I have run the Davis UV sensor for some years, have done cleaning, and worked through some degradation concerns.
The first question is what comparison you are making that leads you to suppose the reading is too low. Cloudless day UV is surprisingly variable from year to year. So just comparing your reading this year to years past in a single snapshot is not good enough.
I don't know of any source around here which provides current UV readings from actual observation other than amateur postings from equipment like ours, which are quite variable. The national weather database that supports most other observations seems to limit itself to forecast numbers, and not actual current UV observations. One source that I have found reports the forecast from that source to 0.1 index number resolution, and allows specification to a specific lat/lon. It is WillyWeather.
I have compiled a spreadsheet with many days of observations of my Davis UV versus the WillyWeather forecast averaging an hour of readings at solar noon on cloudless days.
If others here can suggest a better way of "calibrating" one's UV reporting I am very interested to hear it.
Here in New Mexico, the perhaps half dozen times over the years that I have cleaned my solar and UV sensors, I have sometimes seen as much as a 5% increase in reported solar intensity, but I have not seen a clear change in reported UV. Attempting to respect the user manual requirement to use ethyl alcohol, I formerly employed Everclear, and recently have used some that I sourced on eBay from a Bulgarian vendor. But my types of sensor contamination may not match those at your location, so I think cleaning is worth a try.
About a year ago I believed that my UV sensor had lost sensitivity and as a temporary workaround (after cleaning failed to resolve the discrepancy) instituted a large calibration multiplier in Cumulus (40%--which was probably a bit too high).
About six months ago I resolved to buy a new sensor, figuring that I would then alternate sensors, sending each one in for refurbishment as needed, without having a gap in my reporting. But I learned that Davis in my case would carry out the refurbishment on a swap basis, sending out someone else's refurbished sensor for me to install, then crediting me when I returned mine, with the total cost considerably less than that of a new sensor. (they asked the manufacturing date of mine before telling me the deal, so it may not be the same in all cases).
So I did that. The replacement indeed reads higher than the previous one did in the degraded state it had reached, though with continued comparison to WillyWeather I have settled in on using a 1.15 multiplier.