As the study was an effectiveness trial, all couples received
couple therapy. Data were missing on approximately 6% of the
weekly questionnaires. However, therapist reports indicated that
these responses were generally missing because of staff omission or
late commencement of a session, and therefore these data were
treated as missing at random. Mean treatment length at the San
Diego VAMC was 12.3 sessions (SD ¼ 7.75), whereas at the
Charleston VAMC, the mean number of sessions was 4.49
(SD ¼ 3.37). Therapy was provided by 40 co-therapy teams who, on
average, provided therapy to approximately two couples. Treatment
at the Charleston site was a brief, problem-focussed theoretical
orientation, in which communication skill training was a
central component of the therapy. Treatment at the San Diego site
was a largely behavioural intervention. A detailed description of the
therapy and treatment fidelity is provided by Doss et al. (2012).
There were no significant differences in the effects of treatment
across the two sites. In order to optimise power we evaluated the
predictability of therapy outcome from therapy progress across the
two treatment sites combined, and then examined potential differences
between sites.