Reading Time: 3 minutes

By Ed Simcox, Healthcare Practice Leader at Logicalis Healthcare Solutions

In my last blog post, I defined healthcare analytics as the access, manipulation, reporting and examination of data to drive business performance and decision-making. In this post, I want to focus on important steps in preparing a healthcare organization for an analytics initiative.

No matter how much time and money is spent building a house, the end result will be sub-standard, unfit for use, and possibly unsafe without a proper foundation. Building a healthcare analytics solution is a lot like building a house. The foundation must be built correctly.

Here are five necessary steps to build a solid healthcare analytics foundation for your organization:

Step 1 – Determine the Objectives

First, identify the business and clinical objectives: what do you want to measure and improve? Do the objectives align across operations, IT and clinical constituents? It’s important to know if your organization is benefitting from analytics initiatives. After prioritizing each objective so you know which ones to tackle first, determine how you will measure the success of each. Measures of success could include quicker administrative decision-making, better patient survey scores, and cost reductions.

Step 2 – Identify Your Data Sources

Once you figure out what to measure, it’s time to determine from where the required data will come. You may need to tap into multiple systems containing disparate or overlapping data sets. Can you locate all the data you need? Will you need tools or resources to access data and to develop new data sources, and if so, can you afford to build them? There’s no sense in trying make business or clinical decisions using data sets if they don’t fully support the business objectives.

Step 3 – Determine the Appropriate Types of Analytics to Use

Depending on your objectives, you may require “prescriptive” analytics—what has happened in the past; “descriptive” analytics—what’s going on right now; or “predictive” analytics—what will happen in the future. In some cases, you need to employ two types or even all three. Each type requires a different approach. Prescriptive analytics is typically easiest to implement while the other two are more complex and costly. Ensure you understand the complexity, cost and operational impact on your organization for implementing and running the solution to make sure the investment matches the possible reward.

Step 4 – Identify the Solution Components

Identify any additional software and hardware components necessary to complete the solution. Time should also be spent analyzing whether the internal staff is capable of implementing the solution, or if outside help would improve or accelerate the project’s outcomes. The need for outside help will vary widely depending on a particular organization’s capability and maturity. Many organizations find it helps to collaborate with an IT partner who has healthcare expertise and knowledge of industry best practices.

Step 5 – Define IT Processes and Draft a Run Book

An analytics initiative is not a project with a discrete beginning and end, but rather an ongoing initiative, which when done correctly, can significantly benefit a healthcare organization. IT processes are needed to operate the solution and to produce the desired outputs. Documentation must be created on how to gather, sort, clean and curate the data. You also need to document IT workflows and on-going staff roles and responsibilities. All of the information gathered in this step should be compiled into a draft “run book,” and staff should be consulted on the proper “care and feeding” of the analytics initiative. The run book should be revisited right before the analytics project goes live and then on a recurring basis.

To find out more about building a healthcare analytics solution to support your big data initiatives, visit the Logicalis Healthcare Solutions website. Then check out this infographic that illustrates how healthcare IT changes are happening faster than you think.