Package {SEQTaRget}

Type:	Package
Title:	Sequential Trial Emulation
Version:	1.4.4
Description:	Implementation of sequential trial emulation for the analysis of observational databases. The 'SEQTaRget' software accommodates time-varying treatments and confounders, as well as binary and failure time outcomes. 'SEQTaRget' allows to compare both static and dynamic strategies, can be used to estimate observational analogs of intention-to-treat and per-protocol effects, and can adjust for potential selection bias induced by losses-to-follow-up. (Paper to come).
License:	MIT + file LICENSE
Encoding:	UTF-8
LazyData:	true
Suggests:	rmarkdown, testthat (≥ 3.0.0)
Imports:	data.table, parglm, doFuture, doRNG, fastglm, future, future.apply, ggplot2, knitr, methods, stringr, survival, parallelly, utils
Config/roxygen2/markdown:	TRUE
Config/roxygen2/version:	8.1.0
Config/testthat/edition:	3
Depends:	R (≥ 4.1)
URL:	https://causalinference.github.io/SEQTaRget/, https://github.com/CausalInference/SEQTaRget
VignetteBuilder:	knitr
NeedsCompilation:	no
Packaged:	2026-08-28 06:45:33 UTC; runner
Author:	Ryan O'Dea [aut, cre], Tom Palmer [aut], Paul Madley-Dowd [aut], Alejandro Szmulewicz [aut], Miguel A. Hernán [aut], The President and Fellows of Harvard College [cph]
Maintainer:	Ryan O'Dea <ryan.odea@psi.ch>
Repository:	CRAN
Date/Publication:	2026-08-28 07:50:02 UTC

SEQTaRget: Sequential Trial Emulation

Description

Implementation of sequential trial emulation for the analysis of observational databases. The 'SEQTaRget' software accommodates time-varying treatments and confounders, as well as binary and failure time outcomes. 'SEQTaRget' allows to compare both static and dynamic strategies, can be used to estimate observational analogs of intention-to-treat and per-protocol effects, and can adjust for potential selection bias induced by losses-to-follow-up. (Paper to come).

Author(s)

Maintainer: Ryan O'Dea ryan.odea@psi.ch (ORCID)

Authors:

• Ryan O'Dea ryan.odea@psi.ch (ORCID)

• Tom Palmer remlapmot@hotmail.com (ORCID) (ROR)

• Paul Madley-Dowd p.madley-dowd@bristol.ac.uk (ORCID)

• Alejandro Szmulewicz aszmulewicz@hsph.harvard.edu (ORCID)

• Miguel A. Hernán mhernan@hsph.harvard.edu (ORCID)

Other contributors:

• The President and Fellows of Harvard College (ROR) [copyright holder]

See Also

Useful links:

• https://causalinference.github.io/SEQTaRget/

• https://github.com/CausalInference/SEQTaRget

Print a coefficient table from a cleaned fastglm object

Description

Print a coefficient table from a cleaned fastglm object

Usage

.coef_table(model)

Arguments

Function to return the internal data from a SEQuential object

Description

Function to return the internal data from a SEQuential object

Usage

SEQ_data(object)

Arguments

Value

data.table

Simulated observational example data for SEQuential

Description

Simulated observational example data for SEQuential()

Usage

SEQdata

Format

A data frame with 12,180 rows and 11 columns:

ID

Integer: Unique ID emulating individual patients

time

Integer: Time of observation, always begins at 0, max time of 59. Should be continuous

eligible

Binary: eligibility criteria for timepoints

outcome

Binary: If an outcome is observed at this time point

tx_init

Binary: If treatment is observed at this time point

sex

Binary: Sex of the emulated patient

N

Numeric: Normal random variable from N(10,5)

L

Numeric: 4% continuously increase from U(0, 1)

P

Numeric: 2% continuously decrease from U(9, 10)

excusedOne

Binary: Once one, always one variable emulating an excuse for treatment switch

excusedZero

Binary: Once one, always one variable emulating an excuse for treatment switch

Simulated lost-to-followup example data for SEQuential()

