Discussion on improvements for simple/complex notes

Hi,

As discussed on the SHR call, I though i’d get the conversation on OpenMRS notes started, and see where we can take it from here.

So my two cents are,

As Ryan suggested, there are two types of notes,

  1. Simple Notes - Store text notes that links to the patient record in the OpenMRS data model

  2. Complex Notes - documents (PDFs, scanned docs) could be stored as a complex note rather than as an observation (a PDF-MIME type)

As per my previous discussion on the OpenMRS mailing list, an OpenMRS Note is tied to an encounter. A Note is not intended to store an HL7 NTE segment - we have obs.comment for that.

Does this correspond with what you had in mind ?

And also, is it our intention to store complex notes using the same mechanism used to store unstructured complex obs ? if so, we may have some trouble getting our changes into the OpenMRS core…

···


Best Regards,

Suranga

Hi Suranga,

Thanks for getting this started. I agree with those two types of notes.

So what do people think of having an note attached to an encounter? Should this always be the case or should this be optional? Is there a case where we would not want a note to be attached to an encounter. I can think of summary documents that may span multiple encounters. In that case we may not want to associate with any particular encounter. So should we link a note to 0…1 encounter?

For the mechanism to store complex notes, I think we should implement this the same way as complex obs. Basically copying the interfaces required for obs and altering them for the notes functionality. The default handler for the complex obs and complex notes could be the same it would just need to implement both the complex notes handler and the complex obs handler interfaces.

Cheers,

Ryan

···

On Wed, Nov 6, 2013 at 12:12 AM, Suranga Kasthurirathne surangakas@gmail.com wrote:

Hi,

As discussed on the SHR call, I though i’d get the conversation on OpenMRS notes started, and see where we can take it from here.

So my two cents are,

As Ryan suggested, there are two types of notes,

  1. Simple Notes - Store text notes that links to the patient record in the OpenMRS data model
  1. Complex Notes - documents (PDFs, scanned docs) could be stored as a complex note rather than as an observation (a PDF-MIME type)

As per my previous discussion on the OpenMRS mailing list, an OpenMRS Note is tied to an encounter. A Note is not intended to store an HL7 NTE segment - we have obs.comment for that.

Does this correspond with what you had in mind ?

And also, is it our intention to store complex notes using the same mechanism used to store unstructured complex obs ? if so, we may have some trouble getting our changes into the OpenMRS core…

Suranga


Best Regards,

You received this message because you are subscribed to the Google Groups “Shared Health Record (OpenHIE)” group.

To unsubscribe from this group and stop receiving emails from it, send an email to openhie-shr+unsubscribe@googlegroups.com.

For more options, visit https://groups.google.com/groups/opt_out.


Ryan Crichton

Software Developer, Jembi Health Systems | SOUTH AFRICA

Mobile: +27845829934 | Skype: ryan.graham.crichton
E-mail: ryan@jembi.org

Hi

Just to clear things up... in the current data model there is no link between a Note and an Encounter. We would have to alter the Encounter table as well to have a Notes field. I'm still not quite sure what functionality a Note would provide to an Encounter that an Observation can't. You can have a Text or a Complex Concept that is a called "Visit Note" and just store whatever you want against that.
I can see a functionality for a Note that is completely separate from Encounters and only tied to a patient. If a doctor wants to add something about a patient that he wants other doctors to be aware of, that is not part of an encounter. Basically the digital version of putting a Post-it on a medical ledger?

Kari Schoonbee
Software Engineer - Jembi Health Systems | SOUTH AFRICA
Mobile: +27 83 488 3025 | Office: +27 21 701 0939
E-mail: kari@jembi.org

···

On 06 Nov 2013, at 9:16 AM, Ryan Crichton <ryan@jembi.org> wrote:

Hi Suranga,

Thanks for getting this started. I agree with those two types of notes.

So what do people think of having an note attached to an encounter? Should this always be the case or should this be optional? Is there a case where we would not want a note to be attached to an encounter. I can think of summary documents that may span multiple encounters. In that case we may not want to associate with any particular encounter. So should we link a note to 0..1 encounter?

For the mechanism to store complex notes, I think we should implement this the same way as complex obs. Basically copying the interfaces required for obs and altering them for the notes functionality. The default handler for the complex obs and complex notes could be the same it would just need to implement both the complex notes handler and the complex obs handler interfaces.

Cheers,
Ryan

On Wed, Nov 6, 2013 at 12:12 AM, Suranga Kasthurirathne <surangakas@gmail.com> wrote:

Hi,

As discussed on the SHR call, I though i'd get the conversation on OpenMRS notes started, and see where we can take it from here.
So my two cents are,

As Ryan suggested, there are two types of notes,
1) Simple Notes - Store text notes that links to the patient record in the OpenMRS data model
2) Complex Notes - documents (PDFs, scanned docs) could be stored as a complex note rather than as an observation (a PDF-MIME type)

As per my previous discussion on the OpenMRS mailing list, an OpenMRS Note is tied to an encounter. A Note is not intended to store an HL7 NTE segment - we have obs.comment for that.
Does this correspond with what you had in mind ?
And also, is it our intention to store complex notes using the same mechanism used to store unstructured complex obs ? if so, we may have some trouble getting our changes into the OpenMRS core...

--
Best Regards,

Suranga

--
You received this message because you are subscribed to the Google Groups "Shared Health Record (OpenHIE)" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openhie-shr+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

--
Ryan Crichton
Software Developer, Jembi Health Systems | SOUTH AFRICA
Mobile: +27845829934 | Skype: ryan.graham.crichton
E-mail: ryan@jembi.org

--
You received this message because you are subscribed to the Google Groups "Shared Health Record (OpenHIE)" group.
To unsubscribe from this group and stop receiving emails from it, send an email to openhie-shr+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

I must admit... I'm having a déjà vu on this topic. May I ask, have we had a
chance to look at the FHIR models regarding this? FHIR is based on HL7v3 (as
its information model) and I know that these exact issues have been modeled
using HL7v3; we use them in our Canadian ERHS information models.

V3's modeling convention is very weird and colourful (and idiosyncratic!)...
that's why I'm wondering if it has been described on the FHIR site, which I
think would be more digestible.

Just a thought...

Derek.

Derek Ritz, P.Eng., CPHIMS-CA
ecGroup Inc.
+1 (905) 515-0045
www.ecgroupinc.com

Let me offer a different perspective.

I think that, in HL7 V3, a note would simply be captured as an observation
whose value was text.

The distinction between note and observation, which I do not think is worth
making, is - I think - an artifact of HL7 Version 2 which has an NTE segment
to allow sender to put whatever in various places, and an OBX segment that
allows senders to put laboratory observations and/or whatever in several
places.

So, the question is, is there an intrinsic value is distinguishing note and
observation? To put it differently, what is that value?

Mead

···

-----Original Message-----
From: openhie-shr@googlegroups.com [mailto:openhie-shr@googlegroups.com] On
Behalf Of Derek Ritz (ecGroup)
Sent: Wednesday, November 06, 2013 9:22 AM
To: 'Kari Schoonbee'; 'Ryan Crichton'
Cc: 'Suranga Kasthurirathne'; 'openhie-shr'
Subject: RE: Discussion on improvements for simple/complex notes

I must admit... I'm having a déjà vu on this topic. May I ask, have we had a
chance to look at the FHIR models regarding this? FHIR is based on HL7v3 (as
its information model) and I know that these exact issues have been modeled
using HL7v3; we use them in our Canadian ERHS information models.

V3's modeling convention is very weird and colourful (and idiosyncratic!)...
that's why I'm wondering if it has been described on the FHIR site, which I
think would be more digestible.

Just a thought...

Derek.

Derek Ritz, P.Eng., CPHIMS-CA
ecGroup Inc.
+1 (905) 515-0045
www.ecgroupinc.com

--
You received this message because you are subscribed to the Google Groups
"Shared Health Record (OpenHIE)" group.
To unsubscribe from this group and stop receiving emails from it, send an
email to openhie-shr+unsubscribe@googlegroups.com.
For more options, visit https://groups.google.com/groups/opt_out.

