apportion_prior_healthcare: Apportion CDI cases on basis of prior healthcare interactions

In PublicHealthEngland/hcaidcs: R functions for the mandatory surveillance of bacteraemia and C. difficile infection

Apportion CDI cases on basis of prior healthcare interactions

Description

The mandatory surveillance launched a new apportioning algorithm on 01/04/2017. This new apportioning algorithm takes into account prior healthcare interactions as recorded by trust users to apportion to one of four categories:

• Healthcare onset - healthcare associated

• Community onset - healthcare associated

• Community onset - indeterminate association

• Community onset - community associated

Records created prior to 01/04/2017 should not be apportioned.

Usage

apportion_prior_healthcare(
  patient_location,
  patient_category,
  adm_date,
  spec_date,
  adm_3_mo,
  adm_4_weeks,
  adm_12_weeks,
  date_record_created
)

Arguments

patient_location	The patient's location at time of sample
patient_category	The patient's category at time of sample
adm_date	Date of admission
spec_date	Date of specimen
adm_3_mo	Was the patient an inpatient at the trust in previous three months. String yes, no or don't know
adm_4_weeks	Was the patient admitted at any point 4 weeks prior to specimen. String yes or no
adm_12_weeks	Was the patient admitted at any point 12 weeks prior to specimen. String yes or no
date_record_created	Date record entered onto DCS

Value

A string variable giving the apportioning type; one of hoha, coha, coia or coca, all_blank if all of the prior healthcare exposures were NA, unknown_3_mo if prior health care in past three months was "Don't know" or NA if date record created < 01/04/2017

Examples

testdat <- data.frame(
  patient_location = "NHS Acute Trust",
  patient_category = "In-patient",
  date_admitted = lubridate::dmy("01-01-2017"),
  specimen_date = lubridate::dmy("05-01-2017"), admission_in_3 = "yes",
  adm_4_wks = "yes", adm_12_wks = "yes",
  date_entered = lubridate::dmy("01-04-2017"),
  stringsAsFactors = FALSE)
testdat

testdat$apportioned_prior_hc <- apportion_prior_healthcare(
  patient_location = testdat$patient_location,
  patient_category = testdat$patient_category,
  adm_date = testdat$date_admitted,
  spec_date = testdat$specimen_date,
  adm_3_mo = testdat$admission_in_3,
  adm_4_weeks = testdat$adm_4_wks,
  adm_12_weeks = testdat$adm_12_wks,
  date_record_created = testdat$date_entered
)
testdat
## Not run: 
testdat$apportioned_prior_hc <- NULL
testdat2 <- data.frame(
  patient_location = rep("NHS Acute trust", 3),
  patient_category = rep("In-patient", 3),
  date_admitted = rep(lubridate::dmy("01-01-2017"), 3),
  specimen_date = rep(lubridate::dmy("01-01-2017"), 3),
  admission_in_3 = c("yes", "yes", "yes"),
  adm_4_wks = c("yes", "no", "no"),
  adm_12_wks = c("no", "yes", "no"),
  date_entered = lubridate::dmy("01-04-2017"),
  stringsAsFactors = FALSE)
testdat <- dplyr::bind_rows(testdat, testdat2)
testdat

testdat$apportioned_prior_hc <- apportion_prior_healthcare(
  patient_location = testdat$patient_location,
  patient_category = testdat$patient_category,
  adm_date = testdat$date_admitted,
  spec_date = testdat$specimen_date,
  adm_3_mo = testdat$admission_in_3,
  adm_4_weeks = testdat$adm_4_wks,
  adm_12_weeks = testdat$adm_12_wks,
  date_record_created = testdat$date_entered
)
testdat

## End(Not run)

PublicHealthEngland/hcaidcs documentation built on Jan. 19, 2024, 8:38 a.m.