Description

Simulated lost-to-followup example data for SEQuential()

Usage

SEQdata.LTFU

Format

A dataframe with 54,687 rows and 13 columns:

ID

Integer: Unique ID emulating individual patients

time

Integer: Time of observation, always begins at 0, max time of 59; however, if lost-to-followup, time is truncated at a random point

eligible

Binary: eligibility criteria for timepoints

outcome

Binary: If an outcome is observed at this time point

tx_init

Binary: If treatment is observed at this time point

sex

Binary: Sex of the emulated patient

N

Numeric: Normal random variable from N(10,5)

L

Numeric: 4% continuously increase from U(0, 1)

P

Numeric: 2% continuously decrease from U(9, 10)

excusedOne

Binary: Once one, always one variable emulating an excuse for treatment switch

excusedZero

Binary: Once one, always one variable emulating an excuse for treatment switch

LTFU

Binary: Flag for losing a simulated ID to followup, if 1 there are no more records of the ID afterwards

eligible_cense

Binary: emulates columns which are eligible to entering into censoring models (e.g. if you want to limit columns for the LTFU model)

Simulated multitreatment example data for SEQuential() multinomial models

Description

Simulated multitreatment example data for SEQuential() multinomial models

Usage

SEQdata.multitreatment

Format

A dataframe with 5,976 rows and 11 columns:

ID

Integer: Unique ID emulating individual patients

time

Integer: Time of observation, always begins at 0, max time of 59; however, if lost-to-followup, time is truncated at a random point

eligible

Binary: eligibility criteria for timepoints

outcome

Binary: If an outcome is observed at this time point

tx_init

Integer: Which treatment is observed at this time point

sex

Binary: Sex of the emulated patient

N

Numeric: Normal random variable from N(10,5)

L

Numeric: 4% continuously increase from U(0, 1)

P

Numeric: 2% continuously decrease from U(9, 10)

excusedOne

Binary: Once one, always one variable emulating an excuse for treatment switch

excusedZero

Binary: Once one, always one variable emulating an excuse for treatment switch

Estimate the (very rough) time to run SEQuential analysis on current machine

Description

Estimate the (very rough) time to run SEQuential analysis on current machine

Usage

SEQestimate(
  data,
  id.col,
  time.col,
  eligible.col,
  treatment.col,
  outcome.col,
  time_varying.cols = list(),
  fixed.cols = list(),
  method,
  options,
  verbose = TRUE
)

Arguments

Details

This is a rough heuristic only. Empirically its predictions land within roughly a factor of two of observed run time, and the direction of the error depends on the configuration (e.g. it can over- or under-estimate as followup.max changes), so the result should be treated as an order-of-magnitude guide rather than a reliable timing.

Value

A list of (very rough) estimates for the time required for SEQuential containing:

• modelTime estimated time used when running models

• expansionTime estimated time used when expanding data

• totalTime sum of model and expansion time

Creates an expanded dataset for use with SEQuential()

Description

Creates an expanded dataset for use with SEQuential()

Usage

SEQexpand(params)

