Statistical implication analysis: a novel approach to understand the reciprocal relationships between outcomes in early psychosis.

Background: Patients can respond differently to intervention in the early phase of psychosis. Diverse symptomatic and functional outcomes can be distinguished and achieving one outcome may mean achieving another, but not necessarily the other way round, which is difficult to disentangle with cross-sectional data. The present study's goal was to evaluate implicative relationships between diverse functional outcomes to better understand their reciprocal dependencies in a cross-sectional design, by using statistical implication analysis (SIA).

Methods: Early psychosis patients of an early intervention program were evaluated for different outcomes (symptomatic response, functional recovery, and working/living independently) after 36 months of treatment. To determine which positive outcomes implied other positive outcomes, SIA was conducted by using the Iota statistical implication index, a newly developed approach allowing to measure asymmetrical bidirectional relationships between outcomes.

Results: Two hundred and nineteen recent onset patients with early psychosis were assessed. Results at the end of the three-years in TIPP showed that working independently statistically implied achieving all other outcomes. Symptomatic and functional recovery reciprocally implied one another. Living independently weakly implied symptomatic and functional recovery and did not imply independent working.

Conclusions: The concept of implication is an interesting way of evaluating dependencies between outcomes as it allows us to overcome the tendency to presume symmetrical relationships between them. We argue that a better understanding of reciprocal dependencies within psychopathology can provide an impetus to tailormade treatments and SIA is a useful tool to address this issue in cross-sectional designs.