The All of Us Data and Statistics Dissemination Policy was enacted to safeguard and protect participant privacy while promoting and/or publishing research completed using the All of Us Research Program’s dataset. This policy includes what information a researcher can and cannot reveal through text, tables, or figures about their dataset and the analyses conducted so that participant privacy protection isn't at risk.
Under the All of Us Data and Statistics Dissemination Policy, researchers:
- CANNOT reveal participant-level data in any way, including but not limited to: taking screenshots, making copies, downloading, or in written text.
- CANNOT display any participant count that is less than 20.
- A count of zero is permitted, however.
- Options for avoiding violating this requirement include collapsing data across groups to increase participant counts or coarsening data.
- CANNOT report a participant count that would allow another count value to be derived revealing a count less than 20.
- This includes the use of percentages or other mathematical formulas that in combination could allow for reverse calculation revealing participant count less than 20.
- This includes the use of percentages or other mathematical formulas that in combination could allow for reverse calculation revealing participant count less than 20.
- CAN provide the total number of participants (N) included in a study that includes a group count of less than 20 if it is done in one of the following ways:
- The group count of less than 20 is displayed as less than or equal to 20 (i.e., <=20) and the total N count is preceded by a less than or equal to symbol as well (<=N); see example table B below
- The total N is reported but the participant counts of the subgroups are suppressed; see example table C below
If data users have a compelling reason to justify directly publishing or disseminating data or aggregate statistics that correspond to a participant count of less than 20, they may ask for an exception using this form. Exceptions will be rare.
TABLE EXAMPLES:
(fake data are reported below)
(A)Gender Differences in HbA1c % Levels
Does not comply with the All of Us Data and Statistics Dissemination Policy
n | Mean | SD | |
Female | 17 | 5.15 | 0.29 |
Male | 24 | 5.36 | 0.30 |
Total | 41 |
Vs.
(B)Gender Differences in HbA1c % Levels
Complies with the All of Us Data and Statistics Dissemination Policy
n | Mean | SD | |
Female | ≤20 | 5.15 | 0.29 |
Male | 24 | 5.36 | 0.30 |
Total | ≤44 |
OR
(C)Gender Differences in HbA1c % Levels
Complies with the All of Us Data and Statistics Dissemination Policy
n | Mean | SD | |
Female | * | 5.15 | 0.29 |
Male | * | 5.36 | 0.30 |
Total | 41 |
* Counts suppressed per the All of Us Data and Statistics Dissemination Policy
FIGURE EXAMPLE:
Figure 1A. Does not comply with the All of Us Data and Statistics Dissemination Policy because it's clear that the participant count for females is less than 20 (purple box).
Figure 1B. Complies with the All of Us Data and Statistics Dissemination Policy because the participant count for females has been changed so that it's at least 20 (green box).
Please review our All of Us Research Program Data and Statistics Dissemination Policy, All of Us Research Program Publication and Presentation Policy and Egress Alert Policy for further information.