Arguments

Parameter Builder for SEQuential Model and Estimates

Description

Parameter Builder for SEQuential Model and Estimates

Usage

SEQopts(
  bootstrap = FALSE,
  bootstrap.nboot = 100,
  bootstrap.sample = 0.8,
  bootstrap.CI = 0.95,
  bootstrap.CI_method = "se",
  cense = NA,
  cense.denominator = NA,
  cense.eligible = NA,
  cense.numerator = NA,
  compevent = NA,
  covariates = NA,
  data.return = FALSE,
  denominator = NA,
  deviation = FALSE,
  deviation.col = NA,
  deviation.conditions = c(NA, NA),
  deviation.excused = FALSE,
  deviation.excused_cols = c(NA, NA),
  end_of_fup = FALSE,
  end_of_fup.time = NA,
  end_of_fup.type = "binary",
  end_of_fup.window = 0,
  excused = FALSE,
  excused.cols = c(NA, NA),
  expand.only = FALSE,
  fastglm.method = 2L,
  followup.class = FALSE,
  followup.include = TRUE,
  followup.max = Inf,
  followup.min = 0,
  followup.spline = FALSE,
  followup.spline.df = 4L,
  glm.package = "fastglm",
  hazard = FALSE,
  indicator.baseline = "_bas",
  indicator.squared = "_sq",
  km.curves = FALSE,
  multinomial = FALSE,
  ncores = availableCores(omit = 1L),
  nthreads = getDTthreads(),
  numerator = NA,
  parallel = FALSE,
  parglm.control = NULL,
  plot.colors = c("#F8766D", "#00BFC4", "#555555"),
  plot.labels = NA,
  plot.subtitle = NA,
  plot.title = NA,
  plot.type = "survival",
  risk.times = NA,
  seed = NULL,
  selection.first_trial = FALSE,
  selection.prob = 0.8,
  selection.random = FALSE,
  subgroup = NA,
  survival.max = Inf,
  treat.level = c(0, 1),
  trial.include = TRUE,
  visit = NA,
  visit.denominator = NA,
  visit.numerator = NA,
  weight.eligible_cols = c(),
  weight.lower = 0,
  weight.lag_condition = TRUE,
  weight.p99 = FALSE,
  weight.preexpansion = TRUE,
  weight.upper = Inf,
  weighted = FALSE
)

Arguments

Value

An object of class 'SEQopts'

An S4 class of user options to feed into the SEQuential processes and estimates This class should match SEQopts in file SEQopts.R

Description

An S4 class of user options to feed into the SEQuential processes and estimates This class should match SEQopts in file SEQopts.R

An S4 class used to hold the outputs for the SEQuential process

Description

An S4 class used to hold the outputs for the SEQuential process

Slots

params

SEQparams object

outcome

outcome covariates

numerator

numerator covariates

denominator

denominator covariates

outcome.model

list of length bootstrap.nboot containing outcome coefficients

hazard

hazard ratio

survival.curve

ggplot object for the survival curves

survival.data

data.table of survival data

risk.difference

risk difference calculated from survival data

risk.ratio

risk ratio calculated from survival data

time

time used for the SEQuential process

eof.data

end-of-follow-up estimates per treatment arm when end_of_fup = TRUE, empty otherwise. See end_of_fup()

eof.comparison

pairwise between-arm contrasts of the end-of-follow-up estimates when end_of_fup = TRUE, empty otherwise

weight.statistics

information from the weighting process, containing weight coefficients and weight statistics

info

list of diagnostic tables (outcome, follow-up, switch, and competing-event counts where applicable), each split by baseline treatment arm. The "unique" tables count distinct subjects; the "non-unique" tables count rows: total outcome events for the outcome tables, and total person-time intervals for the follow-up tables. The outcome tables are NA for a continuous end-of-follow-up outcome, which has no events to count. See diagnostics().

ce.model

list of competing event models if compevent is specified, NA otherwise

An internal S4 class to carry around parameters during the SEQuential process - inherits user facing parameters from SEQopts

Description

An internal S4 class to carry around parameters during the SEQuential process - inherits user facing parameters from SEQopts

Slots

data

pre expansion data

DT

post expansion data

id

id column as defined by the user

time

time column as defined by the user

eligible

eligible column as defined by the user

treatment

treatment column as defined by the user

time_varying

list of time varying columns as defined by the user

fixed

list of fixed columns as defined by the user

method

method of analysis as defined by the user

SEQuential trial emulation

Description

SEQuential is an all-in-one API to SEQuential analysis, returning a SEQoutput object of results. More specific examples can be found on pages at https://causalinference.github.io/SEQTaRget/

Usage

SEQuential(
  data,
  id.col,
  time.col,
  eligible.col,
  treatment.col,
  outcome.col,
  time_varying.cols = list(),
  fixed.cols = list(),
  method,
  options,
  verbose = TRUE
)

Arguments

Details

Implementation of sequential trial emulation for the analysis of observational databases. The SEQuential software accommodates time-varying treatments and confounders, as well as binary and failure time outcomes. SEQuential allows to compare both static and dynamic strategies, can be used to estimate observational analogs of intention-to-treat and per-protocol effects, and can adjust for potential selection bias induced by losses-to-follow-up.

