A higher sport-related reinjury risk does not mean inadequate rehabilitation: the methodological challenge of choosing the correct comparison group