01_DEMOG.Rnw

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%% DEMOGRAPHICS - BEGIN
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

\clearpage
\newpage
\section{Statistical Methods}
Statistical analyses of this non-interventional study were descriptive and exploratory in nature.
For continuous variables, summary statistics, e.g., the mean, standard deviation (SD), median, 
lower (Q1) and upper quartiles (Q3) and minimum and maximum values were calculated. 
For dichotomous and categorical variables, the number and percentage of participants in each
category were reported. 95\% confidence intervals (CIs) for categorical variables were
calculated according to the Wilson score method, without continuity correction. 95\% CIs for 
approximately normally distributed continuous variables were calculated using 
the \textit{t}-distribution. For the body weight, the geometric mean was presented 
togehther with 95\% CIs. The geometric mean was calculated as the anti-logarithm of the 
mean of the log-transformed values. \\
The Fisher's exact test or Pearson's $\chi^2$ test was used to compare categorical
variables, the Wilcoxon test or Kruskal-Wallis test was used to compare differences in continuous
variables between two or more groups. \\
Shift tables from Baseline to week 24 were presented for HGB and WBC based on changes in the 
normal ranges and tested with the McNemar’s test/Bowker’s symmetry test. \\
Log body weight was analyzed in a multiple linear regression model with sex, age and 
height as independent variables. An ANOVA table was constructed with Type I, II and III
sum of squares.

Overall survival (time from baseline to death or censoring) was estimated
by the univariate Kaplan-Meier method and the median time to event was presented 
together with 95\% CIs, overall and stratified by age group <65 years vs $\ge$ 65 years
Survival curves were compared by the log-rank test. \\

All p-values were descriptive and no adjustments for multiplicity were made. \\

All analyses were based on the Full Analysis Set (FAS), consisting of all 
subjects that received at least on dose of WonderDrug\texttrademark. Analysis was based 
on available data at each timepoint (i.e. missing data were not imputed), with exception
of the analysis of WBC over time: the last observation carried 
forward (LOCF) method was used to input missing data and was presented - as sensitivity
analysis - in addition to the analysis of the non-imputed data. \\
Subgroup analyses were performed by age group (<65 years, $\ge$ 65 years). \\

R version 3.4.2 (The R Foundation for Statistical Computing, Vienna, Austria) was
used for statistical analysis. \\

All data were \textit{fictitious} and the statistical methods that were applied served mainly
as examples to show the usage of \emph{R code} within different \emph{R packages}.
Data were generated with \href{https://www.sas.com}{SAS} v. 9.4 
(SAS Institue Inc., Cary, NC, USA) using e.g. following functions: 
\textit{rand("BINOMIAL", p, n)}, \textit{rand("NORMAL", mu, sigma)}, 
\textit{exp(rand("NORMAL", mu, sigma))} and \textit{-log(ranuni(x))}. \emph{CDISC ADaM}
standards were partly applied. 
\href{https://github.com/clindocu/sasxpt-r}{clindocu/sasxpt-r} was used to 
generate the data definition table and to import SAS .xpt files into R with associated
user-defined formats (i.e. R \emph{factor(s),
levels and label(s)}). \\
This report was generated with \href{http://www.tug.org/texlive}{Tex Live 2017},
\href{https://github.com/rstudio/rstudio}{RStudio},
\href{https://www.r-project.org}{R} and \href{https://github.com/yihui/knitr}{knitr}. \\

The data definition table (DDT) of all analysis datasets can be found here: 
\attachfile[appearance=true, icon=Tag, description=Data Definition Table Analysis Datasets, 
color=0.254902 0.411765 0.882353, mimetype=application/pdf,subject=DDT]{DDT.pdf}
\tiny (double click to open) \\
\normalsize

Following \emph{R packages} were used:

\begin{itemize}
\item library(knitr)
\item library(foreign)
\item library(tidyverse)
\item library(Hmisc)
\item library(xtable)
\item library(gridExtra)
\item library(survminer)
\item library(ggthemes)
\item library(reporttools)
\item library(texreg)
\item library(car)
\end{itemize}

All code for generating this report can be found in appendix \ref{AllCode}.