Value

An S4 object of class SEQoutput. If options = SEQopts(expand.only = TRUE), returns the expanded data.table directly, with analysis steps skipped.

Examples

data <- SEQdata

# Intention-to-treat (ITT) effect: subjects are assigned to the treatment
# arm defined by their baseline treatment and followed regardless of any later
# treatment changes, so no weighting is required.
SEQuential(data, id.col = "ID",
           time.col = "time",
           eligible.col = "eligible",
           treatment.col = "tx_init",
           outcome.col = "outcome",
           time_varying.cols = c("N", "L", "P"),
           fixed.cols = "sex",
           method = "ITT",
           options = SEQopts())

# Per-protocol effect via artificial censoring: subjects are censored when they
# deviate from their assigned strategy, and inverse-probability-of-censoring
# weights adjust for the resulting selection bias. The denominator models the
# probability of remaining uncensored given the time-varying confounders, while
# the numerator uses only the baseline covariates to stabilize the weights (so
# the two formulas must differ - identical formulas give weights of 1).
SEQuential(data, id.col = "ID",
           time.col = "time",
           eligible.col = "eligible",
           treatment.col = "tx_init",
           outcome.col = "outcome",
           time_varying.cols = c("N", "L", "P"),
           fixed.cols = "sex",
           method = "censoring",
           options = SEQopts(weighted = TRUE,
                             numerator = "sex",
                             denominator = "N + L + P + sex"))

# End-of-follow-up outcome: rather than a time-to-event, the outcome is read
# once per trial-period at follow-up time 12 and averaged within each baseline
# arm, weighted by the weight at that time. Trial-periods with no measurement
# at exactly 12 fall back to the nearest one within 12 +/- 3; any with none
# in that window are censored out of the average. Use end_of_fup.type =
# "continuous" for a continuous outcome, which is reported as a mean.
eof <- SEQuential(data, id.col = "ID",
                  time.col = "time",
                  eligible.col = "eligible",
                  treatment.col = "tx_init",
                  outcome.col = "outcome",
                  time_varying.cols = c("N", "L", "P"),
                  fixed.cols = "sex",
                  method = "ITT",
                  options = SEQopts(end_of_fup = TRUE,
                                    end_of_fup.time = 12,
                                    end_of_fup.type = "binary",
                                    end_of_fup.window = 3,
                                    bootstrap = TRUE,
                                    bootstrap.nboot = 5))
end_of_fup(eof)

An internal S4 class to help transfer weight statistics out of internal_weights

Description

An internal S4 class to help transfer weight statistics out of internal_weights

Slots

weights

a data.table containing the estimated weights, either pre or post expansion

coef.n0

numerator zero model

coef.n1

numerator one model

coef.d0

denominator zero model

coef.d1

denominator one model

coef.ncense

numerator censoring model

coef.dcense

denominator censoring model

Bake fixed knots into any ns(followup, df = N) term in params@covariates

Description

splines::ns() recomputes knots from whatever data it sees, so without fixed knots the basis at prediction time differs from the basis used at fit time. This helper rewrites every ns(followup, df = N) token in params@covariates to an explicit ns(followup, knots = c(...), Boundary.knots = c(...)) computed once from the full expanded followup column. The result is a formula whose model.matrix output is invariant to the row subset passed in.

Usage

bake_followup_spline(params)

Details

Returns the original covariates string unchanged when no ns(followup, df = ...) token is present (e.g. user supplied custom covariates that already specify knots).

