How would I calculate the convergent validity of two diagnostic tests where both outcome measures are binary?