The HSLS Data Services team has provided feedback on dozens of Data Management and Sharing Plans (DMSPs) since the new NIH Data Management and Sharing Policy went into effect in January 2023. We’ve also read comments from NIH program officers (POs) to the researchers who submitted their plans, usually when we’ve been asked to help revise a plan in response to PO comments. These tips are intended to supplement our NIH Data Management and Sharing Guide to help you craft a comprehensive yet concise plan that will meet POs’ expectations.
1. Address each element of a Data Management and Sharing Plan completely, in its own section of the plan
The NIH notice “Supplemental Information to the NIH Policy for Data Management and Sharing: Elements of an NIH Data Management and Sharing Plan” outlines six specific elements that should be addressed in a DMSP: data type; related tools, software, and code; standards; data preservation, access, and associated timelines; access, reuse, or distribution considerations; and oversight. We recommend addressing these elements one at a time, taking into account each subcomponent of the element. Refer back to the Supplemental Information document as you write to make sure each subcomponent is addressed.
Ask yourself: Have I answered every aspect of the prompt for this element? Have I answered only the prompt for this element? Don’t put down more information than the specific element is asking for, or you might confuse your reader.
We have occasionally read comments from POs asking questions that are answered in another part of the plan. For example, if you mention in Element 1 (on data types) that you will share your data in a public repository, the PO might leave a comment asking you to specify which repository, even though you’ve named the repository in Element 4 (on data access and preservation). To avoid this, feel free to reference other plan elements by number. For example, you might write in Element 1, “We will share all cleaned data in a public repository (see Element 4 below).”
2. Use a template
We have observed that DMSPs that don’t follow a standardized template come back from the NIH POs with many more comments and questions than plans that do use one. We recommend the templates available through DMP Tool, which include explanations and examples for each required element. To use a template, log into DMP Tool with your Pitt email address, choose NIH as your funder, and select any of the customized templates. The “NIH-Default DMSP” template has more guidance than the two “FDP Pilot” templates. Once you write your plan within DMP Tool, you can export it as a PDF or Word document to attach to your submission.
The NIH has created a format page that outlines each required element. This isn’t as complete as the DMP Tool templates, but it’s a good structure if you’re writing a second or third plan and already feel comfortable with the required elements.
3. Request feedback from HSLS librarians
When you write your plan in DMP Tool, you can send it to HSLS librarians for comments by clicking the “Request Feedback” tab. When we receive your submission, we’ll reach out to you for information about your deadline, then give you an estimated date by which we’ll return comments. You can send plans to us for comments at any stage of the writing process, but please allow extra time for discussion and revision if you’re contacting us at an early stage.
4. Try not to leave “Element 3: Standards” blank, and don’t forget about metadata
Element 3, about common data standards to be applied to the data and associated metadata, seems to be the most troublesome element required in the plan. That may be because “standards” are only vaguely defined in the official notices. While the NIH indicates that applicants may write that no consensus standards exist for their field or data type, in practice it seems like POs find that insufficient. So think expansively. Are you using any standardized data collection instruments, normalization techniques, or file organization formats? Will you use Common Data Elements (CDEs) for any variables? Those are all standards that can be discussed in this section.
If you’re sharing your data in a repository, that repository will usually organize and encode each deposit’s metadata according to a standard like DataCite or Dublin Core. Check the repository website’s help pages or FAQ for information about metadata schemas and standards. If you’re drawing upon an ontology like SNOMED CT or NCBI Taxonomy for your metadata, that would fit here, too.
5. The Data Management and Sharing Plan and the Resource Sharing Plan are two separate documents
The DMSP covers how researchers will organize and disseminate their scientific data, including genomic data that used to be described in a Genomic Data Sharing (GDS) plan. As of 2023, the GDS plan has been folded into the DMSP. However, model organisms and all other non-data resources should still be described in a separate Resource Sharing Plan. Keep your mice out of the DMSP! For more information, see the NIH’s Data Management and Sharing Plan vs. Resource Sharing Plan. (Note that Element 2 does ask about related tools, software, and code needed to access and manipulate the data, and that includes researcher-developed software. If you are sharing transformation or analysis code with your data, discuss it here; you probably don’t need to submit a Resource Sharing Plan. If you are creating an entire software suite of data analysis tools, describe it in a Resource Sharing Plan and mention it in Element 2.)
For more strategies on writing a DMSP, contact HSLS Data Services, attend our upcoming classes, or send us a plan to review.
~Helenmary Sheridan