In ohdsi-studies/MentalHealthEquity: Assessing Health Equity in Mental Healthcare Delivery Using a Federated Network Research Model

Baseline Characterization of Bipolar Disorder, Depresssion, and Suicidality

Executive Summary

Date: 2023-02-20

Summary: Characterization to assess prevalence of Bipolar Disorder, Depresssion, and Suicidality

Running Characterization

Packages

The following packages will be loaded to conduct the characterization:

library(DatabaseConnector)
library(dplyr)
library(lubridate)
library(readr)
library(SqlRender)
library(tibble)

To learn more about these packages, see the Appendix.

Defining Connection Details

Here, we need to set-up connection to the OMOP CDM database we will assess. To do so, we need to define some constants that will be used for the connection. The following list of constants:

• dbms - the database management system that is used to host your database; common options include (see all options here):
  • "postgresql"
  • "sql server"
• server - name of the server; could be localhost, an address like 123.0.1.5, etc.
• user - your username to access the server
• password - the password you use to access the server
• port - the port where the database is hosted
• schema - name of the database schema used

Must be defined in this code block:

dbms <- "Fill in here"
server <- "Fill in here"
user <- "Fill in here"
password <- "Fill in here"
port <- "Fill in here"
schema <- "Fill in here"

An additional step needed is to configure the required driver to connect to the database as follows:

1. Determine the name of your database management system based on the list here
2. Download the drivers by running the following code block after following instructions:

This is accomplished in the following codeblock (change eval = FALSE to eval = FALSE when you have set these variables correctly):

pathToDriver <- "/location/that/you/want"
downloadJdbcDrivers(dbms = dbms, pathToDriver = pathToDriver, method = "auto")

Once this is done, we can create the connection to the database (change eval = FALSE to eval = FALSE when you have set these variables correctly):

connectionDetails <- createConnectionDetails(dbms=dbms, 
                                             server=server,
                                             user=user,
                         password=password,
                         port=port,
                         pathToDriver=pathToDriver)

connection <- connect(connectionDetails)

If there were no errors, then we should be able to continue with the analysis!

WARN: As you proceed with this analysis, if you encounter a Java issue like this: "Insufficient java heap memory", please run the following code block:

r options(java.parameters = c("-XX:+UseConcMarkSweepGC", "-Xmx8192m")) This is only an emergency work around and should be removed when a better solution is found.

Example Connection

If any of this was confusing, here is an example of how to fill out the above connection information:

dbms <- "postgresql"
server <- "test.data.americus.edu/mimic_omop"
user <- "mimic"
password <- "omoprocks"
port <- 5042
schema <- "mimic.omop"

pathToDriver = "utils"
downloadJdbcDrivers(dbms = dbms, pathToDriver = pathToDriver, method = "auto")

connectionDetails <- createConnectionDetails(dbms=dbms, 
                                             server=server,
                                             user=user,
                                             password=password,
                         port=port,
                         pathToDriver=pathToDriver)

connection <- connect(connectionDetails)

Load condition_concept_ids for a Given Condition

For this task, we want to load the concept set and get all of the concept IDs for the diseases being studied:

Bipolar Disorder

pathToPhenotype_bipolar <- "../phenotypes/bipolar_concept_set.csv"
bipolar_concept_ids <- read.csv(pathToPhenotype_bipolar)$CONCEPT_ID

Depression

pathToPhenotype_depression <- "../phenotypes/depression_concept_set.csv"
depression_concept_ids <- read.csv(pathToPhenotype_depression)$CONCEPT_ID

Suicidality

pathToPhenotype_suicidality <- "../phenotypes/suicidality_concept_set.csv"
suicidality_concept_ids <- read.csv(pathToPhenotype_suicidality)$CONCEPT_ID

Filter Database Patients To Those Captured by a Concept Set

Here, we find every patient that has a history of one of the diseases we are studying (at least one diagnosis):

source("./sql/condition_filter.R")

bipolar_patients <- condition_filter(schema, bipolar_concept_ids, dbms, connection)

depression_patients <- condition_filter(schema, depression_concept_ids, dbms, connection)

suicidality_patients <- condition_filter(schema, suicidality_concept_ids, dbms, connection)

Characterize Each Person by Race, Gender, and Age Group

Now we run a query for these patient populations to get their stratification by the axes of race, gender, and age group:

source("./sql/stratified_person.R")

bipolar_patients <- stratify_persons(schema, bipolar_patients, dbms, connection)

depression_patients <- stratify_persons(schema, depression_patients, dbms, connection)

suicidality_patients <- stratify_persons(schema, suicidality_patients, dbms, connection)

Filter Patient Counts

To maintain patient privacy and security, we will follow HITECH standards to filter out any patient subpopulation that has a count less than 11 from this dataset.

Bipolar Disorder

bipolar_patients <- bipolar_patients %>%
  filter(COUNTS > 10)

Depression

depression_patients <- depression_patients %>%
  filter(COUNTS > 10)

Suicidality

suicidality_patients <- suicidality_patients %>%
  filter(COUNTS > 10)

Finalize Outputs

Here we now finally save this data to the data directory in the root of the study:

exportFolder <- "../data/baseline"

write.csv(bipolar_patients, 
          file = file.path(exportFolder, "bipolar_person_stratified_breakdown.csv"), 
          row.names = FALSE)

write.csv(depression_final_df, 
          file = file.path(exportFolder, "depression_person_stratified_breakdown.csv"), 
          row.names = FALSE)

write.csv(suicidality_final_df, 
          file = file.path(exportFolder, "suicidality_person_stratified_breakdown.csv"), 
          row.names = FALSE)

Conclusion

This concludes the steps required for generating the baseline of our characterization study. Please take the folder named baseline located in the root of the study in the data folder and send it to us at GTRI at [e-mail]. Thank you very much for participating in this study!

Appendix

Packages Details

• renv - create reproducible environments for R projects
• dplyr - grammar for data manipulation
• tibble - improved data.frame functionality
• SqlRender - package for rendering parameterized SQL
• DatabaseConnector - package for connecting to databases using JDBC
• lubridate - makes it easier to work with date times
• readr - a fast and friendly way to read rectangular data

ohdsi-studies/MentalHealthEquity documentation built on March 9, 2023, 12:39 p.m.