Package ‘parmsurvfit’

December 7, 2018
Title Parametric Models for Survival Data
Version 0.1.0
Description Executes simple parametric models for right-censored
survival data. Functionality emulates capabilities in 'Minitab', including
fitting right-censored data, assessing fit, plotting survival functions,
and summary statistics and probabilities.
Depends R (>= 3.5.0)
License GPL-2

URL https://github.com/apjacobson/parmsurvfit

BugReports https://github.com/apjacobson/parmsurvfit/issues
Encoding UTF-8
LazyData true
RoxygenNote 6.1.1
Imports fitdistrplus, flexsurv, ggplot2, stats, graphics
Suggests survival, utils, knitr, rmarkdown
VignetteBuilder utils, knitr
NeedsCompilation no
Author Ashley Jacobson [aut, cre],
Victor Wilson [aut],
Shannon Pileggi [aut]
Maintainer Ashley Jacobson <[email protected]>
Repository CRAN
Date/Publication 2018-12-07 05:30:03 UTC

R topics documented:
aggressive . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2
compute_AD . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
firstdrink . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3

1
2 aggressive

fit_data . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
graduate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
oscars . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
parmsurvfit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
plot_cumhaz . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
plot_density . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
plot_haz . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
plot_ppsurv . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
plot_surv . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
rearrest . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
surv_prob . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
surv_summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12

Index 13

aggressive Data on time until drivers honked their horn when being blocked from
an intersection

Description

Diekmann et al. (1996) investigated the association between driver characteristics and social sta-
tus of cars to aggressive driver responses by measuring the time that elapsed between the being
blocked and honking the horn. Researchers intentionally blocked 57 motorists at a green light by a
Volkswagen Jetta, and recorded the time it took for motorists to show signs of aggression. Signs of
aggression included honking their horn or beaming the headlights at the Jetta

Usage

aggressive

Format

A data frame with 57 rows and 2 variables:

seconds Number of seconds until showing signs of aggression

censor censoring status indicator variable (0 = censored event time, 1 = complete event time)

Source

https://stats.idre.ucla.edu/other/examples/alda/
compute_AD 3

compute_AD Anderson-Darling goodness of fit test statistic

Description
Computes Anderson-Darling goodness of fit test statistic given that the data follows a specified
parametric distribution.

Usage
compute_AD(data, dist, time = "time", censor = "censor")

Arguments
data A dataframe containing a time column and a censor column.
dist A string name for a distribution that has a corresponding density function and
a distribution function. Examples include "norm", "lnorm", "exp", "weibull",
"logis", "llogis", "gompertz", etc.
time The string name of the time column of the dataframe. Defaults to "time".
censor The string name of the censor column of the dataframe. Defaults to "censor".
The censor column must be a numeric indicator variable where complete times
correspond to a value of 1 and incomplete times correspond to 0.

Examples
data("rearrest")
compute_AD(rearrest, "lnorm", time = "months")
compute_AD(rearrest, "weibull", time = "months")

firstdrink Data on age at first drink of alcohol.

Description
Data on age at first drink of alcohol.

Usage
firstdrink

Format
A data frame with 1000 rows and 3 variables:
age the age at which the survey respondent had their first drink of alcohol
censor censoring status indicator variable (0 = censored event time, 1 = complete event time)
gender a dichotomous variable identifying gender (1 = male, 2 = female)
4 fit_data

Source

"National Comorbidity Survey (1990-1992)"

fit_data Fitting right censored survival data to distribution

Description

Fits right censored data to a distribution using maximum likelihood estimates.

Usage

fit_data(data, dist, time = "time", censor = "censor", by = "")

Arguments

data A dataframe containing a time column and a censor column.

dist A string name for a distribution that has a corresponding density function and
a distribution function. Examples include "norm", "lnorm", "exp", "weibull",
"logis", "llogis", "gompertz", etc.
time The string name of the time column of the dataframe. Defaults to "time".
censor The string name of the censor column of the dataframe. Defaults to "censor".
The censor column must be a numeric indicator variable where complete times
correspond to a value of 1 and incomplete times correspond to 0.
by The string name of a grouping variable. If specified, the function returns a list.
The list will be in alphabetic order of the values in the by column. Variable can
contain logical, string, character, or numeric data.

