1 of 7

AI-READI Survey Mapping Pilot Phase Summary

Davera Gabriel, RN FHL7 FAMIA

1

September 29, 2023

2 of 7

AI-READI Survey Mapping Pilot Phase Summary

  • Summary Statistics (to-date)
  • Mapping Metadata overview
  • Mapping Exclusions
  • Issues for Discussion

​

3 of 7

Survey Mapping Summary

  • Survey Element Distribution (total n=2382)
    • Questions: 25% (n=606)
    • Answers: 57% (n=1350)
    • Not mapping to OMOP: 18% (n=426)
  • Target Vocabulary Achieved
    • OMOP Target Concept_id: 95% (n=1850)
    • OMOP Extension Required: 5% (n=92)
  • Map precision indices using SSSOM metadata elements
    • Predicate values
  • skos: exact match
  • skos: close match
  • skos: broad match
  • skos: narrow match
  • skos: related match
    • Mappers Confidence Score

https://mapping-commons.github.io/sssom/spec/

4 of 7

OMOP Target_id Mapping Metadata

SSSOM Predicate

100%

75%

50%

25%

0%

Totals

skos: exact match

1071

3

1

0

0

1075

skos: close match

213

96

18

0

0

327

skos: broad match

20

54

52

43

7

176

skos: narrow match

39

7

4

3

0

53

skos: related match

16

28

165

1

3

213

​

​

​

​

​

​

​

Totals

1359

188

240

47

10

​

5 of 7

OMOP Target_id Mapping Metadata

SSSOM Predicate

100%

75%

50%

25%

0%

Totals

skos: exact match

1071

3

1

0

0

1075

skos: close match

213

96

18

0

0

327

skos: broad match

20

54

52

43

7

176

skos: narrow match

39

7

4

3

0

53

skos: related match

16

28

165

1

3

213

​

​

​

​

​

​

​

Totals

1359

188

240

47

10

​

75% of all maps (n=1383)

6 of 7

Pilot Mapping Exclusions

  • Operational Variables
    • Interviewer Code / Technician ID
    • Survey instructions
      • “If yes, please choose between:”
      • “If yes, please check all that apply:”
    • Specimen shipped
  • OHDSI Eye Care & Vision Research Workgroup Elements
    • Imaging
    • MARS
  • Extensions Required
    • PhenX instruments
      • SDoH (combined) variables
      • Neighborhood / Environment Survey
      • Racial Ethnic Discrimination Survey

​

7 of 7

Issues for discussion

  • Prospective data collection vs observational data alignment
  • Survey design factors
    • Semantics of questions with Yes / No answers
    • Component question values versus compiled assessments
      • Example: PAID-5 DM survey
  • Ordinal scales
    • Examples: Novel time intervals
  • New expressions related to prior, similar instruments
    • Examples: Alcohol / Tobacco use questions

​