This tutorial describes how to build longitudinal medication dose data from the raw output of an NLP system using the Pro-Med-NLP module.
EHR package.This tutorial describes how to use the Pro-Med-NLP module to generate longitudinal medication dose data from the raw output of an NLP system. The module is divided into two parts. Part I parses the raw output and pairs entities together, and Part II calculates dose intake and daily dose and removes redundant information (see Choi et al.\(^{1}\) and McNeer et al.\(^{2}\) for details).
To begin we load the EHR package.
Two main functions are used in this part of the module, a parse function and buildDose.
Parse functions are available for the medExtractR, MedXN, CLAMP, and MedEx systems (parseMedExtractR, parseMedXN, parseCLAMP, parseMedEx). If the user has output from another NLP system, the user can write code to standardize the output before calling the buildDose function.
In this tutorial, we will demonstrate the process using the parseMedExtractR function. See “1. EHR Vignette for Extract-Med and Pro-Med-NLP” in the EHR package for details on the other parse functions.
The parse functions output a standardized form of the data that includes a row for each drug mention and columns for all entities anchored to that drug mention.
parseMedExtractRFirst, we create variables for the filenames of our raw NLP system data. In the EHR package, we have example medExtractR data for tacrolimus and lamotrigine.
tac_mxr_fn <- system.file("examples", "tac_mxr.csv", package = "EHR")
lam_mxr_fn <- system.file("examples", "lam_mxr.csv", package = "EHR")Below is the example medExtractR output for tacrolimus:
filename entity expr pos
1 tacpid1_2008-06-26_note1_1.txt DrugName Prograf 930:937
2 tacpid1_2008-06-26_note1_1.txt Strength 1 mg 951:955
3 tacpid1_2008-06-26_note1_1.txt DoseAmt 3 956:957
4 tacpid1_2008-06-26_note1_1.txt Frequency twice a day 976:987
5 tacpid1_2008-06-26_note1_1.txt DrugName prograf 2709:2716
6 tacpid1_2008-06-26_note1_1.txt Dose 3mg 2717:2720
7 tacpid1_2008-06-26_note1_1.txt Frequency bid 2721:2724
8 tacpid1_2008-06-26_note1_1.txt LastDose 8:30pm 2740:2746
9 tacpid1_2008-06-26_note2_1.txt DrugName Prograf 618:625
10 tacpid1_2008-06-26_note2_1.txt Strength 1 mg 639:643
11 tacpid1_2008-06-26_note2_1.txt DoseAmt 3 644:645
12 tacpid1_2008-06-26_note2_1.txt Frequency twice a day 664:675
13 tacpid1_2008-06-26_note2_1.txt LastDose 14 hr 678:683
14 tacpid1_2008-12-16_note3_1.txt DrugName Tacrolimus 722:732
15 tacpid1_2008-12-16_note3_1.txt DrugName Prograf 761:768
16 tacpid1_2008-12-16_note3_1.txt Strength 1 mg 770:774
17 tacpid1_2008-12-16_note3_1.txt DoseAmt 3 775:776
18 tacpid1_2008-12-16_note3_1.txt Frequency twice a day 795:806
19 tacpid1_2008-12-16_note3_1.txt DoseChange decrease 2170:2178
20 tacpid1_2008-12-16_note3_1.txt DrugName Prograf 2179:2186
21 tacpid1_2008-12-16_note3_1.txt Dose 2mg 2190:2193
22 tacpid1_2008-12-16_note3_1.txt Frequency bid 2194:2197
23 tacpid1_2008-12-16_note3_1.txt DrugName Prograf 2205:2212
24 tacpid1_2008-12-16_note3_1.txt LastDose 10:30 pm 2231:2239Here we demonstrate how to run parseMedExtractR using the raw medExtractR output from above.
tac_mxr_parsed <- parseMedExtractR(tac_mxr_fn)
lam_mxr_parsed <- parseMedExtractR(lam_mxr_fn)The output from all systems, once parsed, has the same structure as the example parsed medExtractR output below.
