Map OMOP data to inpatient_microbiology table

[razekmh]

Extract data from the example OMOP data to fill the inpatient_microbiology table from the validate article. This is a split from #17

Please feel free to assign yourself to the issue. Please the respective branch for development.

[AngharadGreen]

• It has been difficult to map the simulated OMOP data I have been working with to the RAMSES inpatient_microbiology table.
• There isn't a clear way of storing lab microbiology results in the OMOP CDM.
• I have found this paper useful - https://pubmed.ncbi.nlm.nih.gov/35612079/
• When we receive OMOP extract data from UCLH it will be useful to see how lab microbiology results are stored within the CDM.
• I have been working on issue Map OMOP data to inpatient_episodes table  #28 instead

[zsenousy]

• It has been difficult to map the simulated OMOP data I have been working with to the RAMSES inpatient_microbiology table.
• There isn't a clear way of storing lab microbiology results in the OMOP CDM.
• I have found this paper useful - https://pubmed.ncbi.nlm.nih.gov/35612079/
• When we receive OMOP extract data from UCLH it will be useful to see how lab microbiology results are stored within the CDM.
• I have been working on issue Map OMOP data to inpatient_episodes table  #28 instead

@AngharadGreen, Would you like to discuss this further in a call?

[AngharadGreen]

• It has been difficult to map the simulated OMOP data I have been working with to the RAMSES inpatient_microbiology table.
• There isn't a clear way of storing lab microbiology results in the OMOP CDM.
• I have found this paper useful - https://pubmed.ncbi.nlm.nih.gov/35612079/
• When we receive OMOP extract data from UCLH it will be useful to see how lab microbiology results are stored within the CDM.
• I have been working on issue Map OMOP data to inpatient_episodes table  #28 instead

@AngharadGreen, Would you like to discuss this further in a call?

Hi @zsenousy I have found some useful resources to help me with this mapping, I am just working through those today. I can meet tomorrow if there is a time that works for you?

[AngharadGreen]

Following this paper https://pubmed.ncbi.nlm.nih.gov/35612079/ I will try to map the OMOP CDM to the inpatient_microbiology ramses table following this figure -

I have put together a summary presentation going over how the OMOP CDM can be mapped to the Ramses fata frame
Mapping of OMOP to RAMSES.pdf

[AngharadGreen]

I now want to just pull out the relevant columns from the Merged_df_1 table created in code from #45 that represent the Ramses Inpatient_microbiology table to see if I can use the OMOP data to create the data frames required by Ramses

Ramses	OMOP
patient_id	person_id (Observation table)
specimen_id	specimen_id (Specimen table)
status	All entries are NA so will ignore
specimen_type_display	concept_name_specimen (Concept table)
specimen_datetime	specimen_datetime (Specimen table)
organism_display_name	concept_name_observation (Concept table)
isolate_id	observation_id (Observation Table)
agent_display_name	Still need to look into this
rsi_code	Still need to look into this
isolation_datetime	observation_datetime (Observastion table)

## This is following on from the code in issue 45: 

Merged_df_2 <- subset(Merged_df_1, select = c("person_id","specimen_id","specimen_type_display","specimen_datetime","organism_display_name","observation_id","observation_datetime"))

## lets rename these columns to match the Ramses Inpatient_micrbiology table 

Merged_df_3 <- Merged_df_2%>% rename(patient_id = person_id,
                                     isolate_id = observation_id,
                                     isolation_datetime = observation_datetime)
## Lets add in the columns status, agent_display_name and rsi_code and fill them with NA as these have not been mapped to the OMOP data yet

Merged_df_3['status'] <- NA
Merged_df_3['agent_display_name'] <- NA
Merged_df_3['rsi_code'] <- NA

I will attempt to load this table into Ramses

library(Ramses)
library(ggplot2)
library(dplyr)

microbiology_data <- list()


> microbiology_data <- list()
> microbiology_data$raw <- Merged_df_3
> microbiology_data$raw <- microbiology_data$raw %>% 
+   mutate(
+     organism_code = AMR::as.mo(if_else(
+       organism_display_name == "No growth",
+       NA_character_,
+       organism_display_name)),
+     agent_code = AMR::as.ab(agent_display_name)) %>% 
+   mutate(organism_name = AMR::mo_name(organism_code),
+          agent_name = AMR::ab_name(agent_code))
i Microorganism translation was uncertain for "Anaerobic bacteria" (assumed
Anaerococcus lactolyticus) and "Mycobacterium, avium-intracellulare Group"
(assumed Mycobacterium avium-intracellulare complex). Run mo_uncertainties() to
review these uncertainties, or use add_custom_microorganisms() to add custom
entries.
i The following microorganisms were taxonomically renamed (use keep_synonyms =
TRUE to leave uncorrected):
  * Bacillus subtilis subtilis (Nakamura et al., 1999)  ->  Bacillus subtilis (Cohn, 1872)
  * Mycobacteroides abscessus (Gupta et al., 2018)  ->  Mycobacterium abscessus (Tortoli et al., 2016)
Error in `mutate()`:
ℹ In argument: `agent_code = AMR::as.ab(agent_display_name)`.
Caused by error:
! in AMR::as.ab(): argument x must be a text string, a (whole) number, or of
class 'factor', i.e. not be TRUE or `FALSE
Run `rlang::last_trace()` to see where the error occurred.