See Also

fitdistcens

Examples
data("rearrest")
fit_data(rearrest, "lnorm", time = "months")

fit_data(rearrest, "weibull", time = "months", by = "personal")

graduate 5

graduate Data on time until graduation for 1000 college students.

Description
A dataset that contains the time (in years) that 1000 students (472 males and 528 females) took to
graduate (obtain a bachelor’s degree) from college (measured from the time they entered a post-
secondary institution, i.e. either a junior college or four year degree granting institution). The
Gender column contains the gender of each student (1 = male, 2 = female), and Censor contains the
values of the censoring status variable.

Usage
graduate

Format
A data frame with 1000 rows and 3 variables:

years years until graduation

censor censoring status indicator variable (0 = censored event time, 1 = complete event time)
gender a dichotomous variable identifying gender (1 = male, 2 = female)

Source
National Educational Longitudinal Survey (NELS) from 1988-2002

oscars Data on time until actors receive their first Academy Award nomina-
tion

Description
The dataset contains data for the top 128 grossing actors up to 2017 as listed on Box Office Mojo.
The data for the first film appearance and for the first Oscar nomination was taken from IMDb. It
should be noted that of the 128 observations in the data set, 48 were right-censored. Right-censored
observations represent actors who have not received an Oscar nomination by the year 2017 or actors
that died before 2017 without ever receiving an Oscar nomination. For the censor variable "1"
represents complete observations, actors who received an Oscar nomination by the year 2017, and
"0" represents right-censored observations.

Usage
oscars
6 parmsurvfit

Format
A data frame with 128 rows and 12 variables:
obs observation number
name name of actor
adj_gross actor’s total adjusted gross earnings (in millions)
num_movies number of movies actor received credit for
avg_gross actor’s average gross earnings per movie
top_movie title of actor’s movie with the top gross earnings
top_gross actor’s top gross earnings from a single movie
gender actor’s gender
years_until_nom number of years between actor’s first full film appearance and first Oscar nomi-
nation
censor censoring status indicator variable (0 = censored event time, 1 = complete event time)
first_film_appearance year of actor’s first full film appearance
first_oscar_nom year of actor’s first Oscar nomination

Source
https://github.com/shannonpileggi/SP--Pablo--RProgramming

parmsurvfit parmsurvfit: Fitting right censored data to parametric distributions.

Description
Executes parametric survival analysis techniques similar to those in ’Minitab’. Fits right censored
data to a given parametric distribution, produces summary statistics of the fitted distribution, and
plots parametric survival, hazard, and cumulative hazard plots. Produces Anderson-Darling test
statistic and probability plots to assess goodness of fit of right censored data to a distribution.

Details
Functions
• fit_data
• surv_summary
• surv_prob
• plot_surv
• plot_haz
• plot_cumhaz
• plot_density
plot_cumhaz 7

• plot_ppsurv
• compute_AD

Datasets

• aggressive
• firstdrink
• graduate
• oscars
• rearrest

plot_cumhaz Plotting parametric cumulative hazard curves

Description
Plots cumulative hazard curve of right censored data given that it follows a specified parametric
distribution.

Usage
plot_cumhaz(data, dist, time = "time", censor = "censor", by = "")

Arguments
data A dataframe containing a time column and a censor column.
dist A string name for a distribution that has a corresponding density function and
distribution function. Examples include "norm", "lnorm", "exp", "weibull", "lo-
gis", "llogis", "gompertz", etc.
time The string name of the time column of the dataframe. Defaults to "time".
censor The string name of the censor column of the dataframe. Defaults to "censor".
The censor column must be a numeric indicator variable where complete times
correspond to a value of 1 and incomplete times correspond to 0.
by The string name of a grouping variable. If specified, multiple lines will be plot-
ted. Variable can contain logical, string, character, or numeric data.