tac_mxr_parsed filename drugname strength dose route freq dosestr
1: tacpid1_2008-06-26_note1_1.txt Prograf::930::937 1 mg::951::955 3::956::957 twice a day::976::987
2: tacpid1_2008-06-26_note1_1.txt prograf::2709::2716 bid::2721::2724 3mg::2717::2720
3: tacpid1_2008-06-26_note2_1.txt Prograf::618::625 1 mg::639::643 3::644::645 twice a day::664::675
4: tacpid1_2008-12-16_note3_1.txt Tacrolimus::722::732
5: tacpid1_2008-12-16_note3_1.txt Prograf::761::768 1 mg::770::774 3::775::776 twice a day::795::806
6: tacpid1_2008-12-16_note3_1.txt Prograf::2179::2186 bid::2194::2197 2mg::2190::2193
7: tacpid1_2008-12-16_note3_1.txt Prograf::2205::2212
dosechange lastdose
1:
2: 8:30pm::2740::2746
3: 14 hr::678::683
4:
5:
6: decrease::2170::2178
7: 10:30 pm::2231::2239lam_mxr_parsed filename drugname strength dose route
1: lampid1_2016-02-05_note4_1.txt Lamictal::810::818
2: lampid1_2016-02-05_note4_1.txt Lamotrigine::847::858 200mg::859::864 1.5::865::868
3: lampid1_2016-02-05_note4_1.txt Lamotrigine XR::954::968 100 mg::969::975 3::1000::1001`2::1037::1038
4: lampid1_2016-02-05_note5_1.txt ltg::442::445 200 mg::446::452 1.5::454::457
5: lampid1_2016-02-05_note5_1.txt ltg xr::465::471 100 mg::472::478 3::479::480`2::488::489
6: lampid2_2008-07-20_note6_1.txt lamotrigine::1267::1278
7: lampid2_2008-07-20_note6_1.txt lamictal::1280::1288
8: lampid2_2008-07-20_note6_1.txt Lamictal::2273::2281
9: lampid2_2012-04-15_note7_1.txt lamotrigine::103::114 150 mg::115::121
10: lampid2_2012-04-15_note7_1.txt Lamictal::141::149 1::151::152
freq dosestr dosechange lastdose
1: BID::826::829 300 mg::819::825
2: twice daily::873::884
3: morning::1025::1032`evening::1062::1069
4: daily::459::464
5: in am::481::486`in pm::490::495
6:
7: q12h::1299::1303 150 mg::1289::1295
8: BID::2294::2297 200mg::2285::2290 Increase::2264::2272
9:
10: twice a day::169::180 We now have a single row for each drug mention, but we need to pair appropriate entities together. For example, the third lamotrigine mention has two different doses and two different frequencies. The dose of “3” should be paired with the frequency “morning”, and the dose of “2” should be paired with the frequency “evening”. In the next section, we describe how to use the buildDose function to process this parsed data.
buildDoseAfter the NLP output is parsed, the buildDose function is run to pair the parsed entities. The main buildDose function arguments are as follows:
dat: standardized form of the data (i.e., output from one of the parse functions)dn: argument to specify drug names of interest (generally not used with medExtractR since medExtractR is a targeted system and only includes the drug of interest in the raw output)The output of the buildDose function is a dataset with a column for each entity and a row for each pairing.
buildDoseIn our medExtractR example from above, the output of the buildDose function is the following:
(tac_part_i_out <- buildDose(tac_mxr_parsed)) filename drugname strength dose route freq dosestr dosechange lastdose drugname_start
1 tacpid1_2008-06-26_note1_1.txt Prograf 1 mg 3 NA twice a day <NA> <NA> <NA> 930
2 tacpid1_2008-06-26_note1_1.txt prograf <NA> <NA> NA bid 3mg <NA> 8:30pm 2709
3 tacpid1_2008-06-26_note2_1.txt Prograf 1 mg 3 NA twice a day <NA> <NA> 14 hr 618
4 tacpid1_2008-12-16_note3_1.txt Tacrolimus <NA> <NA> NA <NA> <NA> <NA> <NA> 722
5 tacpid1_2008-12-16_note3_1.txt Prograf 1 mg 3 NA twice a day <NA> <NA> <NA> 761
6 tacpid1_2008-12-16_note3_1.txt Prograf <NA> <NA> NA bid 2mg decrease <NA> 2179
7 tacpid1_2008-12-16_note3_1.txt Prograf <NA> <NA> NA <NA> <NA> <NA> 10:30 pm 2205(lam_part_i_out <- buildDose(lam_mxr_parsed)) filename drugname strength dose route freq dosestr dosechange lastdose drugname_start
1 lampid1_2016-02-05_note4_1.txt Lamictal <NA> <NA> <NA> BID 300 mg <NA> <NA> 810
2 lampid1_2016-02-05_note4_1.txt Lamotrigine 200mg 1.5 <NA> twice daily <NA> <NA> <NA> 847
3 lampid1_2016-02-05_note4_1.txt Lamotrigine XR 100 mg 3 <NA> morning <NA> <NA> <NA> 954
