In their re-examination of data from the UK700 study of standard v. intensive case management, Burns et al (Reference Burns, Yiend and Doll2007) conclude that although case-load size can affect the quantity and possibly the type and quality of community care delivered to people with severe psychotic disorders, there is no overall clinical advantage associated with any particular case-load size within the approximate range 1:10 to 1:20.

The results of another randomised controlled trial involving 193 ‘heavy users’ of psychiatric in-patient services (Reference Harrison–Read, Lucas and TyrerHarrison-Read et al, 2002) can also throw light on this issue. In this study, intensive case management was delivered to ‘enhance’ the care already provided by the standard locality mental health service in a socially deprived outer-London borough. The case-load size of the standard service was 1:20 or more, and although the case-load size of the intensive case management team varied between only 1:8 and 1:15, and achieved a mean 2.4-fold increase in community contacts compared with the standard service alone, overall the intervention produced no statistically significant benefits on hospital bed use, direct costs of care or clinical outcomes.

In about one-fifth of the ‘heavy users’ in the study group, needs were already being adequately met by the standard service, and the study team did little more than serve a care coordination role, with very low rates of community contact. These users were mainly managed by the standard service, as of course were all those in the control group. However, after excluding this subgroup of users in a post hoc analysis, there were still no measureable benefits from the study intervention. Since minimal intervention corresponds to greater ‘virtual’ case-load size (Reference Burns, Yiend and DollBurns et al, 2007), the implication of this finding is that case-load size in the approximate range 1:10 to 1:20 does not have a major impact on health and cost outcomes of intensive case management in a sample of this type.

By contrast, when the impact of the study intervention was re-examined in a subgroup of ‘very heavy users’ representing the upper quartile of the study sample (n=23), the healthcare costs were nearly halved in comparison with controls (n=25, P<0.001). These ‘very heavy users’ tended to receive the most intensive care from the study team, with particular attention being paid to addressing vulnerability factors determining relapse and presentations which were likely to lead to hospital readmission. This provisional finding implies that for this particular subgroup, targeted intensive case management which achieves improved costs and clinical outcomes might indeed benefit from or require small case-loads of around 1:10. This adds to the conclusion of Burns et al (Reference Burns, Yiend and Doll2007) that as well as care structures such as case-load size, and the type and quality of care, it is also the targeted patient population that determines the impact of intensive case management.