Aloha,

So to discuss the parts that i’m familiar with,

@Ryan, based on your comment that a note can be tied to multiple encounters, can we say multiple encounters == visit, or do you mean multiple encounters spanning across different visits ?

I’m pretty sure we want to make notes optional, especially if we want it to go into the core.

In terms of why we cant use Obs to store notes, Burke’s original comments on a thread I had on the dev list were "While observations could be used to store narrative notes and text-based observations and some implementations have started down this path, our intent is to eventually store narrative text separately, since a different table structure (or even non-SQL options) can better handle narrative documents than an obs table designed to handle discrete data. "

Considering that we’re now looking at non-sql support for complex obs, i’m sure we can re-discuss this with him…

Best regards,

Suranga

···

On Wed, Nov 6, 2013 at 9:21 AM, Derek Ritz (ecGroup) derek.ritz@ecgroupinc.com wrote:

I must admit… I’m having a déjà vu on this topic. May I ask, have we had a

chance to look at the FHIR models regarding this? FHIR is based on HL7v3 (as

its information model) and I know that these exact issues have been modeled

using HL7v3; we use them in our Canadian ERHS information models.

V3’s modeling convention is very weird and colourful (and idiosyncratic!)…

that’s why I’m wondering if it has been described on the FHIR site, which I

think would be more digestible.

Just a thought…

Derek.

Derek Ritz, P.Eng., CPHIMS-CA

ecGroup Inc.

+1 (905) 515-0045

www.ecgroupinc.com

You received this message because you are subscribed to the Google Groups “Shared Health Record (OpenHIE)” group.

To unsubscribe from this group and stop receiving emails from it, send an email to openhie-shr+unsubscribe@googlegroups.com.

For more options, visit https://groups.google.com/groups/opt_out.


Best Regards,

Suranga

Hello,
This is a good discussion, and one where my experience and views are bit orthogonal to the comments made. This is how I see notes in general terms.

I define a note as a unit of unstructured text. I do not include data types other than text as being notes, they are other wonderful thing. Notes comes in a wide range of character lengths and how they are recorded/captured, stored, processed and retrieved is highly dependent on their size and the IS design.

In healthcare, notes come from several sources including transcribed dictation (ex: discharge summary or path report), provider typed short answers (ex: sig in an eprescription or chief complaint or description of an observation), fabrication from template and picklist data entry (ex: the sig or PE findings), and clerical entry of written provider notes. In my work we generally refer to large pieces of text, typically dictations, as ‘notes’ to help distinguish from short pieces of unstructured text, that we generally call ‘text’.

All notes and text are linked to other data by identifiers and datetime stamps. Some notes are values of attributes, some are (in essence) the entities themselves, and some are entity-centric, some encounter-centric, and others are patient-centric, because for me, notes means chunks of text.

Handling of notes is size and IS design dependent - short text (< 128 char, or < 256 char, whatever) is like coded or numeric data, just plopped into the correct db cell, while big text is specially stored, sometimes in tables by themselves, in unique structures such as CLOBs or the like. Either way, the identifiers of the entity or encounter or patient (or other objects) and their datetime stamps allow access, processing and retrieval of the text data. That processing might involve sql ‘like’ expressions or regular expression, or serious NLP, or just retrieval as the unstructured text itself.

On the encounter level, when speaking of specific notes, we try to be explicit in the name, such as Encounter Note, or Radiology Note. These are examples of encounters {{little ‘e’ encounter = is an interaction, event or activity involving a healthcare organization
or its agent, and a patient or their health record that generates data to be
stored in a database. Each
encounter occurs at a specific place and time, with a specific part or agent of
the healthcare organization, and for a specific reason.}} where the valuable data is embedded in a large amount of unstructured text, whose structure, as well as encounter attributes other than the note, depend on the Note type.

On the entity level, we often add a short note or comment attribute in the design of an entity - always optional and rarely populated, but available for anything the data creation or entry person wants to say to annotate the otherwise highly structured entity.

Thanks,

David

···

On Tuesday, November 5, 2013 5:12:57 PM UTC-5, Suranga Kasthurirathne wrote:

Hi,

As discussed on the SHR call, I though i’d get the conversation on OpenMRS notes started, and see where we can take it from here.

So my two cents are,

As Ryan suggested, there are two types of notes,

  1. Simple Notes - Store text notes that links to the patient record in the OpenMRS data model
  1. Complex Notes - documents (PDFs, scanned docs) could be stored as a complex note rather than as an observation (a PDF-MIME type)

As per my previous discussion on the OpenMRS mailing list, an OpenMRS Note is tied to an encounter. A Note is not intended to store an HL7 NTE segment - we have obs.comment for that.

Does this correspond with what you had in mind ?

And also, is it our intention to store complex notes using the same mechanism used to store unstructured complex obs ? if so, we may have some trouble getting our changes into the OpenMRS core…


Best Regards,

Suranga

Hi All,

There has been some very valuable discussion here. I think it has come to a stage where we should be discussing this more closely with the OpenMRS community. I’m going to post our discussion there in a few mins time so we can continue discussing this with the wider community. I’d urge you to join the OpenMRS developers mailing lists to be able to continue to contribute to this discussion without your email bouncing. You can do that here: https://wiki.openmrs.org/display/RES/Mailing+Lists if you are not already part of that mailing list list.

Cheers,

Ryan

···

On Sun, Nov 10, 2013 at 6:30 PM, David Aronow uncle0wacko@gmail.com wrote:

Hello,
This is a good discussion, and one where my experience and views are bit orthogonal to the comments made. This is how I see notes in general terms.

I define a note as a unit of unstructured text. I do not include data types other than text as being notes, they are other wonderful thing. Notes comes in a wide range of character lengths and how they are recorded/captured, stored, processed and retrieved is highly dependent on their size and the IS design.

In healthcare, notes come from several sources including transcribed dictation (ex: discharge summary or path report), provider typed short answers (ex: sig in an eprescription or chief complaint or description of an observation), fabrication from template and picklist data entry (ex: the sig or PE findings), and clerical entry of written provider notes. In my work we generally refer to large pieces of text, typically dictations, as ‘notes’ to help distinguish from short pieces of unstructured text, that we generally call ‘text’.

All notes and text are linked to other data by identifiers and datetime stamps. Some notes are values of attributes, some are (in essence) the entities themselves, and some are entity-centric, some encounter-centric, and others are patient-centric, because for me, notes means chunks of text.

Handling of notes is size and IS design dependent - short text (< 128 char, or < 256 char, whatever) is like coded or numeric data, just plopped into the correct db cell, while big text is specially stored, sometimes in tables by themselves, in unique structures such as CLOBs or the like. Either way, the identifiers of the entity or encounter or patient (or other objects) and their datetime stamps allow access, processing and retrieval of the text data. That processing might involve sql ‘like’ expressions or regular expression, or serious NLP, or just retrieval as the unstructured text itself.

On the encounter level, when speaking of specific notes, we try to be explicit in the name, such as Encounter Note, or Radiology Note. These are examples of encounters {{little ‘e’ encounter = is an interaction, event or activity involving a healthcare organization
or its agent, and a patient or their health record that generates data to be
stored in a database. Each
encounter occurs at a specific place and time, with a specific part or agent of
the healthcare organization, and for a specific reason.}} where the valuable data is embedded in a large amount of unstructured text, whose structure, as well as encounter attributes other than the note, depend on the Note type.

On the entity level, we often add a short note or comment attribute in the design of an entity - always optional and rarely populated, but available for anything the data creation or entry person wants to say to annotate the otherwise highly structured entity.

Thanks,

David

On Tuesday, November 5, 2013 5:12:57 PM UTC-5, Suranga Kasthurirathne wrote:

Hi,

As discussed on the SHR call, I though i’d get the conversation on OpenMRS notes started, and see where we can take it from here.

