Policies

Having sufficient time to engage in science is critical for adhering to principles of rigor and reproducibility. To that end, QSU scientists work with other members of the research team to jointly define shared goals and reasonable timelines that the team can meet.

• Notify QSU scientists at least 3 months before the submission deadline.

• Early contact is crucial for shaping questions and hypotheses and designing a study (i.e., power calculations).

• QSU scientists must be integral to shaping and refining questions and hypotheses (and are less effective if the study has already been designed).

• Develop a Statistical Analysis Plan (SAP) and register the study with ample time.

• The process (internal review, data cleaning, analysis, interpretation) takes at least 2-3 months.

To uphold the highest levels of data integrity and out of respect and fairness to our already existing collaborators, QSU scientists do not engage in last minute collaborations.

The QSU's scientific and intellectual activities in study design, data management, and analysis must be recognized through authorship. Data science -- including data cleaning and wrangling -- are critical scientific and intellectual activities necessary for achieving rigor, reproducibility, and importantly, replicability.

• Contributions to conception/design, data acquisition/analysis/interpretation.
• Drafting/revising work for intellectual content.
• Final approval of the version to be published.
• Accountability for all aspects of the work.

Projects often involve multiple QSU members, and their contributions must be reflected in authorship. We follow the International Committee of Medical Journal Editors (ICMJE) criteria for authorship. The QSU will not agree to LIMIT THE NUMBER OF AUTHORS WHO HAVE STATISTICAL EXPERTISE, and instead will follow criteria driven by intellectual and scientific contributions, just as is done with other investigators with expertise in other domain areas. 

The QSU is a collaborative group, not a fee-for-service center, and we collaborate only with those who hold similar collaborative principles as reflected in our guidelines.

1International Committee of Medical Journal Editors. Recommendations for the conduct, reporting, editing, and publication of scholarly work in medical journals.

For questions regarding our publication and authorship policy, please contact Dr. Manisha Desai at manishad@stanford.edu.

Note: QSU members should be listed as co-authors if they meet ICMJE criteria and typically do not appear in acknowledgment sections.

• The Principal Investigator (PI) is responsible for all data sharing.

• If the QSU team does not include the PI, QSU will not share any data with anyone other than the PI.

• QSU will facilitate data sharing on behalf of the PI.

• A secure Stanford Medicine Box folder will be set up for data exchange between the PI and QSU member only.

• Only the QSU member and the PI will have access to this Box

• QSU will not share data with the PI's colleagues, post-doctoral fellows, lab members, or sponsors.

• QSU will share the code used to create derived data or analytic files to ensure reproducibility.

• Data sharing must adhere to the HIPAA Minimum Necessary Rule.

• Use Stanford Medicine Box or “secure” Stanford email for sharing data with QSU members.

For more information, please contact your QSU collaborator.

These guidelines were inspired by Jeff Leek's data sharing guide at GitHub.

As a Data Coordinating Center, the QSU follows this policy for sharing findings and results with Stanford collaborators and study teams:

• Post-database lock, sharing findings takes time to ensure accuracy and rigor.

• Rushing the Statistical Analysis Plan post-database lock can lead to errors and compromise rigor and reproducibility.

• Findings will not be shared with the internal Stanford clinical team for at least fourteen (14) days post-database lock.

• This period allows for thorough validation of results before any internal dissemination.

• In most cases, a longer period may be required before sharing results with collaborators.

• Clinical validation of findings must continue for at least an additional fourteen (14) days internally at Stanford prior to being shared externally.

• Only after this period can results be shared externally (e.g., with sponsors, regulatory entities).

This timeline ensures thorough validation and reproducibility of findings before dissemination.

In order to adhere to rigor & reproducibility of our studies, the QSU has established guidelines that all projects must be pre-registered and have an accompanying Statistical Analysis Plan that has been internally reviewed before any analyses are conducted. This is the case for presentations, abstracts, manuscripts or any form of disseminating findings. 

Statistical Analysis Plan Template (PDF)  

Statistical Analysis Plan Template (WORD) 

Pre-register your study