changes at all (see, for example, Glasby and Lester, 2004, 2005; Lester et al, 2004 for products from the broader study).