Build the function that gives each bootstrap copy of a subject a unique ID

Description

Numeric IDs are relabeled arithmetically (orig_id * multiplier + copy index), but only while every relabeled value stays within the 2^53 exact-integer range of doubles and the IDs are non-negative integers: beyond 2^53 consecutive copy indices round to the same double, so distinct bootstrap copies of a subject silently merge under by-ID grouping (e.g. 10-digit IDs). All other cases relabel by string concatenation (orig_id_b<copy index>).

Usage

bootstrap_id_relabeler(UIDs, n_sample)

Arguments

Value

a function(id_col, idx) returning unique relabeled IDs

Columns referenced by a (possibly per-treatment-level) cache entry

Description

Columns referenced by a (possibly per-treatment-level) cache entry

Usage

cached_cols(cached)

Check a fitted fastglm model for signs of perfect or quasi-complete separation

Description

Issues a warning when separation is detected. Because fastglm uses IWLS and stops at its iteration limit rather than diverging to Inf, separation is identified by two complementary signals:

1. Any coefficient with |coef| > 25 (logit > 25 implies P > 1 - 1e-11, unreachable without separation).

2. Non-finite coefficients (Inf/-Inf/NaN), which can occur with other solvers.

Usage

check_separation(model, label = "logistic regression")

Arguments

Strip large components from a model object returned by fit_glm

Description

Strip large components from a model object returned by fit_glm

Usage

clean_fastglm(model)

Arguments

Strip fitted-model bulk from a bootstrap iteration's result

Description

No-op in end_of_fup mode, which fits no outcome model.

Usage

clean_models(out, params)

Function to return competing event models from a SEQuential object

Description

Function to return competing event models from a SEQuential object

Usage

compevent(object)

Arguments

Value

A fastglm object, or a named list of fastglm objects when subgroups are specified

Retrieves Outcome, Numerator, and Denominator Covariates

Description

Retrieves Outcome, Numerator, and Denominator Covariates

Usage

covariates(object)

Arguments

Value

List of SEQuential covariates

Internal Function to create 'default' loss-to-followup formula

Description

Internal Function to create 'default' loss-to-followup formula

Usage

create.default.LTFU.covariates(params, type)

Internal Function to create 'default' formula

Description

Internal Function to create 'default' formula

Usage

create.default.covariates(params)

Internal Function to create 'default' weighting formula

Description

Internal Function to create 'default' weighting formula

Usage

create.default.weight.covariates(params, type)

Assemble end-of-follow-up estimates and bootstrap confidence intervals

Description

Mirrors create.risk(): eof.data holds the per-arm estimate and eof.comparison the pairwise between-arm contrasts, both with bootstrap confidence intervals when bootstrap = TRUE. Contrasts are paired by bootstrap iteration, so the interval accounts for the correlation between arms.

Usage

create.endoffup(full, boots, params)

Arguments

Value

list with eof.data and eof.comparison data.tables

Internal function to pull Risk Ratio and Risk Difference from data when km.curves = TRUE

Description

Internal function to pull Risk Ratio and Risk Difference from data when km.curves = TRUE

Usage

create.risk(data, params, boot_risks = NULL)

Retrieves Denominator Models from SEQuential object

Description

Retrieves Denominator Models from SEQuential object

Usage

denominator(object)

Arguments

Value

List of both denominator models

Function to return diagnostic tables from a SEQuential object

Description

Function to return diagnostic tables from a SEQuential object

Usage

diagnostics(object)

Arguments

Value

A named list of diagnostic tables, each broken down by baseline treatment arm. The "unique" and "non-unique" variants count different things:

• outcome.unique / outcome.nonunique: distinct subjects who had the outcome vs. the total number of outcome events. These coincide for a one-time (terminal) outcome, since each subject contributes at most one event row. Both are NA for a continuous end-of-follow-up outcome, which has no events to count.

