|
|
|
Index - Major Sections
Site Map
Product and Services _______________ Index - Same Level Subject
Index - Child Subjects |
Introduction
"Once outcomes become measurable, they become manageable." “Once decision support becomes predictive, they become proactive instead of reactive.”
Healthcare Intelligence is what it is all about. We will cover this topic over a number of sections. We feel that the main part should be the "Uses" of Healthcare Intelligence.
Note: This section has been rewritten a number of times. It seems that there are advances and hopefully, we will continue to have to rewrite these sections. These sections are also very long. We have created short summaries. A full report can be downloaded as a pdf file. Requirements In order to analysis data (and mine data) the right data must be available and in the right structured format. This data is NOT currently available. (see Prediction) Healthcare (Business) Intelligence is the name given to the technology that enables the management of data for the purpose of analysis, exploration, reporting, mining and visualization (Nielson, White & Parui "Mcrosoft SQL Server 2008 Bible", Wiley Publishing, Inc. , 2009) Healthcare Intelligence systems can facilitate change in an organization by helping to make better-informed decisions. By having better information, managers can identify trends, problem areas, and new opportunities. By having all the data in one place, the whole picture can be seen at once. Decisions do not have to rely on parts of reports gather from many different sources. There is the rather trite story of 5 blind people describing what an elephant is by feeling only one area. Here at InHCc, Data and How to use data is at the core of our mission, and our website is a vital part of our effort to make this information work for you. Our site is designed to provide you with a reference to the wide range of information and data available. Problems with Health Care SystemToday, in most organizations (or countries), there is a lack of sufficient data in which to perform any type of analysis much less Healthcare Intelligence. These is no one place where the data stores contains records that have the potential to reveal patterns. This may be for several reasons, not the least of which is the belief by some professionals in healthcare that they know it all and they do not need to be told how to do things or record data for others and of the non-professional that they do not want the management to know what they are doing. Only extremely aggregated data is collected with little or no value (billing codes) for management. This data is, in general, collected only to be sent to a state or national data collection center to be used to produce the annual reports or third party billing. Healthcare data is not only transactional but also historical. What this means is that normally in a non-health care organization individual sales data may be "rolled" up into summaries after a certain period of time. There is no need to keep the details of any one transaction and this reduces the size of the data warehouse. However, in health care the client's history must be kept "forever." This history is needed at any time that the client returns to the clinic. Over time, the size of the data warehouse can grow to an enormous size. The good news is that the cost of storage is so cheap that this is no longer a problem. After the birth of the PC, it was realized that tasks could be distributed to the computer that best could perform the work function required. Application presentation could be performed on the PC and the server did the heavy work of storage and managing the data. This model was called the client/server model. However, now data became scattered all over the organization, creating the “islands of data” problem. A different type of problem arose when the user needed information that came from different sources. Questions about the relationships between data from different locations were impossible to gather. Data had to be combined from every system in the organization. To solve this problem, at attempt was made to copy needed data to the location that needed the data. Soon, the need to copy all the data from all the locations to all the users computers because impossible to manage. Most systems are still trying to develop a "network" of scared data....It will not work.
Healthcare Intelligence
The term "Business Intelligence" (BI) are fundamentally about providing users of information with the right information, in the right format, at the right time, in the right place. BI should support both strategic planning and operational planning. In the InHCc HMIS, Business Intelligence is called Healthcare Intelligence. Good information requires historical data as well as integrated data from multiple sources, both internal and external AND the ability to use predictive modeling. This section will attempt to explain the concepts of Healthcare Intelligence, it purpose and its benefits. Today's technology pyramidTransactional Processing is the day-to-day recording of information. In health care, this system is the Clinical System. Data WarehouseExtremely large collections of data. Data Warehouses are distinct and separate from operating data and are generally accessed in Read Only mode. Governmental organizations are under enormous pressure to provide additional services to their population. These organizations must react quickly to changes in the needs of their population if they are to function efficiently. Data Warehousing will become an important tool as governmental agencies seek to gain more information in order to improve health care and to reduce health care costs. “Adoption of data warehousing in health care has been slowed by the lack of understanding of the benefits offered by the technology" (Pedersen and Jensen, 1998). What is important is not what data is in the data warehouse, but rather what type of data and how that data is structured. Analysis Services (OLAP)On-Line-Analytical-Processing (OLAP) is the use of summarized data in the form of "dimensions" for easy analysis. It consists of a set of predefined indicators used to measure the performance of the organization. It provides methods for "slice and dice," "drill down", "drill up" and "drill through." OLAP systems generally depend on detail information provided by relational databases to form the "Cubes" that are part of its structure. OLAP provides for easy query and search methods most commonly used by users without a lot of training.On-Line Analytical Processing (OLAP) makes it possible to solve market analysis and forecasting by using database designs that are array-oriented and multidimensional in nature. These problems are characterized by the need to retrieve large numbers of records from very large data sets and summarize them in real time (Berson and Smith, 1997). On-Line Analytical Processing (OLAP) builds on the work that has already been laid out in the discussion of Relational databases and Data Warehousing. While OLAP uses these technologies, it is essentially a different application architecture. While Excel is an excellent tool for analysis trends, it is limited in its ability to product the data to make this analysis. OLAP is "data aware", in that it understands the hierarchical structure of the data. While data warehouses are usually based on relational technology, OLAP uses a multidimensional view of aggregate data to provide quick access to important information for further analysis. OLAP enables analysts, managers, and executives to gain insight into data through fast, consistent, interactive access to a wide variety of possible views of information. OLAP transforms raw data to useful information so that it reflects the real factors affecting or enhancing the line of business of the enterprise.
By moving the analyses to the clinical level, the organization empowers the operational managers to makes decisions based on their own environments. Data MiningWe will define data mining as:
Data mining reports the non-trivial, and previously unknown relationships between variables. No prior assumptions need be made. If the databases were not large, then no new techniques would be needed. (see note on use of terms Data Mining vs Knowledge discovery). Data mining is the use of artificial intelligence and advanced statistical analysis for searching of data for patterns. Once patterns are founded, in sometimes apparently unrelated variables, relationships can be formed and trends can be predicted. This however, requires very, very large quantities of historic and detailed data. It cannot used summarized data. Data mining tools can be applied to client-centric services. One goal might be to define the attributes of clients who are likely to continue with a Disease Management program. Once these attributes are identified, resources can be used to target other likely groups. Research Healthcare Intelligence and Research needs massive databases. Today it is possible. Research is reaching a point where the hypothesis is not needed. Applications are programmed to automatically mine masses databases. Any little "relationship", groups of "similar" data, "predictions", Sequence events, can be automatically detected. It may not even matter what the rules are...the only important thing, is that it occurs. Decision SupportThe ability to collect large amounts of data and perform complex calculations does not tell us what questions to ask. Information is a level above traditional data collection and is generally derived from an act of comparisons. OLAP makes these comparisons relatively easy. Expert SystemsExpert Systems are not normally classified under data analysis. Expert Systems are collections of knowledge databases and rules on how to use this know. An Example would be, if the client had a diagnosis of "Hypertension", what would be their treatment? However, it is Data Analysis that is used to provide these collections of knowledge. Why Now?Until recently data mining activities were limited by computer capabilities and were very expensive. Recently, several factors have come together to make the investment in data mining an attractive one to ordinary organizations through the following:
Links and Download
|
|
|