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%% Text with \Sexpr{}
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

\clearpage
\newpage
\section{Data Sets Analysed}
Data were collected from \Sexpr{length(which(ADSL[["FASFL"]] == "Yes"))} subjects
from \Sexpr{sum(table(table(ADSL$SITEID)))} study sites, all meeting the criteria for
the FAS. \Sexpr{length(which(ADSL[["AGEGR1"]] == "<65"))} subjects were younger 
than 65 years, \Sexpr{length(which(ADSL[["AGEGR1"]] == ">=65"))} subjects were older 
than or equal 65 years. \\

The disposition of subjects is shown in figure \ref{SubjDis} and is tabulated in
table \ref{Dispos}.  Baseline characteristics (Age, Sex, Race, Weight, Height, BMI, 
ECOG, HGB) are summarized in table \ref{ADemo2}.

\subsection{Study Sites, by Age Category}

<<results='asis'>>=
# Hmisc:::summary.formula options;
# overall= TRUE      - column with overall statistics;
# test = FALSE       - no test statistics;
# na.include = FALSE - if TRUE, exclude missing values from being counted as category;
# continuous = 8     - threshold for when a numeric variable is considered to be continuous;
s6 <- summary(AGEGR1 ~ SITEID, data = ADSL, method = 'reverse', 
              overall = TRUE, test = FALSE, na.include = FALSE, continuous = 8)

# Hmisc:::latex() options:;
# npct = "both"       - numerator and denominator;
# landscape = FALSE   - portrait;
# long = TRUE         - print the results for the first category on its own line;
# prmsd = TRUE        - print mean and SD after the three quantiles;
# here = TRUE         - table will appear here;
# prtest = 'P'        - P value when test = TRUE (can be c('P','stat','df','name'));
# dotchart = FALSE    - no dotcharts in output;
# longtable = FALSE:  - does not allow multiple pages;
# caption = ""        - text string to use as a caption;
# caption.loc = "top" - top or bottom;
# middle.bold = TRUE  - bold face for the middle quantile;
# file = ""           - LaTeX code will just printed to standard output;
# digits = 2          - digits for printing (continuous variables);
# exclude1 = FALSE    - e.g. prints Female/Male, when TRUE: Female only;
# label = ""          - unique label, can be refferenced with \ref{myLabel};
latex(s6, npct = "both", landscape = FALSE, long = TRUE, prmsd = TRUE, here = TRUE, 
  prtest = "P", dotchart = FALSE, longtable = FALSE, 
  caption = "Study Sites by Age Category - Hmisc:::summary.formula", 
  caption.loc = "top", 
  middle.bold = TRUE, file = "", digits = 2, exclude1 = FALSE, label = "Site01")
@

<<results='asis'>>=
s6 <- summary(AGEGR1 ~ SITEID, data = ADSL, method = 'reverse', 
              overall = TRUE, test = FALSE, na.include = FALSE, continuous = 8)

latex(s6, npct = "both", landscape = FALSE, long = TRUE, prmsd = TRUE, here = TRUE, 
  prtest = "P", dotchart = TRUE, longtable = FALSE,
  caption = "Study Sites by Age Category - Hmisc.summary with dotchart", 
  caption.loc = "top", 
  middle.bold = TRUE, file = "", digits = 2, exclude1 = FALSE, label = "Site02")
@

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%% Include a PDF File
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\clearpage
\newpage
\subsection{Disposition of Subjects}
\begin{figure}[htp!]
  \centering
  \includegraphics[width=15cm]{Disposition.pdf}
  \caption{Disposition of Subjects}
  \label{SubjDis} 
\end{figure}

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\clearpage
\newpage
\subsection{Disposition of Subjects, by Age Category}
<<results='asis', echo=TRUE>>=
s6 <- summary(AGEGR1 ~ EOSSTT + DCSREAS, data = ADSL, method = 'reverse', 
              overall = TRUE, test = FALSE, na.include = FALSE, continuous = 8)

latex(s6, npct = "both", landscape = FALSE, long = TRUE, prmsd = TRUE, here = TRUE, 
  prtest = "P", dotchart = FALSE,
  caption = "Disposition of Subjects - Hmisc.summary with na.include = F", 
  caption.loc = "top", 
  middle.bold = TRUE, file = "", digits = 2, exclude1 = FALSE, label = "Dispos")
@

\subsubsection{Listing Reason for Non-Completion}

<<results='asis', echo=TRUE>>=
df1 <- ADSL %>%
  select(USUBJID, SEX, AGE, DCSREAS, DSTERM) %>%
  filter(DCSREAS != "") %>%
  arrange(DCSREAS)

my_table <- xtable(df1, caption = "Listing Reason for Non-Completion - xtable", 
                   align = "lrrrlp{6cm}", caption.placement = "top")

digits(my_table)[4] <- 0 # Age (column 4 with 0 digits);
print(my_table, caption.placement = "top", size = "small", include.rownames = TRUE, 
  table.placement = "htb", tabular.environment = "longtable", floating = FALSE)
@

