When it comes to Business Intelligence (BI) there is an early on choice for some healthcare enterprises. Do we build our own BI solution, after all, we have lots of tools in MS SQL and other DB programs and then there is always Crystal Reports Enterprise, or, do we buy a pre-made solution for the area we are working on within Decision Support? That is the question.

When building your own there are some clear benefits and drawbacks:

1. First, cost in terms of out of pocket expense might be less. This is often considered as a key factor saying we already have the staff in place and we are paying them to do this. However, is this really true? Yes, you have the staff in place, however, is their job really building and testing a system or is it deploying the system and teaching users how to get the most from real-time dashboards when changing the culture of decision making in the hospital?
2. Building permits IT to take increased ownership of the data and how they support the organization. This is very accurate and is likely most effective in large teaching settings where staff is plentiful and budgets are approaching adequate. For most community and cottage hospitals the exact opposite is true. Staff is scarce and over extended and budgets do not permit over runs or delays in achieving benefits from IT initiatives.
3. A package like Crystal Reports Enterprise is all we really need as we have been using them if various forms for years. This may be true, and certainly a level of comfort with existing software does help. The alternative to this is engaging a firm that is dedicated to true BI analytics and not just report generation. The influence from outside experts is often what DI BI companies can bring, along with the best practices and evidence based real-time dashboards learned at other sites that are now well oiled processes and templates. In other words not having to re-invent the wheel at your own house.

The BI Decision Support community can be a great source of stimulating information and practices learned from each other and taken to new heights through the exchange of peers. An internal build does not necessarily facilitate an inter-collegial atmosphere of sharing insights and knowledge.

Some vendors market their solutions using a Software as a Service model thus facilitating a continual research and development of BI analytics in Diagnostic Imaging and providing this at no additional cost. This constant upgrading and enhancement provides a culture of improvement and constant change, which will ultimately provide a higher quality of decision support in order to make better and better decisions over the long run. These same vendors also tend to foster a community of users for the very purpose of learning from each other and having the ability to raise the overall standard of healthcare by sharing best practices, inter-hospital templates and regular on-line meetings.

There is no perfect decision here. For some, they will think building their own solution to meet their own custom problems is the right answer, and this may be the best solution for them. For others the ability to leverage from a vendor who applies a SaaS model and constant and continuous paradigm to the development of analytics, best practices in DI and fosters the inter-collegial approach to raising the standards in DI may be the right solution. Each case is different and each decision maker will need to make their own decision based on the best information they have in front of them at the time of the decision.

Readings in various BI books indicate over 55% of all business intelligence initiatives fail. Other publishers suggest as high at 70% fail. Why is this?

There are a large number of components that go into making a successful BI project and each one of the components brings with it its own risks. Here are just a few of the reasons and pitfalls to watch out for:

