Twenty-one days at room temperature is about 71 refrigerated days of equivalent exposure. Room temperature is not a number, so take the pharmacopoeial 20–25 °C and its 22.5 °C midpoint: 17.5 kelvin above the 5 °C middle of a 2–8 °C refrigerator. The ten-degree rule of thumb — degradation rate roughly doubling per 10 K — puts that at about 3.4 times the refrigerated rate. It is an order-of-magnitude statement about a rate, not a shelf life, and the top of the 20–25 °C band runs about 1.4 times faster than the bottom of it. At 71 equivalent days a test is more likely to reassure you than to tell you something, which is a legitimate reason to run it and a poor reason to expect a finding. Either way, record the twenty-one days and the temperature now, while you still know them; an excursion you did not write down is an excursion you cannot interpret later.
The relevant physics is latent heat: a phase-change pack holds temperature only while it is changing phase, and once it has melted it is a warm mass.
Asking a supplier to hold a shipment during a heatwave is reasonable and the better ones agree; the cost is a week and the benefit is a shorter warm exposure for anything in solution.
Specifically, anything shipped in solution is a different risk category, because hydrolysis and deamidation proceed in the aqueous phase and are strongly temperature-dependent.
Solutions are a different question and deserve the worry the solids get.