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\clearpage
\newpage
\section{Demographic and Other Baseline Characteristics}
\subsection{Demographic and Other Baseline Characteristics - Hmisc}

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%\clearpage
%\newpage

<<results='asis'>>=
# Add unit to Age in ADSL;
units(ADSL$AGE) <- "years"

s6 <- summary(AGEGR1 ~ AGE + SEX + RACE + WEIGHTBL + HEIGHT + BMI + ECOGBL + HGBBL, 
            data = ADSL, method = 'reverse', overall = TRUE, test = FALSE,
            na.include = FALSE, na.rm = TRUE, continuous = 8)

latex(s6, npct = "both", landscape = TRUE, long = TRUE, prmsd = TRUE, here = T, 
  prtest = "P", dotchart = FALSE,
  caption = "Demographic and Other Baseline Characteristics - Hmisc.summary", 
  caption.loc = "top", 
  middle.bold = TRUE, file = "", digits = 2, exclude1 = FALSE, label = "ADemo1")
@

<<results='asis'>>=
# Create new variables;
ADSL$WBL <- ADSL$WEIGHTBL
ADSL$HBL <- ADSL$HEIGHT
ADSL$BML <- ADSL$BMI
ADSL$HML <- ADSL$HGBBL

# Add Labels;
label(ADSL$WBL) <- "Weight at Baseline"
label(ADSL$HBL) <- "Height"
label(ADSL$BML) <- "Body Mass Index"
label(ADSL$HML) <- "Hemoglobin at Baseline"

### Add uniits;
units(ADSL$WBL) <- "kg"
units(ADSL$HBL) <- "cm"
units(ADSL$BML) <- "kg/m2"
units(ADSL$HML) <- "g/dL"

s6 <- summary(AGEGR1 ~ AGE + SEX + RACE + WBL + HBL + BML + ECOGBL + HML, data = ADSL, 
              method = 'reverse', overall = TRUE, test = TRUE, 
              na.include = FALSE, na.rm = TRUE, continuous = 8)

latex(s6, npct = "both", landscape = TRUE, long = TRUE, prmsd = TRUE, here = T, 
  prtest = "P", dotchart = FALSE,
  caption = "Demographic and Other Baseline Characteristics - Hmisc.summary with units and P-value",
  caption.loc = "top", middle.bold = TRUE, file = "", digits = 2, exclude1 = FALSE, label = "ADemo2")
@

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\clearpage
\newpage
\subsection{Demographic and Other Baseline Characteristics - reporttools:::tableContinuous}

<<01_BaseChar, results='asis', echo=TRUE>>=
# Generate dataframe with variables;
vars <- with(ADSL, data.frame(
  "Age years" = AGE,
  "Weight kg" = WEIGHTBL,
  "Height cm" = HEIGHT,
  "BMI kg.m2" = BMI,
  "Hemoglobin g.dL" = HGBBL))

cap <- "Baseline Characteristics - reporttools:::tableContinuous - Wilcoxon test"
lab <- "tab:BaseChar"
bye <- ADSL$AGEGR1
@

<<ref.label="tabContBy", results='asis', echo=TRUE>>=
@


%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
\clearpage
\newpage
\subsection{Demographic and Other Baseline Characteristics - reporttools:::tableNominal}

<<01_BaseCategor, results='asis', echo=TRUE>>=
vars <- with(ADSL, data.frame(
  "Sex" = SEX,
  "Race" = RACE,
  "ECOG" = ECOGBL))

cap <- "Baseline Char. - tableNominal - NAs excluded - Fisher's exact test"
lab <- "tab:01_BaseCategor"
bye <- ADSL$AGEGR1
@

<<ref.label="tabNomBy", results='asis', echo=TRUE>>=
@

<<01_BaseCatMiss, results='asis'>>=
cap <- "Baseline Characteristics - tableNominal - NAs as category"
lab <- "tab:01_BaseCatMiss"
bye <- ADSL$AGEGR1
@

<<ref.label="tabNomBy2", results='asis', echo=TRUE>>=
@

%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%%% DEMOGRAPHICS - END
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%