What patient characteristics predict treatment success or failure in short- and long-term psychotherapies?

Abstract

Background

Patient factors make a strong contribution to psychotherapy outcome. Pre-treatment patient characteristics – demographic, pathology and personal – may either facilitate achievement of recovery and lasting benefits or constitute a risk for different kinds of treatment failure.

Methods

A narrative review focused on evaluating the role of pre-treatment patient characteristics on the success or failure of short- and long-term psychotherapy for mood and anxiety disorders. Success was conceptualized as significant improvement or recovery and failure as non-improvement, deterioration, or non-attendance/dropout.

Results

There is no consistent evidence of demographic variables as predictors of treatment failure, except for lower socioeconomic status, being male and ethnic minority status for therapy non-attendance. Patients’ pathology, i.e. severity of psychiatric symptoms, greater functional impairment, personality disorder and other comorbidities, have been shown to be mostly associated with lesser recovery across different types of disorders, especially in short-term therapies, but it does not consistently predict treatment failure. Some pre-treatment personal characteristics (e.g. secure attachment style) and capacities (e.g. motivation, self-observing capacity and good overall psychological suitability) mostly increase the likelihood of therapy success whereas some vulnerabilities and poor psychological suitability, or overall complexity (i.e. combination of demographic, clinical and personality factors), increase the risk of poor outcome mostly in short-term but not in long-term therapy, or moderate dropout. Likewise, the subsequent match between therapy type and patient’s pre-treatment expectations and preferences may account for different aspects of treatment failure.

Conclusions

The relative importance of pre-treatment patient characteristics on treatment failure is not known. Meta-analyses on this issue are needed.