README.Rmd

---
output: github_document
---

```{r, include = FALSE}
knitr::opts_chunk$set(
  warning = FALSE, message = FALSE,
  collapse = TRUE,
  comment = "#>",
  fig.path = "man/figures/README-",
  out.width = "100%"
)
```

# PatientProfiles <img src="man/figures/logo.png" align="right" height="200"/>

[![CRAN status](https://www.r-pkg.org/badges/version/PatientProfiles)](https://CRAN.R-project.org/package=PatientProfiles)
[![codecov.io](https://codecov.io/github/darwin-eu/PatientProfiles/coverage.svg?branch=main)](https://app.codecov.io/github/darwin-eu/PatientProfiles?branch=main)
[![R-CMD-check](https://github.com/darwin-eu/PatientProfiles/workflows/R-CMD-check/badge.svg)](https://github.com/darwin-eu/PatientProfiles/actions)
[![Lifecycle: stable](https://img.shields.io/badge/lifecycle-stable-brightgreen.svg)](https://lifecycle.r-lib.org/articles/stages.html#stable)
[![metacran downloads](https://cranlogs.r-pkg.org/badges/PatientProfiles)](https://cran.r-project.org/package=PatientProfiles)
[![metacran downloads](https://cranlogs.r-pkg.org/badges/grand-total/PatientProfiles)](https://cran.r-project.org/package=PatientProfiles)

## Package overview

PatientProfiles contains functions for adding characteristics to OMOP CDM tables containing patient level data (e.g. condition occurrence, drug exposure, and so on) and OMOP CDM cohort tables. The characteristics that can be added include an individual´s sex, age, and days of prior observation Time varying characteristics, such as age, can be estimated relative to any date in the corresponding table. In addition, PatientProfiles also provides functionality for identifying intersections between a cohort table and OMOP CDM tables containing patient level data or other cohort tables. 

## Package installation

You can install the latest version of PatientProfiles like so:

```{r, eval=FALSE}
install.packages("PatientProfiles")
```

## Citation

```{r}
citation("PatientProfiles")
```

## Example usage
### Create a reference to data in the OMOP CDM format 
The PatientProfiles package is designed to work with data in the OMOP CDM format, so our first step is to create a reference to the data using the CDMConnector package. 

```{r, message=FALSE}
library(CDMConnector)
library(PatientProfiles)
library(dplyr)
```

Creating a connection to a Postgres database would for example look like:
```{r, eval=FALSE}
con <- DBI::dbConnect(
  RPostgres::Postgres(),
  dbname = Sys.getenv("CDM5_POSTGRESQL_DBNAME"),
  host = Sys.getenv("CDM5_POSTGRESQL_HOST"),
  user = Sys.getenv("CDM5_POSTGRESQL_USER"),
  password = Sys.getenv("CDM5_POSTGRESQL_PASSWORD")
)

cdm <- cdm_from_con(
  con,
  cdm_schema = Sys.getenv("CDM5_POSTGRESQL_CDM_SCHEMA"),
  write_schema = Sys.getenv("CDM5_POSTGRESQL_RESULT_SCHEMA")
)
```
To see how you would create a reference to your database please consult the CDMConnector package [documentation](https://darwin-eu.github.io/CDMConnector/articles/a04_DBI_connection_examples.html). For this example though we'll work with simulated data, and we'll generate an example cdm reference like so:

```{r}
cdm <- mockPatientProfiles(numberIndividuals = 1000)
```

### Adding individuals´ characteristics
#### Adding characteristics to patient-level data
Say we wanted to get individuals´sex and age at condition start date for records in the condition occurrence table. We can use the `addAge` and `addSex` functions to do this:

```{r}
cdm$condition_occurrence |>
  glimpse()

cdm$condition_occurrence <- cdm$condition_occurrence |>
  addAge(indexDate = "condition_start_date") |>
  addSex()

cdm$condition_occurrence |>
  glimpse()
```

We could, for example, then limit our data to only males aged between 18 and 65
```{r}
cdm$condition_occurrence |>
  filter(age >= 18 & age <= 65) |>
  filter(sex == "Male")
```

