episode_calculation: Calculates Hypo/Hyperglycemic episodes with summary statistics

Description

The function determines episodes or events, calculates summary statistics, and optionally returns data with episode label columns added

Usage

episode_calculation(
  data,
  lv1_hypo = 70,
  lv2_hypo = 54,
  lv1_hyper = 180,
  lv2_hyper = 250,
  dur_length = 15,
  return_data = FALSE,
  dt0 = NULL,
  inter_gap = 45,
  tz = ""
)

Value

If return_data is FALSE, a single dataframe with columns:

id: Subject id
type: Type of episode - either hypoglycemia or hyperglycemia
level: Level of episode - one of lv1, lv2, or extended
avg_ep_per_day: Average number of episodes per day calculated as (total # episodes)/(recording time in days (24hrs))
avg_ep_duration: Average duration of episodes in minutes
avg_ep_gl: Average glucose in the episode in mg/dL

If return_data is TRUE, returns a list where the first entry is the episode summary dataframe (see above) and the second entry is the input data with episode labels added. Mostly for use with epicalc_profile function. Format of the second list entry is:

id: Subject id
time: Interpolated timestamps
gl: glucose in mg/dL
**episode_label**: One column per episode label - i.e. lv1_hypo, lv2_hypo, lv1_hyper, lv2_hyper, ext_hypo. 0 means not this type of episode, a positive integer label is assigned to each episode. Note the labels are *not* unique by subject only unique by segment

Arguments

data: DataFrame object with column names "id", "time", and "gl". Should only be data for 1 subject. In case multiple subject ids are detected, the warning is produced and only 1st subject is used.
lv1_hypo: Numeric value specifying a hypoglycemia threshold for level 1
lv2_hypo: Numeric value specifying a hypoglycemia threshold for level 2
lv1_hyper: Numeric value specifying a hyperglycemia threshold for level 1
lv2_hyper: Numeric value specifying a hyperglycemia threshold for level 2
dur_length: Numeric value specifying the minimum duration in minutes to be considered an episode. Note dur_length should be a multiple of the data recording interval otherwise the function will round up to the nearest multiple.
return_data: Boolean indicating whether to also return data with episode labels. Defaults to FALSE which means only episode summary statistics will be returned
dt0: The time frequency for interpolation in minutes, the default will match the CGM meter's frequency (e.g. 5 min for Dexcom).
inter_gap: The maximum allowable gap (in minutes) for interpolation. The values will not be interpolated between the glucose measurements that are more than inter_gap minutes apart. The default value is 45 min.
tz: A character string specifying the time zone to be used. System-specific (see as.POSIXct), but " " is the current time zone, and "GMT" is UTC (Universal Time, Coordinated). Invalid values are most commonly treated as UTC, on some platforms with a warning.

Author

Johnathan Shih, Jung Hoon Seo, Elizabeth Chun

Details

We follow the definition of episodes given in the 2023 consensus by Battelino et al. Note we have classified lv2 as a subset of lv1 since we find the consensus to be slightly ambiguous.

References

Battelino et al. (2023): Continuous glucose monitoring and metrics for clinical trials: an international consensus statement Lancet Diabetes & Endocrinology 11(1) .42-57, tools:::Rd_expr_doi("10.1016/s2213-8587(22)00319-9").

Examples

Run this code

episode_calculation(example_data_5_subject, lv1_hypo=100, lv1_hyper= 120)

Run the code above in your browser using DataLab