So my two cents are,

As Ryan suggested, there are two types of notes,

  1. Simple Notes - Store text notes that links to the patient record in the OpenMRS data model
  1. Complex Notes - documents (PDFs, scanned docs) could be stored as a complex note rather than as an observation (a PDF-MIME type)

As per my previous discussion on the OpenMRS mailing list, an OpenMRS Note is tied to an encounter. A Note is not intended to store an HL7 NTE segment - we have obs.comment for that.

Does this correspond with what you had in mind ?

And also, is it our intention to store complex notes using the same mechanism used to store unstructured complex obs ? if so, we may have some trouble getting our changes into the OpenMRS core…


Best Regards,

Suranga

You received this message because you are subscribed to the Google Groups “Shared Health Record (OpenHIE)” group.

To unsubscribe from this group and stop receiving emails from it, send an email to openhie-shr+unsubscribe@googlegroups.com.

For more options, visit https://groups.google.com/groups/opt_out.


Ryan Crichton

Software Developer, Jembi Health Systems | SOUTH AFRICA

Mobile: +27845829934 | Skype: ryan.graham.crichton
E-mail: ryan@jembi.org

Hi everyone,

While we have been thinking about creating a Shared Health Record out of OpenMRS we have begun thinking about how notes should be implemented in OpenMRS. We have had a bit of discussion around this topic and you can read below for some more details of that discussion.

Basically, there are 2 main use cases for storing notes:

  1. Storing narrative text in a patient medical record such as discharge summaries, clinical notes etc.
  2. The second is to store other unstructured information about a patient that could be classified as a note. Such as a scanned document (image), CDA document (XML), etc.
    We have so far been thinking of implementing notes functionality with the addition of “complex notes” similar to the complex obs functionality to store unstructured data. We have faced a few difficulties in thinking about notes. I’ve listed some here but refer to the below discussion for more details.
  • At what point should a note be attached in the OpenMRS data model? - ie, does it belong to an encounter? Does it belong to a vists? How should it be linked to a patients record?
  • Does it make sense to store unstructured content such as images and CDA content in this sort of structure?
    To those who were apart of the original discussion, perhaps you could do a better job of expanding upon these issues and your thoughts on notes?

What are other people thoughts on storing notes in OpenMRS.

Cheers,

Ryan

···

On Mon, Nov 11, 2013 at 12:42 PM, Ryan Crichton ryan@jembi.org wrote:

Hi All,

There has been some very valuable discussion here. I think it has come to a stage where we should be discussing this more closely with the OpenMRS community. I’m going to post our discussion there in a few mins time so we can continue discussing this with the wider community. I’d urge you to join the OpenMRS developers mailing lists to be able to continue to contribute to this discussion without your email bouncing. You can do that here: https://wiki.openmrs.org/display/RES/Mailing+Lists if you are not already part of that mailing list list.

Cheers,

Ryan


Ryan Crichton

Software Developer, Jembi Health Systems | SOUTH AFRICA

Mobile: +27845829934 | Skype: ryan.graham.crichton
E-mail: ryan@jembi.org

On Sun, Nov 10, 2013 at 6:30 PM, David Aronow uncle0wacko@gmail.com wrote:

Hello,
This is a good discussion, and one where my experience and views are bit orthogonal to the comments made. This is how I see notes in general terms.

I define a note as a unit of unstructured text. I do not include data types other than text as being notes, they are other wonderful thing. Notes comes in a wide range of character lengths and how they are recorded/captured, stored, processed and retrieved is highly dependent on their size and the IS design.

In healthcare, notes come from several sources including transcribed dictation (ex: discharge summary or path report), provider typed short answers (ex: sig in an eprescription or chief complaint or description of an observation), fabrication from template and picklist data entry (ex: the sig or PE findings), and clerical entry of written provider notes. In my work we generally refer to large pieces of text, typically dictations, as ‘notes’ to help distinguish from short pieces of unstructured text, that we generally call ‘text’.

All notes and text are linked to other data by identifiers and datetime stamps. Some notes are values of attributes, some are (in essence) the entities themselves, and some are entity-centric, some encounter-centric, and others are patient-centric, because for me, notes means chunks of text.

Handling of notes is size and IS design dependent - short text (< 128 char, or < 256 char, whatever) is like coded or numeric data, just plopped into the correct db cell, while big text is specially stored, sometimes in tables by themselves, in unique structures such as CLOBs or the like. Either way, the identifiers of the entity or encounter or patient (or other objects) and their datetime stamps allow access, processing and retrieval of the text data. That processing might involve sql ‘like’ expressions or regular expression, or serious NLP, or just retrieval as the unstructured text itself.

On the encounter level, when speaking of specific notes, we try to be explicit in the name, such as Encounter Note, or Radiology Note. These are examples of encounters {{little ‘e’ encounter = is an interaction, event or activity involving a healthcare organization
or its agent, and a patient or their health record that generates data to be
stored in a database. Each
encounter occurs at a specific place and time, with a specific part or agent of
the healthcare organization, and for a specific reason.}} where the valuable data is embedded in a large amount of unstructured text, whose structure, as well as encounter attributes other than the note, depend on the Note type.

On the entity level, we often add a short note or comment attribute in the design of an entity - always optional and rarely populated, but available for anything the data creation or entry person wants to say to annotate the otherwise highly structured entity.

Thanks,

David

On Tuesday, November 5, 2013 5:12:57 PM UTC-5, Suranga Kasthurirathne wrote:

Hi,

As discussed on the SHR call, I though i’d get the conversation on OpenMRS notes started, and see where we can take it from here.

So my two cents are,

As Ryan suggested, there are two types of notes,

  1. Simple Notes - Store text notes that links to the patient record in the OpenMRS data model
  1. Complex Notes - documents (PDFs, scanned docs) could be stored as a complex note rather than as an observation (a PDF-MIME type)

As per my previous discussion on the OpenMRS mailing list, an OpenMRS Note is tied to an encounter. A Note is not intended to store an HL7 NTE segment - we have obs.comment for that.

Does this correspond with what you had in mind ?

And also, is it our intention to store complex notes using the same mechanism used to store unstructured complex obs ? if so, we may have some trouble getting our changes into the OpenMRS core…


Best Regards,

Suranga

You received this message because you are subscribed to the Google Groups “Shared Health Record (OpenHIE)” group.

To unsubscribe from this group and stop receiving emails from it, send an email to openhie-shr+unsubscribe@googlegroups.com.

For more options, visit https://groups.google.com/groups/opt_out.


Ryan Crichton

Software Developer, Jembi Health Systems | SOUTH AFRICA

Mobile: +27845829934 | Skype: ryan.graham.crichton

E-mail: ryan@jembi.org

Aloha,

Now that we have a good idea of what we want, shall we discuss how we want to complete this ?

  1. I’m assuming this work will be fully incorporated into the core.

  2. Our complex note will reflect existing complex obs design (complex note handlers etc.).

  3. As per Burke’s recommendation, an encounter should have zero to many notes.

Do we have enough info to begin work on this ?

···

On Mon, Nov 11, 2013 at 1:06 PM, Burke Mamlin burke@openmrs.org wrote:

-Burke

Encounters as clinical transactions can have any combination of notes, observations, and/or orders, which means 0-to-n of each. So, an encounter that comprises just one narrative text (note) would be fine/expected (e.g., a nutritionist recording a consult note for the patient in the hospital, a specialist completing a consult note for a referral, etc.)

OpenMRS Developers: http://go.openmrs.org/dev

Post: dev@openmrs.org | Unsubscribe: dev+unsubscribe@openmrs.org

Manage your OpenMRS subscriptions at https://id.openmrs.org/

To unsubscribe from this group and stop receiving emails from it, send an email to dev+unsubscribe@openmrs.org.


Best Regards,

Suranga

On Mon, Nov 11, 2013 at 11:25 AM, Jeremy Keiper jeremy@openmrs.org wrote:

Burke, at the minimum, would you say that encounter could house zero observations but one large narrative text (note)? I think so, but want to be sure that’s acceptable. In fact, I don’t think any observations are currently required for encounters.

