In Part II of this IRB submission blog series, we focus on some common errors seen during IRB review. While each study is unique, there are some errors that are commonly seen. Below are examples of these common mistakes based on the IRB experience of the Learn eCORE founders and are supported from informal discussions with other IRB members.
In these blogs, the term” IRB Application” refers to an IRB-specific form, completed by an investigator that describes the study for which IRB approval is being sought. This may be done within a larger electronic IRB management system or may be a simple document template. We use the term “IRB Submission” to refer to all the study materials that must be submitted for IRB review, such as the IRB application, consent documents, and materials to advertise the study.
Informed Consent Issues:
1. The consent form does not begin with Key Information. This is a new requirement starting with the implementation of the 2018 revised common rule. Unfortunately, some investigators are either unaware of this requirement or are unsure how to implement it. While the regulations do not include details about what should be included in the Key Information section, suggestions for information that could satisfy this requirement can be found in the Preamble to the Revised Common Rule.
2. The consent form is overly complex. While the informed consent process strives to accurately describe the research, it must be understandable to the prospective participants. Suggestions include:
a. Avoid the use of technical, medical, or scientific jargon and instead using plain English to describe procedures. See the federal government’s plain language website for more information. For example, using teaspoons instead of milliliters to describe the amount of blood that will be taken from U.S.-based research participants.
b. Aim for about an eighth-grade reading level for most consent forms. MS Word has a Readability Function option to estimate reading level. In addition, here are some other commonly used tools to estimate grade level of text:
i. SMOG (Simple Measure of Gobbledygook)- a formula that focuses on multi-syllable words
ii. Flesch-Kinkaid formula uses both syllables per word and words per sentence to determine grade level.
iii. The National Cancer Institute’s Pink Book, “Making Health Communications Programs Work, available at https://www.cancer.gov/publications/health-communication/pinkbook.pdf, has additional suggestions for estimating readability.
c. Consider the use of graphics, pictures, or even short videos to help explain the research to potential participants.
Inconsistencies across various components of the IRB application:
This is another common issue noted during IRB review. Basically, the information contained in the IRB application should match the information in other parts of an IRB submission, like the protocol and the consent document. Two common inconsistencies:
1. Numeric, such as the number of subjects, number of procedures or visits required, time required for procedures, how long identifiable data will be retained.
2. The order or type of research procedures
In our experience, inconsistencies are often due to inappropriate cutting and pasting from other submissions, difficulty in managing multiple versions as the submission and associated documents evolve, and/or using different terms to refer to the same procedure. Even seemingly simple inconsistencies such as variation across IRB submission components in the number of participants or the time estimated for specific research procedures will likely get flagged by the IRB and require a revision. Having a system for version management and someone to proof-read across all components are two suggestions for cutting down on these common errors.
Learn eCORE has created an IRB submission Toolkit that describes these and other tips and includes a useful checklist. Click here to get your copy of this useful tool.
Both versions of the Learn eCORE human subjects research training courses teach about IRB’s and the regulatory requirements for IRB approval.