Table 1: Potential confounders used in implementing TMLE and PSM.

Covariate set Model covariates Description
Simple set W1 Age, sex, race, type of insurance, primary disease category, and apache score
Moderate set W2 W1 + cancer status, comorbidities, SUPPORT model estimate
of the probability of surviving 2 months, mean blood pressure, white blood cell count, heart rate, respiratory rate, temperature,
albumin, sodium, pH levels, do-not-resuscitate (DNR) status on day 1
Full set W3 W2 + years of education, income, secondary disease category,
respiratory diagnosis, cardiovascular diagnosis, neurological diagnosis, gastrointestinal diagnosis, renal diagnosis, metabolic diagnosis,
hematologic diagnosis, sepsis diagnosis, trauma diagnosis, orthopedic diagnosis, Duke Activity Status Index (DASI),
death by date 30, Glasgow Coma Score, weight, PaO2/FlO2 ratio, PaCo2,
hematocrit, potassium, creatinine, bilirubin