Measure for predicting dropping out of psychotherapy.

Abstract

Scored content of transcriptions of recordings of initial interviews with 70 patients at a psychiatric outpatient clinic was used to predict whether a patient would come back for at least 3 more interviews if offered treatment. Each sentence unit or 5 sec. of silence was assigned a rating of + (favorable to continuation), or -(unfavorable), or 0 (irrelevant). Excess of plus over minus units predicted continuation; excess of minus over plus units predicted dropping out. Coefficients of reliability of ratings by 3 pairs of scores were significant at .001. Individual scorers predicted correctly 73%, 81%, and 80% of cases; averages of paired scorers predicted 81% correctly; averages including scores of mediator predicted 83% correctly.

Similarity, missed sessions, and early termination.

“…Whatever the limitations of this study in terms of sample and personality variables, the findings point to the great importance of the initial stages, and particularly the first interview, in the clinical process. (In this regard see also Overall &Aronson, 1963 andWhite, Fichtenbaum, &Dollard, 1964. ) The report of the Outpatient Studies Section of the National Institute of Mental Health indicates that approximately onequarter of patients seeking treatment at outpatient psychiatric clinics fail to return after the first session, and in a number of studies, the majority of patients terminated by the third or fourth session (Brandt, 1965).…”

Measured relationships between sexual motivation and anxiety.

“…All the therapy sessions were recorded, with peimission, and the transcripts edited to remove peisonally identifying references The talk of the patient and therapist was divided into sentence units (Auld &. White, 1956) Two independent scorers assigned a category to each unit, an arbiter choosing the final score when the scorera disagreed The reliability of the four categories used was checked on one session randomlj selected from each of nine casts The two scorers agreed on 90.07% of 3,421 units (Phi coefficients for individual scoring categories have been reported m previous studies'- White et al, 1964White et al, , 1966 The frequency of each category was divided b\ thr total number of units of the patient's speech in each session. These percentage figures were then used to compute means foi groups of sessions or groups of patienti…”

Content-analysis studies of therapeutic interviews: 1954 to 1964.

“…His results strongly confirmed theirs. White, Fichtenbaum, and Dollard (1964) successfully used a modification of this system to predict continuation in treatment of psychiatric outpatients from an analysis of initial interviews.…”