A note on the asymptotic optimality of a two population sequential probability ratio test