Transforming OLTP relational data base to Business Intelligence solution in a healthcare environment

TRANSFORMING OLTP RELATIONAL DATA BASE TO BUSINESS INTELLIGENCE SOLUTION IN A HEALTHCARE ENVIRONMENT

Transforming OLTP relational data base to Business Intelligence solution in a healthcare environment

Jean Rannou

UC Berkeley Extension

EE X404.8 Business Intelligence with SQL Server

Erick Ahrens

Abstract

This paper aims to present the operation of Healthcare Business Intelligence and introduces a case study consisting of an OLTP relational data base into Business Intelligence solution transformation project at the London Health Sciences Centre (LHSC), in order to tackle its high infection rate issue. It will also describe the usually found data sources in a healthcare environment, since the data sources for this particular centre are not available to the public.

Introduction

Nosocomial infections, otherwise known as hospital-acquired infections, are infections acquired in hospital or healthcare service unit that first appear 48 hours or more after hospital admission or within 30 days after discharge following in patient care. These infections affect one in ten patients admitted to hospital[1].

The London Health Sciences Centre (LHSC) in Ontario is facing a major challenge related to nosomical rate, which is way above the national mean. It is allocating an extreme amount of time, energy, supplies and equipment resources trying to lower it, unsuccessfully. As a result, the LHSC is considered as one of the poorest performing teaching centres in Ontario. For the centre, lowering the infection rate has hence become a strategic priority.

A Business Intelligence solution could be considered in order to solve this issue. In a healthcare environment, we talk of “Clinical Business Intelligence” or “Healthcare BI”. The uses of Business Intelligence in clinical settings are numerous: treatment planning based on symptoms, facility supply chain management, identifying patients who need follow up care, volume prediction, etc[1]. It has already been successfully implemented in several hospitals, like in Yuan’s General Hospital, in Taiwan[2].

Objective

Transforming OLTP relational data base to Business Intelligence solution, data sources presentation in a healthcare environment

Materials and methods

At all level of a clinical administration, the goal of Healthcare Business Intelligence is to use analysis service tools in an efficient way. To do so, there are many components that are used, such as data warehouses, OLAP, data mining, etc.

We can categorize Business Intelligence applications in healthcare in two sets of solution: the Business Solutions, which is the business focus analytical applications, and the Technology Solutions, which are the Data and Information Tools and services [3].

[pic 1]

Fig 1: Business Solutions and Technology Solutions

The case we saw in class, “Identifying Patterns in Medical Records through Latent Semantic Analysis”, can be considered as a data mining technology solution.

For our study case, the need will be to gather all of the data required for the “Business Solutions”, and to implement the “Technology Solutions” in order to create efficient Business Intelligence.

According to our course, the three keys of effective decision making are an agreed on set of goals, a way to measure if these goals are getting further away or closer, and a way to provide the information based on these measures to the right decision maker in a timely manner [4].

Here, our agreed-on goal in to lower the infection rate.

For measuring and information providing, this case study presents an “Infection safety solution”, aiming to define and recommend an appropriate data management and reporting structure. This solution is composed of three major actions:

1. Hand Hygiene Audit Technology: providing a way to audit the hand-washing rate within the centre
2. Corporate Data Warehouse with Data Integration and Analysis Tools: creation of a solid data infrastructure, ETL automation, multi-dimensional cube buildings for different level of users
3. Infection Prevention and Control (IPaC) Reporting Tool / Dashboard: review of the existing reporting processes and KPIs, benchmarking of existing dashboarding tools

We will here focus on the data sources that are required for the corporate data warehouse implementation.

Data sources

A data warehouse can, in a healthcare environment, be called “Clinical Data Warehouse”. It allows “rapid querying and reporting across patients, and is used to support the discovery of new relationships between the cause and effects of diseases, and to find specific patients that qualify for research studies”[5]. In order to build this data warehouse, we need data sources.