In a town with an epidemic, the probability that a randomly chosen person is infected is $p$. A test for the disease has sensitivity $s_1$ and specificity $s_2$. If a person tests positive, they are retested with a different test having sensitivity $s_3$ and specificity $s_4$. Assuming the two tests are conditionally independent, what is the probability the person is actually infected given both tests returned positive results?