OpenMRS Developers: http://go.openmrs.org/dev

Post: dev@openmrs.org | Unsubscribe: dev+unsubscribe@openmrs.org

Manage your OpenMRS subscriptions at https://id.openmrs.org/

Jeremy Keiper
OpenMRS Core Developer
AMPATH / IU-Kenya Support

On Mon, Nov 11, 2013 at 8:49 AM, Burke Mamlin burke@openmrs.org wrote:

For OpenMRS:

  • An encounter represents a clinical transaction – a provider sitting down and entering any combination of notes, orders, observations for a patient.
  • A visit represents a collection of one or more encounters into a real-world patient visit (e.g., a clinic visit or a hospitalization).
    We see a note as part of an encounter. See Notes within OpenMRS (Design Page) for some initial design ideas for notes within OpenMRS. I didn’t have notes from our conversation at OMRS13, so I may have dis-remembered some points. :slight_smile:

-Burke

OpenMRS Developers: http://go.openmrs.org/dev

Post: dev@openmrs.org | Unsubscribe: dev+unsubscribe@openmrs.org

Manage your OpenMRS subscriptions at https://id.openmrs.org/

On Mon, Nov 11, 2013 at 6:00 AM, Ryan Crichton ryan@jembi.org wrote:

Hi everyone,

While we have been thinking about creating a Shared Health Record out of OpenMRS we have begun thinking about how notes should be implemented in OpenMRS. We have had a bit of discussion around this topic and you can read below for some more details of that discussion.

Basically, there are 2 main use cases for storing notes:

  1. Storing narrative text in a patient medical record such as discharge summaries, clinical notes etc.
  2. The second is to store other unstructured information about a patient that could be classified as a note. Such as a scanned document (image), CDA document (XML), etc.
    We have so far been thinking of implementing notes functionality with the addition of “complex notes” similar to the complex obs functionality to store unstructured data. We have faced a few difficulties in thinking about notes. I’ve listed some here but refer to the below discussion for more details.
  • At what point should a note be attached in the OpenMRS data model? - ie, does it belong to an encounter? Does it belong to a vists? How should it be linked to a patients record?
  • Does it make sense to store unstructured content such as images and CDA content in this sort of structure?
    To those who were apart of the original discussion, perhaps you could do a better job of expanding upon these issues and your thoughts on notes?

What are other people thoughts on storing notes in OpenMRS.

Cheers,

Ryan

OpenMRS Developers: http://go.openmrs.org/dev

Post: dev@openmrs.org | Unsubscribe: dev+unsubscribe@openmrs.org

Manage your OpenMRS subscriptions at https://id.openmrs.org/

On Mon, Nov 11, 2013 at 12:42 PM, Ryan Crichton ryan@jembi.org wrote:

Hi All,

There has been some very valuable discussion here. I think it has come to a stage where we should be discussing this more closely with the OpenMRS community. I’m going to post our discussion there in a few mins time so we can continue discussing this with the wider community. I’d urge you to join the OpenMRS developers mailing lists to be able to continue to contribute to this discussion without your email bouncing. You can do that here: https://wiki.openmrs.org/display/RES/Mailing+Lists if you are not already part of that mailing list list.

Cheers,

Ryan

Ryan Crichton

Software Developer, Jembi Health Systems | SOUTH AFRICA

Mobile: +27845829934 | Skype: ryan.graham.crichton

E-mail: ryan@jembi.org

On Sun, Nov 10, 2013 at 6:30 PM, David Aronow uncle0wacko@gmail.com wrote:

Hello,
This is a good discussion, and one where my experience and views are bit orthogonal to the comments made. This is how I see notes in general terms.

I define a note as a unit of unstructured text. I do not include data types other than text as being notes, they are other wonderful thing. Notes comes in a wide range of character lengths and how they are recorded/captured, stored, processed and retrieved is highly dependent on their size and the IS design.

In healthcare, notes come from several sources including transcribed dictation (ex: discharge summary or path report), provider typed short answers (ex: sig in an eprescription or chief complaint or description of an observation), fabrication from template and picklist data entry (ex: the sig or PE findings), and clerical entry of written provider notes. In my work we generally refer to large pieces of text, typically dictations, as ‘notes’ to help distinguish from short pieces of unstructured text, that we generally call ‘text’.

All notes and text are linked to other data by identifiers and datetime stamps. Some notes are values of attributes, some are (in essence) the entities themselves, and some are entity-centric, some encounter-centric, and others are patient-centric, because for me, notes means chunks of text.

Handling of notes is size and IS design dependent - short text (< 128 char, or < 256 char, whatever) is like coded or numeric data, just plopped into the correct db cell, while big text is specially stored, sometimes in tables by themselves, in unique structures such as CLOBs or the like. Either way, the identifiers of the entity or encounter or patient (or other objects) and their datetime stamps allow access, processing and retrieval of the text data. That processing might involve sql ‘like’ expressions or regular expression, or serious NLP, or just retrieval as the unstructured text itself.

On the encounter level, when speaking of specific notes, we try to be explicit in the name, such as Encounter Note, or Radiology Note. These are examples of encounters {{little ‘e’ encounter = is an interaction, event or activity involving a healthcare organization
or its agent, and a patient or their health record that generates data to be
stored in a database. Each
encounter occurs at a specific place and time, with a specific part or agent of
the healthcare organization, and for a specific reason.}} where the valuable data is embedded in a large amount of unstructured text, whose structure, as well as encounter attributes other than the note, depend on the Note type.

On the entity level, we often add a short note or comment attribute in the design of an entity - always optional and rarely populated, but available for anything the data creation or entry person wants to say to annotate the otherwise highly structured entity.

Thanks,

David

On Tuesday, November 5, 2013 5:12:57 PM UTC-5, Suranga Kasthurirathne wrote:

Hi,

As discussed on the SHR call, I though i’d get the conversation on OpenMRS notes started, and see where we can take it from here.

So my two cents are,

As Ryan suggested, there are two types of notes,

  1. Simple Notes - Store text notes that links to the patient record in the OpenMRS data model
  1. Complex Notes - documents (PDFs, scanned docs) could be stored as a complex note rather than as an observation (a PDF-MIME type)

As per my previous discussion on the OpenMRS mailing list, an OpenMRS Note is tied to an encounter. A Note is not intended to store an HL7 NTE segment - we have obs.comment for that.

Does this correspond with what you had in mind ?

And also, is it our intention to store complex notes using the same mechanism used to store unstructured complex obs ? if so, we may have some trouble getting our changes into the OpenMRS core…


Best Regards,

Suranga

You received this message because you are subscribed to the Google Groups “Shared Health Record (OpenHIE)” group.

To unsubscribe from this group and stop receiving emails from it, send an email to openhie-shr+unsubscribe@googlegroups.com.

For more options, visit https://groups.google.com/groups/opt_out.


Ryan Crichton

Software Developer, Jembi Health Systems | SOUTH AFRICA

Mobile: +27845829934 | Skype: ryan.graham.crichton

E-mail: ryan@jembi.org

Hi Suranga,

While I agree that from an OpenMRS general use perspective, the designs seem to be sound.

What I’m left wondering after thinking around this for a while is whether CDA document should really be stored as clinical notes. From reading the descriptions of clinical notes on the design doc page it seems like notes are really shorter narrative text about a patient where as CDA documents are just that; documents. I’m starting to think that a document is quite different from a note. Document may even contain notes and observations. Perhaps a document should be a separate data structure within OpenMRS (perhaps supplied by a module).

What do others think? Is storing of CDA document as notes in OpenMRS appropriate?

Cheers,
Ryan

···

On Tue, Nov 12, 2013 at 10:49 PM, Suranga Kasthurirathne surangakas@gmail.com wrote:

Aloha,

Now that we have a good idea of what we want, shall we discuss how we want to complete this ?

  1. I’m assuming this work will be fully incorporated into the core.
  1. Our complex note will reflect existing complex obs design (complex note handlers etc.).
  1. As per Burke’s recommendation, an encounter should have zero to many notes.

Do we have enough info to begin work on this ?