• eof.unique / eof.nonunique: present only when end_of_fup = TRUE, accounting for every trial-period at the end-of-follow-up time across four mutually exclusive categories - measured At k, measured In window, Excluded (outside window) and Excluded (no measurement) - against the Eligible total. The non-unique (trial-period) counts partition Eligible, and At k plus In window equals the trial-periods contributing to the estimate. The unique (subject) counts need not sum to Eligible, since one subject can fall into different categories for different trials. For a continuous outcome eof.summary additionally reports the N, mean and SD of the raw analysed measurements per arm, standing in for the suppressed outcome count tables.

• followup.unique / followup.nonunique: distinct subjects contributing follow-up vs. the total number of person-time intervals (expanded rows). The non-unique count is much larger because each subject contributes one row per follow-up period; it is the denominator that, with outcome.nonunique, gives the per-arm event rate.

Extract the end-of-follow-up outcome estimates

Description

Available when SEQuential() was run with end_of_fup = TRUE. The estimate in each arm is the weighted average of the outcome measured at end_of_fup.time, weighted by the period-trial-specific weight at the time the measurement was taken.

Usage

end_of_fup(object)

Arguments

Value

A named list, one element per subgroup, each a list of two data.tables:

• estimates: the weighted proportion (binary) or mean (continuous) in each baseline treatment arm, with its bootstrap confidence interval, the trial-periods eligible in each arm, partitioned into those analysed, those censored for want of a measurement in the window - measured at some point but not within [k - window, k + window] - and those never measured at all, with the censoring also as a percentage of the eligible total. Subjects counts the distinct contributing subjects.

• comparison: the pairwise between-arm difference - in proportions for a binary outcome, in means for a continuous one - with its bootstrap standard error and confidence interval, paired by iteration so the interval accounts for the correlation between arms. For a binary outcome the ratio of proportions is also given, with a confidence interval computed on the log scale and a log(Ratio) SE for inverse-variance pooling. A ratio is not reported for a continuous outcome, where the outcome need not be bounded away from zero.

Account for every trial-period at the end-of-follow-up time

Description

Classifies each trial-period in the analysis data into exactly one of four mutually exclusive categories, so that the trial-periods contributing to the estimate can be reconciled against those excluded:

• measured at k - contributes, using the measurement at exactly end_of_fup.time;

• measured in the window - contributes, having no measurement at k but one within [k - window, k + window];

• excluded, outside the window - has a measurement somewhere, but none within the window;

• excluded, no measurement - has no non-missing outcome at any follow-up time. Under method = "censoring" this includes trial-periods artificially censored before any measurement was taken.

Usage

endoffup.counts(DT, params, type)

Arguments

Details

Counted over the same data the estimate is computed from, so the first two categories always sum to the contributing trial-periods reported by end_of_fup().

Value

named list of data.tables, one element per subgroup, each with a row per baseline treatment arm

Weighted end-of-follow-up average within each treatment arm

Description

Weight truncation (weight.lower / weight.upper, including the bounds weight.p99 resolves to) is applied here as it is for the outcome model, since this average is the estimator in end_of_fup mode.

Usage

endoffup.estimate(DT, params)

Arguments

Details

Alongside the estimate this counts the trial-periods censored for want of a measurement in the window - those measured at some point but not within [k - window, k + window] - so that share can be reported next to the estimate they were dropped from. Trial-periods never measured at all are counted separately rather than folded in, so the analysed, censored and never-measured counts partition the eligible total. Trial-periods are the unit because one subject can be analysed in one trial and censored in another.

Value

named list of per-arm data.tables, one element per subgroup

Select the end-of-follow-up outcome measurement for each trial-period

Description

An end-of-follow-up outcome is measured once, at follow-up time end_of_fup.time (k), rather than as a time-to-event. For each (id, trial) this returns the single row the estimate is read from: the measurement at exactly k when one exists, otherwise - if end_of_fup.window is non-zero - the measurement nearest to k within [k - window, k + window], with ties (measurements equally far either side of k) broken toward the later one, so that at least k of follow-up has elapsed. Trial-periods with no measurement anywhere in the window contribute no row, i.e. they are censored out of the estimate.

