Here, we’ll go over some examples of using dose-response. First we need to load the library before getting in to some sample use cases.
Currently, dose-response analysis through SEQuential only supports binary treatment values. Therefore; running multinomial models will lead to errors.
Dose-response With 5 bootstrap samples
options <- SEQopts(# tells SEQuential to create Kaplan-Meier curves
km.curves = TRUE,
# tells SEQuential to bootstrap
bootstrap = TRUE,
seed = 1636,
# tells SEQuential to run bootstraps 5 times
bootstrap.nboot = 5)
# use example data
data <- SEQdata
model <- 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 = "dose-response",
options = options)
#>
#> Full dataset: 12,180 observations, 11 variables
#>
#> Non-required columns provided, pruning for efficiency
#>
#> Pruned
#>
#> Original dataset (eligible subjects): 9,203 observations, 9 variables
#>
#> Expanding Data...
#>
#> Pre-filter expansion: 310,080 observations
#>
#> Expanded dataset: 248,485 observations, 15 variables
#>
#> Expansion Successful
#>
#> Final analysis dataset: 248,485 observations, 15 variables
#>
#> Moving forward with dose-response analysis
#>
#> Bootstrapping with 80% of 300 subjects (240 subjects, ~198,788 observations per resample) 5 times
#>
#> dose-response model created successfully
#>
#> Creating Survival curves
#>
#> Completed
km_curve(model, plot.type = "risk") # retrieve risk plot
risk_data(model)
#> Index: <Followup>
#> Method Followup A Risk 95% LCI 95% UCI SE
#> <char> <num> <char> <num> <num> <num> <num>
#> 1: dose-response 60 0 0.5282782 0.03851259 1.0000000 0.24988501
#> 2: dose-response 60 1 0.8949096 0.85228977 0.9375295 0.02174522
risk_comparison(model)
#> Followup A_x A_y Risk Ratio RR 95% LCI RR 95% UCI log(RR) SE
#> <num> <fctr> <fctr> <num> <num> <num> <num>
#> 1: 60 risk_0 risk_1 1.6940120 0.7612927 3.769479 0.4080876
#> 2: 60 risk_1 risk_0 0.5903146 0.2652887 1.313555 0.4080876
#> Risk Difference RD 95% LCI RD 95% UCI RD SE
#> <num> <num> <num> <num>
#> 1: 0.3666314 -0.1084919 0.8417547 0.2424143
#> 2: -0.3666314 -0.8417547 0.1084919 0.2424143Dose-response with 5 bootstrap samples and losses-to-followup
options <- SEQopts(km.curves = TRUE,
bootstrap = TRUE,
seed = 1636,
bootstrap.nboot = 5,
# tells SEQuential to expect LTFU as the censoring column
cense = "LTFU",
# tells SEQuential to treat this column as the
# censoring eligibility column
cense.eligible = "eligible_cense")
# use example data for LTFU
data <- SEQdata.LTFU
model <- 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 = "dose-response",
options = options)
#>
#> Full dataset: 54,687 observations, 13 variables
#>
#> Non-required columns provided, pruning for efficiency
#>
#> Pruned
#>
#> Original dataset (eligible subjects): 29,624 observations, 11 variables
#>
#> Expanding Data...
