find_scores {nphRCT}

Calculate scores

Description

Weighted log-rank tests can also be thought in terms of assigning a score to the each of the events (including censoring) and comparing the average score on each arm, see Magirr (2021) doi:10.1002/pst.2091. This function calculates the scores for different types of weighted log-rank test, the modestly-weighted log-rank test and the Fleming-Harrington (\rho,\gamma) test, in addition to the standard log-rank test.

Usage

find_scores(
  formula,
  data,
  method,
  t_star = NULL,
  s_star = NULL,
  rho = NULL,
  gamma = NULL,
  tau = NULL
)

Arguments

formula

Formula object. The response (on the left of the ~ operator) must be a survival object as returned by the Surv function. The terms (on the right of the ~ operator) must include the treatment arm indicator, and additionally can include strata using the strata function.

data

Data frame containing time-to-event data.

method

Character string specifying type of method to calculate scores. Either one of the weighted log-rank tests (log-rank "lr", Fleming-Harrington "fh", modestly weighted "mw") or pseudovalue-based scores (restricted mean survival time "rmst", milestone analysis "ms")

t_star

Parameter t^* in the modestly weighted ("mw") test, see Details.

s_star

Parameter s^* in the modestly weighted ("mw") test, see Details.

rho

Parameter \rho in the Fleming-Harrington ("fh") test, see Details.

gamma

Parameter \gamma in the Fleming-Harrington ("fh") test, see Details.

tau

Parameter \tau in the RMST ("rmst") or milestone analysis ("ms") test.

Details

Select which of the tests to perform using argument method. For the weighted log-rank tests, the output is calculated as outlined in vignette("weighted_log_rank_tests", package="nphRCT").

Value

Data frame. Each row corresponds to an event or censoring time. At each time specified in t_j the columns indicate

• event the event indicator

• group the treatment arm indicator

• score the assigned score at time t_j

• standardized_score the value of score standardized to be between -1 and 1

References

Magirr, D. (2021). Non-proportional hazards in immuno-oncology: Is an old perspective needed?. Pharmaceutical Statistics, 20(3), 512-527. doi:10.1002/pst.2091

Magirr, D. and Burman, C.F., 2019. Modestly weighted logrank tests. Statistics in medicine, 38(20), 3782-3790.

Examples

library(nphRCT)
set.seed(1)
sim_data <- sim_events_delay(
  event_model=list(
    duration_c = 36,
    duration_e = c(6,30),
    lambda_c = log(2)/9,
    lambda_e = c(log(2)/9,log(2)/18)
  ),
  recruitment_model=list(
    rec_model="power",
    rec_period = 12,
    rec_power = 1
  ),
  n_c=50,
  n_e=50,
  max_cal_t = 36
)
df_scores<-find_scores(formula=Surv(event_time,event_status)~group,
  data=sim_data,
  method="mw",
  t_star = 4
)
plot(df_scores)