Usage

endoffup.measure(DT, params)

Arguments

Details

Rows carrying a missing outcome are not measurements: under method = "censoring" these are the artificially censored (treatment switch) rows, so a subject who deviates before k is correctly excluded rather than contributing a carried-forward value.

Value

data.table with one row per contributing (id, trial): the outcome value, the weight at that time, the baseline treatment arm, and the follow-up time the measurement was taken at

Per-arm summary of the analysed end-of-follow-up measurements

Description

N, mean and SD of the raw selected measurements per baseline arm - the unweighted analogue of the outcome count tables, reported for continuous outcomes where event counts have no meaning.

Usage

endoffup.summary(DT, params)

Arguments

Value

named list of data.tables, one element per subgroup

Fit a GLM using the package specified in params@glm.package

Description

Fit a GLM using the package specified in params@glm.package

Usage

fit_glm(X, y, family, weights = NULL, params, start = NULL)

Arguments

Count follow-up per treatment arm

Description

Summarises follow-up in each treatment arm, grouped by the baseline treatment value, restricted to expanded-data rows with an observed (non-NA) outcome - the person-time the outcome/survival model is fit on. Rows censored under method = "censoring" (which carry outcome = NA) are excluded, matching the analysis.

Usage

followup.table(params, type, filter = NA)

Arguments

Details

type = "nonunique" counts follow-up intervals (rows), i.e. total person-time; dividing the non-unique outcome counts by these gives the per-arm event rate. type = "unique" counts the distinct subjects contributing follow-up to the arm (a subject contributing several trials to the same arm is counted once; a subject who appears in both arms is counted in each).

Value

data.table with the baseline-treatment column and n (count)

Nicely cleans time for readability

Description

Nicely cleans time for readability

Usage

format_time(seconds)

Extract underlying column names from RHS formula strings

Description

Wraps all.vars(as.formula(...)) so that function-wrapped terms in formula strings (e.g. ns(followup, df = 4), I(x^2), factor(grp)) resolve to their underlying variable names rather than being treated as raw column names.

Usage

formula_vars(x)

Arguments

Value

character vector of unique variable names referenced by x.

Function to return hazard ratios from a SEQuential object

Description

Function to return hazard ratios from a SEQuential object

Usage

hazard_ratio(object)

Arguments

Value

A named vector of hazard ratios, or a named list of vectors when subgroups are specified

Create formula cache

Description

Create formula cache

Usage

init_formula_cache(params)

Arguments

Helper Function to inline predict a fastglm object

Description

Helper Function to inline predict a fastglm object

Usage

inline.pred(
  model,
  newdata,
  params,
  type = NULL,
  case = "default",
  multi = FALSE,
  target = NULL,
  cache = NULL
)

Arguments

Internal analysis tool for handling parallelization/bootstrapping on multiple OS types

Description

Internal analysis tool for handling parallelization/bootstrapping on multiple OS types

Usage

internal.analysis(params)

Generic function to format a dataset for hazard ratio calculation

Description

Generic function to format a dataset for hazard ratio calculation

Usage

internal.hazard(model, params, cache)

Internal function for fitting outcome models

Description

Internal function for fitting outcome models

Usage

internal.model(data, params, start = NULL)

Plotting for survival curves

Description

Plotting for survival curves

Usage

internal.plot(survival.data, params)

Arguments

Internal function for creating survival curves

Description

Internal function for creating survival curves

Usage

internal.survival(params, outcome)

Internal function for defining weights

Description

Internal function for defining weights

Usage

internal.weights(DT, data, params, cache)

Arguments

Function to print Kaplan-Meier curves

Description

Function to print Kaplan-Meier curves

Usage

km_curve(
  object,
  plot.type = "survival",
  plot.title,
  plot.subtitle,
  plot.labels,
  plot.colors
)

