There are a number of limitations of this study. Because it was designed as a pilot study, the
sample size was small and thus statistical power was weak. Although some effect sizes were
large, they were not significant due to the lack of power. For example, effect sizes for theexercise and sertraline interventions on clinician-rated depressive symptoms (38.7% – 46.0%
change) and on mental health quality of life (26.6% – 42.1% change) were substantial and there
were small to large effects of each intervention on domains of physical functioning. The
variance of the effect sizes suggests heterogeneity in participants’ responses to the intervention:
the intervention may have produced substantial changes in outcomes for some participants
while the average effect size may only be moderate. Furthermore, even with the small sample
size, there were indeed trends and significant differences for multiple outcomes. Thus, it
appears that exercise and sertraline have the potential to produce significant changes in
depression and physical functioning among older adults with minor depression. The
generalizability of these findings may be limited by sample bias due to the recruitment
strategies employed. Another limitation is that the diagnosis of minor depression was based
on self-reported symptoms rather than a full clinical interview. Participant satisfaction with
treatment was not assessed; however, it is likely that people were satisfied with treatment as
adherence to the exercise intervention averaged 80% and only 9% of participants in the
sertraline condition dropped out of the study, less than the drop out rate of a large study of
sertraline for the treatment of late-life major depression (Schneider et al., 2003). Finally, the
exercise intervention was center-based and findings may not generalize to home-based exercise
programs.