You received this message because you are subscribed to the Google Groups “Shared Health Record (OpenHIE)” group.

To unsubscribe from this group and stop receiving emails from it, send an email to openhie-shr+unsubscribe@googlegroups.com.

For more options, visit https://groups.google.com/groups/opt_out.


Ryan Crichton

Software Developer, Jembi Health Systems | SOUTH AFRICA

Mobile: +27845829934 | Skype: ryan.graham.crichton
E-mail: ryan@jembi.org

On Mon, Nov 11, 2013 at 1:06 PM, Burke Mamlin burke@openmrs.org wrote:

-Burke

Encounters as clinical transactions can have any combination of notes, observations, and/or orders, which means 0-to-n of each. So, an encounter that comprises just one narrative text (note) would be fine/expected (e.g., a nutritionist recording a consult note for the patient in the hospital, a specialist completing a consult note for a referral, etc.)

OpenMRS Developers: http://go.openmrs.org/dev

Post: dev@openmrs.org | Unsubscribe: dev+unsubscribe@openmrs.org

Manage your OpenMRS subscriptions at https://id.openmrs.org/
To unsubscribe from this group and stop receiving emails from it, send an email to dev+unsubscribe@openmrs.org.

Suranga


Best Regards,

On Mon, Nov 11, 2013 at 11:25 AM, Jeremy Keiper jeremy@openmrs.org wrote:

Burke, at the minimum, would you say that encounter could house zero observations but one large narrative text (note)? I think so, but want to be sure that’s acceptable. In fact, I don’t think any observations are currently required for encounters.

OpenMRS Developers: http://go.openmrs.org/dev

Post: dev@openmrs.org | Unsubscribe: dev+unsubscribe@openmrs.org

Manage your OpenMRS subscriptions at https://id.openmrs.org/

Jeremy Keiper
OpenMRS Core Developer
AMPATH / IU-Kenya Support

On Mon, Nov 11, 2013 at 8:49 AM, Burke Mamlin burke@openmrs.org wrote:

For OpenMRS:

  • An encounter represents a clinical transaction – a provider sitting down and entering any combination of notes, orders, observations for a patient.
  • A visit represents a collection of one or more encounters into a real-world patient visit (e.g., a clinic visit or a hospitalization).
    We see a note as part of an encounter. See Notes within OpenMRS (Design Page) for some initial design ideas for notes within OpenMRS. I didn’t have notes from our conversation at OMRS13, so I may have dis-remembered some points. :slight_smile:

-Burke

OpenMRS Developers: http://go.openmrs.org/dev

Post: dev@openmrs.org | Unsubscribe: dev+unsubscribe@openmrs.org

Manage your OpenMRS subscriptions at https://id.openmrs.org/

On Mon, Nov 11, 2013 at 6:00 AM, Ryan Crichton ryan@jembi.org wrote:

Hi everyone,

While we have been thinking about creating a Shared Health Record out of OpenMRS we have begun thinking about how notes should be implemented in OpenMRS. We have had a bit of discussion around this topic and you can read below for some more details of that discussion.

Basically, there are 2 main use cases for storing notes:

  1. Storing narrative text in a patient medical record such as discharge summaries, clinical notes etc.
  2. The second is to store other unstructured information about a patient that could be classified as a note. Such as a scanned document (image), CDA document (XML), etc.
    We have so far been thinking of implementing notes functionality with the addition of “complex notes” similar to the complex obs functionality to store unstructured data. We have faced a few difficulties in thinking about notes. I’ve listed some here but refer to the below discussion for more details.
  • At what point should a note be attached in the OpenMRS data model? - ie, does it belong to an encounter? Does it belong to a vists? How should it be linked to a patients record?
  • Does it make sense to store unstructured content such as images and CDA content in this sort of structure?
    To those who were apart of the original discussion, perhaps you could do a better job of expanding upon these issues and your thoughts on notes?

What are other people thoughts on storing notes in OpenMRS.

Cheers,

Ryan

OpenMRS Developers: http://go.openmrs.org/dev

Post: dev@openmrs.org | Unsubscribe: dev+unsubscribe@openmrs.org

Manage your OpenMRS subscriptions at https://id.openmrs.org/

On Mon, Nov 11, 2013 at 12:42 PM, Ryan Crichton ryan@jembi.org wrote:

Hi All,

There has been some very valuable discussion here. I think it has come to a stage where we should be discussing this more closely with the OpenMRS community. I’m going to post our discussion there in a few mins time so we can continue discussing this with the wider community. I’d urge you to join the OpenMRS developers mailing lists to be able to continue to contribute to this discussion without your email bouncing. You can do that here: https://wiki.openmrs.org/display/RES/Mailing+Lists if you are not already part of that mailing list list.

Cheers,

Ryan

Ryan Crichton

Software Developer, Jembi Health Systems | SOUTH AFRICA

Mobile: +27845829934 | Skype: ryan.graham.crichton

E-mail: ryan@jembi.org

On Sun, Nov 10, 2013 at 6:30 PM, David Aronow uncle0wacko@gmail.com wrote:

Hello,
This is a good discussion, and one where my experience and views are bit orthogonal to the comments made. This is how I see notes in general terms.

I define a note as a unit of unstructured text. I do not include data types other than text as being notes, they are other wonderful thing. Notes comes in a wide range of character lengths and how they are recorded/captured, stored, processed and retrieved is highly dependent on their size and the IS design.

In healthcare, notes come from several sources including transcribed dictation (ex: discharge summary or path report), provider typed short answers (ex: sig in an eprescription or chief complaint or description of an observation), fabrication from template and picklist data entry (ex: the sig or PE findings), and clerical entry of written provider notes. In my work we generally refer to large pieces of text, typically dictations, as ‘notes’ to help distinguish from short pieces of unstructured text, that we generally call ‘text’.

All notes and text are linked to other data by identifiers and datetime stamps. Some notes are values of attributes, some are (in essence) the entities themselves, and some are entity-centric, some encounter-centric, and others are patient-centric, because for me, notes means chunks of text.

Handling of notes is size and IS design dependent - short text (< 128 char, or < 256 char, whatever) is like coded or numeric data, just plopped into the correct db cell, while big text is specially stored, sometimes in tables by themselves, in unique structures such as CLOBs or the like. Either way, the identifiers of the entity or encounter or patient (or other objects) and their datetime stamps allow access, processing and retrieval of the text data. That processing might involve sql ‘like’ expressions or regular expression, or serious NLP, or just retrieval as the unstructured text itself.

On the encounter level, when speaking of specific notes, we try to be explicit in the name, such as Encounter Note, or Radiology Note. These are examples of encounters {{little ‘e’ encounter = is an interaction, event or activity involving a healthcare organization
or its agent, and a patient or their health record that generates data to be
stored in a database. Each
encounter occurs at a specific place and time, with a specific part or agent of
the healthcare organization, and for a specific reason.}} where the valuable data is embedded in a large amount of unstructured text, whose structure, as well as encounter attributes other than the note, depend on the Note type.

On the entity level, we often add a short note or comment attribute in the design of an entity - always optional and rarely populated, but available for anything the data creation or entry person wants to say to annotate the otherwise highly structured entity.

Thanks,

David

On Tuesday, November 5, 2013 5:12:57 PM UTC-5, Suranga Kasthurirathne wrote:

Hi,

As discussed on the SHR call, I though i’d get the conversation on OpenMRS notes started, and see where we can take it from here.

So my two cents are,

As Ryan suggested, there are two types of notes,

  1. Simple Notes - Store text notes that links to the patient record in the OpenMRS data model
  1. Complex Notes - documents (PDFs, scanned docs) could be stored as a complex note rather than as an observation (a PDF-MIME type)

As per my previous discussion on the OpenMRS mailing list, an OpenMRS Note is tied to an encounter. A Note is not intended to store an HL7 NTE segment - we have obs.comment for that.

Does this correspond with what you had in mind ?

