Digital Support for Addiction Recovery
An app for young people, in five languages, developed in PREPARE — a five-year BMBF-funded research project led by DISuP at the Katholische Hochschule NRW.
German was the condition for getting help
Evidence-based addiction prevention existed, and it existed in German. Young people who did not read German were left outside it.
BePrepared supports young people through alcohol and cannabis addiction in Arabic, German, English, Farsi and Pashto. It tracks consumption, sets personal goals, teaches relaxation techniques and holds the reference material. Free, anonymous, available anytime.
Designing inside five years of addiction-prevention research
The team was one user researcher, two engineers, two psychologists and an addiction-prevention expert. I was the sole designer. Before design started, the researchers spent about a year on the domain report everything else was built on: where risk happens, what stops people from seeking help, and when a digital intervention can be trusted.
Five languages, one consistent product
The hard part was not the feature set. The same product had to work in five languages without forking into five products, so the design system had to absorb what each language demanded.
Right to left. Layout, navigation order, directional icons and progress indicators mirror for Arabic. Components were built bidirectionally from the start, not retrofitted, so the Arabic version is not a translated afterthought.
Text length. The same string changes width and height across the five scripts. Every component was sized for the worst case, so the hierarchy holds everywhere.
Tone. Addiction carries different stigma in each group. Language and imagery stay measured, with no metaphors that travel badly or imply judgement. The Arabic version was reviewed with Arabic-speaking colleagues; user testing with the target group ran through the project partners as part of the clinical protocol.
It held across the hardest audience to keep
Keeping participants in a multilingual study over six months is materially harder than a general-population sample. The design held across five languages and that timeframe.