In harrelfe/hreport: Interactive Graphical HTML Reports for Clinical Trials

require(hreport)  # add to above: results='hide'
knitrSet(lang='markdown', fig.path='figure/')
options(prType='html')
mu   <- markupSpecs$html   # in Hmisc - HTML markups
frac <- mu$frac

mu$styles()              # define HTML styles, functions

# Function to generate data for one study
gendat <- function(study, n) {
f <- function(p) rbinom(n, 1, p)
d <-
  data.frame(study=study,
               subjid=1:n,
             site=sample(1:10, n, TRUE),
             tx=sample(c('A','B'), n, TRUE),
                         age=rnorm(n, 50, 10),
                         sex=sample(c('female', 'male'), n, TRUE),
                         edate=as.Date('2019-10-01') + round(runif(n, 0, 100)),
                         sbp=rnorm(n, 120, 7),
             dbp=rnorm(n,  80, 6),
             race=sample(c('Asian', 'Black/AA', 'White'), n, TRUE),
             meda=f(0.3),
             medb=f(0.2),
                         td=round(runif(n, 0, 30)),
                         death=rbinom(n, 1, 0.4),
                         e1     = f(.03),
             e2     = f(.03),
             e3     = f(.03),
             e4     = f(0.05),
             tested = rbinom(n, 1, .75),
             e6     = f(.03),
                 e7     = f(0.04)
                         )
d <-
  upData(d,
        e5     = ifelse(tested, f(.06), NA),    # don't randomize if exclusion present:
        rdate  = ifelse(e1|e2|e3|e4|ifelse(is.na(e5), 0, e5)|e6|e7, NA,
                    edate + round(runif(n, 0, 30))),
        rdate  = as.Date(rdate, origin='1970-01-01'),   # ifelse loses Date class
        tx     = ifelse(is.na(rdate), NA, tx),
        labels=c(age='Age',
                 rdate='Randomization Date',
                         edate='Enrollment Date',
                   e1='Prior MI', e2='History of Asthma',
             e3='History of Upper GI Bleeding',
             e4='No Significant CAD', e5='Inadequate Renal Function',
             e6='Pneumonia within 6 weeks', e7='Hostile',
             race='Race', sex='Sex', tx='treatment',
             sbp='Systolic BP', dbp='Diastolic BP',
             meda='Comcomitant medication A',
                         medb='Concomitant medication B',
                         td='Follow-up Time', death='Death'),
             units=c(sbp='mmHg', dbp='mmHg', age='years', td='day'), 
       print=FALSE)

denom <- c(enrolled=nrow(d), randomized=sum(! is.na(d$rdate)),
           table(subset(d, ! is.na(rdate))$tx))
sethreportOption(tx.var='tx', denom=denom, study=study)
d
}
set.seed(1)
d1 <- gendat('Study 1', 500)
d2 <- gendat('Study 2', 250)
d3 <- rbind(d1, d2)
denom <- c(enrolled=nrow(d3), randomized=sum(! is.na(d3$rdate)),
           table(subset(d3, ! is.na(rdate))$tx))
sethreportOption(tx.var='tx', denom=denom, study='Study 1+2')

Introduction

This is an example report using simulated data. Two studies are simulated. Variable names in the two datasets are the same here, though they needn't be. The study IDs are Study 1 and Study 2. These IDs are used to label the tabs that can be clicked to allow the alternate study result to be viewed.

Since the variable names are the same, the two studies have also been combined to produce a third "study" called e.g. Study 1+2. For certain results, study is used as a stratification variable for the Study 1+2 tab. Since patient accrual is unique to each study, a combined accrual report is not presented. The exclusion criteria report is also not included for the combined trials.

For accrual and exclusion reports, the entire database is used. After that point, datasets are subsetted to include only randomized participants.

Interactive Graphs

Most of the graphs produced here are semi-interactive. One can hover over elements of graphs with the mouse to have detailed information pop up.

Figure Captions

Needles represent the fraction of observations used in the current analysis. The first needle (red) shows the fraction of enrolled patients used. If randomization was taken into account, a second needle (green) represents the fraction of randomized subjects included in the analysis. When the analyses consider treatment assignment, two more needles may be added to the display, showing, respectively, the fraction of subjects randomized to treatment A used in the analysis and the fraction of subjects on treatment B who were analyzed. The colors of these last two needles are the colors used for the two treatments throughout the report. The following table shows some examples. dNeedle uses colors in sethreportOption(tx.col=, er.col=).

