Dear Editor:

Tibbo and others (1) report significant improvement in functioning in a large cohort of individuals with serious mental illness who received care in an assertive community treatment (ACT) program in Edmonton. They base their conclusions on statistically significant improvement in Global Assessment of Functioning (GAF) scores over a 3-year period following admission to the program. Their results suggest that patients with lower baseline GAF and those with diagnoses of schizophrenia, schizoaffective disorder, and dysthymia are more likely to benefit from ACT than those with diagnoses of delusional disorder, bipolar disorder, and depression.

There are several problems in interpreting these results. GAF is not designed to measure functioning in any domain that is meaningful to individuals living in the community. GAF ratings are influenced more by symptoms than by social and occupational functioning. Several other measures could have more accurately provided information on social, community, and occupational functioning (for example, the Life Skills Profile, several Quality of Life measures, or the Social and Occupational Functioning Assessment Scale). If symptoms did in fact improve, that is significant in itself and needs to be reported as such: controlled ACT trials have generally failed to report any significant improvement in psychopathology. Hopefully, the authors have data on symptom ratings to report such results.

It is unclear whether the raters had received any specific training in the use of GAF. While consensus ratings more likely reflect a true picture of the patient, establishing reliability among raters across teams is essential, because the data reported are being aggregated from disparate sources.

The authors correctly question whether statistical significance is clinically meaningful. The fact that the most significant improvements showed a mean change within the same category (for example, for the schizophrenia group, from 51.4 to 57.9) is particularly important. It would be more meaningful to know what proportion of patients moved from a lower category to a higher category (in units of 10). Further, the results might have been more meaningful if they had been reported for patient clusters based not so much on diagnosis as on other variables known to influence community functioning (for example, sex, length and number of previous hospitalizations, and the severity and type of residual symptoms). A more complex data analysis could have provided more informative results regarding the relative influence of other variables. For example, we do not know whether the patients received different medications (such as novel vs typical antipsychotics) during the ACT phase.

Last, but not least, the results suggest that ACT may be an “overkill” for patients with a higher GAF at baseline, because they do not appear to benefit at any time during the 3 years of treatment. It raises the important question whether a single approach in the form of ACT, with full fidelity to all its components (including a high staff-to-patient

ratio of around 1 to 10) may be unnecessary for a substantial proportion of patients with severe mental illness, including those with schizophrenia—as suggested by some recent studies (2). An opportunity exists to design care delivery methods incorporating the elements of intensive and assertive community treatment that are relevant to a particular group of patients with severe mental illness. This would allow a larger number of patients to have greater access to effective methods of care, because the ratio of case managers to patients can be significantly more economical than is the case in a traditional ACT program; there is, indeed, evidence to support this (3).

References

1. Tibbo P, Joffe K, Chue P, Metelitsa A, Wright E. Global assessment of functioning following assertive community treatment in Edmonton, Alberta: a longitudinal study. Can J Psychiatry 2001;46:144–8.
2. Burns T, Creed F, Fahy T, Thompson S, Tyrer P, White I. Intensive versus standard case management for severe psychotic illness: a randomized trial. Lancet 1999;353:2185–9.
3. Malla AK, Norman RMG, McLean T, Cheng S, Rickwood A, McIntosh E, and others. An integrated medical and psychosocial treatment program for psychotic disorders: patient characteristics and outcome. Can J Psychiatry 1998;43:698–705.

Ashok Malla,
MB, FRCP
London, Ontario

Dear Editor:

Dr Malla correctly observes that Global Assessment of Functioning (GAF) is not the best measure of functional change; however, when our program was originally designed in late 1992, limited resources, together with its already frequent use and familiarity (in hospital discharge summaries, for example) led us to choose it as the primary program outcome measure.