4 lampid1_2016-02-05_note4_1.txt Lamotrigine XR 100 mg 2 <NA> evening <NA> <NA> <NA> 954
5 lampid1_2016-02-05_note5_1.txt ltg 200 mg 1.5 <NA> daily <NA> <NA> <NA> 442
6 lampid1_2016-02-05_note5_1.txt ltg xr 100 mg 3 <NA> in am <NA> <NA> <NA> 465
7 lampid1_2016-02-05_note5_1.txt ltg xr 100 mg 2 <NA> in pm <NA> <NA> <NA> 465
8 lampid2_2008-07-20_note6_1.txt lamotrigine <NA> <NA> <NA> <NA> <NA> <NA> <NA> 1267
9 lampid2_2008-07-20_note6_1.txt lamictal <NA> <NA> <NA> q12h 150 mg <NA> <NA> 1280
10 lampid2_2008-07-20_note6_1.txt Lamictal <NA> <NA> <NA> BID 200mg Increase <NA> 2273
11 lampid2_2012-04-15_note7_1.txt lamotrigine 150 mg <NA> <NA> <NA> <NA> <NA> <NA> 103
12 lampid2_2012-04-15_note7_1.txt Lamictal <NA> 1 <NA> twice a day <NA> <NA> <NA> 141We see that the third mention in the lamotrigine dataset now has two rows (rows 3 and 4), one with a dose of “3” and a frequency of “morning” and one with a dose of “2” and a frequency of “evening”. The “drugname_start” column gives us the start position for the drug name, which tells us that these two rows come from the same mention.
If the checkForRare argument is set to TRUE, any extracted expressions with a proportion of occurrence less than 0.2 are returned as rare values. When rare values are identified, a warning is printed to notify the user. The var column indicates the entity (note that dose in this output refers to dose amount, while dosestr would indicate dose given intake). This can be used as a quick check for potentially inaccurate information and allow the user to remove incorrect extractions before applying the Pro-Med-NLP module as incorrect extractions would reduce accuracy of the dose building data. Note that these values may still be correct extractions even though they are rare, as is the case for our output below.
lam_checkForRare <- buildDose(lam_mxr_parsed, checkForRare=TRUE) var val Freq Prop
1 drugname ltg 1 0.08333333
2 strength 150mg 1 0.14285714
3 dose 1 1 0.14285714In Part II of the module, we form the final analysis datasets containing computed dosing information at the note and date level for each patient. This process requires more detailed meta data associated with each clinical note file, the format of which is described below.
noteMetaDataThe meta data argument is required by the functions collapseDose and processLastDose, and requires four columns: filename, pid, date, note. In our example data, pid (patient ID), date, and note can all be extracted from the filename. Take the filename “tacpid1_2008-06-26_note1_1.txt” for example. It contains information in the form “[PID]_[date]_[note]”, where PID = “tacpid1”, date = “2008-06-26” and note = “note1”. The function below can build our meta data from each of the filenames.
bmd <- function(x) {
fns <- strsplit(x, '_')
pid <- sapply(fns, `[`, 1)
date <- as.Date(sapply(fns, `[`, 2), format = '%Y-%m-%d')
note <- sapply(fns, `[`, 3)
data.frame(filename = x, pid, date, note, stringsAsFactors = FALSE)
}
bmd("tacpid1_2008-06-26_note1_1.txt") filename pid date note
1 tacpid1_2008-06-26_note1_1.txt tacpid1 2008-06-26 note1(tac_metadata <- bmd(tac_part_i_out[['filename']])) filename pid date note
1 tacpid1_2008-06-26_note1_1.txt tacpid1 2008-06-26 note1
2 tacpid1_2008-06-26_note1_1.txt tacpid1 2008-06-26 note1
3 tacpid1_2008-06-26_note2_1.txt tacpid1 2008-06-26 note2
4 tacpid1_2008-12-16_note3_1.txt tacpid1 2008-12-16 note3
5 tacpid1_2008-12-16_note3_1.txt tacpid1 2008-12-16 note3
6 tacpid1_2008-12-16_note3_1.txt tacpid1 2008-12-16 note3
7 tacpid1_2008-12-16_note3_1.txt tacpid1 2008-12-16 note3collapseDoseThe main function used in Part II of the module is the collapseDose function. The output of this function is the final dose data with entities standardized, missing values imputed, dose intake and daily dose calculated, and redundancies removed. Two data.frames are generated one with redundancies removed at the note level and one at the date level (see McNeer et al. (2020) for details).