# Store using short variable names so Rmarkdown table column
# width will not be wider than actually needed
dned <- function(x) dNeedle(x, study='Study 1')
n1 <- dned(1)
n2 <- dned((3:4)/4)
n3 <- dned((1:2)/4)
n4 <- dned(c(1,2,3,1)/4)

|Signpost | Interpretation | |------- | -------------------------------------------------| | r n1 | All enrolled subjects analyzed, randomization not considered| | r n2 | Analysis uses r frac(3,4) of enrolled subjects, and all randomized subjects| | r n3 | Analysis uses r frac(1,4) of enrolled subjects, and r frac(1,2) of randomized subjects| | r n4 | Same as previous example, and in addition the analysis utilized treatment assignment, analyzing r frac(3,4) of those randomized to A and r frac(1,4) of those randomized to B|

Survival Curves

Graphs containing pairs of Kaplan-Meier survival curves show a shaded region centered at the midpoint of the two survival estimates and having a height equal to the half-width of the approximate 0.95 pointwise confidence interval for the difference of the two survival probabilities. Time points at which the two survival estimates do not touch the shaded region denote approximately significantly different survival estimates, without any multiplicity correction.

Accrual {.tabset .tabset-fade}

Study 1

accrualReport(enroll(edate) + randomize(rdate) ~ site(site),
              data=d1,
              dateRange=c('2019-10-01', '2020-03-01'),
              targetN=
                data.frame(edate=c(250, 500), rdate=c(125, 250)),
              targetDate=c('2020-01-15', '2020-03-01'),
              closeDate='2020-06-01', study='Study 1')

Study 2

accrualReport(enroll(edate) + randomize(rdate) ~ site(site),
              data=d2,
              dateRange=c('2019-10-01', '2020-03-01'),
              targetN=
                data.frame(edate=c(125, 250), rdate=c(75, 125)),
              targetDate=c('2020-01-15', '2020-03-01'),
              closeDate='2020-06-01', study='Study 2')

Exclusions {.tabset .tabset-fade}

Study 1

exReport(~ e1 + e2 + e3 + e4 + e5 + e6 + e7 +
         id(subjid) + cond(e5, 'Tested', tested),
         whenapp= c(e4='CCTA done'), study='Study 1', data=d1)

Study 2

exReport(~ e1 + e2 + e3 + e4 + e5 + e6 + e7 +
         id(subjid) + cond(e5, 'Tested', tested),
         whenapp= c(e4='CCTA done'), study='Study 2', data=d2)

Baseline Variables {.tabset .tabset-fade}

Study 1

# From here on, only randomized participants will be considered
d1 <- subset(d1, ! is.na(rdate))
d2 <- subset(d2, ! is.na(rdate))
d3 <- subset(d3, ! is.na(rdate))
dReport(race + sex ~ 1, head='Overall frequencies of categorical demographic variables',
        data=d1, study='Study 1')

# Show continuous variables stratified by treatment
dReport(dbp + sbp + age ~ tx, data=d1, study='Study 1', sopts=list(width=800))

Study 2

dReport(race + sex ~ 1, head='Overall frequencies of categorical demographic variables',
        data=d2, study='Study 2')
dReport(dbp + sbp + age ~ tx, data=d2, study='Study 2', sopts=list(width=800))

Study 1+2

When combining studies we don't need to examine baseline variables stratified by treatment but rather stratified by study.

dReport(race + sex ~ study,
        head='Overall frequencies of categorical demographic variables',
        data=addMarginal(d3, study),
        study='Study 1+2')

dReport(dbp + sbp + age ~ study,
        data=addMarginal(d3, study),
        study='Study 1+2', sopts=list(width=800, ncols=2))

Survival {.tabset .tabset-fade}

Study 1

survReport(Surv(td, death) ~ tx, data=d1, study='Study 1')

Study 2

survReport(Surv(td, death) ~ tx, data=d2, study='Study 2')

Study 1+2

survReport(Surv(td, death) ~ tx,
           data=d3, study='Study 1+2')

harrelfe/hreport documentation built on July 26, 2021, 9:09 a.m.