Little is known about the stabilization phase of schizophrenia. To address this deficiency, we studied the prescribing patterns of fluphenazine decanoate (FZD) for a cohort of schizophrenic patients recovering from an acute psychotic episode who were being followed in the Treatment Strategies in Schizophrenia (TSS) multi-site maintenance study. Dosage data from the first 208 successfully-stabilized TSS subjects were examined. We hypothesized five groups on the basis of dosage prescribed and dosage stability over time: (1) STABLE HIGH, (2) STABLE INTERMEDIATE, (3) STABLE LOW, (4) UNSTABLE RAISING, and (5) UNSTABLE LOWERING. The dosage patterns were rated and analyzed against predictive and outcome variables. Most of the subjects (n = 168, or 82%) fit one of the five hypothesized dosage pattern groups. Our preliminary findings are as follows: (1) all groups improved during stabilization; (2) baseline symptom severity did not distinguish among dosage pattern groups; (3) STABLE dose groups needed less time to stabilize than did the UNSTABLE groups, but the STABLE groups did not have a better medication response; (4) many of the UNSTABLE LOWERING subjects apparently received a "loading dose" strategy (e.g., initial FZD dose > or = 25 mg), which was paradoxically associated with a longer stabilization time.