collapseDose allows the user to split the data using drug names given by regular expressions (...). For example, if the data includes multiple drugs, regular expressions can be specified for each drug. Another use of this function is to split the data by different formulations of the drug, such as separating immediate release formulations from extended release formulations, which are often written using “XR” or “ER” in the drug name.
The collapseDose function requires the following arguments:
x: output from the buildDose function, or from addLastDose if last dose information is incorporated (see Handling lastDose section below)noteMetaData: a data.frame with columns for filename, pid (patient id), date, and notenaFreq: method to use when assigning missing frequencies; the default is to assign the most common frequencycollapseDoseBelow, we demonstrate collapseDose using our lamotrigine example. In the function call, we supply an additional argument 'xr|er' to indicate that we want to separately consider extended release formulations of lamotrigine, (usually denoted by “XR” or “ER”). This prevents regular lamotrigine mentions from being collapsed with lamotrigine XR mentions, even if the dosage is identical.
data(lam_metadata, package = 'EHR')
lam_part_ii <- collapseDose(lam_part_i_out, lam_metadata, naFreq = 'most', 'xr|er')Note level collapsing:
lam_part_ii$note filename drugname strength dose route freq dosestr dosechange lastdose drugname_start dosestr.num
1 lampid1_2016-02-05_note4_1.txt Lamictal <NA> <NA> orally bid 300 mg <NA> <NA> 810 300
2 lampid1_2016-02-05_note4_1.txt Lamotrigine XR 100 mg 3 orally am <NA> <NA> <NA> 954 NA
3 lampid1_2016-02-05_note4_1.txt Lamotrigine XR 100 mg 2 orally pm <NA> <NA> <NA> 954 NA
4 lampid1_2016-02-05_note5_1.txt ltg 200 mg 1.5 orally daily <NA> <NA> <NA> 442 NA
5 lampid1_2016-02-05_note5_1.txt ltg xr 100 mg 3 orally am <NA> <NA> <NA> 465 NA
6 lampid1_2016-02-05_note5_1.txt ltg xr 100 mg 2 orally pm <NA> <NA> <NA> 465 NA
7 lampid2_2008-07-20_note6_1.txt lamictal <NA> <NA> orally bid 150 mg <NA> <NA> 1280 150
8 lampid2_2008-07-20_note6_1.txt Lamictal <NA> <NA> orally bid 200mg Increase <NA> 2273 200
9 lampid2_2012-04-15_note7_1.txt Lamictal <NA> 1 orally bid <NA> <NA> <NA> 141 NA
strength.num doseamt.num freq.num dose.intake intaketime dose.seq dose.daily
1 NA NA 2 300 <NA> NA 600
2 100 3.0 1 300 am 1 500
3 100 2.0 1 200 pm 2 500
4 200 1.5 1 300 <NA> NA 300
5 100 3.0 1 300 am 1 500
6 100 2.0 1 200 pm 2 500
7 NA NA 2 150 <NA> NA 300
8 NA NA 2 200 <NA> NA 400
9 150 1.0 2 150 <NA> NA 300Date level collapsing:
lam_part_ii$date filename drugname strength dose route freq dosestr dosechange lastdose drugname_start dosestr.num
1 lampid1_2016-02-05_note4_1.txt Lamictal <NA> <NA> orally bid 300 mg <NA> <NA> 810 300
2 lampid1_2016-02-05_note4_1.txt Lamotrigine XR 100 mg 3 orally am <NA> <NA> <NA> 954 NA
3 lampid1_2016-02-05_note4_1.txt Lamotrigine XR 100 mg 2 orally pm <NA> <NA> <NA> 954 NA
4 lampid1_2016-02-05_note5_1.txt ltg 200 mg 1.5 orally daily <NA> <NA> <NA> 442 NA
5 lampid2_2008-07-20_note6_1.txt lamictal <NA> <NA> orally bid 150 mg <NA> <NA> 1280 150
6 lampid2_2008-07-20_note6_1.txt Lamictal <NA> <NA> orally bid 200mg Increase <NA> 2273 200
7 lampid2_2012-04-15_note7_1.txt Lamictal <NA> 1 orally bid <NA> <NA> <NA> 141 NA
strength.num doseamt.num freq.num dose.intake intaketime dose.seq dose.daily
1 NA NA 2 300 <NA> NA 600
2 100 3.0 1 300 am 1 500
3 100 2.0 1 200 pm 2 500
4 200 1.5 1 300 <NA> NA 300
5 NA NA 2 150 <NA> NA 300
6 NA NA 2 200 <NA> NA 400
7 150 1.0 2 150 <NA> NA 300In the final datasets above, we see that a daily dose has been calculated and redundant daily doses have been removed at either the note or date level.