And also, is it our intention to store complex notes using the same mechanism used to store unstructured complex obs ? if so, we may have some trouble getting our changes into the OpenMRS core…


Best Regards,

Suranga

You received this message because you are subscribed to the Google Groups “Shared Health Record (OpenHIE)” group.

To unsubscribe from this group and stop receiving emails from it, send an email to openhie-shr+unsubscribe@googlegroups.com.

For more options, visit https://groups.google.com/groups/opt_out.


Ryan Crichton

Software Developer, Jembi Health Systems | SOUTH AFRICA

Mobile: +27845829934 | Skype: ryan.graham.crichton

E-mail: ryan@jembi.org

To clarify further how these terms are used in OpenMRS, could the group tell me:

  • Encounter as a clinical transaction: this makes sense to me, although I prefer ‘transaction’ be changed to ‘interaction’, which to me does not imply a patient is present. Burke’s examples seem to be provider face to patient face, while I have considered activities like provider review of a patient chart or a patient having a follow-up mammogram to be encounters for documentation purposes in the medical record. What are these patient-not-present events called in OpenMRS?

  • Visit has been an Encounter Type in my prior work, not a bundler of encounters, but that type of Encouner in which the provider is face to face with the patient (or their representatives). For the bundler, I have used Admission for inpatients and Episode for outpatients - are these part of the OpenMRS jargon?

  • Notes - I don’t understand why there is a desire to call things ‘notes’ in the medical record that are not ASCII or what is the benefit in a ‘complex note’. Would not Ryan’s second use case, that of the images and CDAs, be easier to handle if images are called, well, images and no longer confounders to how to manage text data type items? They should not be thought of as unstructured, but as other-structured. This might permit a clearer future connection to PACS. CDAs are not exactly unstructured, either. They are structured at a document level, and whether they are treated as attachments or text notes or parsed into individual attributes, the contents of the sections themselves can be treated as structured or unstructured text.

  • How does this discussion relate to storage of URLs as data?

Thanks,

David

···

On Sunday, November 17, 2013 10:54:14 PM UTC-5, burke wrote:

Discharge summaries, Admission history & physicals, progress notes, clinic notes, pathology reports, letters, scanned versions of these documents, etc. would be notes within OpenMRS. Pretty much any patient-specific document. Any patient-specific document that is entirely or primarily a text/narrative (or, in the case of scanned documents, would be if OCR were magical). Any or all of these could be transmitted using CDA documents.

-Burke

On Wed, Nov 13, 2013 at 2:09 AM, Ryan Crichton ry...@jembi.org wrote:

Hi Suranga,

While I agree that from an OpenMRS general use perspective, the designs seem to be sound.

What I’m left wondering after thinking around this for a while is whether CDA document should really be stored as clinical notes. From reading the descriptions of clinical notes on the design doc page it seems like notes are really shorter narrative text about a patient where as CDA documents are just that; documents. I’m starting to think that a document is quite different from a note. Document may even contain notes and observations. Perhaps a document should be a separate data structure within OpenMRS (perhaps supplied by a module).

What do others think? Is storing of CDA document as notes in OpenMRS appropriate?

Cheers,
Ryan

On Tue, Nov 12, 2013 at 10:49 PM, Suranga Kasthurirathne suran...@gmail.com wrote:

Aloha,

Now that we have a good idea of what we want, shall we discuss how we want to complete this ?

  1. I’m assuming this work will be fully incorporated into the core.
  1. Our complex note will reflect existing complex obs design (complex note handlers etc.).
  1. As per Burke’s recommendation, an encounter should have zero to many notes.

Do we have enough info to begin work on this ?

On Mon, Nov 11, 2013 at 1:06 PM, Burke Mamlin bu...@openmrs.org wrote:

-Burke

Encounters as clinical transactions can have any combination of notes, observations, and/or orders, which means 0-to-n of each. So, an encounter that comprises just one narrative text (note) would be fine/expected (e.g., a nutritionist recording a consult note for the patient in the hospital, a specialist completing a consult note for a referral, etc.)

On Mon, Nov 11, 2013 at 11:25 AM, Jeremy Keiper jer...@openmrs.org wrote:

Burke, at the minimum, would you say that encounter could house zero observations but one large narrative text (note)? I think so, but want to be sure that’s acceptable. In fact, I don’t think any observations are currently required for encounters.

Jeremy Keiper
OpenMRS Core Developer
AMPATH / IU-Kenya Support

On Mon, Nov 11, 2013 at 8:49 AM, Burke Mamlin bu...@openmrs.org wrote:

For OpenMRS:

  • An encounter represents a clinical transaction – a provider sitting down and entering any combination of notes, orders, observations for a patient.
  • A visit represents a collection of one or more encounters into a real-world patient visit (e.g., a clinic visit or a hospitalization).
    We see a note as part of an encounter. See Notes within OpenMRS (Design Page) for some initial design ideas for notes within OpenMRS. I didn’t have notes from our conversation at OMRS13, so I may have dis-remembered some points. :slight_smile:

-Burke

On Mon, Nov 11, 2013 at 6:00 AM, Ryan Crichton ry...@jembi.org wrote:

Hi everyone,

While we have been thinking about creating a Shared Health Record out of OpenMRS we have begun thinking about how notes should be implemented in OpenMRS. We have had a bit of discussion around this topic and you can read below for some more details of that discussion.

Basically, there are 2 main use cases for storing notes:

  1. Storing narrative text in a patient medical record such as discharge summaries, clinical notes etc.
  2. The second is to store other unstructured information about a patient that could be classified as a note. Such as a scanned document (image), CDA document (XML), etc.
    We have so far been thinking of implementing notes functionality with the addition of “complex notes” similar to the complex obs functionality to store unstructured data. We have faced a few difficulties in thinking about notes. I’ve listed some here but refer to the below discussion for more details.
  • At what point should a note be attached in the OpenMRS data model? - ie, does it belong to an encounter? Does it belong to a vists? How should it be linked to a patients record?
  • Does it make sense to store unstructured content such as images and CDA content in this sort of structure?
    To those who were apart of the original discussion, perhaps you could do a better job of expanding upon these issues and your thoughts on notes?

What are other people thoughts on storing notes in OpenMRS.

Cheers,

Ryan

On Mon, Nov 11, 2013 at 12:42 PM, Ryan Crichton ry...@jembi.org wrote:

Hi All,

There has been some very valuable discussion here. I think it has come to a stage where we should be discussing this more closely with the OpenMRS community. I’m going to post our discussion there in a few mins time so we can continue discussing this with the wider community. I’d urge you to join the OpenMRS developers mailing lists to be able to continue to contribute to this discussion without your email bouncing. You can do that here: https://wiki.openmrs.org/display/RES/Mailing+Lists if you are not already part of that mailing list list.

Cheers,

Ryan

On Sun, Nov 10, 2013 at 6:30 PM, David Aronow uncle...@gmail.com wrote:

Hello,
This is a good discussion, and one where my experience and views are bit orthogonal to the comments made. This is how I see notes in general terms.

I define a note as a unit of unstructured text. I do not include data types other than text as being notes, they are other wonderful thing. Notes comes in a wide range of character lengths and how they are recorded/captured, stored, processed and retrieved is highly dependent on their size and the IS design.

In healthcare, notes come from several sources including transcribed dictation (ex: discharge summary or path report), provider typed short answers (ex: sig in an eprescription or chief complaint or description of an observation), fabrication from template and picklist data entry (ex: the sig or PE findings), and clerical entry of written provider notes. In my work we generally refer to large pieces of text, typically dictations, as ‘notes’ to help distinguish from short pieces of unstructured text, that we generally call ‘text’.

All notes and text are linked to other data by identifiers and datetime stamps. Some notes are values of attributes, some are (in essence) the entities themselves, and some are entity-centric, some encounter-centric, and others are patient-centric, because for me, notes means chunks of text.

Handling of notes is size and IS design dependent - short text (< 128 char, or < 256 char, whatever) is like coded or numeric data, just plopped into the correct db cell, while big text is specially stored, sometimes in tables by themselves, in unique structures such as CLOBs or the like. Either way, the identifiers of the entity or encounter or patient (or other objects) and their datetime stamps allow access, processing and retrieval of the text data. That processing might involve sql ‘like’ expressions or regular expression, or serious NLP, or just retrieval as the unstructured text itself.

