Merge pull request #1688 from opensafely/iaindillingham/datapast

Add "The Clinical DATAPAST Function"

docs/access-policies.md

@@ -2,3 +2,4 @@ The following policies control access to the OpenSAFELY platform.
 
 * [Policies for Researchers](https://www.opensafely.org/policies-for-researchers/)
 * [Data Access Policy](data-access-policy.md)
+* [The Clinical DATAPAST Function](datapast.md)

docs/datapast.md

@@ -0,0 +1,95 @@
+# The Clinical DATAPAST Function
+
+The Clinical DATAPAST (**Data** **P**ublication, **A**dvice and **S**upport **T**eam) function will act as a distinct but integral part of the publication approval service which was piloted during the pandemic.
+
+**Publication approval = Clinical DATAPAST advice + NHS England approval decision**
+
+*What does the Clinical DATAPAST do?*
+
+The Clinical DATAPAST function will provide advisory feedback to users (and the study leads) from clinical advisors[^1] on any work submitted for publication approval prior to NHS England receiving the work for an approval decision.
+
+Each submitted piece of work will ordinarily receive feedback from two advisors.
+
+*Why does the Clinical DATAPAST exist?*
+
+The Clinical DATAPAST forms one of several mitigations established by the OpenSAFELY service to address the GP profession’s requirements:
+
+- to maintain alignment with the project’s approved purpose;
+- to maintain the quality of the publication content by advising project leads regarding how clinical interactions are coded within GP EHR systems;
+- to help ensure conclusions on clinical policies and services are supported by the data.
+
+*Who is part of the Clinical DATAPAST?*
+
+Prior to full implementation of the Clinical DATAPAST (expected sometime shortly after April 2025), we will roll out an interim process where the regular clinical advisor will be Dr Amir Mehrkar, an NHS GP, Clinical Informatician, and Director of IG and External Relations at OpenSAFELY. He will be supported by other clinicians (such as clinical pharmacists and clinical informaticians), as well as data scientists and epidemiologists with strong experience in GP data.
+
+It is expected that when the Clinical DATAPAST is fully implemented, all advisors will be clinicians or clinical informaticians (or another member of the general practice team appropriate for the study in question); all will have some knowledge of Electronic Health Record data and quantitative research methods, and be independent from the OpenSAFELY service and the Bennett Institute for Applied Data Science (University of Oxford).
+
+The Joint GP IT Committee of the British Medical Association and Royal College of GPs are in support of the interim process, pending full implementation.
+
+*What does the NHS England approval decision assess?*
+
+The primary assessment of NHS England is to determine that the study leads did in fact do what they said they would do as was written in their initial application. Put simply, was the material submitted in-line with the approved project purpose? If not, NHS England will provide advice to the study leads.
+
+To make their judgement, NHS England will use the actual material submitted, including the lay summary, and the information from the DATAPAST templates. As data controller, NHS England may exercise their judgement on how to constructively act on (both internally and externally) any information that could be deemed contentious, for example, information that could materially affect patient safety. NHS England will promptly communicate with study leads if there are any matters of concern.
+
+*How will the interim process work?*
+
+A summary
+
+* There will be a structured DATAPAST template that advisors use to provide advice on your submitted work; academic users may recognise some parallels with the peer review process for journal submission.
+* The template includes sections for users to provide responses.
+* The completed templates, submitted work (with any amendments following users' consideration of the advice), and the lay summary will then be submitted to NHS England for an approval decision.
+* The completed templates, independent of the final NHS England approval decision, will be placed in the public domain (most likely attached to the project on our jobserver, but the exact location is yet to be determined).
+
+The detail
+The [DATAPAST template](https://docs.google.com/document/d/10derqFbS3c0SbZxGnBb1xQ0R7EgcxHYOGxnujTT-4QY/edit?tab=t.0) has four sections:
+
+* Compliance with study purpose and policies for researchers
+* Quality
+* Discussion and conclusions
+* Any other comments
+
+The workflow for submission for publication approval
+
+1. You must ensure your co-pilot is aware of your intent to submit as a list of co-pilot checks need to be actioned prior to submission.
+
+2. All submissions must be read and approved by the study leads, sponsor and line manager (for service evaluation/audit work only) and Principal Investigator and line manager (for research work only).
+
+3. All submissions must be roughly “90%” finalised, with the results and conclusions final.
+
+4. After steps 1-3 are complete, please submit your work (alongside the Project ID), to the [[email protected]](mailto:[email protected]) inbox. Copy in your co-pilot, study leads, and sponsor and line manager or Principal Investigator and line manager, where appropriate. You must specify that study leads, sponsor, Principal Investigator and line manager have read and approved the submission.
+
+5. There is no longer a specific regular Wednesday deadline to submit your work. You can submit at any time. However, over specific holiday periods (such as Christmas, Easter and the summer break, the review approval periods may be longer).
+
+6. Two DATAPAST advisors will be selected to review the submission independently using the DATAPAST template; when completed, they will arrange a moderation discussion.
+
+7. On the DATAPAST template, advisors may, for example, ask questions, point out areas that they consider require further clarification, or make suggestions. Advisors will capture key important content in the template; however, they may also make comments on the actual document (typically less critical comments). Advisors will not copy large bodies of content or tables, graphs or data into the template.
+
+8. Following moderation, users, including the study sponsor, Principal Investigator, line manager and co-pilot, will be emailed the DATAPAST templates. Users are expected to provide written responses in each part of the template where appropriate; if no responses are provided where required, templates will still be published in the public domain and can be forwarded to relevant journals if there are concerns. In addition, the DATAPAST advice may result in users updating the content of their work. Users may reasonably ask their co-pilot for support in addressing any advice.
+
+9. The completed DATAPAST templates, alongside any updated content must be re-submitted for approval. The updated parts must be clearly highlighted. The amendments must also be approved by the study leads, sponsors / line managers (for service evaluation/audit work only) and Principal Investigator (for research work only).
+
+10. The advisors will read the users’ responses and any updated content where appropriate. There may be further discussion between the DATAPAST advisors and users, but it is intended that this to-and-fro is kept to a minimum, and ideally not required following the first round of advice.
+
+11. Lastly, the completed DATAPAST templates, alongside any updated work, and the lay summary of work, will be submitted to NHS England’s publication approval team.
+
+12. NHS England will provide an approval response, or may request a further discussion regarding the submitted work.
+
+13. The completed templates, independent of the final NHS England approval decision, will be placed in the public domain (most likely attached to the project on our jobserver, but the exact location is yet to be determined).
+
+The indicative timelines
+
+14. We anticipate that users will receive the DATAPAST templates from advisors within 4 weeks of users submitting their work for publication approval (subject to annual leave and other service pressures).
+
+15. After the DATAPAST templates have been reviewed and completed by users (alongside any amendments to work, where appropriate), and assuming there is no further round of discussion, the work will be submitted to NHS England for an approval decision.
+
+16. NHS England typically responds within 2 weeks ( subject to annual leave and service pressures).
+
+We will shortly update the [policies for researchers page](https://www.opensafely.org/policies-for-researchers/#details-all-datasets) regarding the new Clinical DATAPAST function, and links to the above information.
+
+Please reach out to your co-pilot in the meantime if you have any questions, or post them in the user slack channel.
+
+The OpenSAFELY Leadership Team.
+
+
+[^1]:  GPs, including clinicians from other relevant professions with the necessary experience and skills to provide project-specific advice