The problem statement poses two research
questions: "Does inpatient length of stay differ between patients
with and without AKI?" and “How does the likelihood of
having AKI change with length of stay?” The three questions
below will guide the selection of appropriate statistical methods.
-
What is the response
(Y) of interest and how is it measured? For the first research question,
the length of stay is of interest. It is a continuous variable.
For the second research question the outcome variable is a diagnosis
of AKI, which is a nominal variable.
-
Are predictor variables
mentioned in the problem statement? If so, how many and what are
their measurement levels? Outcome (absence or presence of AKI) is
in the role of predictor in the first research question. Lengthy
of stay takes on the role of predictor in the second research question.
-
What are you being asked
to deliver? A data description, an interval estimate, an answer to
a question, or a predictive model? For the first research question,
a test of hypothesis is appropriate to determine if statistically
significant difference in length of stay exists based on the presence
of AKI. For the second research question, a statistical model that
describes the relationship between the likelihood of AKI and length
of hospital stay will address the research question.