One of the main principles when it comes to patient privacy, to protect individuals, while at the same time preserving the integrity of the information, is to come up with good de-identification schemes. On one hand you want the freedom to collect data, because we all need data to analyze - especially health data, but people have to feel re-assured that the data can't be used against them at some point (a minimal level of control over their data). So you have competing entities, all who can benefit from analyzing data. Patients need doctors to make decisions about their treatment, researchers need data make discoveries about correlations. De-identification addresses this - but has to be intelligently implemented. I've attended seminars one some disease, and there is patient info in the figures! Not necessary! I don't know that PHP, HTML and CSS are very relevant here, but design of SQL for separating data from patient information - there's some meat there. You should read up on SQL.

This topic also touches on data in bioinformatics environments in general I think. Often there are research groups in an organization working on similar, or related problems, who may use a bioinformatics core, and the systems can be set up to share information or hide information from each other - yet large data sets provide information for everyone usually beyond the immediate goals of the particular experiment. Thus finding good ways to get access to data, or know about the existence of data, while at the same time protecting individuals from having to share all information is relevant to many organizations.

It would also help if you could learn to ask questions better. No one wants to do your homework for you.