lastdoseIn this section, we cover how incorporation of the last dose entity should be handled if it was extracted using medExtractR. In the Running buildDose section above, we see the raw last dose time extractions for the tacrolimus dataset. Using the functions processLastDose and addLastDose, we convert the extracted times into a processed and standardized datetime variable, and add the processed times to the buildDose output.
The processLastDose function requires the following arguments:
mxrData: raw output from the extractMed functionnoteMetaData: note meta data for each file name in mxrDatalabData: a data frame containing lab dates and times associated with the file names within mxrData. This must contain at a minimum the columns pid and date (in the same format as noteMetaData), as well as labtime, a POSIXct variable indicating the date and time of a laboratory drug measurementExtracted last dose times can fall into two categories: a time expression (e.g., “10am”, “22:00”, “7 last night”) or a duration expression (e.g. “14 hour” level), where the “time” of last dose indicates the number of hours since the last dose was taken relative to the time of the clinical visit. In the latter case, the lab time (from the labData argument) is needed in order to convert the extracted duration expression into a datetime variable. Below is an example lab dataset for our sample tacrolimus data.
data(tac_lab, package = 'EHR')
tac_lab pid date labtime
1 tacpid1 2008-06-26 2008-06-26 10:42:00
2 tacpid1 2008-12-16 2008-12-16 12:11:00Within processLastDose, extracted times are converted to time expressions of the format “HH:MM:SS” and assigned a date based on the date of the corresponding note. When the last dose time is after 12pm, it is assumed to have been taken on the previous date.
(tac_ld <- processLastDose(mxrData = tac_mxr, noteMetaData = tac_metadata, labData = tac_lab)) filename lastdose ld_pos pid date raw_time ld_start time_type labtime
1 tacpid1_2008-06-26_note1_1.txt 2008-06-25 20:30:00 2740:2746 tacpid1 2008-06-26 8:30pm 2740 time 2008-06-26 10:42:00
2 tacpid1_2008-06-26_note2_1.txt 2008-06-25 20:42:00 678:683 tacpid1 2008-06-26 14 hr 678 duration 2008-06-26 10:42:00
3 tacpid1_2008-12-16_note3_1.txt 2008-12-15 22:30:00 2231:2239 tacpid1 2008-12-16 10:30 pm 2231 time 2008-12-16 12:11:00The function output contains the processed and standardized last dose time (lastdose), the original extracted expression (raw_time), whether the raw expression was a time or duration (time_type), as well as position information for the last dose time (ld_start) for appropriate pairing with dosing information in addLastDose. The labtime column in the output above corresponds to the information provided in the labData argument.
The addLastDose function requires the following arguments:
buildData: output from buildDoselastdoseData: dataset containing last dose time information for the file names in buildData. This should include columns for filename and lastdose, with lastdose being a processed POSIXct datetime variable.In the case where last dose information was extracted from clinical notes using medExtractR, the lastdoseData input should be output from the processLastDose function containing the last dose start positions, as demonstrated below. It is possible for multiple times to be extracted from a clinical note. For extracted times within a 2 hour window of one another, addLastDose treats these as equivalent and extracts the last dose time. Note that this may be context-dependent, and this rule was determined based on drugs administered every 12 hours and assuming a trough drug level. For time differences of more than two hours, the last dose start position is used to pair the extracted time with the closest drug mention. Alternatively, if the user has a separate dataset with validated last dose times, they can provide their own dataset. When providing a validated dataset, there should be only one last dose time per patient ID and date.