Examples
data("rearrest")
plot_cumhaz(rearrest, "lnorm", time = "months")
plot_cumhaz(rearrest, "weibull", time = "months", by = "personal")
8 plot_haz

plot_density Plotting density function overlayed on top of a histogram of data

Description
Creates histogram of right censored data with the density function of a fitted parametric distribution
overlayed.

Usage
plot_density(data, dist, time = "time", censor = "censor", by = "")

Arguments
data A dataframe containing a time column and a censor column.
dist A string name for a distribution that has a corresponding density function and
distribution function. Examples include "norm", "lnorm", "exp", "weibull", "lo-
gis", "llogis", "gompertz", etc.
time The string name of the time column of the dataframe. Defaults to "time".
censor The string name of the censor column of the dataframe. Defaults to "censor".
The censor column must be a numeric indicator variable where complete times
correspond to a value of 1 and incomplete times correspond to 0.
by The string name of a grouping variable. If specified, the function plots each
group individually along with the plot for all groups together. Variable can con-
tain logical, string, character, or numeric data.

Examples
data("rearrest")
plot_density(rearrest, "exp", time = "months")
plot_density(rearrest, "weibull", time = "months", by = "personal")

plot_haz Plotting parametric hazard curves

Description
Plots hazard curve of right censored data given that it follows a specified parametric distribution.

Usage
plot_haz(data, dist, time = "time", censor = "censor", by = "")
plot_ppsurv 9

Arguments
data A dataframe containing a time column and a censor column.
dist A string name for a distribution that has a corresponding density function and
distribution function. Examples include "norm", "lnorm", "exp", "weibull", "lo-
gis", "llogis", "gompertz", etc.
time The string name of the time column of the dataframe. Defaults to "time".
censor The string name of the censor column of the dataframe. Defaults to "censor".
The censor column must be a numeric indicator variable where complete times
correspond to a value of 1 and incomplete times correspond to 0.
by The string name of a grouping variable. If specified, multiple lines will be plot-
ted. Variable can contain logical, string, character, or numeric data.

Examples
data("rearrest")
plot_haz(rearrest, "logis", time = "months")
plot_haz(rearrest, "weibull", time = "months", by = "personal")

plot_ppsurv Plotting percent-percent plots for parametric fitting of data

Description
Creates percent-percent plot of right censored data given that it follows a specified parametric dis-
tribution.

Usage
plot_ppsurv(data, dist, time = "time", censor = "censor")

Arguments
data A dataframe containing a time column and a censor column.
dist A string name for a distribution that has a corresponding density function and
distribution function. Examples include "norm", "lnorm", "exp", "weibull", "lo-
gis", "llogis", "gompertz", etc.
time The string name of the time column of the dataframe. Defaults to "time".
censor The string name of the censor column of the dataframe. Defaults to "censor".
The censor column must be a numeric indicator variable where complete times
correspond to a value of 1 and incomplete times correspond to 0.

Examples
data("rearrest")
plot_ppsurv(rearrest, "weibull", time = "months")
plot_ppsurv(rearrest, "exp", time = "months")
10 rearrest

plot_surv Plotting parametric survival curves

Description
Plots survival curve of right censored data given that it follows a specified parametric distribution.

Usage
plot_surv(data, dist, time = "time", censor = "censor", by = "")

Arguments
data A dataframe containing a time column and a censor column.
dist A string name for a distribution that has a corresponding density function and
distribution function. Examples include "norm", "lnorm", "exp", "weibull", "lo-
gis", "llogis", "gompertz", etc.
time The string name of the time column of the dataframe. Defaults to "time".
censor The string name of the censor column of the dataframe. Defaults to "censor".
The censor column must be a numeric indicator variable where complete times
correspond to a value of 1 and incomplete times correspond to 0.
by The string name of a grouping variable. If specified, multiple lines will be plot-
ted. Variable can contain logical, string, character, or numeric data.

Examples
data("rearrest")
plot_surv(rearrest, "lnorm", time = "months")
plot_surv(rearrest, "weibull", time = "months", by = "personal")

rearrest Data on time until re-incarceration for 194 inmates.

