owned this note
owned this note
Published
Linked with GitHub
# Med Adherence Extension Proposed Updates
## Current State
### USCDI
![image](https://hackmd.io/_uploads/H1fcm43op.png)
### Design Goals
- Lightweight/Simple
- Bridge to US Core's planned transition to MedicationStatement when move to FHIR R6
### Solution
Based on FHIR R6 MedicationStatement created a Complex Extension with 3 sub-extensions to be used on MedicationRequest:
1. Med Adherence Code 1..1 MS CodeableConcept
2. dateAsserted 1..1 MS dateTime
3. informationSource 0..* MS Reference
![image](https://hackmd.io/_uploads/S1dd6zio6.png)
## 1. Change informationSource from References to Codes
based on comments [FHIR-43664](https://jira.hl7.org/browse/FHIR-43664), [FHIR-43466](https://jira.hl7.org/browse/FHIR-43466) from EHR Vendors.
- TL;DR: 2 major vendors want actor codes instead of Resource References!
:::info
FHIR R6 MedicationStatement suffers same deficit!
:::
### OPTION 1: Change to Reference Sub-Extension to Code:
1. Med Adherence Code 1..1 MS CodeableConcept
2. dateAsserted 1..1 MS dateTime
:::warning
3. informationSource 0..* MS CodeableConcept with an *example* binding of codes
:::
![image](https://hackmd.io/_uploads/r11LGQsoa.png)
#### Benefits
1. Simplifies the conformance rules since removes need to document the support reference choice.
2. Implies but does not directly communicate method used ( e.g., interview patient vs review dispense record.)
3. Starter set of codes promotes interoperability
#### Issues
1. No single existing value set to use, either
- too large
- too abstract
1. Does not follow FHIR R6 MedicationStatement Content Modeling (will need "top cover")
- Need to give and get feedback from Pharmacy before FHIR R6 publication (e.g., propose update` MedicationStatement.informationSource` type to codeable concept or codeableReference)
3. Implies but does not directly communicate method used ( e.g., interview patient vs review dispense record.)
### OPTION 2 Remove Information Source ( 2 extensions ) and Use FHIR Provenance for Provenance
Remove provenance from the extensions and provide guidance to use Provenance on the element for source and method if needed.
1. Med Adherence Code 1..1 MS CodeableConcept
2. dateAsserted 1..1 MS dateTime
![image](https://hackmd.io/_uploads/S1-TdVnj6.png)
#### Benefits
1. Simplifies the extension
2. Removes the unused references extension
3. Avoid the issues of defining codes
4. Provenance provides more provenance information
#### Issues
1. Using Provenance is considered to heavy
2. Avoid the issues raised by commenters
### OPTION 3 Add Codes and Keep references (4 extensions)
1. Med Adherence Code 1..1 MS CodeableConcept
2. dateAsserted 1..1 MS dateTime
3. informationSource 0..* MS Reference
:::warning
4. informationSource 0..* MS CodeableConcept with a small starter set of codes
:::
![image](https://hackmd.io/_uploads/rkIm543jp.png)
#### Benefits
1. See Option 1
#### Issues
1. See Option 1
2. Extension getting heavier
3. Commenter indicate that References not that useful or too hard to support so why keep them?
### OPTION 4: Change Reference Sub Extension to note
1. Med Adherence Code 1..1 MS CodeableConcept
2. dateAsserted 1..1 MS dateTime
:::warning
3. note 0..* MS Annotation
:::
![image](https://hackmd.io/_uploads/HkW0pVniT.png)
#### Benefits
1. Simplifies the conformance rules since removes need to document the support reference choice.
2. Able to identify source and or method or anything else
#### Issues
1. Unstructured
### OPTION 5 pre-adopt Codeableconcept
#### Only Issues no Benefits
### :new: OPTION 6 Use PractitionerRole
Keep same structure and reference PractitionerRole which may only contain a role code
#### Benefits
- consistent with Base FHIR model
- no changes to current model
#### Issues
- Last time we checked - not every server supports PR
- not every role is a practitioner (patient,family-member,MAR?)
