Data Driven Clinical Care - Dynamic Graphing Website

January 16, 2013 1:56 PM

I would like advice on creating a dynamic graphing website to display clinical data on a intranet website to better aide clinicians. What my superiors are seeking is a "Data Mall." My programming/server-admin. experience is minimal, but my Linux-hobbyist mentality hates the MS Access solution that is being built.

The scene:

So, our administrators have fallen in love with a concept that they have seen at a conference where clinicians could quickly and easily see a short list of their patients who have certain 'abnormal' clinical values/conditions, or to rank themselves amongst other providers according to certain metrics. Etc. Some call it a "Data Mall."

Our senior data coordinator has done a great job of leveraging an MS Access database (pulled from our true clinical database) with a VB front-end to be accessed by all providers. But, it still kind of sucks. It takes a long time to load, and is just unwieldy to ask clinicians to use the Access front-end.

Me, and where I have gotten so far:

I have previously setup my own LAMP CentOS server running MediaWiki just for fun, and was basically able to create a similar setup on a Windows PC using "EasyPHP", on top of which I installed Mediawiki and found a script for making calls to "Ploticus" from Mediawiki. I was able to display the simplest-of-simplest pie-charts on my wiki.

The heart of the matter:

The problem is that this solution seems complex, especially since my PHP and SQL knowledge is limited. Would the best solution just to be to dig in and master this option? Any recommendations for starting points?

Or is there another tool, method, etc. that I should be looking at?

I realize that building this could almost be a full-time job, but like I said, part of me is just intrigued by the challenge, and part of me wants to impress the higher-ups. Still, maybe it is best left to the pros...?

posted by rosswald to computers & internet (4 answers total)

I cannot imagine a business scenario where your intervention in this scenario would not be seen by the organisation as you -- at best! -- meddling in someone else's job, and at its worst, a direct assault on the technical qualifications and decision making abilities of the senior data coordinator.

The problem isn't whether or not Access is the right or wrong tool for the task. The problem is that you are not being paid by this organisation to manage or administer or implement or develop or maintain that solution (no matter what the technology is).

So, for the sake of all that is good and decent in your job environment, butt out. Officially.

However, if you have a good, friendly relationship with the tech person already responsible for this project, then maybe take that person aside, over coffee, outside of the office, and have a chat about the project. Maybe you can make some interesting suggestions by way of a fun conversation, but if you detect any resistance at all to alternate approaches, then you should simply back off and never ever mention it again.
posted by seanmpuckett at 2:15 PM on January 16 [1 favorite]

Thanks seanmpuckett! You are right that there is somewhat of a "political" element to this, but everyone understands that the MS Access solution is not ideal, just the best we can do with our staff/abilities.

It might not be the "safest" play, but that isn't stopping me (yet, at least).
posted by rosswald at 2:19 PM on January 16

I'm glad you understand it is a political issue. Navigating those waters is often difficult for technical people! Your best bet to make a change without rocking the boat is to gain the ear of the tech people working on the current solution and make a convincing argument that they will look like absolute heroes to their bosses if they use a different solution.

This must be done unofficially and with the greatest of tact. You have to make sure that any big process changes will make them look GREAT. And it's got to come from THEM, not from you.

So, sure, make your pitch. Make it subtly, make it possible for it to be a huge win for them, and be prepared at any moment to forget the whole thing.
posted by seanmpuckett at 2:28 PM on January 16

My first suggestion is to learn VB(A; whatever it's built in) and help improve the existing one. I'm sure the loading times could be improved fairly simply if you know your way around the code. Starting over from scratch should only be done if necessary.

Secondly, from your description you want to look into what are called, "search facets" in many circles, which is the "categories and checkboxes in the sidebar to winnow the search results," thing we've all seen around. This would require running a search engine that indexes your current database without modifying anything on it. This will get you a way to compile listings based on criteria, which are then displayed using the website language of your choice (PHP, etc.).

Thirdly, forget about Mediawiki for this, and, particularly, forget about Ploticus entirely.

Lastly, if you have actual graphical graphing needs, as well as providing "live update" type behavior for the search facets, javascript is in your future.
posted by rhizome at 3:55 PM on January 16 [1 favorite]