I-90 preliminary median falls 38 days, but the 8-week trend still points upward

The week of August 24 produced a preliminary 66-day median for the final-stage wait on I-90 applications, down from 104 days the prior week and below the 98-day reading from the same week last year. Both comparisons are encouraging on their face. But this is an early read: at 4,136 cases approved, the sample sits at roughly 50% to 70% of a typical week's volume, which means the figure can shift as more approvals post. At the same time, the 8-week regression slope of +7.9 days per week runs counter to this week's apparent improvement, indicating the broader trend has been moving in the opposite direction since late June.

A welcome dip inside a rising trend

The week-over-week swing of 38 days (-36.9%) and the year-over-year improvement of 33% relative to last August's 98-day reading are the two most visible data points this week. Both are genuine in the sense that the numbers are real, but neither settles the question of direction. The 12-week trend tells a more complicated story: medians were as low as 15 days in the week of June 15 and have generally climbed since late July, touching 104 days in W34 before this week's pullback.

The spread between the 25th percentile (p25) and 75th percentile (p75) this week runs from 41 to 68 days, an interquartile range of 27 days. That is noticeably wider than earlier low-volatility weeks: in W26 (June 22), the same spread was just 10 days (p25 = 12, p75 = 22). The widening may reflect greater mix in the types of cases resolving at this point in the cycle. The 3-year same-week average of 98 days provides further context: at 66 days, this week's preliminary figure sits well below that historical benchmark.

The only anomaly flag fired this week is `rfe_fast`. Cases where a response to an RFE (Request for Evidence, USCIS's standard follow-up letter asking for missing documents) had been received cleared at a median of 20 days, while cases that proceeded through plain processing without an RFE cleared at a median of 35 days. That inversion is counterintuitive: RFE-response cases ordinarily take longer because the adjudication (USCIS's case-decision step) must resume after the applicant provides additional material. One structural explanation, though the data does not confirm it directly, is that a concentrated batch of RFE-response cases that had been pending for a similar period resolved together this week, pulling the RFE median below the non-RFE figure. This does not signal a policy change, and it does not mean that receiving an RFE has become advantageous.

Most I-90 RFEs arise from identity or documentation gaps: a missing police report for a lost-card replacement, a photo issue, or a signature discrepancy. The historical RFE rate for I-90 has sat well below 10%, reflecting the form's relatively straightforward eligibility (the applicant is already a lawful permanent resident). With 559 RFE-response cases in the weighted-average approval cohort (a group of cases) over the last 12 weeks, the group is small relative to the 80,607 cases in plain processing. A single week's result for that smaller group can fluctuate materially.

One week of faster RFE clearing is a statistical footnote, not a signal that the adjudication process has changed.