- heavy solution for a role code
## 2. Provide definition for Medication Adherence adherence codes
Based on [FHIR-43757](https://jira.hl7.org/browse/FHIR-43664)
- TL;DR; Snomed Codes need definitions
1. Remove 457831000124109 and add 715036001
2. If possible in VSAC represent the ValueSet as a hierarchy same as SNOMED
3. Add *Profile Specific Implementation Guidance* to the extension to supplying some US Core definition to this starter set of codes.
> Note that this valueset is extensible, it is comprised of a set of basic Medication Adherence SNOMED CT and NullValue codes. More specific codes or other codes can be used from SNOMED CT or other code system including local codes. Refer to the section on [Extensible Bindings](#.html)
|Rank|Code (SCT ID)|Preferred Term|US Core Definition|
|---|---|---|---|
|0|1156699004|Complies with drug therapy (finding)| The medication is being taken as directed.|
|0|702565001|Non-compliance of drug therapy (finding)| The medication is being taken, but not as directed.( e.g., occasionaly forgets to take or 2 pills instead of 3))|
|1|~~275928001~~|~~Drugs - partial non-compliance (finding)~~| |
|1|~~275927006~~|~~Drugs - total non-compliance (finding)~~| |
|0|~~457831000124109~~|~~Medication therapy compliance variable (finding)~~ |
|0|715036001 | Does not take medication (finding) | The medication is not being taken ( e.g., Brett never picks up his meds :eyes:)
Armed with these definitions, we can map these scenarios from the tracker:
-Patient is supposed to take 30 mg twice daily and has reported they're only taking their morning dose -> 702565001 or 275927006*
-Patient is supposed to take 30 mg twice daily and has reported they're taking 20 mg twice daily zc -> 702565001 or 275927006*
-Patient is supposed to take 30 mg twice daily. They are following the prescribed instructions, but data from a pill tracker indicates they're regularly forgetting doses. -> 702565001 or 275928001*
-Patient is supposed to take 30 mg twice daily. They are very compliant when they have supply, but can't always afford to refill the medication on time. -> 702565001 or 275928001*
-Patient received adherence coaching a few months ago and is currently complying with their medications. Their PDC score for the year hasn't fully caught up and is under 50%. -> 1156699004 ? require clinical interpretation which outside the scope of US Core.
\* The second code is a child of the first in SNOMED CT
## 3. Add Method
See [FHIR-43756](https://jira.hl7.org/browse/FHIR-43756)
- TL;DR. add a method sub extension
:::info
FHIR R6 MedicationStatement suffers same deficit!
:::
We recognize the value of patient adherence to therapy and that various methods exist to measure adherence. The Medication Adherence Extension is designed to represent a statement of adherence as captured by the provider regardless of the method used. As a result of other ballot comments we have updated this Extension to support information source codes that we think will reflect the different metrics as well. Therefore, a sub-extension for the method is not a minimum expectation for the Medication Adherence Extension. However, other profiles and implementers, can introduce it. If there is sufficient adoption by the implementation community, we can reconsider adding it to US Core.
## 4 Med-Adherance Does Not Work
see [FHIR-44076](https://jira.hl7.org/browse/FHIR-44076)
- TL;DR Concern that indirect methods are not considered
1) Regarding the comment:
>To identify the "medication adherence" a query could be done comparing the MedicationRequest (aka orders) to the MedicationDispense. This will provide data on if indeed the patient is adherent to the medication (indirectly). This current query would require an assessment of the adherence at the point of the order. This is not possible (in my opinion).
We recognize the value of patient adherence to therapy and that various methods exist to measure adherence. Based on the USCDI defintion, the Medication Adherence Extension is designed to represent a *statement* of adherence as captured by the provider regardless of the method used. As a result of other ballot comments we have updated this Extension to support information source codes that we think will reflect the different metrics as well.
2. In responding to this tracker we discovered that the Extension.extension:medicationAdherence slice element's definition contains typos and editorial errors.
Change:
"Code stating whether the patient is consuming or being administered the medication has according to instructions. For example, the patient in not taking the medication as prescribed, or the patient is taking the medication as prescribed."
to:
"Code stating whether the patient is consuming medications according to instructions."