Survival Analysis from 50 years forward

Pre-analysis

Create first exposure_diagnoses and response_diagnoses -datasets. Also we need to have a dataset of population including dates DATE_BIRTH, DATE_DEATH, DATE_MIGRATION and DATE_50. This can be created by classify_population()-function.

First let’s define a case

exposure_icd9 = ""
exposure_icd8 = ""
exposure_src = c("")
response_icd10 = ""
response_icd9 = ""
response_icd8 = ""
response_src = c("")

Then we create study population

dpop <- healthpopR::classify_population(exposure_icd10 = exposure_icd10,
                                exposure_icd9 = exposure_icd9,
                                exposure_icd8 = exposure_icd8,
                                exposure_src = exposure_src,
                                response_icd10 = response_icd10,
                                response_icd9 = response_icd9,
                                response_icd8 = response_icd8,
                                response_src = response_src,
                                data_population = population,
                                data_diagnoses = diagnoses)

Next we create datasets exposure_diagnoses and response_diagnoses respectly:

exposure_diagnoses <- healthpopR::search_diagnoses(regex_icd10 = exposure_icd10,
                                                   regex_icd9 = exposure_icd9,
                                                   regex_icd8 = exposure_icd8,
                                                   registry_source = exposure_src)
response_diagnoses <- healthpopR::search_diagnoses(regex_icd10 = response_icd10,
                                                   regex_icd9 = response_icd9,
                                                   regex_icd8 = response_icd8,
                                                   registry_source = response_src)

Survival Analysis Function

We can define which is starting point to analysis. Choices are Exposure, Response or Follow up start date (DATE_50). This defined by argument start:

start = "DATE_EXPOSURE"
start = "DATE_RESPONSE"
start = "DATE_50"

We can define how to handle pre entry diagnoses by argument pre_entry_handling.

• truncate: Diagnosis date is set to entry date.
• skip: Diagnoses before entry are ignored; first post-entry diagnosis is used.
• asis: Diagnosis date is used as recorded.

In this example we choose to start survival analysis from Exposure date and we truncate dates before this entry date:

surv_results <- healthpopR::survival_analysis(exposure_diagnoses = exposure_diagnoses, 
                              response_diagnoses = exposure_diagnoses,
                              dpop = dpop,
                              start = "DATE_EXPOSURE",
                              censoring_date = as.Date("2024-12-31"),
                              pre_entry_handling = "truncate")

We can access the plots and data by created object:

surv_results$plot_years
surv_results$plot_mortality

Case of starting DATE_50

Actually, you can redefien starting date by mutating “DATE_50”. DATE_50 is picked as variable name, because we used that for our specific dataset.

• In future maybe we change this variable name to DATE_CUSTOM. But this needs to tested first.

dpop <- dpop %>% 
  mutate(DATE_50 = as.Date("2001-03-15"))

And then create results from this date:

surv_results <- healthpopR::survival_analysis(exposure_diagnoses = exposure_diagnoses, 
                              response_diagnoses = exposure_diagnoses,
                              dpop = dpop,
                              start = "DATE_50",
                              censoring_date = as.Date("2024-12-31"),
                              pre_entry_handling = "truncate")