#### Adding characteristics of a cohort
As with other tables in the OMOP CDM, we can work in a similar way with cohort tables. For example, say we have the below cohort table

```{r}
cdm$cohort1 |>
  glimpse()
```

We can add age, age groups, sex, and days of prior observation to a cohort like so
```{r}
cdm$cohort1 <- cdm$cohort1 |>
  addAge(
    indexDate = "cohort_start_date",
    ageGroup = list(c(0, 18), c(19, 65), c(66, 100))
  ) |>
  addSex() |>
  addPriorObservation()

cdm$cohort1 |>
  glimpse()
```

We could use this information to subset the cohort. For example limiting to those with at least 365 days of prior observation available before their cohort start date like so
```{r}
cdm$cohort1 |>
  filter(prior_observation >= 365)
```

### Cohort intersections

#### Detect the presence of another cohort in a certain window
We can use `addCohortIntersectFlag` to add a flag for the presence (or not) of a cohort in a certain window.

```{r, echo = FALSE}
cdm <- mockPatientProfiles()
```
```{r}
cdm$cohort1 |>
  glimpse()

cdm$cohort1 <- cdm$cohort1 |>
  addCohortIntersectFlag(
    targetCohortTable = "cohort2",
    window = c(-Inf, -1)
  )

cdm$cohort1 |>
  glimpse()
```

#### Count appearances of a certain cohort in a certain window
If we wanted the number of appearances, we could instead use the `addCohortIntersectCount` function

```{r, echo = FALSE}
cdm <- mockPatientProfiles()
```
```{r}
cdm$cohort1 |>
  glimpse()

cdm$cohort1 <- cdm$cohort1 |>
  addCohortIntersectCount(
    targetCohortTable = "cohort2",
    targetCohortId = 1,
    window = list("short_term" = c(1, 30), "mid_term" = c(31, 180))
  )

cdm$cohort1 |>
  glimpse()
```

#### Add a column with the first/last event in a certain window
Say we wanted the date at which an individual was in another cohort then we can use the `addCohortIntersectDate` function. As there might be multiple records for the other cohort, we can also choose the first or the last appearance in that cohort.

First occurrence:
```{r, echo = FALSE}
cdm <- mockPatientProfiles()
```
```{r}
cdm$cohort1 |>
  glimpse()

cdm$cohort1 <- cdm$cohort1 |>
  addCohortIntersectDate(
    targetCohortTable = "cohort2",
    targetCohortId = 1,
    order = "first",
    window = c(-Inf, Inf)
  )

cdm$cohort1 |>
  glimpse()
```

Last occurrence:
```{r, echo = FALSE}
cdm <- mockPatientProfiles()
```
```{r}
cdm$cohort1 |>
  glimpse()

cdm$cohort1 <- cdm$cohort1 |>
  addCohortIntersectDate(
    targetCohortTable = "cohort2",
    targetCohortId = 1,
    order = "last",
    window = c(-Inf, Inf)
  )

cdm$cohort1 |>
  glimpse()
```

#### Add the number of days instead of the date
Instead of returning a date, we could return the days to the intersection by using `addCohortIntersectDays`

```{r, echo = FALSE}
cdm <- mockPatientProfiles()
```
```{r}
cdm$cohort1 |>
  glimpse()

cdm$cohort1 <- cdm$cohort1 |>
  addCohortIntersectDays(
    targetCohortTable = "cohort2",
    targetCohortId = 1,
    order = "last",
    window = c(-Inf, Inf)
  )

cdm$cohort1 |>
  glimpse()
```


#### Combine multiple cohort intersects
If we want to combine multiple cohort intersects we can concatenate the operations using the `pipe` operator:

```{r, echo = FALSE}
cdm <- mockPatientProfiles()
```
```{r}
cdm$cohort1 |>
  glimpse()

cdm$cohort1 <- cdm$cohort1 |>
  addCohortIntersectDate(
    targetCohortTable = "cohort2",
    targetCohortId = 1,
    order = "last",
    window = c(-Inf, Inf)
  ) |>
  addCohortIntersectCount(
    targetCohortTable = "cohort2",
    targetCohortId = 1,
    window = c(-Inf, Inf)
  )

cdm$cohort1 |>
  glimpse()
```

```{r}
mockDisconnect(cdm)
```