Assessing Therapist Success

To the Editor.—  Luborsky et al¹ reported in the Archives on the determinants of therapist success in treating a sample of opiate addicts. Their results are intriguing, but there are some methodologic and statistical problems that call their conclusions into question.The authors state that their sample included 110 patients. However, the number of subjects falls to 78 in Table 1, 41 in Table 7, and 28 in Table 8. The excluded subjects may have had certain characteristics that promoted their departure, thus leaving a non-representative subsample. Further, selective subject loss from some therapist caseloads, but not from others, may have resulted in groups heterogeneous with respect to some unmeasured factor that influenced outcome. Thus, the observed intertherapist outcome differences may have been affected by factors unrelated to the treatment that the patients received.The number of therapy sessions varied between three and 24 among patients. The differences among