R/clinical_features.R
clinical_feature_ols_trend.RdAdd a clinical feature (variable) to a therapy or encounter longitudinal table. The feature corresponds to the Ordinary Least Squares (OLS) intercept and slope of clinical observations of interest
clinical_feature_ols_trend(
x,
observation_code,
hours,
observation_code_system = NULL,
compute = TRUE
)
# S4 method for RamsesObject
clinical_feature_ols_trend(
x,
observation_code,
hours,
observation_code_system = NULL,
compute = TRUE
)an object of class TherapyEpisode or Encounter
a character vector of clinical investigation codes
matching the observation_code field in the inpatient_investigation
table (see validate_investigations()
the maximum number of hours the observations should date back from
t_start, the starting time of every row in the longitudinal table
(optional, reserved to situations where
observation_code is ambiguous across code systems/vocabularies) a single
character string specifying the code system identifier of observation_code
(for example: "http://snomed.info/sct").
The default (NULL) filters observations using the observation_code only.
if TRUE (the default), the remote therapy table will
be computed on the remote server. This is generally faster.
an object of class TherapyEpisode or
Encounter
The feature will be computed exclusively on numeric investigations
marked with status "final", "preliminary", "corrected",
or "amended".
The returned regression slope coefficient corresponds to the mean change associated with a 1-hour time increment.
The returned regression intercept is defined with respect to time equals
zero at t_start. It thus corresponds to the value of the linear
(straight line) extrapolation of the trend to t_start.
if (FALSE) {
fake_db <- create_mock_database("example.duckdb")
temperature_check <- clinical_feature_ols_trend(
TherapyEpisode(fake_db, "4d611fc8886c23ab047ad5f74e5080d7"),
observation_code = "8310-5",
hours = 24
)
str(longitudinal_table(temperature_check, collect = TRUE))
}