You are here: Home Blog MediFormatica's BLOG Medical Informatics I: Principles of Database Design

Mediformatica - The Medical Informatics Portal

Mediformatica - The Medical Informatics Portal

Latest Blog Articles

Medical Informatics I: Principles of Database Design

E-mail Print PDF
Article Index
Medical Informatics I: Principles of Database Design
Page 2
Page 3
Page 4
All Pages

Principles of Database DesignThis lecture was delivered as part of a week-long survey course designed to familiarize individuals with the application of computer technologies and information science in medicine. The course a National Library of Medicine fellowship program directed at medical educators, medical librarians, medical administrators, and young faculty who are not currently knowledgeable but can become agents of change in their institutions.

"[Medical Informatics] is the science of organizing information to make it useful, to make it retrievable, so people can use it to solve health problems and understand health and disease better. It is the technology for implementing that science, such as databases, communication networks, and other forms of digital tools.

We are increasingly surrounded by so much information that only computers provide a plausible way to keep it under control and find out the facts that we need. That is what medical informatics is about."
Dan Masys, interview at MBL, May 2001

In practice, medical informatics is the application of technology to all aspects of health care information. It involves both the art and science of organizing medical knowledge and applying such knowledge for the purpose of preventing human disease and suffering. Technology is pervasive and necessary to handle and manipulate the growing body of medical knowledge. Technology winds its way throughout health care--from processing results of medical research to applying knowledge in clinical practice; from accessing and processing patient records to making decisions in evidence based practice; from telemedicine to knowledge-based and decision-support systems.

Below you will find the transcript for this lecture, or you can view the video here

Transcript Part I: Introduction:

"Principles of Database Design" has been a staple of this course for the last 7 years, for a reason that endures, and is increasingly important in our ubiquitous computing environment, where all of you probably have desk top applications that allow you to easily create databases for your own purpose, and that of your organization.

The outline of what we are going to cover in the next 90 minutes is, first of all, our explanation to you of why we think it is important to learn this.

Secondly, the principles and the paradigms. "Paradigm"...Ted Nelson has described it as an idea that is too big to get through the door. The paradigms of the generations of database systems that have evolved over the last 3 decades. The principles of the most widely used database modeling these days, relational databases, are important so we are going to focus in some detail upon how they are designed and built, look at some of the methods that are used for creating systems of various sizes, and then lastly we will have a classroom exercise where here in real time on the screen, we will convert an old-style MEDLINE record into a relational database equivalent of it, so you can understand some of the principles of both entity relationship modeling and one-to-many parent/child relationships.

Why this topic is important/Historical Perspective:

So why is this important? In our health care environments and in laboratory and research environments, you are constantly besieged by vendors who sell you very attractive looking user interfaces. But the principle message of this session is that the durability of systems--their flexibility, their power--is really in the data model and not so much in the user interface. And the data model, if well-designed, will outlast and support uses of the data that nobody even imagined when the first system was developed.

Now when I said this before, actually the people from Columbia University were here and they said, "You don't understand. The data model outlives the original system design even if it is a bad one. They persist abnormally long in some cases regardless of whether they are well done or badly done, so it is important to try to do them well if you can." The evaluation and comparison of vendor products, again which is often done just on the way they look and feel to users, is actually much more powerfully done by comparing data models if those are available, and it also gives you a vocabulary for communicating with vendors and people who are helping you to build systems. Not the least, however, is the ability to make your own databases, and most people do, in fact, with tools such as Microsoft Office, that includes Microsoft Access.

How many have built at least one database in something like Microsoft Access? So, it looks like about two-thirds, maybe three-quarters of the class has done that. Then you realize how easy it is. Any fool can build a database and many do, just like Web pages. The entry level skill set is very, very low but in fact you quickly confront the issues of data representation and modeling whenever the task gets more than just the equivalent of recipe cards.

What is a database? Well, the term is associated extricably with computers, but in fact in its generic sense, it is really any organized collection of information, though these days computer-based representation is implicit in the idea of database. The general notion is that they are systematic. That is, they have a structure and a functionality that supports automated retrieval beyond what it is capable to do with just manual record keeping, and also support automated symbol manipulation. The idea of being able to reason on the data, make conclusions based on values that are present in the database, is an important aspect of things such as clinical decision support and other forms of rule-based systems design that are increasingly built, not with coded rules but rather with reasoning using the values of data in databases.

Historically, the first computer systems were, by and large, mainframe systems from the 50's and 60's, created with a software program that wrote a file, that is, a sequence of characters on the disc that were the sole province of the program that created them. That is, they were owned by the application program. And by and large the files in those early days were what were called sequential files. If you think of a document as just a long string of characters, maybe a million characters long, the sequential file would simply have them head to tail in the sequence in which it would be read from the beginning to the end.

Sequential files gave way to what are called "random access files" where, instead of trying to find a particular value half way through a million bytes, one could look up in a table where a particular record was in the middle of that very large collection of bytes, and go directly to that location on the disc to retrieve it. Those systems were brittle in a sense that if one changed the definition of how the file was laid out, all programs that accessed the data had to be rewritten and many would stop returning the correct values, if someone for example, inserted a few extra characters into each record in the sequential file. And as well, the general problem of a program owning the data did not allow other programs to have simultaneous access to it. So they were good for single applications that were generally referred to as stovepipe applications. They didn't talk to one another. They were stand-alone applications and such programs are commonly written and still existing in the world today. It is a situation where an application has the hubris of believing that it is the entire universe of importance and does not have any way of either exchanging data with other applications or changing the structure of their contents.


Last Updated ( Saturday, 01 June 2013 23:23 )  

Google Translate


DrugFormatica provides a simple drug database providing information about FDA approved medications.
ProcFormatica provides a simple medical procedures database including information about the latest ICD-10-PCS codes.
PsychFormatica provides a simple medical procedures database including information about the latest DSM IV codes.

Healthcare IT in Egypt

In this section you'll find a list of companies and organizations working in the healthcare IT sector in Egypt.

Read more here...


Who's Online

We have 72 guests online


Members : 1
Content : 543
Web Links : 68
Content View Hits : 2845393




Medical Informatics

Medical informatics has to do with all aspects of understanding and promoting the effective organization, analysis, management, and use of information in health care.


Hospital Information Systems

A Hospital Management Information Systems (HMIS) is a comprehensive, integrated information system designed to manage the administrative, financial and clinical aspects of a hospital.



Mediformatica's videos section is rapidly becoming one of the largest health informatics video libraries linking to over 250 videos to this date...


Mediformatica's Blog

Mediformatica's Blog is frequently updated with the latest and most exciting news and reviews related to the subject of Medical Informatics.