Aren't you just normalizing the question type out of the client-question-response ?
consider finding a client that has
have no Risk of a fall
have low Risk to client or caregivers from pets, livestock, etc.
have high home odors (gas, rotting food, etc.)
have no skin condition
have low access to adequate medical supplies
has no recall recent of events
Would you look for clients by their responses across survey's ?
If they didn't answer that question should that mean that they are still eligable to be returned if your search was only looking for clients with no risk of fall ?
I would not be scared of this type of normalisation with client,question and client-question-response covered in appropriate indexes and only a couple of million responses.
If you were proposing to have client as an normalised type out of say person I would still probably not.
But if Question,Person and survey were normalized types out of some entity table, and person-has-a-survey, survey-has-a-question and Person-takes-a-survey-question where normalized types out of some entity-relationship table then I would start to squirm.
If you know that the client takes a survey not a supervisor or an operator and that the survey is the only thing that has questions that you need to search for clients by. Then denormalising into the database foreign key relationship the survey-question-response increases the scalability possibilities, specifically the case of clients that have no risk of fall type search. And has low, if not no risk to the flexibility of the application.
Anyone else ?