Skip to content

Latest commit

 

History

History
132 lines (92 loc) · 5.09 KB

demo.md

File metadata and controls

132 lines (92 loc) · 5.09 KB
Raw

Using R to derive robust insights from real-world healthcare data

Load data

  • Delivery = 02
  • ETL version = 02
tables_02 <- receive_delivery_02()

tables_02_etl_02 <- tables_02 %>%
  etl_02()

QC

Patients

## ✔ All Pass No failures across all 13 Tests!
fail_n test column any_fail
0 Key columns exist birth_dt FALSE
0 Everyone born after 01-01-1900 birth_dt FALSE
0 Key columns exist cncr FALSE
0 Dataset contains patients with and without cancer cncr FALSE
0 Key columns exist death_dt FALSE
0 Everyone born after 01-01-1900 death_dt FALSE
0 Dataset contains alive and dead patients death_observed FALSE
0 Key columns exist desynpuf_id FALSE
0 Dataset has only One Row Per Patient (OORP) desynpuf_id FALSE
0 Key columns exist diabetes FALSE
0 Key columns exist race_cd FALSE
0 Key columns exist sex_ident_cd FALSE
0 Dataset contains both Male and Female patients sex_ident_cd FALSE 
## ✔ All Pass No failures across all 13 Tests!

Inpatient

## ✔ All Pass No failures across all 4 Tests!
fail_n test column any_fail
0 Claim admin is a date clm_from_dt FALSE
0 Claim payment amount is positive clm_pmt_amt FALSE
0 Claim admin is a date clm_thru_dt FALSE
0 Key columns exist desynpuf_id FALSE 
## ✔ All Pass No failures across all 4 Tests!

ORPP

## ℹ ORPP | Added 5 var(s) from inpatient:

## • has_inpatient_claims = Flag if patient has any inpatient claims

## • inpatient_claims_n = Number of inpatient claims

## • inpatient_payment_median = Median dollar amount of inpatient claims

## • inpatient_payment_min = Minimum dollar amount of inpatient claims

## • inpatient_payment_max = Maximum dollar amount of inpatient claims

## ℹ ORPP | Added 5 var(s) from prescription:

## • has_prescription = Flag if patient has any prescriptions

## • prescription_n = Number of prescription events

## • prescription_patient_payment_median = Median patient prescription payment
##   amount

## • prescription_rx_cost_median = Median prescription total cost amount

## • prescription_patient_payment_percentage_median = Median percentage of total
##   prescription cost paid by patient

## ✔ All Pass No failures across all 3 Tests!
fail_n test column any_fail
0 Patients have non-negative median inpatient payments inpatient_payment_median FALSE
0 Patients have non-negative median prescription costs inpatient_payment_median FALSE
0 Patients do not have negative survival time survival_years FALSE 
## ✔ All Pass No failures across all 3 Tests!

Cohort

Attrition Table

description n n_dropped
Everyone 10000 NA
Race is white or black 9323 -677
Has Diabetes 3605 -5718
Is Male 1526 -2079
Has at least 1000 inpatient costs 268 -1258
Median prescription cost is > 50 83 -185
18 years of age or older 83 0

Attrition Plot

Analyses

Survival

# Fit the survival model
#  Follow-up time is survival_years
#  Status indicator is death_observed
#  Stratify results by sex_cd

fit <- survfit(Surv(survival_years, death_observed) ~ race_cd,
  data = cohort_tbl
)

# Create the final Kaplan-Meier curve
ggsurvplot(fit,
           conf.int = TRUE,
           surv.median.line = "hv") +
  labs(title = "Survival of patients from birth to death",
       subtitle = "Using CMS Medicare claims synthetic public use files",
       caption = glue::glue("Patients = {nrow(cohort_tbl)}\nPatients checks pass = {patients_checks_pass}\nInpatient checks pass = {inpatient_checks_pass}\nORPP checks pass = {orpp_checks_pass}"))