On the encounter level, when speaking of specific notes, we try to be explicit in the name, such as Encounter Note, or Radiology Note. These are examples of encounters {{little ‘e’ encounter = is an interaction, event or activity involving a healthcare organization
or its agent, and a patient or their health record that generates data to be
stored in a database. Each
encounter occurs at a specific place and time, with a specific part or agent of
the healthcare organization, and for a specific reason.}} where the valuable data is embedded in a large amount of unstructured text, whose structure, as well as encounter attributes other than the note, depend on the Note type.

On the entity level, we often add a short note or comment attribute in the design of an entity - always optional and rarely populated, but available for anything the data creation or entry person wants to say to annotate the otherwise highly structured entity.

Thanks,

David

On Tuesday, November 5, 2013 5:12:57 PM UTC-5, Suranga Kasthurirathne wrote:

Hi,

As discussed on the SHR call, I though i’d get the conversation on OpenMRS notes started, and see where we can take it from here.

So my two cents are,

As Ryan suggested, there are two types of notes,

  1. Simple Notes - Store text notes that links to the patient record in the OpenMRS data model
  1. Complex Notes - documents (PDFs, scanned docs) could be stored as a complex note rather than as an observation (a PDF-MIME type)

As per my previous discussion on the OpenMRS mailing list, an OpenMRS Note is tied to an encounter. A Note is not intended to store an HL7 NTE segment - we have obs.comment for that.

Does this correspond with what you had in mind ?

And also, is it our intention to store complex notes using the same mechanism used to store unstructured complex obs ? if so, we may have some trouble getting our changes into the OpenMRS core…


Best Regards,

Suranga

You received this message because you are subscribed to the Google Groups “Shared Health Record (OpenHIE)” group.

To unsubscribe from this group and stop receiving emails from it, send an email to openhie-shr...@googlegroups.com.

For more options, visit https://groups.google.com/groups/opt_out.


Ryan Crichton

Software Developer, Jembi Health Systems | SOUTH AFRICA

Mobile: +27845829934 | Skype: ryan.graham.crichton

E-mail: ry...@jembi.org

Ryan Crichton

Software Developer, Jembi Health Systems | SOUTH AFRICA

Mobile: +27845829934 | Skype: ryan.graham.crichton

E-mail: ry...@jembi.org

OpenMRS Developers: http://go.openmrs.org/dev

Post: d...@openmrs.org | Unsubscribe: dev+uns...@openmrs.org

Manage your OpenMRS subscriptions at https://id.openmrs.org/

OpenMRS Developers: http://go.openmrs.org/dev

Post: d...@openmrs.org | Unsubscribe: dev+uns...@openmrs.org

Manage your OpenMRS subscriptions at https://id.openmrs.org/

OpenMRS Developers: http://go.openmrs.org/dev

Post: d...@openmrs.org | Unsubscribe: dev+uns...@openmrs.org

Manage your OpenMRS subscriptions at https://id.openmrs.org/

OpenMRS Developers: http://go.openmrs.org/dev

Post: d...@openmrs.org | Unsubscribe: dev+uns...@openmrs.org

Manage your OpenMRS subscriptions at https://id.openmrs.org/
To unsubscribe from this group and stop receiving emails from it, send an email to dev+uns...@openmrs.org.

Suranga


Best Regards,

You received this message because you are subscribed to the Google Groups “Shared Health Record (OpenHIE)” group.

To unsubscribe from this group and stop receiving emails from it, send an email to openhie-shr...@googlegroups.com.

For more options, visit https://groups.google.com/groups/opt_out.


Ryan Crichton

Software Developer, Jembi Health Systems | SOUTH AFRICA

Mobile: +27845829934 | Skype: ryan.graham.crichton

E-mail: ry...@jembi.org

OpenMRS Developers: http://go.openmrs.org/dev

Post: d...@openmrs.org | Unsubscribe: dev+uns...@openmrs.org

Manage your OpenMRS subscriptions at https://id.openmrs.org/

Thanks Burke - answers are clear and helpful.
David

···

On Monday, November 25, 2013 3:37:40 PM UTC-5, burke wrote:

On Sun, Nov 24, 2013 at 8:11 PM, David Aronow uncle...@gmail.com wrote:

To clarify further how these terms are used in OpenMRS, could the group tell me:

  • Encounter as a clinical transaction: this makes sense to me, although I prefer ‘transaction’ be changed to ‘interaction’, which to me does not imply a patient is present. Burke’s examples seem to be provider face to patient face, while I have considered activities like provider review of a patient chart or a patient having a follow-up mammogram to be encounters for documentation purposes in the medical record. What are these patient-not-present events called in OpenMRS?

We have used the definition of “clinical transaction” to imply that we’re grouping information according to what is useful clinically rather than at the level of database transactions. There is no assumption that the patient is present for these transactions (e.g., phone encounters, between visits documentation, chart review, etc. are all valid encounters). Any set of orders, observations, notes, and/or data that represent a clinical event (an order session, writing a progress note, performing a chart review, etc.) would be an encounter within OpenMRS. I suppose “clinical interaction” could be used; however, it’s likely more people would infer that a “clinical interaction” directly involves the patient compared to a “clinical transaction”. Regardless, visit, encounter, interaction, transaction, etc. are all subject to interpretation, so we just need to do the best we can to communicate & document our intent. We chose “transaction” because developers understood it as information being recorded at a point in time.

  • Visit has been an Encounter Type in my prior work, not a bundler of encounters, but that type of Encouner in which the provider is face to face with the patient (or their representatives). For the bundler, I have used Admission for inpatients and Episode for outpatients - are these part of the OpenMRS jargon?

We choose “Visit” to bundle one or more encounters, since the terms “Clinic Visit” and “Inpatient Visit” (or “Hospital Visit”) were easy to understand and the same term can apply to any setting (inpatient, clinic, emergency room, etc.). We would consider “Admission” to be one of the first encounters in an inpatient visit. We have defined, but not implemented yet, the notion of “episodes of care” as something that would be focused on a diagnosis or a treatment program and could potentially group encounters across a series of visits. For example, an episode of care for tuberculosis could comprise many clinic visits and one or more hospital stays; an episode of care for pneumonia might include specific encounters during a hospital stay along with a few clinic visits that follow; etc.).

  • Notes - I don’t understand why there is a desire to call things ‘notes’ in the medical record that are not ASCII or what is the benefit in a ‘complex note’. Would not Ryan’s second use case, that of the images and CDAs, be easier to handle if images are called, well, images and no longer confounders to how to manage text data type items? They should not be thought of as unstructured, but as other-structured. This might permit a clearer future connection to PACS. CDAs are not exactly unstructured, either. They are structured at a document level, and whether they are treated as attachments or text notes or parsed into individual attributes, the contents of the sections themselves can be treated as structured or unstructured text.

We presume that all notes would be ASCII, but need to acknowledge that some “ASCII” may come in non-ASCII formats (for example, a simple text or handwritten note that is digitially scanned to be included in the record or a text note that is in a PDF or Word document). We define notes as a narrative that ideally would be in text format – i.e., if we could magically OCR all scanned text and reliably extract meaningful text from DOC and PDF files, then all notes would be text.

  • How does this discussion relate to storage of URLs as data?

Not sure I’m answering the right question, but our model for complex observations, which we figured we could borrow for notes, allows a handler to store whatever it needs to get to the actual payload and this might be a URI. For example, if an implementation had a server that stored scanned documents, a complex note handler could be written that would take a scanned image of a note via the OpenMRS API, send the image to scanned document server and store a complex note entry that contained a brief summary or OCR’d content along with a URI to locate the original scanned image in the remote server. When someone asked for the note within OpenMRS, the task of rendering would be given to the handler and it would show the simple text along with a link (created from the URI) that could get the user to the original scanned image.

