Page 124 - Age of onset of disruptive behavior of residentially treated adolescents -Sjoukje de Boer

not include a control group, we cannot conclude about the effectivity of our interventions. Further, adolescents received a diversity of interventions, therefore a possible effect cannot be related to specific components. Unlike in RCTs, in which the effect of a single intervention is studied, the treatment of De Fjord consists of a combination of practice-based and evidence-based interventions, which have often been adjusted on the basis of experience and insight to make them suitable for the specific target group of De Fjord and utmost, for an individual. Also, in clinical institutions like De Fjord, no uniform sample without comorbidities exists. Thus, although evidence-based interventions were used, we are not sure to what extent the specific combination of interventions actually work for our heterogeneous target group.
Because of the naturalistic design, we cannot conclude about the representativeness and replicability of the study. Nevertheless, we believe that the results of the present thesis are relevant for a specific subgroup of youths, since several of our findings were in line with literature.
Next, the long period of research has likely affected the representativeness and replicability. During seventeen years, we conducted program evaluation research in a naturalistic context, which provided a wealth of valuable information. At the same time, the clinical practice was undoubtedly subject to substantial change, due to changes in policy, management, therapeutic climate, therapists and adolescents. We do not know to what extent changes over time in the influx of youths in the clinic has influenced our results. Changes in therapists obviously must have affected the treatment and hence the treatment outcome as well as the attrition. Besides, due to experience and new insights, the approach to subjects that could potentially drop out may have evolved over time.
Other limitations concern the restrictions attached to the way in which the classification in EO and AO was made. In the present thesis the EO and AO division was based on specific information that was obtained from the youngsters, parents and professionals, as well as information available in files. Of course, this information is likely to differ in other studies and in other circumstances. Because the EO and AO constructs can be operationalized in different ways, the outcomes may vary (Colins & Vermeiren, 2013). It is not only unclear to what extent the results of the retrospective
122