FREE DM Review Site Registration!
Sign-up today and access DM Review on the Web!

Your FREE registration entitles you to:
FREE email newsletters

FREE access to all DM Review content

FREE access to web seminars, resource portals, our white paper library and more!

   

Publisher reserves the right to serve qualified requesters only.

Data Warehousing Makes Way for Health Analytics

Data Strategy Adviser

A modern medical operating room bears little resemblance to one from 50 or a 100 years ago. Technology has worked wonders in transforming the practice of medicine. Yet the back office of the medical center or hospital admissions room is hardly modern. There are tons of paper and manual procedures that remain unintegrated with the actual medical events taking place once a visit or admission has occurred. These unconnected silos of data contribute significantly to the overall lack of transparency in prices and clinical outcomes. Even when costs are available, they are not meaningful without other parameters such as patient demographics, symptoms and related variables. Obviously a single silver bullet will not be sufficient to transform this situation. Many silver bullets will be needed. This article will look at three of them - data warehousing, electronic healthcare records (EHR) and pay-for-performance. The first is a technology, the second is an enabler of operational efficiency (among other things) and the third is a business solution. What knits the three of these together is the common theme of health analytics.

Since this discussion uses only information that is already publicly available, a scan of the Internet using Google reveals what is a matter of public record. The Blue Cross Blue Shield Association is building a data warehouse called Blue Health Intelligence or BHI. "BHI will play a central role in transforming the present healthcare system into a focused, knowledge-based healthcare system," said BCBSA President and CEO Scott P. Serota. "In the near term, BHI will support the efforts of employers to better manage the healthcare benefits they offer their employees. Eventually, BHI will provide consumers with the information they need to make informed healthcare decisions and will heighten collaboration with providers as they deliver high-quality, evidence-based care to their patients." 1

While not necessarily an example of the health analytics available from the BHI data warehouse, Figure 1 shows a basic health analytics data cube. Health analytics data cubes tend to include a large number of dimensions, data sparseness, non-additive facts, volatile attributes and dimensions that change over time (time variant). Such analytics enable doctors contemplating treatment for a form of skin cancer to know what procedures and medications were employed in the last 50 occurrences with anonymized patients fitting a profile similar to the current subject. If searching for patients to join a drug trial, once patient permission is granted, an algorithm will undo the anonymization to arrange the necessary communications. If financial details are required about the therapy, the claim dimension can be resolved to provide the particular about costs. It is clear that aggregations of data will make possible the identification, testing, and improvement of healthcare best practices based on factual evidence, information-based medicine.

Figure 1: Health Data Analytics Cube

Meanwhile, when one looks at the amount of paperwork in the hospital, doctor's office and outpatient clinics, it is obvious to visual inspection that basic automation and workflow can increase productivity in significant ways. One form of automation that has gained support across providers, payers and many consumers as well as governmental policy makers is the electronic healthcare record (EHR). This is where data warehouses need to be complemented with the network effect of widespread adoption and buy-in.

Today the electronic healthcare record (EHR) is known in the healthcare world as being a transformational technology for reducing costs and improving clinical outcomes, benefiting patient well being. According to a Wall Street Journal article,2 Intel Corp., Wal-Mart Stores Inc. and British Petroleum have data warehousing initiatives to allow consumers to coordinate their own healthcare based on an EHR. (Along similar lines, in October 2005, IBM committed to rolling out an online personal health record for its U.S.-based employees.3) In another example, Geisinger Health System has been digitizing its medical records since 1997 according to widely circulated press reports (see footnotes). It is now in a position to exploit the first-mover advantage of its foundational work in EHRs to perform data mining. Mining and analyzing this information will improve treatment outcomes, best practices and genomic profiles as well as patient preferences based on finances and life style. "Most of the world is just getting to the point where they're saying, 'We should install electronic medical record systems,' says Ronald Paulus, M.D., Geiginger's chief health information technology officer, "We're moving up to the next level."4 This is an outgrowth of our commitment to health IT," said Paulus, in a separate conversation with Health IT News. Geisinger has had EHRs for 10 years, "but we wanted to be able to organize our data in a way that would allow us to gain more clinical insight. We didn't have the resources to develop this internally, but during discussions with IBM we recognized we shared a vision about the need to integrate real-time clinical data with historical data."5

Today Geisinger is in a leadership position, but also figuratively in the position of the owner of the first fax machine ever made. The value of the fax function grows explosively as more and more people acquire and use them. As additional payers, providers, employers, and consumers come on stream, the probability is that the value of Geisinger's investments will be increasingly rewarded.

