Maybe do a usability interview with the users of the previous version of the site. Ask them some pretty deliberate questions as they’re moving through the site. Maybe - What content do you find yourself looking for the most? Why? Something which gets them moving to content and talking about it. This will no doubt provide deeper qualitative insights.
Is there a way for users to “bookmark” content from the previous version? You could analyze which content was most bookmarked, which would give you a sense of which was most useful.
Also, you could survey as many users from the previous site as possible. Maybe something like: “What 3 pieces of content do you find most useful from the previous version of the website?” “Why?”
Finally, if the Information Architecture of the previous version of the site lends itself to some sort of arrangement of content into categories, maybe you could synthesize the reams of content into main content categories. Then you could do something of a card sort with the doctors to put the main categories into groups. Then those groupings could be ranked by priority from “Most helpful” to “Least helpful” in the context of when the doctor is engaging with a patient. Then you’d know at least that certain categories of content would be best brought into the new website.
At the end of the day, you will always be able to go back and further analyze the content from the previous version of the website. Just make sure you don’t lose that content. Then, over time, as you migrate users over to the new version, keep your ear to the ground. If you build a mechanism into the design of the new site which allows your users to provide you with immediate feedback when the content they’re looking for isn’t available, then you can tack more quickly and add the content that they want.
It’s the difference between migrating everything to the new site in a “flip the switch” method verses something more gradual.
Of course, most all of this depends on how technologically robust the back-end architecture of the sites are, in particular the search function.
The new solution is going to have problems. It won’t be perfect. But, as long as it’s better than the old solution, you’ve hit designer gold, my friend.
Kudos to you for taking on this project from (what appears to be) someone else’s initial work. A mid-project handoff is never fun, but you seem to be on the right track.
I hope this helps! Keep us posted!