Great question!
This kind of display is probably possible depending on how the program calculates those things. Real orbits are not actually perfect Keplerian orbits because Earth's gravity field deviates from spherically symmetric by about 1 part in 1000. Earth's $J_2$ is about $1.08 \times 10^{-3}$.
This is off by maybe 1 km out of 7,000 km (distance to center of Earth). It's hard to tell exactly because the apogee and perigee are rounded to the nearest km.
If the program did something fancy and propagated the most recent and complete orbit and used max and min to find them, then there's something wrong here.
However if the program estimates them based on the mean eccentricity and inferred semimajor axis from the mean motion in the TLE (and I'll bet this is what it does) perhaps using only the monopole $GM$, then the current altitude certainly might wander slightly outside those estimates.
The atmosphere is a drag sometimes
At 238 km atmospheric drag is still fairly low and altitude will not drop this much from one orbit to the next, but once it gets a lot lower, say around 120 km, the orbit will be much more of a fast downward spiral and so orbital elements will become meaningless as soon as they're calculated.