TODO - Manuscript Task: Finalize Phenotypes

[callahantiff]

Manuscript: First manuscript on method

Task:

• Phenotypes
• Mapping medication strings to RxNorm ingredients
• Ok to include procedures?
• How to map problem lists (do we treat these like conditions?)

Description: Get Tell's feedback with the list

Due Date: ASAP

[callahantiff]

Updates after today's meeting with @tdbennett , details can be found here.

[callahantiff]

@tdbennett, @mgkahn can you please help me verify/answer the following questions related to the eMERGE phenotypes?

General

1. How do we map broad categories of medications, when specifics are not provided?
2. How to handle situations where a string and not a code is provided in the phenotype documentation (e.g. "Tofranil" instead of RxNORM_835583)?
3. For definitions that include lists of specific covariates, do we also need to ensure that we include these variables when we generate embeddings for the scalability experiments?
4. When definitions include verifying the presence or absence of codes or strings in a patient's "chart" (which we don't have), can we just verify that the patient's record does not include any mentions of the code/string?
5. If two exclusionary criteria are separated by "and/or" can we just treat it as "and"?

Specific phenotype questions have been converted to individual issues that reference this

[callahantiff]

UPDATE:

General

1. How do we map broad categories of medications, when specifics are not provided?
2. How to handle situations where a string and not a code is provided in the phenotype documentation (e.g. "Tofranil" instead of RxNORM_835583)?

For both of the scenarios described above, we will treat a source string as a string and use it to query for matches to an OMOP standard terminology concept. There are several ways we can do this, here is one way, organized by clinical domain:

SV: Source Vocabulary
ST: Standardized Terminology

• Diagnoses/Problem Lists:

  • If an SV code is provided, it will be mapped to SNOMED CT concepts
  • If no SV code is provided, a match between the provided string and an ST code will be will be found by querying the string against ST code labels (i.e. condition_source_value or concept_name)

• Medications:

  • If an SV code is provided, it will be mapped to RxNORM concepts
  • If no SV code is provided, a match between the provided string and an ST code will be will be found by querying the string against ST code labels (i.e. drug_source_value or concept_name)

• Laboratory Tests:

  • If an SV code is provided, it will be mapped to LOINC
  • If no SV code is provided, a match between the provided string and an ST code will be will be found by querying the string against ST code labels (i.e. measurement_source_value or concept_name)

• Procedures:

  • If an SV code is provided, it will be mapped to CPT4
  • If no SV code is provided, a match between the provided string and an ST code will be will be found by querying the string against ST code labels (i.e. procedure_source_value or concept_name)

[callahantiff]

closing this for now, see individual phenotypes issues for details.