The Delphi language is a component-based integrated development environment, and it features the capability to develop applications using VCL, FMX and MFC components. Although some components are not supported within the Delphi community edition, all components can be used with the free community edition of the IDE. Developers must make use of these components, while also accessing any tools that are provided by the IDE. In the event that an application is to be deployed in some other environment, such as the.NET framework, developers can opt to enhance the appearance of the application by using the newer system.
In this study, we have developed a model for the prediction of voluntary shoulder abduction recovery using a subset of the available DELPHI 2010 dataset. The DELPHI 2010 dataset is one of the largest prospectively collected datasets in Europe to develop prediction models for poststroke upper limb motor recovery. The use of this dataset for developing our prediction model allowed us to focus on enhancing this model with high reliability, validity and interindividual variability. In addition, we assessed the reliability of our model across time points poststroke to identify whether the predictive accuracy could be improved through repeated assessments. In addition, our approach is based on standardised prediction procedures to detect and assess the validity of the outcome measures used. During the development of our model, we included known determinants of recovery to assess their added value.
The DELPHI 2010 dataset is a multicenter, prospective, observational registry of consecutive patients with first-ever ischemic or hemorrhagic stroke of the anterior circulation admitted to 14 participating hospitals in Germany. The data were collected between July 2007 and April 2010 by trained medical staff using a standardized protocol. For this study, we included all patients who had at least three assessments post stroke (baseline (last 14 days), week 3 and month 6) with assessments of arm function based on the Upper Extremity subsection of the Stroke Impact Scale (SIS-UE; 16). For our analyses, we considered patients with upper extremity motor impairment at baseline (ARAT 3d9ccd7d82