Write implementation protocol

[pnrobinson]

• [x] draft by Peter
• [x] revisions by Amy Yates and Justin Ramsdill: implementation in FHIR/Epic system
• [x] discuss with Chris Chute.

[pnrobinson]

Link to GoogleDoc https://docs.google.com/document/d/1ydNi9X8E-Fb-xdbEzN7mDq6vhrheXFcZ4xBOo30Qwoc/edit?usp=sharing

[pnrobinson]

@cgchute Chris, we are now finished with the draft version of the protocol (the same one you have already started to comment). Could you please now let me know if the protocol is sufficiently detailed to go on to the next step? Thanks!

[cgchute]

Yes, reviewed your changes. Much better. thanks.

[pnrobinson]

@cgchute @mellybelly -- thanks! So how should we now proceed? Is this document enough to approach IRBs at JHU/OHSU and possibly external collaborators?

[cgchute]

Yes, I can proceed with this. We can probably do this we the deidentified data set to which we already had IRB access. We should write a protocol.

[kingmanzhang]

This is great! Let me know if I can be of any help.

On Tue, Mar 19, 2019 at 9:39 PM Christopher Chute notifications@github.com wrote:

Yes, I can proceed with this. We can probably do this we the deidentified data set to which we already had IRB access. We should write a protocol.

— You are receiving this because you were assigned.

Reply to this email directly, view it on GitHub https://github.com/data2health/ehr2HPO.prj/issues/18#issuecomment-474650056, or mute the thread https://github.com/notifications/unsubscribe-auth/AVfMCwWcGqhXqZELWB4S2BWOXu3Vnkykks5vYZFpgaJpZM4bvede .

--

[mellybelly]

tagging @davedorr9 and @schuffr here too if we need to do anything for OHSU.

[pnrobinson]

@davedorr9 @schuffr Will this be possible at OHSU? Do you need more information from me? If it is not possible, please let me know ASAP!

[schuffr]

@davedorr9 @schuffr Will this be possible at OHSU? Do you need more information from me? If it is not possible, please let me know ASAP!

hi all, I think this is possible but I need to be clear on what the "this" is that we are talking about. Peter can you please elaborate? I have (quickly) read the protocol but not sure what the proposed next step is.

[pnrobinson]

@davedorr9 @aeyates The game plan could look something like this

1. Decide on some dataset that we can get access to. We are going to be looking at a large deidentified dataset at JHU with I believe 2.2 million individuals with a mix of diseases. At a minimum we would need a dataset with a range of diseases that we should discuss by phone.
2. Amy Yates (@aeyates) would work with you to determine how our app would interact with the FHIR and/or clinical warehouse and/or files that would contain the data.
3. Our app would run on the data and produce files that contain summary statistics but no individual level data or PHI. This file would be sent to us. Nothing else leaves your firewall.
4. We would run additional analysis on the files sent to us from multiple institutions and perform the final analysis. The algorithmic details are in the main proposal.

[pnrobinson]

@schuffr Sorry, Robert, I forget to tag you in the above response.

[davedorr9]

Hi!

I was recommended by Melissa to engage with this. Given the current mix of activities I have on CD2H, it is imperative I take a back-seat role, mainly related to the overlap between my CRIO and CD2H roles – appropriate access.

Thus, I would add you need a local, more clinical site PI at OHSU who can help identify appropriate uses, help with the IRB, and advocate for appropriate data access. Perhaps Dave E’s technology can help? The use-case that is indirectly described in the work implies is somewhat focused – I think phenotypic subtypes of common diseases or rare phenotypes where the next step may be omics confirmation. Folks who do that kind of translation may be beneficial.

dd

David Dorr, MD, MS, FACMI Chief Research Information Officer, OHSU Professor & Vice Chair, Medical Informatics; Professor, General Internal Medicine & Geriatrics, OHSU PI, Care Management Plus The mission of Care Management Plus is to better understand how data, information, and knowledge can assist in transforming health for our most vulnerable populations. “We are too far out on the Yerkes-Dodson Curvehttps://en.wikipedia.org/wiki/Yerkes%E2%80%93Dodson_law and we need to back off,” – Stuart Slavin e: dorrd@ohsu.edumailto:dorrd@ohsu.edu v: 503-418-2387 w: http://www.ohsu.edu/cmp

From: Peter Robinson notifications@github.com Sent: Tuesday, April 23, 2019 4:02 PM To: data2health/ehr2HPO.prj ehr2HPO.prj@noreply.github.com Cc: David Dorr dorrd@ohsu.edu; Mention mention@noreply.github.com Subject: Re: [data2health/ehr2HPO.prj] Write implementation protocol (#18)

@davedorr9https://github.com/davedorr9 @aeyateshttps://github.com/aeyates The game plan could look something like this

1. Decide on some dataset that we can get access to. We are going to be looking at a large deidentified dataset at JHU with I believe 2.2 million individuals with a mix of diseases. At a minimum we would need a dataset with a range of diseases that we should discuss by phone.
2. Amy Yates (@aeyateshttps://github.com/aeyates) would work with you to determine how our app would interact with the FHIR and/or clinical warehouse and/or files that would contain the data.
3. Our app would run on the data and produce files that contain summary statistics but no individual level data or PHI. This file would be sent to us. Nothing else leaves your firewall.
4. We would run additional analysis on the files sent to us from multiple institutions and perform the final analysis. The algorithmic details are in the main proposal.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/data2health/ehr2HPO.prj/issues/18#issuecomment-486005084, or mute the threadhttps://github.com/notifications/unsubscribe-auth/AGJX7LXVMNFTJGPCEBSZQD3PR6IOHANCNFSM4G5545PA.

[pnrobinson]

@davedorr9 @schuffr @aeyates Thanks, David. Do you have some suggestion as to where to go from here? I am not familiar with the clinicians at OHSU who might be interested in this project, can you give me a tip please?

[schuffr]

Hi all, We can get access to data through the RDW, but we'll need to write an IRB protocol and identify a PI to do so. I am happy to serve in that role. If de-identified data suffice (including date shifting), we can get this data from our cohort discovery (i2b2 instance), unless there are reasons it will not suffice. In that event we can get data from the RDW directly and deID it as needed, but at additional cost. Definitely should have a phone call to flesh out the plan.

This part of the project is definitely doable now as we do not need to touch the clinical environment in any way. But as David D. rightly points out, when this starts deveopling to the point of touching the clinical environment, we'll need to have a clinical champion identified with a clear clinical problem to solve, and then we'll have to go through some processes to move that part forward. Exciting!

[mellybelly]

That would be terrific Rob!

[aeyates]

@pnrobinson Met with @schuffr and @ramsdill this afternoon to discuss next steps. Let's plan another call with the 4 of us for ironing out IRB details targeting our research data. As we discussed with you the other day, let's finalize the protocol by front-loading with a one-page project plan and some of the technical details to make clear that only aggregate data will be shared. Can we arrange a Zoom meeting next week?

[pnrobinson]

sounds great. I will reply about making the appt per email.

[pnrobinson]

I have moved the link to the protocol to the README.

[schuffr]

Hi @pnrobinson , I may be missing some emails, but the meeting we had scheduled does not appear to have been rescheduled. We should get this rescheduled soon if we are to move this forward.How can we best move this forward? should be with @aeyates and @ramsdill

sounds great. I will reply about making the appt per email.

[pnrobinson]

i will reply via email about the scheduling