#>
#> Pre-filter expansion: 1,609,859 observations
#>
#> Expanded dataset: 1,119,229 observations, 17 variables
#>
#> Expansion Successful
#>
#> Final analysis dataset: 1,119,229 observations, 17 variables
#>
#> Moving forward with dose-response analysis
#>
#> Bootstrapping with 80% of 1,000 subjects (800 subjects, ~895,383 observations per resample) 5 times
#>
#> dose-response model created successfully
#>
#> Creating Survival curves
#>
#> Completed
km_curve(model, plot.type = "risk")
risk_data(model)
#> Index: <Followup>
#> Method Followup A Risk 95% LCI 95% UCI SE
#> <char> <num> <char> <num> <num> <num> <num>
#> 1: dose-response 60 0 0.007952468 0 0.04501433 0.01890946
#> 2: dose-response 60 1 0.018718195 0 0.04142307 0.01158433
risk_comparison(model)
#> Followup A_x A_y Risk Ratio RR 95% LCI RR 95% UCI log(RR) SE
#> <num> <fctr> <fctr> <num> <num> <num> <num>
#> 1: 60 risk_0 risk_1 2.3537592 0.21044019 26.326637 1.231945
#> 2: 60 risk_1 risk_0 0.4248523 0.03798434 4.751944 1.231945
#> Risk Difference RD 95% LCI RD 95% UCI RD SE
#> <num> <num> <num> <num>
#> 1: 0.01076573 -0.01819694 0.03972839 0.01477714
#> 2: -0.01076573 -0.03972839 0.01819694 0.01477714Dose-response with 5 bootstrap samples and competing events
options <- SEQopts(km.curves = TRUE,
bootstrap = TRUE,
seed = 1636,
bootstrap.nboot = 5,
# Using LTFU as our competing event
compevent = "LTFU")
data <- SEQdata.LTFU
model <- 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 = "dose-response",
options = options)
#>
#> Full dataset: 54,687 observations, 13 variables
#>
#> Non-required columns provided, pruning for efficiency
#>
#> Pruned
#>
#> Original dataset (eligible subjects): 29,624 observations, 10 variables
#>
#> Expanding Data...
#>
#> Pre-filter expansion: 1,609,859 observations
#>
#> Expanded dataset: 1,119,229 observations, 16 variables
#>
#> Expansion Successful
#>
#> Final analysis dataset: 1,119,229 observations, 16 variables
#>
#> Moving forward with dose-response analysis
#>
#> Bootstrapping with 80% of 1,000 subjects (800 subjects, ~895,383 observations per resample) 5 times
#>
#> dose-response model created successfully
#>
#> Creating Survival curves
#>
#> Completed
km_curve(model, plot.type = "risk")
risk_data(model)
#> Index: <Followup>
#> Method Followup A Risk 95% LCI 95% UCI SE
#> <char> <num> <char> <num> <num> <num> <num>
#> 1: dose-response 60 0 0.007586789 0 0.04126782 0.01718451
#> 2: dose-response 60 1 0.003641046 0 0.03315852 0.01506021
risk_comparison(model)
#> Followup A_x A_y Risk Ratio RR 95% LCI RR 95% UCI log(RR) SE
#> <num> <fctr> <fctr> <num> <num> <num> <num>
#> 1: 60 inc_0 inc_1 0.4799192 0.03776639 6.098608 1.297064
#> 2: 60 inc_1 inc_0 2.0836841 0.16397183 26.478569 1.297064
#> Risk Difference RD 95% LCI RD 95% UCI RD SE
#> <num> <num> <num> <num>
#> 1: -0.003945743 -0.03928796 0.03139647 0.01803207
#> 2: 0.003945743 -0.03139647 0.03928796 0.01803207Dose-response hazard ratio with 5 bootstrap samples and competing events
options <- SEQopts(# km.curves must be set to FALSE to turn on hazard
# ratio creation
km.curves = FALSE,
# set hazard to TRUE for hazard ratio creation
hazard = TRUE,
bootstrap = TRUE,
seed = 1636,
bootstrap.nboot = 5,
compevent = "LTFU")
data <- SEQdata.LTFU
model <- 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 = "dose-response",
options = options)
#>
#> Full dataset: 54,687 observations, 13 variables
#>
#> Non-required columns provided, pruning for efficiency
#>
#> Pruned
#>
#> Original dataset (eligible subjects): 29,624 observations, 10 variables
#>
#> Expanding Data...
