The focus of the "Instruments and Methods" area is on the German-language adaptation and psychometric validation of existing, well-established survey instruments (mainly from the field of depression) and the development of new instruments to measure treatment satisfaction in the KJPP. Other topics include the process-orientated measurement of emotional well-being, participation and appreciation of therapy as well as the development of goal attainment scales.
In the methodological-statistical area, emphasis is placed on the application of newer analytical procedures (psychometric analyses and analysis of longitudinal data), which in many cases enable a more in-depth and more powerful evaluation of the data.
Several well-established questionnaires are available internationally to assess the severity of depression in self-assessment. For clinical purposes, the Beck Depression Inventory (BDI-II) is frequently used, a German translation of which has been published for adults since 2006 (Hautzinger et al., 2006). Besier et al. (2008) presented the first application in adolescents. The General Depression Scale (ADS) is frequently used in epidemiological studies (school examinations, etc.). The ADS has also been used in our own studies and has contributed to the German ADS test manual (Hautzinger et al., 2012).
As self-assessment and external assessment of depression usually correlate only moderately (cf. Straub et al., 2014), external assessment methods are also important. A very frequently used instrument in the USA is the Children's Depression Rating Scale - Revised (CDRS-R). This semi-structured interview was translated by us, adapted to German conditions and psychometrically evaluated in a clinic sample (Keller et al., 2011) and in a school study (Keller et al., 2012).
Dysfunctional attitudes play an important role in the context of cognitive-behavioural disorder models of depression. With the Dysfunctional Attitudes Scale (DAS), a measurement instrument for adults has been available for some time (Hautzinger et al., 2005) and we have developed a version suitable for adolescents (Keller et al., 2010).
The subjective satisfaction of patients is now regarded as an important evaluation criterion in the context of quality assurance. In order to record treatment satisfaction, the questionnaires for treatmentassessmentin inpatienttherapy(BesT) were developed in a version for children (BesT-K), for adolescents (BesT-J) and for parents (BesT-E). The questionnaires were designed to be quick and easy to answer, but also to cover all the key aspects of child and adolescent psychiatric treatment. In addition, care was taken to ensure that the topics were asked of the child/adolescent and their parents in parallel wherever possible. The questionnaires have been used regularly in several child and adolescent psychiatric clinics for around 10 years. For the children's version, see Keller et al. (2004); a publication on the adolescent and parent versions is currently being submitted.
Scales have also been developed for the daily and weekly recording of well-being, therapy motivation, participation and treatment satisfaction. There are forms for the adolescents, for carers and for therapists. The scales were used in an inpatient context in Ulm (e.g. Frick, 2011; Keller et al., 2006) and in adolescent addiction therapy (JUST) at ZfP Weissenau (Keller et al., submitted).
As part of the development, implementation and evaluation of instruments for educational goal achievement(projects PädZi, MAZ), instruments were developed that, on the one hand, map the general competences of young people in residential care or similar facilities and, on the other hand, contain individually formulated goals (Kleinrahm et al., 2013; Singer et al., 2009).
Other instruments developed include a scale for assessing weight anxiety in anorexia nervosa (Schulze & Keller, 2009). In the forensic field, the Forensic Operationalised Therapy Risk Evaluation System (FOTRES) was tested for its inter-rater reliability (Keller et al., 2011).
In the methodological-statistical field, the focus is often on modelling and testing, e.g. in confirmatory factor analyses (CFA) or more generally in structural equation models, and less on significance testing of many individual variables. Accordingly, some of these methods, especially CFA, were used in the evaluation of the above questionnaires, but also methods of item response theory (IRT), e.g. the Rasch model and its generalisations (Keller et al., 2008; Keller, 2012). Current bifactor models, in which all items load on a general factor and additionally on one of several specific factors, were applied to the BDI-II (Bühler et al., 2014) and the SDQ in the parent version (Keller & Langmeyer, 2017).
Growth curve models are advantageous when analysing longitudinal data. Both known subgroups (e.g. treatment conditions, gender) can be compared in terms of their trajectory (hierarchical linear models - HLM) and unknown (latent) subgroups that show a similar group-specific trajectory can be identified using growth mixture models (GMM). The application to the data of a therapy study on depression can be found in Keller & Hautzinger (2007). Latent subgroups can also be searched for at the level of symptom patterns (latent class analysis - LCA). Applications include the search for subtypes of asthma in children (Weinmayr et al., 2013) and the differentiation of complex PTSD from "simple" PTSD (Sachser et al., in press).