Mulder and Kortrijk analyzed the link between patient characteristics, the Camberwell Assessment of Need Short Appraisal Schedule (cansas), early drop-out and treatment duration. Results showed that patients with severe mental illness (SMI) who remained under treatment by an assertive community treatment team had, both at the start of treatment and at their last assessment, more unfulfilled needs than patients who had been treated for a shorter period.
Significant differences are reported in mean numbers of unfulfilled needs. On average a one-need difference was found over a total of 22 cansas-items. This is a small difference. Moreover, classic analysis of variance (ANOVA) was used which is not robust for an analysis of skewed discrete data and unbalanced groups (Glass et al., 1972).
For 364 patients (36% of all patients) the first assessment indicated over 5 unfulfilled needs on average with a relatively small standard error. Lower scores are reported for all other groups ordered by treatment duration varying from 88 to 191 patients. The largest group with only one assessment differs from some of the other groups, but between these groups the means are close to 4 unfulfilled needs. No post-hoc analyses are reported.
These differences are plotted in Figure 1 (see below) and are interpreted as changes over time. The mean number of needs lowers and than rises according to the authors. However, the confidence intervals of all but the first group overlap by 50% or more indicating that p-values for differences between the means are >.05 (Cumming and Finch, 2005). So statistically this figure shows no trends.
The authors point at the implication for benchmarking. For the correct interpretation of ROM data for patients with SMI, information about patient characteristics and treatment duration are important. Surely case-mix is of interest, but this study showed differences in outcome that are to small to leave a (bench) mark.