Details
-
Change Request
-
Resolution: Persuasive
-
Medium
-
FHIR Core (FHIR)
-
R5
-
Clinical Decision Support
-
Evidence
-
Content
-
-
Brian Alper/Rob McClure: 18-0-0
-
Enhancement
-
Non-compatible
-
R5
Description
In preparation for Evidence Resources to report meta-analysis results we discovered changes to make.
- In the Evidence resource, rename "StatisticCertaintyRating” codesystem to "EvidenceCertaintyRating."
- In the Evidence resource, rename "StatisticCertaintySubcomponentType" codesystem to "CertaintySubcomponentType."
- In the Evidence resource, rename "StatisticCertaintySubcomponentRating" codesystem to "CertaintySubcomponentRating."
- We will add very-serious-concern as an additional code to the codelist for Evidence.certainty.certaintySubcomponent.rating, and we will also change the “critical-concern” code to “extremely-serious-concern” to follow a recent GRADE development
- Create a backbone element Statistic.modelCharacteristic with a cardinality of 0..*. And it will contain:
- Statistic.modelCharacteristic.code which is an extensible codeableConcept 0..1 with a codesystem: statistic-model-code
- Statistic.modelCharacteristic.value[x] which expands to valueCodeableConcept, valueBoolean, valueSimpleQuantity, and valueRange. Where valueCodeableConcept would have the extensible codesystem statistic-model-method
- Change Statistic.attributeEstimate.estimateQualifier to a nested Statistic.attributeEstimate.attributeEstimate, similar to what we learned from creating EvidenceReport.section.section where it says "see section" and it's a nested section.
- Statistic.attributeEstimate.attributeEstimate change short name to "A nested attribute estimate; which is the attribute estimate of an attribute estimate" and have an example in the comments
- Change Statistic.attributeEstimate short name to "An attribute of the Statistic" and definition to something more descriptive.
- Statistic.attributeEstimate.type change short name to "The type of attribute estimate, eg confidence interval or p value"
New Codesystems
statistic-model-code
effectsFixed | Fixed-effects meta-analysis | From a fixed-effects meta-analysis, pair with valueBoolean |
effectsRandom | Random-effects meta-analysis | From a random-effects meta-analysis, pair with valueBoolean |
methodTau | Tau estimation method | Method used for estimating 𝜏2 (used with a random-effects meta-analysis), pair with valueCodeableConcept |
poolMethod | Pooling method | Method used to pool results, pair with valueCodeableConcept |
hartungKnapp | Hartung-Knapp adjustment | Hartung-Knapp/Hartung-Knapp-Sidik-Jonkman adjustment used, pair with valueBoolean |
modifiedHartungKnapp | Modified Hartung-Knapp adjustment | Modified Hartung-Knapp/Hartung-Knapp-Sidik-Jonkman adjustment used, pair with valueBoolean |
zeroCellNumerical | Zero-cell adjustment with constant | Zero-cell adjustment done by adding a constant to all cells of affected studies, pair with valueSimpleQuantity |
zeroCellTreatmentArmContinuityCorrection | Zero-cell adjustment with continuity correction | Zero-cell adjustment done by treatment arm continuity correction, pair with valueBoolean |
statistic-model-method
tauDersimonianLaird | Dersimonian-Laird method | Dersimonian-Laird method for 𝜏2 |
tauPauleMandel | Paule-Mandel method | Paule-Mandel method for 𝜏2 |
tauRestrictedMaximumLikelihood | Restricted Maximum Likelihood method | Restricted Maximum Likelihood method for 𝜏2 |
tauMaximumLikelihood | Maximum Likelihood method | Maximum Likelihood method for 𝜏2 |
tauEmpiricalBayes | Empirical Bayes method | Empirical Bayes method for 𝜏2 |
tauHunterSchmidt | Hunter-Schmidt method | Hunter-Schmidt method for 𝜏2 |
tauSidikJonkman | Sidik-Jonkman method | Sidik-Jonkman method for 𝜏2 |
tauHedges | Hedges method | Hedges method for 𝜏2 |
poolMantelHaenzsel | Mantel-Haenszel method | Mantel-Haenszel method for pooling in meta-analysis |
poolInverseVariance | Inverse variance method | Inverse variance method for pooling in meta-analysis |
poolPeto | Peto method | Peto method for pooling in meta-analysis |
poolGeneralizedLinearMixedModel | Generalized linear mixed model (GLMM) | Generalized linear mixed model (GLMM) method for pooling in meta-analysis |