Data Diagnostics Settings

2023-04-05

Data Diagnostics Settings

These instructions will go through how to use the createDataDiagnosticsSettings function to create your settings based on your clinical question of interest. These settings will then be passed to the executeDbDiagnostics function to determine which databases meet your needs.

Inputs

These are all the available inputs for the settings function:

  • analysisId
    - The identifier for the specific analysis.
  • analysisName
    - The name for the specific analysis.
  • minAge
    - The minimum age for patients included in the analysis. Default is the lowest age available in the database.
  • maxAge
    - The minimum age for patients included in the analysis. Default is the highest age available in the database.
  • genderConceptIds
    - The required sex(es) at birth for patients included in the analysis, expressed as a vector of concepts.List the standard OMOP concepts for the required values as found here. Default is c(8507, 8532) for male and female.
  • raceConceptIds
    - The required races for patients included in the analysis, expressed as a vector of concepts, e.g. c(8515,8527). If no restriction leave blank, else list the standard OMOP concepts for the required values as found here.
  • ethnicityConceptIds
    - The required ethnicities for patients included in the analysis, expressed as a vector of concepts, e.g. c(38003564,38003563). If no restriction leave blank, else list the standard OMOP concepts for the required values as found here.
  • studyStartDate
    - The start date of the analysis. Date format it ‘YYYYMM’. If no restriction leave blank.
  • studyEndDate
    - The end date of the analysis. Date format it ‘YYYYMM’. If no restriction leave blank.
  • requiredDurationDays
    - The minimum required follow-up time in days for patients included in the analysis.
  • requiredDomains
    - The data domains required for ALL patients in the analysis, expressed as a character vector. Valid values are condition, drug, device, measurement, measurementValues, death, procedure, observation. Default is c(“condition”, “drug”)
  • desiredDomains
    - The domains required for SOME patients in the analysis. Meaning you need the dataset to have these data but each person is not required to have a record. This is most often used to identify the presence of outcomes of interest. Valid values are condition, drug, device, measurement, measurementValues, death, procedure, observation. If no restriction leave blank.
  • requiredVisits
    - The visits required for ALL patients in the study, expressed as a character vector, e.g. c(“IP”,“OP”). Valid values are IP,OP,ER. If no restriction leave blank.
  • desiredVisits
    - The visits required for SOME patients in the study, expressed as a character vector, meaning you need the dataset to have these data but each person is not required to have a record. This is most often used to identify the presence of outcomes of interest. Valid values are IP, OP, ER.
  • targetName - The name of the target of interest.
  • targetConceptIds - A vector containing the required target concepts.
  • comparatorName - The name of the comparator of interest.
  • comparatorConceptIds - A vector containing the required comparator concepts.
  • indicationName - The name of the indication of interest.
  • indicationConceptIds - A vector containing the required indication concepts.
  • includeIndicationInCalc - A T/F object indicating whether the proportion of people with the indication concepts should be included in the estimated sample size calculation. Default is FALSE.
  • outcomeName
    - The name of the outcome of interest.
  • outcomeConceptIds - A vector containing the required outcome concepts.

Example Study

We will be investigating whether patients exposed to lisinopril are at a higher risk for acute myocardial infarction compared with patients exposed to hydrochlorothiazide.


# first set your output location
outputFolder <- "/Users/clairblacketer/dbDiagnosticsOutput/Example"

# read in the csv files with the concepts that represent lisinopril and hydrocholorothiazide
# these files can be found in the extras/example_study folder of the github repo

lisinopril <- read.csv(file.path("extras/example_study/lisinoprilConcepts.csv"), stringsAsFactors = FALSE)
hctz <- read.csv(file.path("extras/example_study/hydrocholorothiazideConcepts.csv"), stringsAsFactors = FALSE)

# create the settings for the study

analysisSettings1 <- DbDiagnostics::createDataDiagnosticsSettings(

    analysisId = 1,
    analysisName = "lisinopril v HCTZ for AMI",
    minAge = 18,
    maxAge = 100,
    genderConceptIds = c(8507,8532),
    raceConceptIds = NULL,
    ethnicityConceptIds = NULL,
    studyStartDate = "200501",
    studyEndDate = "201901",
    requiredDurationDays = 365,
    requiredDomains = c("condition","drug"),
    desiredDomains = NULL,
    requiredVisits = NULL,
    desiredVisits = c("IP"),
    targetName = "lisinopril",
    targetConceptIds = lisinopril$lisinopril,
    comparatorName = "hydrochlorothiazide",
    comparatorConceptIds = hzt$hydrocholorothiazide,
    outcomeName = "acute myocardial infarction IP events",
    outcomeConceptIds = c(312327,319039,434376,436706,438170,438438,438447,439693,441579,444406,761736,761737,765132,3189643,3654465,3654466,3654467,3655133,3661502,3661503,3661504,3661520,3661524,3661547,3661641,3661642,3661643,3661644,3661645,3661646,4030582,4051874,4108217,4108218,4108669,4108677,4119456,4119457,4119943,4119944,4119945,4119946,4119947,4119948,4121464,4121465,4121466,4124684,4124685,4124686,4126801,4145721,4151046,4170094,4173632,4178129,4200113,4206867,4207921,4209541,4215259,4243372,4267568,4270024,4275436,4296653,4303359,4323202,4324413,4329847,35610087,35610089,35610091,35610093,35611570,35611571,37309626,43020460,44782712,44782769,45766075,45766076,45766113,45766114,45766115,45766116,45766150,45766151,45766241,45771322,45773170,46270158,46270159,46270160,46270161,46270162,46270163,46270164,46273495,46274044)
    )
    
# IMPORTANT! You need to pass a list of all settings to the executeDbDiagnostics function. It is common for this function to 
# be used to evaluate multiple studies at one time so you need to add them all to one list like below, even if you only have
# one analysis.

settingsList <- list(analysisSettings1)