“…Even so, the sample is not small, and variability is evident, so selection bias for a uniform population does not appear to be operating. A larger study would also allow for fuller exploration of sex-differences, differences in residence prior to admission, and differences in relation to other measures of illness severity [39]. The inter-rater reliability study was not blinded, although this is mitigated by much of the data coming from the health record, entered by people blind to the intent of test-retest reliability.…”