A binomial distribution comes from repeating an experiment many times, where each run takes a fixed number of samples with a binary outcome and counts how many came out one way. Here each day is one run: a fixed number of admissions, each of whom either survives or does not.