Health analytics is also on the critical path to enabling a practical approach to pay for performance (p-for-p) by connecting the dots between the underlying data warehousing technology and the mission of bringing breakthrough pay for performance results to practice groups and healthcare providers. What each of these innovations has in common is an analytic layer devoted to abstracting lessons learned from basic clinical dimensions.

While p-for-p means many different things to different stakeholders, when it works right, it means identifying and reducing the cause of variations in care and the price of care. It can do this because a defined benchmark has been laid down against which performance and cost can be compared going forward. It does not mean requiring a primary care physician who is already seeing 24 patients a day to see 32 going forward.

Pay for performance secures, aggregates, anonymizes and literally converges - meaningfully merges - clinical and financial data for process optimization. It should also be noted that a method or algorithm is required to reidentify patients from the anonymized version of data if permission is given to identify and use them in clinical trials. For those medical practice groups whose data complexity and readiness for analytics is incomplete, whether due to obsolete systems or legacy practices, providers will require data integration services prior to implementing a full blown p-for-p system.

Once in place, p-for-p will provide evidence for other information-based breakthroughs in healthcare. For example, one innovative (and disruptive) hypothesis is that nurse practitioners6 can provide front-line medical care every bit as good, if not better, than that delivered by internists with M.D.s. In many circles, this is heresy. And which healthcare consumers want to be the guinea pig for such an experiment? What are the facts? This is where evidence based on aggregates and related health analytics can offer the meaningful trade-off. The proposal is provide high quality care practitioners to healthcare at a reduced price for the majority of illness for which we seek treatment - colds, flus, aches and pains, and lifestyle ailments. Such an alternative looks attractive to growing families where convenience and cost are at a premium. The betting money is that it would help produce both results consistent with high quality care. A single fact is worth a thousand opinions in taking on the established interests of the stakeholders. Obviously, there is a lot of work to be done. But tough problems call forth tough problem solvers. As healthcare analytics incorporates data warehousing and the electronic healthcare record, innovative solutions such as pay-for-performance in healthcare will make a difference, improving health and well being as well as the underlying economics.

References:

1.http://www.bcbs.com/innovations/bhi/060804-BHI-Release.pdf [site checked on March 14, 2007].
2. McWilliams, Gary. "Big Employers Plan Electronic Health Records." Wall Street Journal. November 29, 2006, page B1.
3.https://w3.ibm.com/jct03001ps/news/w3news/top_stories/2005/10/personal_health_records.html
4. McGee, Marianne Kolbasuk. "IBM, Geisinger Health Deal Aims To Provide More Personalized Patient Care." InformationWeek. October 10, 2006.  http://www.informationweek.com/story/showArticle.jhtml?articleID=193200289
5. Pizzi, Richard. "Geisinger and IBM collaborate on new IT infrastructure."  Healthcare IT News.  October 13, 2006. http://www.healthcareitnews.com/story.cms?id=5723 [checked on March 14, 2007].
6.  Nurse practitioners prescribe medications and have a DEA registration number. Nurse practitioners bill for Medicare and Medicaid and private insurance for services performed as do doctors. Nurse practitioners provide They provide many of the same services as physicians.


Lou Agosta is a business intelligence analyst with IBM. A former industry analyst at Giga Information Group, Agosta has published extensively on industry trends in data warehousing, data mining and data quality. He can be reached at LoAgosta@us.ibm.com.

For more information on related topics, visit the following channels:



Industry Vendors