Rising to the Challenges of Healthcare BI

It’s a popular misconception among healthcare organizations that business intelligence (BI) programs that work for other industries simply don’t work in an environment focused on patient care. While there are clear challenges, the truth is that the concepts, principles and strategies behind a well-structured BI program don’t have to be industry-specific to achieve successful application. This article addresses the challenges faced by organizations that deal directly with patient care, including hospitals, physician groups and networks and health insurance companies.

It’s easy to imagine that hospitals and insurers’ BI needs are as different as night and day, but their challenges are extremely similar—most significantly, a tendency to get mired in the detailed data. From a data standpoint, these healthcare organizations must simply sort these key entities: patients or members, providers (physicians, clinics, etc.), events (visits or encounters, which also include complaints or symptoms, diagnoses, procedures, indicators or test results, prescriptions and revenues and costs), overall patient medical history and payers (with varying plans and statuses). It’s certainly true that operational data warehousing challenges vary from one niche to another, but these generic categories are sufficient to handle the vast majority of healthcare BI data.

A well-structured BI program supports three primary healthcare objectives: improving patient outcomes, optimizing the provision of service and managing costs.

Improving patient outcomes is an obvious key goal of healthcare—and one that BI can help. BI analysis allows users to look closely at patient conditions that serve as indicators of health problems and determine how to intervene. Data can contribute to understanding patients’ current and past diagnoses and treatments in conjunction with their health risks in order to more accurately treat them.

Optimizing the provision of service falls on the operations side of the equation, helping to automate doctors’ and nurses’ activities by offering mobile information via phones or tablets in place of paper charts and files. Putting work at their fingertips improves speed and accuracy.

BI supports healthcare cost management in a couple of ways. Data can clarify whether treatment costs are appropriate and in line with other doctors, and whether treatments can be delivered by a nurse practitioner or physician’s assistant instead of a higher-cost specialist. Such data allows executives to examine how healthcare is delivered in order to identify cost-saving opportunities.

Healthcare Data Challenges

There are six significant issues that healthcare BI systems face:

1. Security: Data and content-level security is extremely important, because privacy and security regulations limit who can access what information, yet that information must be pushed into the field for doctors to perform their jobs effectively and efficiently. HIPAA regulations and ethics both play a role in the importance of safeguarding this information.

2. Data quality: In part because of inadequate source systems and a preponderance of unstructured, non-quantitative data that fails to be reliably structured for data mining, the healthcare industry lags in the quality of available data from its source systems. Even the organizations running top-tier enterprise software often fail to validate simple codes and data matching. Reliance on non-financial data could also contribute to the poor data quality challenging the industry.

3. Retroactive data changes: Changes as simple as a patient’s updated insurance provider can cause problems when a provider isn’t notified and therefore hasn’t changed payment information. Erroneous data written in a patient’s chart can also gum up the system.

4. Complex, many-to-many relationships: This isn’t unique to healthcare, but it’s more prevalent and much more fundamental to the industry. For example, a single patient visit may entail multiple diagnoses and procedures. The doctor notes three problems—headache, sore throat, swollen feet—and prescribes an antibiotic. Multiple providers can add another layer. Such complex data can be difficult to capture and analyze in a meaningful way.

5. Complex set and subset analysis: It’s easy to imagine a reason to want to search data for a specific group of patients, i.e., “all the patients I saw between January 2009 and December 2010 between the ages of 65 and 70 with a particular illness for whom I prescribed a particular medication.” When healthcare organizations purchase enterprise software, they assume it can handle such a query. Unfortunately, most enterprise BI tools don’t handle this type of analysis very well.

6. A lack of technology and focus on business systems: There is a tremendous amount of technology being utilized in healthcare, but most is directed at patient care delivery, not analysis, and the systems aren’t sophisticated enough to tackle big-picture questions or problems. The investment in business-system technology lags significantly behind other industries.

Relevant BI Trends

Four important trends in the realm of BI technology are enabling healthcare companies to make significant strides toward their objectives and to overcome the challenges.

Broad BI platforms meet most of the needs of most of the people.

The long term development of core BI platforms over the past 15 years has reached a maturity that enables delivery of reporting and analysis across the enterprise with feature-rich tools and robust security. Depending on organizational needs, this may be accomplished with a full-service platform deployment from one of the top-tier vendors, a cloud-based BI solution from the more recent contenders, or a best-of-breed solution with a composite of technologies.

Niche tools fill important gaps and improve data handling.

If the platform meets the needs of most people most of the time, what about the rest of the people the rest of the time? That’s where niche tools come in, delivering cost-effective, purpose-fit technology to the users’ desktops. These tools cover a broad range of capabilities: data discovery tools allow integration of enterprise data with external data sets, analytics tools deliver data science capabilities, and data management tools allow cleansing of data and matching of entities across disparate systems.

Mobile BI meets users where they need it most.

Pushing BI content right to the point of care is a current trend that is positively impacting providers’ ability to deliver care and improve patient outcomes. Depending on the needs, these solutions can use standard consumer smartphones and tablet devices or proprietary special-purpose hardware. In either case, the underlying BI solution must be capable of delivering information that is reliable, secure, timely, and relevant in order to leverage the power of mobile BI.

Big data tools push the envelope.

Until recently, it has been impossible to deliver BI with high-volume, fast-arriving data with a long history and in full detail without compromising data with batch processing, summarization, truncated history, etc. While big data analytics is certainly still a maturing technology, it is a reality today and can provide unprecedented richness to BI.

Real-World Steps to Remedy the Problems

How can healthcare organizations use the power of BI to overcome their data challenges and deliver real value? First, get clear on the BI strategy and roadmap. The worst thing a company can do is to throw technology at the problem without clearly defining why.

The fundamentals of BI are the same in any industry in terms of importance of strategy, roadmap and governance. Critical items still revolve around challenges, data quality, security and complexity, and that requires a unique understanding of healthcare challenges and in some cases, different tools than those used outside the healthcare industry because the needs are different. Overall, though, the same principles and design practices apply.

Leaders who want to guarantee BI success must maintain clear executive sponsorship by ensuring a BI strategy that ties back to the company’s overall organizational strategy with a clear roadmap. Clear outputs analysis is another key practice that determines how the requirements are developed for the successful BI system solution. That analysis focuses on answering a few key questions:

1.       Who are the key decision-makers who need better BI tools and content?

2.       What decisions are they trying to make? Where? When? How?

3.       What questions are they asking in the decision-making process?

4.       Are there ways to improve the decision-making process in addition to delivering BI content to support it?

5.       What domains of data are required to answer the questions? What level of detail, history, recency, etc. is required?

6.       How can security issues be confronted from the outset?

7.       How can complex relationships be handled and displayed?

To deliver value with BI, organizations have to find the intersecting points of business outcomes, data, and technology in order to overcome challenges and support key decisions. That process starts with defining the BI strategy and roadmap, then defining the right solution out of the plethora of outstanding options available today.

 

Myron Weber is founder and managing partner of Northwood Advisors. He produces the “Real Time Decisions Webcast” blog and podcast, and was recognized as an IBM Champion for Business Analytics in 2011 and 2012, reflecting his contribution to advancing BI practices and outcomes for business success.  

Image: VILevi/Shutterstock.com

Related