Cheers,

-Burke

Thanks,

David

On Sunday, November 17, 2013 10:54:14 PM UTC-5, burke wrote:

Discharge summaries, Admission history & physicals, progress notes, clinic notes, pathology reports, letters, scanned versions of these documents, etc. would be notes within OpenMRS. Pretty much any patient-specific document. Any patient-specific document that is entirely or primarily a text/narrative (or, in the case of scanned documents, would be if OCR were magical). Any or all of these could be transmitted using CDA documents.

-Burke

On Wed, Nov 13, 2013 at 2:09 AM, Ryan Crichton ry...@jembi.org wrote:

Hi Suranga,

While I agree that from an OpenMRS general use perspective, the designs seem to be sound.

What I’m left wondering after thinking around this for a while is whether CDA document should really be stored as clinical notes. From reading the descriptions of clinical notes on the design doc page it seems like notes are really shorter narrative text about a patient where as CDA documents are just that; documents. I’m starting to think that a document is quite different from a note. Document may even contain notes and observations. Perhaps a document should be a separate data structure within OpenMRS (perhaps supplied by a module).

What do others think? Is storing of CDA document as notes in OpenMRS appropriate?

Cheers,
Ryan

On Tue, Nov 12, 2013 at 10:49 PM, Suranga Kasthurirathne suran...@gmail.com wrote:

Aloha,

Now that we have a good idea of what we want, shall we discuss how we want to complete this ?

  1. I’m assuming this work will be fully incorporated into the core.
  1. Our complex note will reflect existing complex obs design (complex note handlers etc.).
  1. As per Burke’s recommendation, an encounter should have zero to many notes.

Do we have enough info to begin work on this ?

On Mon, Nov 11, 2013 at 1:06 PM, Burke Mamlin bu...@openmrs.org wrote:

-Burke

Encounters as clinical transactions can have any combination of notes, observations, and/or orders, which means 0-to-n of each. So, an encounter that comprises just one narrative text (note) would be fine/expected (e.g., a nutritionist recording a consult note for the patient in the hospital, a specialist completing a consult note for a referral, etc.)

On Mon, Nov 11, 2013 at 11:25 AM, Jeremy Keiper jer...@openmrs.org wrote:

Burke, at the minimum, would you say that encounter could house zero observations but one large narrative text (note)? I think so, but want to be sure that’s acceptable. In fact, I don’t think any observations are currently required for encounters.

Jeremy Keiper
OpenMRS Core Developer
AMPATH / IU-Kenya Support

On Mon, Nov 11, 2013 at 8:49 AM, Burke Mamlin bu...@openmrs.org wrote:

For OpenMRS:

  • An encounter represents a clinical transaction – a provider sitting down and entering any combination of notes, orders, observations for a patient.
  • A visit represents a collection of one or more encounters into a real-world patient visit (e.g., a clinic visit or a hospitalization).
    We see a note as part of an encounter. See Notes within OpenMRS (Design Page) for some initial design ideas for notes within OpenMRS. I didn’t have notes from our conversation at OMRS13, so I may have dis-remembered some points. :slight_smile:

-Burke

On Mon, Nov 11, 2013 at 6:00 AM, Ryan Crichton ry...@jembi.org wrote:

Hi everyone,

While we have been thinking about creating a Shared Health Record out of OpenMRS we have begun thinking about how notes should be implemented in OpenMRS. We have had a bit of discussion around this topic and you can read below for some more details of that discussion.

Basically, there are 2 main use cases for storing notes:

  1. Storing narrative text in a patient medical record such as discharge summaries, clinical notes etc.
  2. The second is to store other unstructured information about a patient that could be classified as a note. Such as a scanned document (image), CDA document (XML), etc.
    We have so far been thinking of implementing notes functionality with the addition of “complex notes” similar to the complex obs functionality to store unstructured data. We have faced a few difficulties in thinking about notes. I’ve listed some here but refer to the below discussion for more details.
  • At what point should a note be attached in the OpenMRS data model? - ie, does it belong to an encounter? Does it belong to a vists? How should it be linked to a patients record?
  • Does it make sense to store unstructured content such as images and CDA content in this sort of structure?
    To those who were apart of the original discussion, perhaps you could do a better job of expanding upon these issues and your thoughts on notes?

What are other people thoughts on storing notes in OpenMRS.

Cheers,

Ryan

On Mon, Nov 11, 2013 at 12:42 PM, Ryan Crichton ry...@jembi.org wrote:

Hi All,

There has been some very valuable discussion here. I think it has come to a stage where we should be discussing this more closely with the OpenMRS community. I’m going to post our discussion there in a few mins time so we can continue discussing this with the wider community. I’d urge you to join the OpenMRS developers mailing lists to be able to continue to contribute to this discussion without your email bouncing. You can do that here: https://wiki.openmrs.org/display/RES/Mailing+Lists if you are not already part of that mailing list list.

Cheers,

Ryan

On Sun, Nov 10, 2013 at 6:30 PM, David Aronow uncle...@gmail.com wrote:

Hello,
This is a good discussion, and one where my experience and views are bit orthogonal to the comments made. This is how I see notes in general terms.

I define a note as a unit of unstructured text. I do not include data types other than text as being notes, they are other wonderful thing. Notes comes in a wide range of character lengths and how they are recorded/captured, stored, processed and retrieved is highly dependent on their size and the IS design.

In healthcare, notes come from several sources including transcribed dictation (ex: discharge summary or path report), provider typed short answers (ex: sig in an eprescription or chief complaint or description of an observation), fabrication from template and picklist data entry (ex: the sig or PE findings), and clerical entry of written provider notes. In my work we generally refer to large pieces of text, typically dictations, as ‘notes’ to help distinguish from short pieces of unstructured text, that we generally call ‘text’.

All notes and text are linked to other data by identifiers and datetime stamps. Some notes are values of attributes, some are (in essence) the entities themselves, and some are entity-centric, some encounter-centric, and others are patient-centric, because for me, notes means chunks of text.

Handling of notes is size and IS design dependent - short text (< 128 char, or < 256 char, whatever) is like coded or numeric data, just plopped into the correct db cell, while big text is specially stored, sometimes in tables by themselves, in unique structures such as CLOBs or the like. Either way, the identifiers of the entity or encounter or patient (or other objects) and their datetime stamps allow access, processing and retrieval of the text data. That processing might involve sql ‘like’ expressions or regular expression, or serious NLP, or just retrieval as the unstructured text itself.

On the encounter level, when speaking of specific notes, we try to be explicit in the name, such as Encounter Note, or Radiology Note. These are examples of encounters {{little ‘e’ encounter = is an interaction, event or activity involving a healthcare organization
or its agent, and a patient or their health record that generates data to be
stored in a database. Each
encounter occurs at a specific place and time, with a specific part or agent of
the healthcare organization, and for a specific reason.}} where the valuable data is embedded in a large amount of unstructured text, whose structure, as well as encounter attributes other than the note, depend on the Note type.

On the entity level, we often add a short note or comment attribute in the design of an entity - always optional and rarely populated, but available for anything the data creation or entry person wants to say to annotate the otherwise highly structured entity.

Thanks,

David

On Tuesday, November 5, 2013 5:12:57 PM UTC-5, Suranga Kasthurirathne wrote:

Hi,

As discussed on the SHR call, I though i’d get the conversation on OpenMRS notes started, and see where we can take it from here.

So my two cents are,

As Ryan suggested, there are two types of notes,

  1. Simple Notes - Store text notes that links to the patient record in the OpenMRS data model
  1. Complex Notes - documents (PDFs, scanned docs) could be stored as a complex note rather than as an observation (a PDF-MIME type)

As per my previous discussion on the OpenMRS mailing list, an OpenMRS Note is tied to an encounter. A Note is not intended to store an HL7 NTE segment - we have obs.comment for that.

Does this correspond with what you had in mind ?

And also, is it our intention to store complex notes using the same mechanism used to store unstructured complex obs ? if so, we may have some trouble getting our changes into the OpenMRS core…


Best Regards,

Suranga