Towards a national, practice based, research data infrastructure for child and adolescent psychiatry

As in other fields of medicine, evidence-based knowledge in child mental health is primarily based on controlled studies conducted in highly selected (academic) populations. Although such studies are of undeniable importance for testing relationships between variables within a study (internal validity), their weakness is a lack of external validity (Rothwell, 2005), thereby limiting generalizability to more complex populations. Thus, clinicians and academics alike have argued the need to increase research on ‘real life’ clinical populations (Von Elm et al., 2008). By doing so, we will substantially increase our knowledge of complex patients who are normally excluded from controlled studies. An important reason for the limited number of practice-based studies is the absence of a research data infrastructure supporting the successful implementation of studies within clinical practice. This is particularly so within the field of (child and adolescent) psychiatry, because diagnostic assessment requires information to be collected from different sources (child, parent, teachers) using a diversity of instruments (interview, self-report questionnaires, tests). When studies are successfully implemented, they all too often have small samples, lacking power to allow the assessment of interactions and the role of moderators (La Greca et al., 2009). To greatly facilitate this pressing need in the field, our consortium is in a unique position to conduct multi-site research studies integrated in everyday clinical care. In 2008, a consortium of 16 clinical and 6 academic institutions (called ROMCKAP) joined forces to implement Routine Outcome Monitoring (ROM) in the Dutch Child and Adolescent Psychiatry field. The aim of ROMCKAP is to improve the quality of care by regularly assessing treatment progress and outcome. This system allows the collection of data for comparing patients as well as institutions on outcome. However, the system currently lacks the flexibility to evaluate specific research protocols that require disorder-specific instruments to be used at specific time points within the system. The initial infrastructure provides an excellent basis to expand this clinician quality measurement system into a research supporting infrastructure. The main aim of this proposal is therefore to develop a research data infrastructure integrated in Routine Outcome Monitoring developed by ROMCKAP to allow for conducting practice based studies. To achieve this aim, we propose to make additional dynamic connections between software packages used in clinical care and a data management system (ProMISe), add research functionalities to ProMISe, build a library and codebook of variables, and enable visualizing data collected for research by clinicians. The research data infrastructure will further be tested by means of a number of technical pilot studies, which will result in implementation manuals and training modules. Although ProMISe is originally designed as a generic research data infrastructure, its main use over the past decade has been in somatic medicine. For optimal application in child and adolescent psychiatry, a number of adaptations are necessary due to the unique characteristics of assessment in child mental health. The data management system will serve as a tool that allows individual researchers to implement studies in (sub)populations in a number of institutions and to adapt instruments and measurement points, while keeping in sync with developments in everyday clinical care. The research data infrastructure will also be able to be used (with relatively minor adaptations) in related fields, such as adult psychiatry, forensic psychiatry and youth care. Furthermore, this system will also facilitate adherence to basic treatment research principles described in CONSORT (Hopewell et al, 2008) and STROBE guidelines (Von Elm et al., 2007), and the Good Clinical Research Practice criteria (WHO, 2002).