(tac_part_i_out_lastdose <- addLastDose(buildData = tac_part_i_out, lastdoseData = tac_ld)) filename drugname strength dose route freq dosestr dosechange lastdose drugname_start
1 tacpid1_2008-06-26_note1_1.txt Prograf 1 mg 3 NA twice a day <NA> <NA> <NA> 930
2 tacpid1_2008-06-26_note1_1.txt prograf <NA> <NA> NA bid 3mg <NA> 2008-06-25 20:30:00 2709
3 tacpid1_2008-06-26_note2_1.txt Prograf 1 mg 3 NA twice a day <NA> <NA> 2008-06-25 20:42:00 618
4 tacpid1_2008-12-16_note3_1.txt Tacrolimus <NA> <NA> NA <NA> <NA> <NA> <NA> 722
5 tacpid1_2008-12-16_note3_1.txt Prograf 1 mg 3 NA twice a day <NA> <NA> <NA> 761
6 tacpid1_2008-12-16_note3_1.txt Prograf <NA> <NA> NA bid 2mg decrease <NA> 2179
7 tacpid1_2008-12-16_note3_1.txt Prograf <NA> <NA> NA <NA> <NA> <NA> 2008-12-15 22:30:00 2205Note that in the lastdose columns, we now have standardized datetime objects instead of the raw extracted expressions.
collapseDose with last dose presentFor our tacrolimus example above, the output of collapseDose is below. Note that we use the output from addLastDose rather than directly from buildDose.
tac_part_ii <- collapseDose(tac_part_i_out_lastdose, tac_metadata, naFreq = 'most')Note level collapsing:
tac_part_ii$note filename drugname strength dose route freq dosestr dosechange lastdose drugname_start dosestr.num
1 tacpid1_2008-06-26_note1_1.txt Prograf 1 mg 3 orally bid <NA> <NA> 2008-06-25 20:30:00 930 NA
2 tacpid1_2008-06-26_note2_1.txt Prograf 1 mg 3 orally bid <NA> <NA> 2008-06-25 20:42:00 618 NA
3 tacpid1_2008-12-16_note3_1.txt Prograf 1 mg 3 orally bid <NA> <NA> 2008-12-15 22:30:00 761 NA
4 tacpid1_2008-12-16_note3_1.txt Prograf <NA> <NA> orally bid 2mg decrease 2008-12-15 22:30:00 2179 2
strength.num doseamt.num freq.num dose.intake intaketime dose.seq dose.daily
1 1 3 2 3 <NA> NA 6
2 1 3 2 3 <NA> NA 6
3 1 3 2 3 <NA> NA 6
4 NA NA 2 2 <NA> NA 4Date level collapsing:
tac_part_ii$date filename drugname strength dose route freq dosestr dosechange lastdose drugname_start dosestr.num
1 tacpid1_2008-06-26_note1_1.txt Prograf 1 mg 3 orally bid <NA> <NA> 2008-06-25 20:30:00 930 NA
2 tacpid1_2008-06-26_note2_1.txt Prograf 1 mg 3 orally bid <NA> <NA> 2008-06-25 20:42:00 618 NA
3 tacpid1_2008-12-16_note3_1.txt Prograf 1 mg 3 orally bid <NA> <NA> 2008-12-15 22:30:00 761 NA
4 tacpid1_2008-12-16_note3_1.txt Prograf <NA> <NA> orally bid 2mg decrease 2008-12-15 22:30:00 2179 2
strength.num doseamt.num freq.num dose.intake intaketime dose.seq dose.daily
1 1 3 2 3 <NA> NA 6
2 1 3 2 3 <NA> NA 6
3 1 3 2 3 <NA> NA 6
4 NA NA 2 2 <NA> NA 4Choi L, Beck C, McNeer E, Weeks HL, Williams ML, James NT, Niu X, Abou-Khalil BW, Birdwell KA, Roden DM, Stein CM. Development of a System for Post-marketing Population Pharmacokinetic and Pharmacodynamic Studies using Real-World Data from Electronic Health Records. Clinical Pharmacology & Therapeutics. 2020 Apr;107(4):934-43. doi: 10.1002/cpt.1787.
McNeer E, Beck C, Weeks HL, Williams ML, James NT, Bejan CA, Choi L. Building Longitudinal Medication Dose Data Using Medication Information Extracted from Clinical Notes in Electronic Health Records. J Am Med Inform Assoc. 2021 Mar 18;28(4):782-790. doi: 10.1093/jamia/ocaa291.
If you see mistakes or want to suggest changes, please create an issue on the source repository.