1. Lack of understanding of Business Intelligence or Decision Support – leads to lack of leadership in the project as to what are the objectives, measurables, and required outcomes;
2. Lack of executive sponsorship. The project is started at a director level or VP level within the healthcare enterprise and along the path a major roadblock is encountered. As this is not a key strategic initiative for the enterprise, senior admin is not prepared to put in the time and dollars to remove the roadblock;
3. The quality of data within the existing database is poor at best and therefore decisions are being based on bad data;
4. The Decision Support team, in its rush to deliver something to show to key stakeholders, shows preliminary dashboards and reports based on non-validated data – the stakeholders see this and lose all trust in the project as they know the data is not accurate. Credibility is paramount. You get one chance to make the first impression here – gaining credibility after showing erroneous data within graphs is something that is not very often achieved;
5. The ability to build interfaces to capture real time data and display in real-time is not a trivial task and getting the right data into the right database field has more issues than initially thought. As data is updated or accession numbers change or are removed, the data can become erroneous and not recognized for months or years;
6. The ability to perform a credible acceptance test ensuring all the data is collected and then all the data is in the right field for analytics to be performed is often overlooked or worse still is not resolved in all the data fields. This leaves users believing they are making decisions based on accurate data when in fact, due to a change in accession number or other data field one to three days after the exam, the data is incorrect at the time of display and the decision being made;
7. The use of existing databases from the HIS and/or RIS to collect data in order to build data cubes for analytics has its own problems. How is the data validated against other sources to confirm it is the single source of truth? Data in the RIS can often have discrepancies with the data in the PACS DB. One hospital who genuinely was concerned about their data quality said they thought they only had 3 maybe 4 different bad data items per day going into their DB based on 280 Accession #s generated per day. Once a data validation tool was turned on between the RIS and the PACS it was found they were missing over 45 pieces of bad data per day;
8. There is no business case to be made for BI. It is like the chicken or the egg. Which comes first, the business analytics so you can perform LEAN, Six Sigma or CQI resulting in improved productivity, lower costs and better real-time decisions, or is it the automation put in place as a result of the re-engineering practice and then measured by the BI to ensure it is sustained? Where are the business benefits attributed?
9. Not understanding the correct formula to establish analytics that represent the condition or problem; inexperienced and well meaning people in Decision Support and BI initiatives often leap in where angels fear to tread. By not understanding the business or the drivers of the business and just applying business analytics methodologies is not enough. Having the subject matter expertise in BI is critical. The ability to ensure everyone knows the drivers behind the problem and then the formula used in the display of the analytics is critical. Not sharing this info will have decision makers making decisions and then guessing at the changed outcomes in the graphics when it does not show the changes they are expecting.

So how can one mitigate the risks of BI when creating a Diagnostic Imaging Decision Support Tool? There are vendors who have predefined and tested BI solutions for DI. They have validated the data in other sites before they come to your site and validate your data. They have pre-defined real-time dashboards and report templates that are tried and tested and interchangeable with other system users for inter-hospital benchmarking. They have removed most of the risk factors by doing this else where first and provide the user with a one stop shop to call for resolution of problems quickly and easily. In short, this reverses the risk from 30% success to 70% success with the 30% remaining risk factor coming from the team themselves and their level of commitment to the project.

Greetings to all of our Blog visitors.

I have an important request for the Diagnostic Imaging community. What are the most important and/or difficult issues facing you today within your DI department(s)?

We believe that by you sharing your thoughts directly with the DI community the result will be an on-line discussion providing assistance not only to you but to the community as a whole.

Help start a community discussion on the issues that concern you today!

To post your comments – simply click on the orange title line of this posting and then enter your comments in the ‘Leave a Reply’ box.

Thank you for visiting and we look forward to hearing from you.

Cheers,

Thomas

Over the past 11 years TNC has worked closely with the Diagnostic Imaging community providing strategic performance improvement solutions accompanied by pragmatic implementation to achieve sustainable results for our clients. We have seen tremendous change within the DI environment over this time and TNC continues to change as well, offering new healthcare technology solutions aligned with the needs of today and the future.

Keeping up with this change is a challenge for any one person, keeping up with the change as a community of people will be easier. Therefore, it is our hope you will join our blog community and share your knowledge and insights, as we will, so we can both benefit from our group knowledge and provide a better quality of healthcare to the patients. After all, ultimately this is why we are all here.

We are very excited about the launch of our new website which will offer the DI community insight into new services and solutions to help them improve their day-to-day operational performance and keep them apprised of the ever changing world of DI.

Like all things new, there are sections of the site we are still working on such as the TEHE section. We are currently collecting information from our clients as to what presentations and professional development courses would be of interest so that we can tailor our courses and presentations to meet your specific needs.

Feel free to post your comments, questions and understandings on our Blog. Tell us what is of interest to you so we, as a community, can help you achieve your goals. Please take a few minutes to look through our site and let us know what you think, as your feedback is important to us and will help guide us going forward.