vendredi 9 juillet 2021

Extracting specific rows from long format dataset by conditions / adopting long format data to survival analysis

Background:

I have a dataset that I am preparing for a survival analysis, it's originally a longitudinal dataset in long format. I have an ID variable separating participants, a time variable (months), and my binary 0/1 event variable (whether or not somebody met a "monthly loss limit" when gambling).

Problem/goal:

I am trying to create the necessary variables for the survival analysis and then remove the excess/unnecessary rows. My event (meeting a loss limit) can technically occur multiple times for each participant across the study period, but I am only interested in the first occurrence for a participant. I have made a time duration variable and attempted to modify it with an if-else statement so that participants that meet a loss limit have that specific month as their endpoint.

The problem is that I can't seem to do the filtering in a way that I only keep the rows that I want. I have attempted some code with an if-else statement but I am getting an error. For participants that have met one or more loss limits I want to extract the row with their first loss limit met because the modified time duration is also contained within this row. For participants that never reach a loss limit I doesn't matter, any row is fine because they all have the necessary information.

How do I accomplish this?

Example data frame and code:

library(dplyr)
# Example variables and data frame in long form
# Includes id variable, time variable and example event variable
id <- c(1, 1, 1, 1, 2, 2, 2, 3, 3, 3, 3, 3 )
time <- c(2, 3, 4, 7, 3, 5, 7, 1, 2, 3, 4, 5)
metLimit <- c(0, 0, 0, 0, 0, 1, 1, 0, 1, 0, 0, 1)

dfLong <- data.frame(id = id, time = time, metLimit = metLimit)

# Making variables, time at start, finish and duration variable 
dfLong <- dfLong %>% 
  group_by(id) %>% 
  mutate(startTime = first(time),
         lastTime = last(time))
dfLong <- dfLong %>% 
  group_by(id) %>% 
  mutate(timeDuration = ifelse(metLimit == "1", c(time - startTime), 
                               lastTime - startTime))
# My failed attempt at solving the problem
dfLong <- dfLong %>% 
  group_by(id) %>% 
  ifelse(metLimit == "1", filter(first(metLimit)), filter(last(time)

Aucun commentaire:

Enregistrer un commentaire