![Errors During Data Mapping
Variable Code Value Unit Date
Glucose 2339-0 130 mmol/l 12.01.2001
Creatinine 2160-0 100.0 mg/dL 07.02.2003
... ... ...
Observation
code
quantity
date
unit
value
text
{
"resourceType": "Observation",
"id": "measurement-1",
"status": "final",
"code": {
"coding": [
{
"system": "http://loinc.org",
"code": "2339-0",
"display": "Glucose [Mass/volume] in Blood"
}
],
"text": "Glucose"
},
"effectiveDateTime": "2001-01-12T00:00:00Z",
"valueQuantity": {
"value": 130.0,
"unit": "mmol/L",
"system": "http://unitsofmeasure.org",
"code": "mmol/L"
}
}](https://image.slidesharecdn.com/healthdatasummit-remzicelebi-250321195948-b1ce1dab/85/Common-Data-Quality-Issues-During-Data-Mapping-and-How-to-Detect-Avoid-Them-3-320.jpg)
More Related Content
Similar to Common Data Quality Issues During Data Mapping and How to Detect/Avoid Them (20)
Recently uploaded (20)
![MICROBIOLOGY FOR NURSES [Autosaved].pptx](https://cdn.slidesharecdn.com/ss_thumbnails/microbiologyautosaved-250403114413-afe32e01-thumbnail.jpg?width=560&fit=bounds)
Common Data Quality Issues During Data Mapping and How to Detect/Avoid Them
- 1. Common Data Quality Issues During Data Mapping and How to Detect/Avoid Them Remzi Celebi remzi.celebi@maastrichtuniversity.nl Maastricht University Technical Coordinator, AIDAVA project 20 March 2025
- 2. Data Quality Issues during Mapping Data quality issues can occur during mapping to a target data model. These errors can be classified into: Syntactic errors, such as incorrect date formatting or mismatched data types. Semantic errors, such as mapping a data element to the wrong category.
- 3. Errors During Data Mapping Variable Code Value Unit Date Glucose 2339-0 130 mmol/l 12.01.2001 Creatinine 2160-0 100.0 mg/dL 07.02.2003 ... ... ... Observation code quantity date unit value text { "resourceType": "Observation", "id": "measurement-1", "status": "final", "code": { "coding": [ { "system": "http://loinc.org", "code": "2339-0", "display": "Glucose [Mass/volume] in Blood" } ], "text": "Glucose" }, "effectiveDateTime": "2001-01-12T00:00:00Z", "valueQuantity": { "value": 130.0, "unit": "mmol/L", "system": "http://unitsofmeasure.org", "code": "mmol/L" } }
- 4. Errors During Data Mapping The date 12.01.2001 could be in the US format. This needs to be converted to the standard format (eg. ISO 8601 format YYYY-MM-DDThh:mm:ss+zz:zz) to ensure the interoperability. Variable Code Value Unit Date Glucose 2339-0 130 mmol/l 12.01.2001 Creatinine 2160-0 100.0 mg/dL 07.02.2003 ... ... ... Observation code quantity date unit value text
- 5. Errors During Data Mapping Either the unit for Creatinine is incorrect, or the value itself is incorrect. The value should be checked for plausibility. Variable Code Value Unit Date Glucose 2339-0 130 mmol/l 12.01.2001 Creatinine 2160-0 100.0 mg/dL 07.02.2003 ... ... ... Observation code quantity date unit value text
- 6. Errors During Data Mapping The identifiers in the Code column can mapped to SNOMED-CT using Variable column, but these are the identifiers for substances being measured, not the actual tests performed. Therefore, mapping these to observation codes would be incorrect. Variable Code Value Unit Date Glucose 67079006 130 mmol/l 12.01.2001 Creatinine 15373003 2.0 mg/dL 07.02.2003 ... ... ... Observation code quantity date unit value text
- 7. RDFCraft -- tool to guide the mapping RDFCraft can be used to semi-automate the mapping process. Nodes are created corresponding to schema classes. Links are created corresponding to relations. Reduce possible errors during the mapping. https://github.com/MaastrichtU-IDS/RDFCraft
- 9. Standard Mapping Definitions Click Save & Map Share standard mappings Multiple formats supported: YARRRML RML RDF/TTL Use mapping engine to execute the mapping
- 10. Data quality check services in AIDAVA 10 Missing Essential variable - Birth date is missing. Multiple values than expected - A measurement must have only one quantity. Wrong data type for property - A hasValue for Quantity should take only string or double. Wrong object type for property - A hasUnit for Quantity must take only Unit type. Not valid code for type - Diagnosis does not use valid code. Data model Based Checks Medical & Common sense Checks Conditional Completeness - Flag If the patient is prescribed an allergy drug, and an allergy is not in their record. Incompatible information (date and time error); - Flag if date of birth > date of admission Incompatible information (gender and diagnosis) - Flag if Gender is equal to Male, Diagnosis Benign neoplasm ovary is present. Incompatible information (age and procedure) - Flag if Age is over 12, Procedure TONSILLECTOMY AND ADENOIDECTOMY; UNDER AGE 12 is present.. Incompatible information (lab measurement and unit) - Flag if Lab LDL Cholesterol is present and Unit is not equal to mg/dL
- 11. Key takeaways Different Errors can occurs Syntactic errors: Mostly caused by incorrect mapping or uncleaned data that does not conform to the target format. Semantic errors: Occur due to misinterpretation of the source data and target model. To minimize errors, use appropriate tools during the mapping process and implement data quality checks after conversion.