> microbiology_data$raw <- microbiology_data$raw %>% 
+   mutate(specimen_type_code = case_when(
+     specimen_type_display == "Blood Culture" ~ 
+       "446131002", # Blood specimen obtained for blood culture
+     specimen_type_display == "Faeces" ~ 
+       "119339001", # Stool specimen
+     specimen_type_display == "MRSA Screen" ~ 
+       "697989009", # Anterior nares swab 
+     specimen_type_display == "Urine" ~ 
+       "122575003", # Urine specimen
+     TRUE ~ NA_character_
+   )) %>% 
+   left_join(transmute(reference_specimen_type,
+                       specimen_type_code = conceptId,
+                       specimen_type_name = pt_term))
Joining with `by = join_by(specimen_type_code)`

## Lets try this again but remove reference to Agent_name and also use SNOMED_microorganism_code to get the organism_name with the AMR package as this column contains the relevant SNOMED codes 
microbiology_data$raw <- Merged_df_3
microbiology_data$raw <- microbiology_data$raw %>% 
  mutate(organism_name = AMR::mo_name(SNOMED_microorganism_code))

microbiology_data$raw <- microbiology_data$raw %>% 
  mutate(specimen_type_code = case_when(
    specimen_type_display == "Blood Culture" ~ 
      "446131002", # Blood specimen obtained for blood culture
    specimen_type_display == "Faeces" ~ 
      "119339001", # Stool specimen
    specimen_type_display == "MRSA Screen" ~ 
      "697989009", # Anterior nares swab 
    specimen_type_display == "Urine" ~ 
      "122575003", # Urine specimen
    TRUE ~ NA_character_
  )) %>% 
  left_join(transmute(reference_specimen_type,
                      specimen_type_code = conceptId,
                      specimen_type_name = pt_term))


microbiology_data$specimens <- microbiology_data$raw %>% 
  transmute(specimen_id,
            patient_id,
            status = "available",
            specimen_datetime,
            specimen_type_code,
            specimen_type_name,
            specimen_type_display) %>% 
  distinct() # Removing duplicates created by multiple isolates and susceptibility testing

microbiology_data$isolates <- microbiology_data$raw %>% 
  transmute(isolate_id,
            specimen_id,
            patient_id,
            organism_code = SNOMED_microorganism_code,
            organism_name,
            organism_display_name,
            isolation_datetime) %>% 
  distinct() # Removing duplicates created by susceptibility testing

The two data frames specimens = microbiology_data$specimens and isolates = microbiology_data$isolates have now been created from the OMOP mapping

I now need to find the relevant data to create the data frame susceptibilities = microbiology_data$susceptibilities)

Information on RSI code:

S - Susceptible, standard dosing regimen: A microorganism is categorised as "Susceptible, standard dosing regimen", when there is a high likelihood of therapeutic success using a standard dosing regimen of the agent.
I - Susceptible, increased exposure*: A microorganism is categorised as "Susceptible, Increased exposure*" when there is a high likelihood of therapeutic success because exposure to the agent is increased by adjusting the dosing regimen or by its concentration at the site of infection.
R - Resistant: A microorganism is categorised as "Resistant" when there is a high likelihood of therapeutic failure even when there is increased exposure.

[AngharadGreen]

• We now have a data frame from OMOP where we have mapped 7 out of the 9 relevant columns for the Ramses Inpatient_microbiology table .
• We now need to find the relevant information in OMOP that relates to agent_display_name and rsi_code

## I now want to identify where I can find information about agent display name and RSI code within the OMOP extract

## I will have a look at the drug_expsure table and identify how we can pull information about the drugs prescribed and map to the information we have in the Inpatient_microbiology table

## Lets read in the drug_exposure table
df_drug_exposure <- read_parquet("~/Ramses_extract_Nov2024/DRUG_EXPOSURE.parquet")

## I have tried to merge the Inpatient_microbiology _OMOP table with the drug_exposure table via the person_id but it keeps causing R to crash
Merge_6 <- merge(x=Inpatient_Microbiology_OMOP_2 , y=df_drug_exposure, by.x = "person_id.x", by.y = "person_id")

## I will look for a way to filter out just those person_id's that are present in the inpatient_microbiology table from the drug exposure table