#>
#> Pre-filter expansion: 1,609,859 observations
#>
#> Expanded dataset: 1,119,229 observations, 16 variables
#>
#> Expansion Successful
#>
#> Final analysis dataset: 1,119,229 observations, 16 variables
#>
#> Moving forward with dose-response analysis
#>
#> Bootstrapping with 80% of 1,000 subjects (800 subjects, ~895,383 observations per resample) 5 times
#>
#> Completed
# retrieve hazard ratios
hazard_ratio(model)
#> Hazard ratio LCI UCI
#> 1.0922586 0.9723626 1.2269383Dose-response with 5 bootstrap samples and competing events in subgroups defined by sex
options <- SEQopts(km.curves = TRUE,
bootstrap = TRUE,
seed = 1636,
bootstrap.nboot = 5,
compevent = "LTFU",
# define the subgroup
subgroup = "sex")
data <- SEQdata.LTFU
model <- 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 = "dose-response",
options = options)
#>
#> Full dataset: 54,687 observations, 13 variables
#>
#> Non-required columns provided, pruning for efficiency
#>
#> Pruned
#>
#> Original dataset (eligible subjects): 29,624 observations, 10 variables
#>
#> Expanding Data...
#>
#> Pre-filter expansion: 1,609,859 observations
#>
#> Expanded dataset: 1,119,229 observations, 16 variables
#>
#> Expansion Successful
#>
#> Final analysis dataset: 1,119,229 observations, 16 variables
#>
#> Moving forward with dose-response analysis
#>
#> Bootstrapping with 80% of 1,000 subjects (800 subjects, ~895,383 observations per resample) 5 times
#>
#> dose-response model created successfully
#>
#> Creating Survival Curves for sex_0
#>
#> Creating Survival Curves for sex_1
#>
#> Completed
km_curve(model, plot.type = "risk")
#> $sex_0
#>
#> $sex_1
risk_data(model)
#> $sex_0
#> Index: <Followup>
#> Method Followup A Risk 95% LCI 95% UCI SE
#> <char> <num> <char> <num> <num> <num> <num>
#> 1: dose-response 60 0 0.01063229 0 0.10408278 0.04767970
#> 2: dose-response 60 1 0.02718814 0 0.07167468 0.02269763
#>
#> $sex_1
#> Index: <Followup>
#> Method Followup A Risk 95% LCI 95% UCI SE
#> <char> <num> <char> <num> <num> <num> <num>
#> 1: dose-response 60 0 0.006494985 0 0.01409034 0.003875252
#> 2: dose-response 60 1 0.005925412 0 0.02610814 0.010297499
risk_comparison(model)
#> $sex_0
#> Followup A_x A_y Risk Ratio RR 95% LCI RR 95% UCI log(RR) SE
#> <num> <fctr> <fctr> <num> <num> <num> <num>
#> 1: 60 inc_0 inc_1 2.5571299 0.006840718 955.8812 3.022376
#> 2: 60 inc_1 inc_0 0.3910634 0.001046155 146.1835 3.022376
#> Risk Difference RD 95% LCI RD 95% UCI RD SE
#> <num> <num> <num> <num>
#> 1: 0.01655585 -0.06846066 0.10157237 0.04337657
#> 2: -0.01655585 -0.10157237 0.06846066 0.04337657
#>
#> $sex_1
#> Followup A_x A_y Risk Ratio RR 95% LCI RR 95% UCI log(RR) SE
#> <num> <fctr> <fctr> <num> <num> <num> <num>
#> 1: 60 inc_0 inc_1 0.9123056 0.1358191 6.128016 0.9717787
#> 2: 60 inc_1 inc_0 1.0961239 0.1631850 7.362734 0.9717787
#> Risk Difference RD 95% LCI RD 95% UCI RD SE
#> <num> <num> <num> <num>
#> 1: -0.0005695735 -0.02407448 0.02293534 0.01199252
#> 2: 0.0005695735 -0.02293534 0.02407448 0.01199252