Arguments

Value

ggplot object of plot plot.type

Function to return survival data from a SEQuential object

Description

Function to return survival data from a SEQuential object

Usage

km_data(object)

Arguments

Value

A data frame of survival values, or a named list of data frames when subgroups are specified

Helper function for nested logistic

Description

Helper function for nested logistic

Usage

multinomial(X, y, family = quasibinomial(), params)

Helper to predict from the nested logistic

Description

Helper to predict from the nested logistic

Usage

multinomial.predict(model, X, target = NULL)

Helper function to get the summary table from multinomial

Description

Helper function to get the summary table from multinomial

Usage

multinomial.summary(model)

Retrieves Numerator Models from SEQuential object

Description

Retrieves Numerator Models from SEQuential object

Usage

numerator(object)

Arguments

Value

List of both numerator models

Retrieves Outcome Models from SEQuential object

Description

Retrieves Outcome Models from SEQuential object

Usage

outcome(object)

Arguments

Value

List of all outcome models

Parameter Helper

Description

Parameter Helper

Usage

parameter.setter(
  data,
  DT,
  id.col,
  time.col,
  eligible.col,
  outcome.col,
  treatment.col,
  time_varying.cols,
  fixed.cols,
  method,
  opts,
  verbose
)

Simplifies parameters down for later use

Description

Simplifies parameters down for later use

Usage

parameter.simplifier(params)

Predict from a model fitted by fit_glm

Description

Predict from a model fitted by fit_glm

Usage

predict_model(model, X, type = "response")

Arguments

Helper function to prepare data for fastglm

Description

Helper function to prepare data for fastglm

Usage

prepare.data_cached(weight, params, type, model, case, cache)

Arguments

Output constructor

Description

Output constructor

Usage

prepare.output(
  params,
  WDT,
  outcome,
  weights,
  hazard,
  survival.data,
  survival.ce,
  risk,
  runtime,
  info,
  eof.data = list(),
  eof.comparison = list()
)

Resolve the follow-up times at which to report risk difference / ratio

Description

Each requested time in risk.times is snapped to the latest available follow-up value at or before it (cumulative incidence is a right-continuous step function). The maximum follow-up time is always included. A NA risk.times reports only the maximum follow-up time.

Usage

resolve_risk_times(grid, risk.times)

Arguments

Value

sorted unique numeric vector of grid follow-up values to report at

Function to return risk information from a SEQuential object

Description

Function to return risk information from a SEQuential object

Usage

risk_comparison(object)

Arguments

Value

A data frame of risk comparison information at the reported follow-up time(s): risk ratios and risk differences, and when bootstrapped their confidence intervals and standard errors. The bootstrap standard errors are reported regardless of bootstrap.CI_method: ⁠RD SE⁠ is the standard error of the risk difference (natural scale) and ⁠log(RR) SE⁠ is the standard error of the log risk ratio. For an inverse-variance-weighted meta-analysis across samples, pool ⁠Risk Difference⁠ with ⁠RD SE⁠, and ⁠log(⁠Risk Ratio⁠)⁠ with ⁠log(RR) SE⁠ (then exponentiate the pooled ratio).

Function to return risk information from a SEQuential object

Description

Function to return risk information from a SEQuential object

Usage

risk_data(object)

Arguments

Value

A data table of risk information at the end of followup

Select the cached weight-model formula for a treatment level

Description

A cache entry made by init_formula_cache() is either a single parsed formula (a list with a $formula element, shared across treatment levels) or an unnamed list of parsed formulas, one per params@treat.level. Returns the entry for level, or the shared entry (or NULL) unchanged.

Usage

select_cached_level(cached, level, params)

Show method for S4 object - SEQoutput.

Description

Show method for S4 object - SEQoutput.

Usage

## S4 method for signature 'SEQoutput'
show(object)

Arguments

Value

No return value, sends information about SEQoutput to the console