Protein folding is a complex thing. There are huge computer algorithms and huge mainframes which are trying to predict the final 3D structure of a protein.
Knowing the tertiary and the quaternary structure of a protein, allows us to understand why diseases happen. In many cases a mutation of the gene provokes an aberrant protein folding. When the protein is misfolded, it is not able to do its function. This particular condition is called "Proteopathy".
If you know the structure of a protein, you are able to design or improve drugs which can directly influence the protein structure and/or function, for example antibody therapies against cancer (Cetuximab-->against EGFR, in colorectal cancer; it is a drug which was very much improved when computers prediction/simulations are used).
It is hard to find out a protein's structure - for some proteins it can require years!
So you first go from the sequence of the healthy people, then you sequence their genomes, and you use this as reference for the people with certain diseases. You might find eventually some mutation is in certain loci in the genomes of people with a certain disease. Then you can study the protein structures, so you can design some new drugs which might tackle this protein. If the protein is misfolded, like the tau protein in the neurodegenerative diseases, then is good to study the structure of this misfolded protein so you can design some drugs for these protein too.
So in the end the structure is extremely important, but often times you have to study many stages of the protein folding process to understand the possible final structure that one protein might have in the body and/or in a certain disease.
I hope this can help!