constellate_criteria: Provide details about individual events within a...
In constellation: Identify Event Sequences Using Time Series Joins

Description Usage Arguments Details Value Imported functions Errors Examples

A function that reads in multiple time series data frames for various events and builds indicator variables for each event. Individual events that occur within a specified number from a timestamp are flagged. The variables for each event can be populated with the time the event took place, a boolean variable (0 or 1) indicating whether or not the event took place, or the result of the variable at the time the event took place.

1 2	constellate_criteria(..., criteria_names, window_hours, join_key, time_var, value = c("boolean", "time", "result"), result_var = NULL)

`...`	An arbitrary number of time series data frames that each include the columns 'join_key' and 'time_var'
`criteria_names`	A vector of strings specifying the name of each event. The order of strings in the vector should align with the order of data frames passed in '...'.
`window_hours`	A single numeric or vector of numerics specifying the number of hours to search for each event. The order of numerics in the vector should align with the order of data frames passed in '...'.
`join_key`	A string name of the column to join all time series data frames
`time_var`	A string name of the time stamp column in all time series data frames. The class of time_var must be POSIXct in all data frames.
`value`	A string specifying the value to be entered within each criteria column. Options include boolean (0 or 1, depdending on whether the criteria event occurred), the time of the criteria event, or the result stored within the criteria event. The default value is boolean.
`result_var`	A string name of the value variable in all data frames. This argument should only be supplied if the "result" option is selected in the value argument.

The user passes an arbitrary number of time series data frames and specifies a name and number of hours to search for each event. The user must also specify a variable to use to join the tables, and the time stamp variable. The timestamps variable in every data frame must be POSIXct class. Finally, the user selects how to populate the individual event variables.

This function extends the constellate function to address a different set of questions, including: 1) at a specific timestamp, which events do and do not occur? 2) what is the sequence of events that trigger the constellation of events that I'm interested in? 3) What are the results of each criteria at the times that each criteria are met? This function can be used to calculate risk scores at any measurement timestamp by building a new variable after the function runs and returns the new data frame. The risk score can add up the crieteria from boolean values (e.g. SIRS Criteria) or can be a linear combination of criteria (e.g., NEWS Score).

A data.frame, data.table with indicator variables for each event. The total number of rows is the unique number of time stamps for all combined measurements.

general data.table syntax

This function returns errors for:

missing arguments (value has a default argument and result_var is missing by default)
passing a window_hours value that is not numeric
passing join_key or time_var values that are not column names in all time series data frames
passing an invalid number of criteria_names (must be equal to number of criteria event data frames)
passing an invalid number of window_hours values (1 or the number of criteria event data frames).

library(data.table)
temp <- as.data.table(vitals[VARIABLE == "TEMPERATURE"])
pulse <- as.data.table(vitals[VARIABLE == "PULSE"])
resp <- as.data.table(vitals[VARIABLE == "RESPIRATORY_RATE"])

temp[, RECORDED_TIME := as.POSIXct(RECORDED_TIME,
  format = "%Y-%m-%dT%H:%M:%SZ", tz = "UTC")]
pulse[, RECORDED_TIME := as.POSIXct(RECORDED_TIME,
  format = "%Y-%m-%dT%H:%M:%SZ", tz = "UTC")]
resp[, RECORDED_TIME := as.POSIXct(RECORDED_TIME,
  format = "%Y-%m-%dT%H:%M:%SZ", tz = "UTC")]

# Pass single window_hours
constellate_criteria(temp, pulse, resp, criteria_names = c("TEMPERATURE",
 "PULSE", "RESPIRATORY_RATE"), window_hours = 6, join_key = "PAT_ID",
 time_var = "RECORDED_TIME", value = "time")
# Pass vector for window_hours
constellate_criteria(temp, pulse, resp, criteria_names = c("TEMPERATURE",
 "PULSE", "RESPIRATORY_RATE"), window_hours = c(6,6,6), join_key = "PAT_ID",
 time_var = "RECORDED_TIME", value = "time")
# Show the value of each criteria at the time the event occurs
constellate_criteria(temp, pulse, resp, criteria_names = c("TEMPERATURE",
 "PULSE", "RESPIRATORY_RATE"), window_hours = c(6,6,6), join_key =
 "PAT_ID", time_var = "RECORDED_TIME", value = "result",
 result_var = "VALUE")