In any healthcare organization, data is vital. An organization’s financial survival and its ability to provide patients with effective health care is directly linked to its data quality and management. For these reasons, finding a Healthcare Data Solution is essential.

Data Warehousing

Data warehousing refers to the use of a systematic approach to the organization and management of data. The goal is to integrate data from various sources into one centralized location. Once centralized, this data can be easily accessed from multiple locations.

Data Warehousing and the Healthcare Industry

Retailers, government agencies, financial institutions and numerous other industries were benefiting from data warehousing long before the healthcare industry. The delay in implementing this practice may have been because the data needs of the healthcare industry vary greatly from the needs of other institutions. According to Bill Inmon, who is considered the father of the data warehouse, the healthcare industry must be able to deal with numbers and text when creating their data warehousing system.

Kalyan Bandarupali, who is a Microsoft® Certified Solutions Expert: Business Intelligence (MCSE), states that to decide which system will work best for your organization, you need to model your data. Your data model has an enormous effect on the time-to-value and adaptability of your enterprise data warehouse (EDW).



Modeling Your Data

    1. Establish your organization’s goals and objectives
    2. Identify your audiences and their necessities
    3. Decide on an extract, transform and load (ETL) tool
    4. Consider methods for accessing the data (reporting and analysis)

Binding Data

The binding of data refers to the mapping of data collected from various source systems to create uniform business rules in the data warehouse. The optimization of this data allows for an accurate analysis.

Data Warehousing Models

The Independent Data Mart Approach

With this approach, you start small and build individual data marts as you need them. You will build a separate data mart for each category within your facility. For example, you would create an independent data mart for Radiology and another one for Medical Records.

The Benefit

You can begin implementing this system rather quickly.

The Drawbacks

        1. May Hinder Your Investigative Abilities
          Because you have individualized data marts, you do not have a comprehensive data warehouse. Generally, transformed data are summarized and located at a higher level. This can be problematic because the lowest level of granularity in data marts does not actually contain data. If information you obtain from a data mart indicates that a particular metric is lacking, you will not have the ability to determine the cause.
        2. Analytics Solution Adaptation Suffers
          With the Independent Data Mart approach, binding of data occurs early on. While being transferred into independent data marts, the data is mapped; this process inhibits analytics solution adaptation.
        3. Redundant Feeds
          This model requires that you build redundant feeds from every source system to essentially feed each of your independent data marts.

The Enterprise Data Model Approach



The goal of this approach is to create the perfect database from the very beginning. This means determining everything you want to analyze in advance.

If you need to build from the ground up-this is the ideal approach; however, in healthcare you are not building a new system, you are creating a secondary system to receive data from the systems you have already created. Taking data from one system and incorporating it into a new system requires skills, patience, time and a lot of money.

Drawbacks of this Approach

        1. Not Adaptable
          Business rules, vocabularies and use cases change rapidly within the healthcare industry. This approach is not ideal for a healthcare organization because it binds data early on and; bound data is very difficult to change.
        2. Actual Data is Disregarded
          This model is inclined to ignore your organization’s data. With the Enterprise Data Model Approach, to obtain the answers you need, you must already have the data you want. A better approach is to build your EDW to your current data, allowing you to move toward your idyllic system.

The Late-Binding™ Approach



This adaptive, practical approach is designed to manage the continuously changing vocabularies and business rules characterizing the healthcare environment.

This approach allows for data to be brought into source marts without transforming it; instead the data remains as raw as possible. Data is not bound to vocabularies or business rules; however, minimal conformance may be performed. For example, the patient address fields will be structured the same way in every relevant source mart. Data is not bound until it is absolutely necessary to do so; when a specific use case needs to be analyzed, the data in that particular data mart is bound.

The Importance of Precise Data in the Healthcare Industry

Data relating to the accuracy of diagnoses, effectiveness of treatments and the variances in the practices of health care providers is vital to health organizations. Especially for those establishments striving to sustain and improve their health care delivery, while reducing the costs to employers and the government.