16 Generate time-to-event data for diabetes related kidney disease

In this chapter, we phenotype right-censored time-to-event data for diabetic kidney disease. An initial time-to-event data are created, and consequent cases and controls are identified following the identical procedure described in chapter 14. The initial time-to-event data are refined by applying the following exclusion criteria:

• case
  • has a complication event after the censoring date
  • has a prior complication event identified before a first evidence of diabetes
  • has an initiation date that is not between the first evidence of diabetes and the date of complication event
  • has less than 5 years of follow-up time since the first evidence of diabetes before a kidney disease event
• control
  • has some event represented in diabetic kidney disease control exclusion event table (generated in chapter 8)
  • has some event represented in kidney disease control exclusion event table (generated in chapter 9)
  • has been found to have microalbuminuria at least once
  • does not have eGFR measurements available
  • has been found to have eGFR level less than 60
  • has a first evidence of diabetes after the censoring date
  • has less than 5 years of follow-up time since the first evidence of diabetes before censoring date
  • has a first evidence of diabetes 6 months after the study initiation date
  • is not represented in the primary care data

We identify subjects who have had microalbuminuria at least once using miicroalbuminuria first occurrence event table generated in chapter 12. Subjects with eGFR level less than 60 are identified using eGFR trajectory data created in chapter 12.

Load pacakges and functions.

library(tidyverse)
library(data.table)
source("functions.R")

Load the following datasets:

• diabetes first occurrence data
• selected demographic data
• first occurrence DKD event table
• eGFR trajectory data
• microalbuminuria first occurrence event table

dm_firstoccur <- readRDS("generated_data/dm_firstoccur.RDS")
demog <- readRDS("generated_data/demog_selected.RDS")
dkd_firstoccur <- readRDS("generated_data/dkd_firstoccur.RDS")
egfr_traj <- readRDS("generated_data/trajectory_egfr.RDS")
microabu_firstoccur <- readRDS("generated_data/microabu_firstoccur.RDS")

Load the following control exclusion event tables:

• DKD control exclusion event table from UKB assessment center data
• DKD control exclusion event table from PCP data

dkd_control_exclusion_events_ukb <- readRDS("generated_data/dkd_control_exclusion_events_ukb.RDS")
kidney_disease_control_exclusion_events_pc <- readRDS("generated_data/kidney_disease_control_exclusion_pc.RDS")

Create additional control exclusion event tables:

• a table containing subjects who have been recorded to have ever had eGFR level less than 60
• a table containing subjects who have been recorded to have ever had microalbuminuria

egfr_lt60_ever_event_tab <- egfr_traj %>% filter(measurement < 60, !is.na(measurement), !is.na(event_dt))
microabu_ever_event_tab <- microabu_firstoccur %>% filter(!is.na(event))

Merge control exclusion event tables.

ckd_control_exclusion_events <- bind_rows(list(dkd_control_exclusion_events_ukb,
            kidney_disease_control_exclusion_events_pc,
            egfr_lt60_ever_event_tab,microabu_ever_event_tab))

Define control exclusion subject ID’s.

ctrl_exclusion_ids <- ckd_control_exclusion_events$f.eid %>% unique

The controls should have linked primary care dataset and at least one eGFR measurement.

pcp_subejct_ids <- readRDS("generated_data/gp_subject_ids.RDS")
eGFR_avail_subject_ids <- (egfr_traj %>% filter(!is.na(measurement) & !is.na(event_dt))) %>% .$f.eid

Define control inclusion subject ID’s.

ctrl_inclusion_ids <- intersect(pcp_subejct_ids,eGFR_avail_subject_ids) %>% unique

dkd_tte <- phenotype_time_to_event(dm_firstoccur,dkd_firstoccur,demog,
                                  control_exclusion_ids = ctrl_exclusion_ids,
                                  control_inclusion_ids = ctrl_inclusion_ids)

We filter out cases that do not have at least 5 years of follow-up time.

dkd_tte <- 
  dkd_tte %>%
  filter(event == 0 | (event == 1 & (lubridate::decimal_date(event_dt_comp) - lubridate::decimal_date(event_dt_dm) >= 5)))

saveRDS(dkd_tte,"generated_data/dkd_tte.RDS")