Description
Henning and Frueh (1996) followed criminal activities of 194 inmates released from a medium
security prison for 36 months. The data from this study can be used to investigate the time until
the former inmates were re-arrested. If the former inmate had been re-arrested for a criminal act
before 36 months (after initial prison release) had passed, then that former inmate’s event time was
complete. If the former inmate had not been re-arrested for a criminal act after 36 months had
passed, or had completely dropped out of the study, then that former inmate’s event time was right
censored.
surv_prob 11

Usage
rearrest

Format
A data frame with 194 rows and 5 variables:

months months until re-arrest

censor censoring status indicator variable (0 = censored event time, 1 = complete event time)
personal a dichotomous variable identifying former inmates who had a history of person-related
crimes (1 = personal), i.e. those with one or more convictions for offenses such as aggravated
assault or kidnapping
property a dichotomous variable indicating whether former inmates were convicted of a property-
related crime (1 = property)
cenage the "centered" age of individual, i.e. the difference between the age of the individual upon
release and the average age of all inmates in the study.

Source
https://stats.idre.ucla.edu/other/examples/alda/

surv_prob Survival probability based on parametric distribution

Description
Computes probability of survival beyond time t given that the data follows a specified parametric
distribution.

Usage
surv_prob(data, dist, x, lower.tail = F, time = "time",
censor = "censor", by = "")

Arguments
data A dataframe containing a time column and a censor column.
dist A string name for a distribution that has a corresponding density function and
a distribution function. Examples include "norm", "lnorm", "exp", "weibull",
"logis", "llogis", "gompertz", etc.
x A scalar quantity, time at which the probability of survival is computed
lower.tail Logical; if F (default), probability is P(T > x), otherwise, P(T < x).
time The string name of the time column of the dataframe. Defaults to "time".
12 surv_summary

censor The string name of the censor column of the dataframe. Defaults to "censor".
The censor column must be a numeric indicator variable where complete times
correspond to a value of 1 and incomplete times correspond to 0.
by The string name of a grouping variable. If specified, the function prints prob-
ability for each group individually along with the overall probability. Variable
can contain logical, string, character, or numeric data.

Examples
data("rearrest")
surv_prob(rearrest, "lnorm", 110, time = "months")
surv_prob(rearrest, "weibull", 90, time = "months", lower.tail = TRUE)

surv_summary Summary statistics based on parametric distribution

Description
Estimates various statistics, including median, mean, standard deviation, and percentiles of survival
time given that the data follows a specified parametric distribution.

Usage
surv_summary(data, dist, time = "time", censor = "censor", by = "")

Arguments
data A dataframe containing a time column and a censor column.
dist A string name for a distribution that has a corresponding density function and
a distribution function. Examples include "norm", "lnorm", "exp", "weibull",
"logis", "llogis", "gompertz", etc.
time The string name of the time column of the dataframe. Defaults to "time".
censor The string name of the censor column of the dataframe. Defaults to "censor".
The censor column must be a numeric indicator variable where complete times
correspond to a value of 1 and incomplete times correspond to 0.
by The string name of a grouping variable. If specified, returns summary statistics
for each group. Variable can contain logical, string, character, or numeric data.

Examples
data("rearrest")
surv_summary(rearrest, "lnorm", time = "months")
surv_summary(rearrest, "weibull", time = "months", by = "personal")
Index

∗Topic datasets
aggressive, 2
firstdrink, 3
graduate, 5
oscars, 5
rearrest, 10

aggressive, 2, 7

compute_AD, 3, 7

firstdrink, 3, 7
fit_data, 4, 6
fitdistcens, 4

graduate, 5, 7

oscars, 5, 7

parmsurvfit, 6
parmsurvfit-package (parmsurvfit), 6
plot_cumhaz, 6, 7
plot_density, 6, 8
plot_haz, 6, 8
plot_ppsurv, 7, 9
plot_surv, 6, 10

rearrest, 7, 10

surv_prob, 6, 11
surv_summary, 6, 12

13