1.4.3.2 Industry-academia: Research as collaboration
Background
The practices outlined below have been developed by a task force of academic and industry members of the EQIPD consortium.
These practices are intended to improve the traceability and integrity of the data obtained from the collaboration between [academic organization] and [industry partner]. They aim to:
- facilitate decision-making,
- minimize bias and errors in the collection, reporting or representation of such information, and
- create reliable scientific and supporting evidence in resulting patents and other types of intellectual property as well as publications.
The experimental record and its thorough description is the ultimate source of information and documentation regarding the experiment. Therefore, the contents of the experimental record must be accurate and thorough enough to be fully traceable to permit the reproduction of the work conducted. The experimental record is the official data record for each experiment and the most important primary source of data. It is expected that the practices outlined in this document will be applied to experimental planning, record-keeping procedures and reporting, to the fullest extent possible. Both partners shall discuss any ambiguities or conflicts regarding these practices or proposals for further refinements prior to the start of the experiments to ensure alignment and understanding.
Expectations summarized below are also available for download as a standalone document that can be used as an appendix to industry-academia collaboration agreements.
Please refer to the Glossary for explanation of specific terms used by EQIPD.
Expectations
Experimental Record
- Unique study identifiers must be used for each experiment (for example, lab notebook page reference could be sufficient)
- Each experimental record should include, directly or by reference, the names of all scientists involved, objectives, procedures, methods, materials, equipment, dates, and any other details considered necessary for reproducibility and reconstruction.
- All raw and any processed data must be retrievable and traceable, directly or by reference. No raw data should be erased.
- An experimental record must describe any changes from the study as planned.
- An experimental record must provide an explanation and justification for exclusion of any data points from analysis.
Rigor in Study Design
- Apply randomization and blinding. If not done or if not maintained throughout the experiment (from subject/sample allocation to analysis), include the reasons in the study protocol, the study report and any publication.
- Define the primary endpoint, justify sample size (e.g., using power analysis) and include the justification in the study protocol, the study report and any publication.
- Ensure inclusion / exclusion criteria and/or acceptance criteria are stated in the study protocol, the study report and any publication.
Analyses of Experimental Data
- Data analysis plan should be described in the study protocol, carried out as described, and reported. Additional analyses are always possible, but should be identified as such.
- Minimize the risks of bias and increase internal validity by including criteria for outlier exclusion, acceptable ranges for standards, reference compound and quality controls upfront in the study protocol.
- Experimental / data analysis record and study report (e.g. publication) should include sufficient detail to reconstruct any analysis performed and record all process steps and calculations used.
Data Storage and Traceability
- Experimental records should be kept in an audit-trailed, version-controlled, safe storage environment such as an appropriate bound-paper laboratory notebook with permanent ink or an electronic laboratory notebook (ELN).
- If bound-paper lab notebooks are used, separate notebooks should be used solely for work under each contract.
- Raw (primary) data must be stored in an un-editable read-only form as soon as it is generated.
- Processed (secondary) data should be clearly labeled as such and should contain a reference to raw data.
Review and Reporting
- Experimental records should be reviewed by a peer for completeness and accuracy and it is advised to document this review
- Report should always provide detailed summaries of all related data, processes, and conclusions, and include justification for excluding any relevant experimental records or individual data points from the summary analyses.
- Reporting should include for each analysis the exact number of biological units for each condition.
- Report should also reference the permanent identifier and/or location where all data are kept.
- Identify all contributing researchers and reference unique identifiers for the experimental records.
- Any external presentation/publication, whether oral or in writing, needs to include a statement of conflict of interest and financial disclosures.
Data Sharing
- A mechanism should be considered to enable the timely sharing of control, reference and other data that do not infringe the IP clauses of the collaboration agreement.
Resources
EQIPD Expectations for Good Research Practice in industry-academia collaboration
back to Toolbox
Next item: 1.4.3.3 Academia-academia: Research as service
