The joint distribution of the starting state, age, and SOFA score of the
1,351 patients in the
VIOLET trial of early high-dose
vitamin D3 versus placebo in critically ill adults. Each row is one
combination of the three characteristics, and n is the number of VIOLET
patients who had it. sim_markov("violet") samples simulated patients from
these rows (see violet_preset).
Format
A data frame with 858 rows and 4 variables:
- yprev
State before the first day of follow-up: 2 = in hospital or another health care facility, 3 = on a ventilator or with acute respiratory distress syndrome (ARDS). The states used by
sim_markov("violet")are 1 = at home, 2 = in hospital or facility, 3 = on a ventilator or with ARDS, and 4 = dead.- age
Age in years.
- sofa
Sequential Organ Failure Assessment (SOFA) score at enrollment, from 0 to 24; higher scores mean more severe organ failure.
- n
Number of VIOLET patients with this starting state, age, and SOFA score. The counts sum to 1,351.
Source
Derived from simlongord, created by Frank Harrell and available through
Hmisc::getHdata(simlongord).
Details
To draw size patients as sim_markov("violet") does, sample rows with
replacement and with probabilities proportional to n:
violet_baseline[sample(nrow(violet_baseline), size, replace = TRUE, prob = violet_baseline$n), ]. Use rep(seq_len(nrow(violet_baseline)), violet_baseline$n) as row indices to get one row per VIOLET patient.
The original VIOLET data are not public. Frank Harrell published a simulated
trial of 250,000 patients, simlongord, whose starting states, ages, and
SOFA scores were drawn with replacement from the 1,351 VIOLET patients with
a record on the first day (Harrell, 2022). Each VIOLET patient was therefore
drawn about 185 times, and dividing how often a combination occurs in
simlongord by 185 recovers how many VIOLET patients had it. The recovered
counts reproduce every summary that Harrell reported for these patients,
including the 474 patients on a ventilator or with ARDS, the number of
patients with each SOFA score, and the mean and percentiles of age. The
script data-raw/violet_baseline.R in the package sources performs these
steps.
References
National Heart, Lung, and Blood Institute PETAL Clinical Trials Network (2019). "Early High-Dose Vitamin D3 for Critically Ill, Vitamin D-Deficient Patients." New England Journal of Medicine, 381(26), 2529-2540. doi:10.1056/NEJMoa1911124
Harrell FE Jr (2022). "Simulating a Large Clinical Trial From VIOLET." https://hbiostat.org/data/repo/simlongord
Examples
head(violet_baseline)
#> yprev age sofa n
#> 1 2 19 1 1
#> 2 2 19 4 1
#> 3 2 20 2 1
#> 4 2 20 3 2
#> 5 2 20 7 1
#> 6 2 21 1 1
# Number of patients starting in each state
tapply(violet_baseline$n, violet_baseline$yprev, sum)
#> 2 3
#> 877 474
# Mean age of the VIOLET patients
weighted.mean(violet_baseline$age, violet_baseline$n)
#> [1] 56.31088