The National Institutes of Health is obligated by statute to make public information concerning the biomedical research grants it funds. There are two main sources of such information that are updated on a weekly basis to reflect newly awarded grants.
One is the NIH RePORTER Web site, which allows one to search a database of all NIH grants awarded over the past thirty or so years via such identifiers as the principal investigator of the grant, substantive keywords, the peer-review panel that reviewed the grant, the institution to whom the grant was awarded, etc. There is also a less well-known Web site that is part of the SILK (Secure Internet LinKed) system, which is actually just a massive comma-separated-values spreadsheet listing all Notices of Grants Award (NGA: the legal document created by NIH that constitutes the grant) issued in the last 90 days.
Each of these sources of information is updated on a weekly basis early Sunday morning, presumably via an automatic population mechanism that sucks information out of the grant award management system used by NIH staff to handle grant applications and awards. There are two dates associated with each NGA: the date of issuance and the date that the grant award officially begins. Sometimes these are the same date, but sometimes the NGA is issued several days or weeks before the start date of the grant.
Well, it turns out that NGAs that have been issued are announced on the SILK Web page even if the start date of the grant hasn’t arrived. In contrast, grants only appear in the NIH RePORTER database after the start date has arrived.
This can be seen, e.g., if you sort the SILK spreadsheet on the date of NGA issuance and look for grants with NGAs issued during the first week of September, but with start dates in the future (a bunch of them have start dates of September 15). None of these grants appear in NIH RePORTER, but grants whose NGAs were issued that same week, but with start dates that have already occurred, do appear.